Zbornik 27. mednarodne multikonference



INFORMACIJSKA DRUŽBA – IS 2024



Proceedings of the 27th International Multiconference



INFORMATION SOCIETY – IS 2024



Slovenska konferenca o umetni inteligenci

Slovenian Conference on Artificial Intelligence



Kognitivna znanost

Cognitive Science



Odkrivanje znanja in podatkovna skladišča - SiKDD

Data Mining and Data Warehouses – SiKDD



Miti in resnice o varovanju okolja

Myths and Truths about Environmental Protection



17. Mednarodna konferenca o prenosu tehnologij

17th International Technology Transfer Conference



Demografske in družinske analize

Demographic and Family Analyses



Vzgoja in izobraževanje v informacijski družbi

Education in Information Society



Digitalna vključenost v informacijski družbi

Digital Inclusion in Information Society



Konferenca o zdravi dolgoživosti

Conference on Healthy Longevity



Legende računalništva in informatike

Legends of Computing and Informatics



ChatGPT v zdravstvu

ChatGPT in Medicine



Digitalna preobrazba zdravstvene nege - SIZN 2024

Digital Transformation of Nursing - NIIG 2024



Uredniki in urednice / Editors



Mitja Luštrek, Matjaž Gams, Rok Piltaver, Anka Slana Ozimič, Borut Trpin, Toma Strle, Olga Markič, Dunja

Mladenić, Marko Grobelnik, Rafael Mihalič, Tomaž Ogrin, Erik Margan, Urška Florjančič, Robert Blatnik,

Špela Stres, Janez Malačič, Mari Jože Osredkar, Uroš Rajkovič, Borut Batagelj, Matjaž Debevc, Ines Kožuh,

Martin Lipovšek, Boštjan Petrič, Tone Stanovnik, Monika Simjanoska Misheva, Stevo Lukić, Franz Wotawa,

Jelena Ficzko, Ema Dornik, Uroš Rajkovič, Tamara Štemberger Kolnik, Gregor Štiglic, Boštjan Žvanut



http://is.ijs.si



7.–11. oktober 2024 / 7–11 October 2024

Ljubljana, Slovenia

Uredniki:



Mitja Luštrek, Odsek za inteligentne sisteme, Institut »Jožef Stefan«, Ljubljana

Matjaž Gams, Odsek za inteligentne sisteme, Institut »Jožef Stefan«, Ljubljana

Rok Piltaver, Outfit7, Ljubljana

Anka Slana Ozimič, Filozofska fakulteta, Univerza v Ljubljani

Borut Trpin, Filozofska fakulteta, Univerza v Ljubljani



Toma Strle, Center za Kognitivno znanost, Pedagoška fakulteta, Univerza v Ljubljani

Olga Markič, Filozofska fakulteta, Univerza v Ljubljani

Dunja Mladenić, Department for Artificial Intelligence, Jožef Stefan Institute, Ljubljana

Marko Grobelnik, Department for Artificial Intelligence, Jožef Stefan Institute, Ljubljana

Rafael Mihalič, Fakulteta za elektrotehniko, Univerza v Ljubljana

Tomaž Ogrin, Oddelek za anorgansko kemijo in tehnologijo, Institut Jožef Stefan, Ljubljana

Erik Margan, Oddelek za eksperimentalno fiziko osnovnih delcev, Institut Jožef Stefan, Ljubljana



Urška Florjančič, Služba za vsebinsko podporo projektom, prenos tehnologij in inovacije, Institut »Jožef Stefan«, Ljubljana

Robert Blatnik, Služba za vsebinsko podporo projektom, prenos tehnologij in inovacije, Institut »Jožef Stefan«, Ljubljana



Špela Stres

Janez Malačič, Ekonomska fakulteta, Ljubljana

Mari Jože Osredkar, Teološka fakulteta Poljanska 4, 1000 Ljubljana

Uroš Rajkovič, Fakulteta za organizacijske vede, Univerza v Mariboru

Borut Batagelj, Fakulteta za računalništvo in informatiko, Univerza v Ljubljani

Matjaž Debevc, Fakulteta za elektrotehniko, računalništvo in informatiko, Univerza v Mariboru

Ines Kožuh, Fakulteta za elektrotehniko, računalništvo in informatiko, Univerza v Mariboru

Martin Lipovšek, Društvo za vitalno podaljševanje življenja Slovenije

Boštjan Petrič, Medicinska fakulteta Univerza v Ljubljani

Tone Stanovnik, Špica, Ljubljana

Monika Simjanoska Misheva, Fakulteta za računalništvo in elektrotehniko, Skopje, Makedonija



Stevo Lukić, Nevrološka klinika Niš, Srbija

Franz Wotawa, Inštitut za tehnologijo programske opreme, Gradec, Avstrija

Jelena Ficzko, Zdravstvena fakulteta, Univerza v Ljubljani

Ema Dornik, Institut informacijskih znanosti



Tamara Štemberger Kolnik, Fakulteta za zdravstvene vede v Celju

Gregor Štiglic, Fakulteta za zdravstvene vede, Univerza v Mariboru

Boštjan Žvanut, Fakulteta za vede o zdravju, Univerza na Primorskem



Založnik: Institut »Jožef Stefan«, Ljubljana

Priprava zbornika: Mitja Lasič, Vesna Lasič, Lana Zemljak

Oblikovanje naslovnice: Vesna Lasič



Dostop do e-publikacije:

http://library.ijs.si/Stacks/Proceedings/InformationSociety



Ljubljana, december 2024



2. objava



Informacijska družba

ISSN 2630-371X



Kataložni zapis o publikaciji (CIP) pripravili v Narodni in univerzitetni knjižnici v Ljubljani

COBISS.SI-ID 217777923

ISBN 978-961-264-314-0 (PDF)





PREDGOVOR MULTIKONFERENCI




INFORMACIJSKA DRUŽBA 2024



Leto 2024 je hkrati udarno in tradicionalno. Že sedaj, še bolj pa v prihodnosti bosta računalništvo, informatika

(RI) in umetna inteligenca (UI) igrali ključno vlogo pri oblikovanju napredne in trajnostne družbe. Smo na

pragu nove dobe, v kateri generativna umetna inteligenca, kot je ChatGPT, in drugi inovativni pristopi utirajo

pot k superinteligenci in singularnosti, ključnim elementom, ki bodo definirali razcvet človeške civilizacije.

Naša konferenca je zato hkrati tradicionalna znanstvena, pa tudi povsem akademsko odprta za nove pogumne

ideje, inkubator novih pogledov in idej.



Letošnja konferenca ne le da analizira področja RI, temveč prinaša tudi osrednje razprave o perečih temah

današnjega časa – ohranjanje okolja, demografski izzivi, zdravstvo in preobrazba družbenih struktur. Razvoj

UI ponuja rešitve za skoraj vse izzive, s katerimi se soočamo, kar poudarja pomen sodelovanja med

strokovnjaki, raziskovalci in odločevalci, da bi skupaj oblikovali strategije za prihodnost. Zavedamo se, da

živimo v času velikih sprememb, kjer je ključno, da s poglobljenim znanjem in inovativnimi pristopi

oblikujemo informacijsko družbo, ki bo varna, vključujoča in trajnostna.



Letos smo ponosni, da smo v okviru multikonference združili dvanajst izjemnih konferenc, ki odražajo širino

in globino informacijskih ved: CHATMED v zdravstvu, Demografske in družinske analize, Digitalna

preobrazba zdravstvene nege, Digitalna vključenost v informacijski družbi – DIGIN 2024, Kognitivna

znanost, Konferenca o zdravi dolgoživosti, Legende računalništva in informatike, Mednarodna konferenca o

prenosu tehnologij, Miti in resnice o varovanju okolja, Odkrivanje znanja in podatkovna skladišča – SIKDD

2024, Slovenska konferenca o umetni inteligenci, Vzgoja in izobraževanje v RI.



Poleg referatov bodo razprave na okroglih mizah in delavnicah omogočile poglobljeno izmenjavo mnenj, ki

bo oblikovala prihodnjo informacijsko družbo. “Legende računalništva in informatike” predstavljajo

slovenski “Hall of Fame” za odlične posameznike s tega področja, razširjeni referati, objavljeni v reviji

Informatica z 48-letno tradicijo odličnosti, in sodelovanje s številnimi akademskimi institucijami in

združenji, kot so ACM Slovenija, SLAIS in Inženirska akademija Slovenije, bodo še naprej spodbujali razvoj

informacijske družbe. Skupaj bomo gradili temelje za prihodnost, ki bo oblikovana s tehnologijami,

osredotočena na človeka in njegove potrebe.



S podelitvijo nagrad, še posebej z nagrado Michie-Turing, se avtonomna RI stroka vsakoletno opredeli do

najbolj izstopajočih dosežkov. Nagrado Michie-Turing za izjemen življenjski prispevek k razvoju in

promociji informacijske družbe je prejel prof. dr. Borut Žalik. Priznanje za dosežek leta pripada prof. dr. Sašu

Džeroskemu za izjemne raziskovalne dosežke. »Informacijsko limono« za najmanj primerno informacijsko

tematiko je prejela nabava in razdeljevanjem osebnih računalnikov ministrstva, »informacijsko jagodo« kot

najboljšo potezo pa so sprejeli organizatorji tekmovanja ACM Slovenija. Čestitke nagrajencem!



Naša vizija je jasna: prepoznati, izkoristiti in oblikovati priložnosti, ki jih prinaša digitalna preobrazba, ter

ustvariti informacijsko družbo, ki bo koristila vsem njenim članom. Vsem sodelujočim se zahvaljujemo za

njihov prispevek k tej viziji in se veselimo prihodnjih dosežkov, ki jih bo oblikovala ta konferenca.



Mojca Ciglarič, predsednica programskega odbora



Matjaž Gams, predsednik organizacijskega odbora





PREFACE TO THE MULTICONFERENCE




INFORMATION SOCIETY 2024



The year 2024 is both ground-breaking and traditional. Now, and even more so in the future, computer

science, informatics (CS/I), and artificial intelligence (AI) will play a crucial role in shaping an advanced and

sustainable society. We are on the brink of a new era where generative artificial intelligence, such as

ChatGPT, and other innovative approaches are paving the way for superintelligence and singularity—key

elements that will define the flourishing of human civilization. Our conference is therefore both a traditional

scientific gathering and an academically open incubator for bold new ideas and perspectives.



This year's conference analyzes key CS/I areas and brings forward central discussions on pressing

contemporary issues—environmental preservation, demographic challenges, healthcare, and the

transformation of social structures. AI development offers solutions to nearly all challenges we face,

emphasizing the importance of collaboration between experts, researchers, and policymakers to shape future

strategies collectively. We recognize that we live in times of significant change, where it is crucial to build

an information society that is safe, inclusive, and sustainable, through deep knowledge and innovative

approaches.



This year, we are proud to have brought together twelve exceptional conferences within the multiconference

framework, reflecting the breadth and depth of information sciences:



• CHATMED in Healthcare

• Demographic and Family Analyses

• Digital Transformation of Healthcare Nursing

• Digital Inclusion in the Information Society – DIGIN 2024

• Cognitive Science

• Conference on Healthy Longevity

• Legends of Computer Science and Informatics

• International Conference on Technology Transfer

• Myths and Facts on Environmental Protection

• Data Mining and Data Warehouses – SIKDD 2024

• Slovenian Conference on Artificial Intelligence

• Education and Training in CS/IS.



In addition to papers, roundtable discussions and workshops will facilitate in-depth exchanges that will help

shape the future information society. The “Legends of Computer Science and Informatics” represents

Slovenia’s “Hall of Fame” for outstanding individuals in this field. At the same time, extended papers

published in the Informatica journal, with over 48 years of excellence, and collaboration with numerous

academic institutions and associations, such as ACM Slovenia, SLAIS, and the Slovenian Academy of

Engineering, will continue to foster the development of the information society. Together, we will build the

foundation for a future shaped by technology, yet focused on human needs.



The autonomous CS/IS community annually recognizes the most outstanding achievements through the

awards ceremony. The Michie-Turing Award for an exceptional lifetime contribution to the development and

promotion of the information society was awarded to Prof. Dr. Borut Žalik. The Achievement of the Year

Award goes to Prof. Dr. Sašo Džeroski. The "Information Lemon" for the least appropriate information topic

was given to the ministry's procurement and distribution of personal computers. At the same time, the

"Information Strawberry" for the best initiative was awarded to the organizers of the ACM Slovenia

competition. Congratulations to all the award winners!



Our vision is clear: to recognize, seize, and shape the opportunities brought by digital transformation and

create an information society that benefits all its members. We thank all participants for their contributions

and look forward to this conference's future achievements.



Mojca Ciglarič, Chair of the Program Committee



Matjaž Gams, Chair of the Organizing Committee

KONFERENČNI ODBORI



CONFERENCE COMMITTEES



International Programme Committee Organizing Committee



Vladimir Bajic, South Africa Matjaž Gams, chair

Heiner Benking, Germany Mitja Luštrek

Se Woo Cheon, South Korea Lana Zemljak

Howie Firth, UK Vesna Koricki

Olga Fomichova, Russia Mitja Lasič

Vladimir Fomichov, Russia Blaž Mahnič

Vesna Hljuz Dobric, Croatia

Alfred Inselberg, Israel

Jay Liebowitz, USA

Huan Liu, Singapore

Henz Martin, Germany

Marcin Paprzycki, USA

Claude Sammut, Australia

Jiri Wiedermann, Czech Republic

Xindong Wu, USA

Yiming Ye, USA

Ning Zhong, USA

Wray Buntine, Australia

Bezalel Gavish, USA

Gal A. Kaminka, Israel

Mike Bain, Australia

Michela Milano, Italy

Derong Liu, Chicago, USA

Toby Walsh, Australia

Sergio Campos-Cordobes, Spain

Shabnam Farahmand, Finland

Sergio Crovella, Italy



Programme Committee



Mojca Ciglarič, chair Marjan Heričko Baldomir Zajc

Bojan Orel Borka Jerman Blažič Džonova Blaž Zupan

Franc Solina Gorazd Kandus Boris Žemva

Viljan Mahnič Urban Kordeš Leon Žlajpah

Cene Bavec Marjan Krisper Niko Zimic

Tomaž Kalin Andrej Kuščer Rok Piltaver

Jozsef Györkös Jadran Lenarčič Toma Strle

Tadej Bajd Borut Likar Tine Kolenik

Jaroslav Berce Janez Malačič Franci Pivec

Mojca Bernik Olga Markič Uroš Rajkovič

Marko Bohanec Dunja Mladenič Borut Batagelj

Ivan Bratko Franc Novak Tomaž Ogrin

Andrej Brodnik Vladislav Rajkovič Aleš Ude

Dušan Caf Grega Repovš Bojan Blažica

Saša Divjak Ivan Rozman Matjaž Kljun

Tomaž Erjavec Niko Schlamberger Robert Blatnik

Bogdan Filipič Stanko Strmčnik Erik Dovgan

Andrej Gams Jurij Šilc Špela Stres

Matjaž Gams Jurij Tasič Anton Gradišek

Mitja Luštrek Denis Trček

Marko Grobelnik Andrej Ule

Nikola Guid Boštjan Vilfan

KAZALO / TABLE OF CONTENTS



Slovenska konferenca o umetni inteligenci / Slovenian Conference on Artificial Intelligence ................ 1

PREDGOVOR / FOREWORD ............................................................................................................................... 3

PROGRAMSKI ODBORI / PROGRAMME COMMITTEES ............................................................................... 5

PandaChat-RAG: Towards the Benchmark for Slovenian RAG Applications / Kuzman Taja, Pavleska Tanja,

Rupnik Urban, Cigoj Primož .............................................................................................................................. 7

Choosing Features for Stress Prediction with Machine Learning / Bengeri Katja, Lukan Junoš, Luštrek Mitja . 11

Predictive Modeling of Football Results in the WWIN League of Bosnia and Herzegovina / Vladić Ervin,

Mehanović Dželila, Avdić Elma ...................................................................................................................... 15

Sarcasm Detection in a Less-Resourced Language / Đoković Lazar, Robnik-Šikonja Marko ............................ 19

Speech-to-Service: Using LLMs to Facilitate Recording of Services in Healthcare / Smerkol Maj, Susič Rok,

Ratajec Mariša, Halbwachs Helena, Gradišek Anton ...................................................................................... 23

Performance Comparison of Axle Weight Prediction Algorithms on Time-Series Data / Kolar Žiga, Susič

David, Konečnik Martin, Prestor Domen, Pejanovič Nosaka Tomo, Kulauzović Bajko, Kalin Jan, Skobir

Matjaž, Gams Matjaž ....................................................................................................................................... 27

Comparison of Feature- and Embedding-based Approaches for Audio and Visual Emotion Classification /

Trojer Sebastijan, Anžur Zoja, Luštrek Mitja, Slapničar Gašper ..................................................................... 31

Multi-modal Data Collection and Preliminary Statistical Analysis for Cognitive Load Assessment / Krstevska

Ana, Kramar Sebastjan, Gjoreski Hristijan, Gjoreski Martin, Lukan Junoš, Trojer Sebastijan, Luštrek Mitja,

Slapničar Gašper .............................................................................................................................................. 35

Predicting Health-Related Absenteeism with Machine Learning: A Case Study / Piciga Aleksander, Kukar

Matjaž ............................................................................................................................................................... 39

Puzzle Generation for Ultimate-Tic-Tac-Toe / Zirkelbach Maj, Sadikov Aleksander ......................................... 43

Ethical Consideration and Sociological Challenges in the Integration of Artificial Intelligence in Mental Health

Services / Poljak Lukek Saša........................................................................................................................... 47

Optimization Problem Inspector: A Tool for Analysis of Industrial Optimization Problems and Their Solutions /

Tušar Tea, Cork Jordan, Andova Andrejaana, Filipič Bogdan ........................................................................ 51

Multi-Agent System for Autonomous Table Football: A Winning Strategy / Založnik Marcel, Šoln Kristjan ... 55

Towards a Decision Support System for Project Planning: Multi-Criteria Evaluation of Past Projects Success /

Hafner Miha, Bohanec Marko .......................................................................................................................... 59

Minimizing Costs and Risks in Demand Response Optimization: Insights from Initial Experiments / Nedić

Mila, Tušar Tea ................................................................................................................................................ 63

Predicting Hydrogen Adsorption Energies on Platinum Nanoparticles and Surfaces With Machine Learning /

Gašparič Lea, Kokalj Anton, Džeroski Sašo .................................................................................................... 67

SmartCHANGE Risk Prediction Tool: Demonstrating Risk Assessment for Children and Youth / Jordan

Marko, Reščič Nina, Kramar Sebastjan, Založnik Marcel, Luštrek Mitja ....................................................... 71

Predicting Mental States During VR Sessions Using Sensor Data and Machine Learning / Kizhevska Emilija,

Luštrek Mitja .................................................................................................................................................... 75

Biomarker Prediction in Colorectal Cancer Using Multiple Instance Learning / Shulajkovska Miljana, Jelenc

Matej, Jonnagaddala Jitenndra, Gradišek Anton .............................................................................................. 79

Feature-Based Emotion Classification Using Eye-Tracking Data / Božak Tomi, Luštrek Mitja, Slapničar Gašper

.......................................................................................................................................................................... 83



Kognitivna znanost / Cognitive Science ..................................................................................................... 87

PREDGOVOR / FOREWORD ............................................................................................................................. 89

PROGRAMSKI ODBORI / PROGRAMME COMMITTEES ............................................................................. 91

The Ontological Shock: What Psychedelics can Teach us about the Nature of the Mind / Sirk Maruša ............. 93

Innovative Supporting Approaches: Integrating Bibliotherapy, Psychodrama and AI as a Therapeutic

Conversational Tool / Mattová Veronika, Lazore Courtney ........................................................................... 98

Usability of Intelligent Assistive Technology Used by People With Dementia and Their Caregivers / Dečman

Klara ............................................................................................................................................................... 102

Open Science and Goodhart's Law / Pisanski Tomaž, Batagelj Vladimir, Pisanski Jan .................................... 106



The Consistency of the Research Field Data: A Case Study of Library and Information Science in Slovenia /

Pisanski Jan .................................................................................................................................................... 110

To Be or Not to Be… a Nahuatl Language Learning App. The Long-Term Survival or Discontinuation of

Indigenous Language Learning Apps on the Example of Nahuatl / Fischer Evelyn .................................... 113

Designing the Flow State Experience Using Modern Digital Technologies / Vidmar Eva................................ 117

The Transparency of Nudging: Evaluating Its Impact on Personal Autonomy / Pajmon Sabina, Strle Toma ... 121

Does the Use of Large Language Models in Scientific Research Bring Us Closer to the Point in Time When

Machines Will Dominate Humans? / Mali Franc .......................................................................................... 125

Comparing Academic Performance Across Course Topics: A Pilot Study / Fink Laura, Cestnik Bojan .......... 130

Linking the Normative and the Descriptive: Bounded Epistemic Rationality / Tomat Nastja ........................... 136

Exploring Human Perception Using Virtual Reality / Zibrek Katja ................................................................... 141

Vpliv generativne umetne inteligence na demokracijo / Košmrlj Lea, Bratko Ivan .......................................... 145

Razložljiva umetna inteligenca: kako naprej? / Farič Ana, Bratko Ivan ............................................................ 150

Exploring Cognitive Science under Analytical Idealism / Rodman Grega ........................................................ 155

Intelligent Revolution – a New Civilization and Cognitive Era / Gams Matjaž ................................................ 158

Cognitive Perspective on Production of Third Person Dative and Accusative Clitic Pronouns in Slovenian

School-Aged Children / Brežnik Dornik Maruša .......................................................................................... 164

Ballot Butts: Nudging towards Pro Environmental Behaviour / Hartmans Anouk, Karnelutti Lucija, Žužek

Leon, Strle Toma, Pajmon Sabina .................................................................................................................. 167

Problem Solving as a Key for Sustainable Future / Štibi Ivana, Gaurina Marija, Katavić Ivana, Stepanić Josip

........................................................................................................................................................................ 171

Mind, the Gap, and Other Cracks / Poljšak Kus Maša, Kordeš Urban ............................................................... 175

Bridging the Challenges in Experience Sampling Research / Seme Barbi, Sirk Maruša, Kordeš Urban .......... 179



Odkrivanje znanja in podatkovna skladišča - SiKDD / Data Mining and Data Warehouses - SiKDD 183

PREDGOVOR / FOREWORD ........................................................................................................................... 185

PROGRAMSKI ODBORI / PROGRAMME COMMITTEES ........................................................................... 187

Integrating Knowledge Graphs and Large Language Models for Querying in an Industrial Environment / Kenda

Klemen, Hočevar Domen ............................................................................................................................... 189

Comparative Analysis of Machine Learning Models for Groundwater Level Forecasting: The Impact of

Contextual Data / Klančič Rok, Kenda Klemen ............................................................................................ 193

Interactive Tool for Tracking Open-source Artificial Intelligence Progress on Hugging Face / Šinik Bogdan,

Vake Domen, Vičić Jernej, Tošić Aleksander................................................................................................ 197

Multilingual Hate Speech Modeling by Leveraging Inter-Annotator Disagreement / Grigor Patricia-Carla, Kralj

Novak Petra, Evkoski Bojan .......................................................................................................................... 201

Predicting Pronunciation Types in the Sloleks Morphological Lexicon of Slovene / Čibej Jaka ...................... 205

Higher-order bibliographic services based on bibliographic networks / Batagelj Vladimir, Pisanski Jan, Pisanski

Tomaž ............................................................................................................................................................. 209

Are papers all that counts? A bibliometric analysis of the Slovenian scientific community / Dupuis Aymeric,

Džeroski Sašo, Koloski Boshko, Martinc Matej ............................................................................................ 213

Empowering Open Education Methodologies with AI-based Strategies for the Customization of Education /

Amiel Tel, Mores Neto Antonio J., Pita Costa Joao, Polajnar Anja, Jermol Mitja ........................................ 217

Addressing Water Sustainability Challenges in North Africa with Artificial Intelligence / Zaouini Mustafa, Pita

Costa Joao, Cherakaoui Manal, Hachimi Hanaa, Abkari M. Wahib, Gourari Kamal, Lachheb Hatim, Tounsi

El Azzoiani Jad ............................................................................................................................................... 221

Predicting poverty using regression / Urbanč Luka, Grobelnik Marko, Pita Costa Joao ................................... 225

Fact Manipulation in News: LLM-Driven Synthesis and Evaluation of Fake News Annotation / Golob Luka,

Sittar Abdul .................................................................................................................................................... 229

Borrowing Words: Transfer Learning for Reported Speech Detection in Slovenian News Texts / Fijavž Zoran

........................................................................................................................................................................ 233

Connecting company performance to ESG terms in financial reports / Andrenšek Luka, Sitar Šuštar Katarina,

Pollak Senja, Purver Matthew ........................................................................................................................ 237

Classification of Patents Into Knowledge Fields: Using a Proposed Knowledge Mapping Taxonomy

(KnowMap) / Motamedi Elham, Novalija Inna, Rei Luis ............................................................................. 241



Enhancing causal graphs with domain knowledge: matching ontology concepts between ontologies and raw text

data / Stegnar Jernej, Rožanec Jože M., Leban Gregor, Mladenić Dunja ..................................................... 245

Measuring and Modeling CO2 Emissions in Machine Learning Processes / Hrib Ivo, Šturm Jan, Topal

Oleksandra, Škrjanc Maja .............................................................................................................................. 249

Enhancing Ontology Engineering with LLMs: From Search to Active Learning Extensions / Kholmska Ganna,

Kenda Klemen, Rožanec Jože M. .................................................................................................................. 255

On the Brazilian Observatory for Artificial Intelligence / Meira Silva Rafael, Godoy Oliveira Cristina, Costa

Luiz, Candia Vieira Joao Paulo, Pita Costa Joao ........................................................................................... 259

Pojavljanje incidentov ob uporabi Umetne Inteligence / Grobelnik Marko, Massri M. Besher, Guček Alenka,

Mladenić Dunja .............................................................................................................................................. 263

Perception of AI in Slovenia / Sittar Abdul, Guček Alenka, Mladenić Dunja ................................................... 267

Naslov / Šker Tesia, Rožanec Jože M., Leban Gregor, Mladenić Dunja ........................................................... 271

Generating Non-English Synthetic Medical Data Sets / Dolinar Lenart, Calcina Erik, Novak Erik ................. 275

LLNewsBias: A Multilingual News Dataset for Lifelong Learning / Swati, Mladenić Dunja .......................... 279

Creating Local World Models using LLMs / Longar Mark David, Novak Erik, Grobelnik Marko .................. 283

Semantic video content search and recommendation / Longar Mark David, Fir Jakob, Pangeršič Bor ............ 287

Continuous Planning of a Fleet of Shuttle Vans as Support for Dynamic Pricing / Stavrov Filip, Stopar Luka 291

Knowledge graph Extraction from Textual data using LLM / Gilliani Khasa, Novak Erik, Kenda Klemen,

Mladenić Dunja .............................................................................................................................................. 295

Solving hard optimization problems of packing, covering, and tiling via clique search / Szabo Sandor, Zavalnij

Bogdan ........................................................................................................................................................... 299



Miti in resnice o varovanju okolja / Myths and Truths About Environmental Protection .................... 303

PREDGOVOR / FOREWORD ........................................................................................................................... 305

PROGRAMSKI ODBORI / PROGRAMME COMMITTEES ........................................................................... 307

LCA data conforms to Benford’s law / Šinik Bogdan, Tošić Aleksander .......................................................... 309

Liquid Neural Network in Modelling Rainfall-Runoff Modelling / Teklu Henok, Gams Matjaž, Wielgosz

Maciej ............................................................................................................................................................. 313

The governance of cultural landscapes in the EU: A Comparative Analysis of the Nature Restoration Law and

the LULUCF Regulations / Schenk Richard ................................................................................................. 317

Serious Errors in the Assessment of the Atmospheric Greenhouse Effect / Margan Erik ................................. 321

Vrednost električne energije intermitirajočih virov in kanibalizem na trgu z električno energijo / Mihalič Rafael

........................................................................................................................................................................ 329

Kako razumeti ogljični odtis? / Radonjič Gregor ............................................................................................... 336

Ali je prispevek slovenske živinoreje k podnebnim spremembam precenjen? / Verbič Jože ............................ 339

Oblikovanje specifičnih habitatov kot ukrepanje za zmanjšanje vplivov hidroelektrarn na okolje / Hribar

Andraž ............................................................................................................................................................ 343

Aksiomi in miti o varovanju okolja in trajnostnem kmetijstvu / Gams Matjaž .................................................. 349

Umetna inteligenca o zeleni agendi / Špiler Jure ............................................................................................... 356

Radon v stavbah in kakovost bivanja / Valenčič Matjaž .................................................................................... 364

Postavljanje sončnih elektrarn na kmetijskih površinah - okoljski kriminal ali trajnostna raba prostora /

Valenčič Matjaž .............................................................................................................................................. 369

Skrivnost ljubljanske pitne vode in kanal C0 / Ogrin Tomaž ............................................................................. 375

Kakšna je cena jedrske energije in obnovljivih virov in kakšen je vpliv na okolje / Trkov Andrej ................... 381

O statusu in prihodnosti jedrske energije / Cizelj Leon ...................................................................................... 384

Natura 2000 in ohranjanje narave, spodbude ali ovire pri razvoju elektroenergetskega omrežja / Kregar Aleš 388

Sprejemljive in „nesprejemljive“ metode žlahtnjenja rastlin kot del boja za racionalno odločanje v okoljevarstvu

/ Bohanec Borut ............................................................................................................................................. 393

No fossil fuels - no food / Senegačnik Andrej .................................................................................................... 395

Nemški energetski prehod ter njegove politične in gospodarske posledice / Povh Dušan ................................. 399

Miti in resnice o elektromagnetnem sevanju / Lipič Karel ................................................................................. 402



17. Mednarodna konferenca o prenosu tehnologij / 17th International Technology Transfer Conference

............................................................................................................................................................... 407

PREDGOVOR / FOREWORD ........................................................................................................................... 409

PROGRAMSKI ODBORI / PROGRAMME COMMITTEES ........................................................................... 411

Intellectual Property as a Success Factor for Startups: Systematic Literature Review / Fortun Novak Maja .... 413

The Reversed European Paradox: do European Patents have a High Market Value but Low Impact? / Hafner

Ana ................................................................................................................................................................. 416

The Importance of Technology Transfer Offices in University Industry Collaboration: KTÜ TTM Example /

İskender Balaban Dilek, Değermenci Beril, Yüksel Harun, Yilmaz Eren, Kalyoncu Sedanur, Ayvaz Emrah,

Yildiz Oktay, Aykut Yalçın, Yildiz İslam, Sağlam Gözde, Ünver Müslüm Serhat, Sabir Hülya, Sönmez

Kerim, Gültekin Güler Tuğba, Aydin Aleyna ................................................................................................ 420

The Impact of International Networks on Grants, R&D, Knowledge and Technology Transfer – Case of COST

Network and KTU / Kalyoncu Sedanur, Yildiz İslam, Ayvaz Emrah, Sağlam Gözde, Ünver Müslüm Serhat,

Yildiz Oktay, Gültekin Güler Tuğba, Yilmaz Eren, Sönmez Kerim, Sabir Hülya, Değermenci Beril, Aykut

Yalçın, Koç Ayhan, İskender Balaban Dilek, Aydin Aleyna, Baş Seda ........................................................ 424

The Effect of Evaluating Graduate Thesis Topics as Invention Notification Form on Industrial and Intellectual

Property Applications: The Case of Karadeniz Technical University / Sönmez Kerim, Ayvaz Emrah, Sabir

Hülya, Değermenci Beril, Kalyoncu Sedanur, Yildiz İslam, Ünver Müslüm Serhat, Aykut Yalçın, Sağlam

Gözde, Yilmaz Eren, Aydin Aleyna, İskender Balaban Dilek, Gültekin Güler Tuğba, Koç Ayhan, Yildiz

Oktay .............................................................................................................................................................. 428

Using Open-Access Resources and Platforms to Create a Technology Transfer Ecosystem / Britchkovski

Viatcheslav ..................................................................................................................................................... 432

Fostering Open Innovation and Technology Transfer: Insights from the Euro-Mediterranean Innovation Camp

(EMIC) / El-Zoheiry Abdelhamid, Gladović Karen ...................................................................................... 436

Research Organisation-Industry Cooperation and State Aid Rules in Slovenia and Europe / Lutman Tomaž,

Florjančič Urška, Fric Urška .......................................................................................................................... 441

Feasibility Analysis for the New Mechanism of Knowledge Transfer within the INDUSAC Project / Odić

Duško, Mrgole Urška, Trobec Marjeta .......................................................................................................... 445

Aproaches to Monitoring and Impact Assessment in Reseach Infrastructures / Plaskan Jure, N. Brečko Barbara

........................................................................................................................................................................ 449

Intellectual Property Valuation in the Cyber Security Sector / E. Wachowicz Marta ........................................ 454

The Challenge of Licensing Artificial Intelligence Technology for Cybersecurity Applications / Rotnicki

Michał ............................................................................................................................................................ 458

Technology Transfer: Revenues Estimation in the Cyber Security Sector / Falkowski Michal J., Kaminski

Jaroslaw, Wachowicz Marta........................................................................................................................... 462

Prospects for the Use of AI Tools in the Republican Center for Technology Transfer Network / Uspenskiy

Alexander, Uspenski Aliaksei, Prybylski Maxim .......................................................................................... 466



Demografske in družinske analize / Demographic and Family Analyses .............................................. 471

PREDGOVOR / FOREWORD ........................................................................................................................... 473

PROGRAMSKI ODBORI / PROGRAMME COMMITTEES ........................................................................... 475

Razvoj empatije, vloga spola in socialne vrednote pri mladostnikih / The development of empathy, the role of

gender and social values in adolescents / Simonič Barbara .......................................................................... 477

Advantages of ICT technology in physical education lessons / Rajher Polona.................................................. 482

E-reading and e-books: specifics and challenges in school practice / Šebjanič Oražem Maja ........................... 486

Zlorabe pri starejših osebah / Repič Slavič Tanja .............................................................................................. 491

Matične knjige pravoslavne župnije sv. Cirila in Metoda v Ljubljani / Osredkar Mari Jože ............................. 496

Emigrations from Kosova in the First Two Decades of the 21st Century- Trend and Challenges / Behrami Sami

........................................................................................................................................................................ 501

The right to grave Possibilities of burying in cemeteries in Slovenia / Čepar Drago ......................................... 507

Physical inactivity burden assessment in Slovenia / Farčnik Daša, Došenović Bonča Petra, Pahor Marko,

Zalaznik Maja, Pruš Daša, Paravlić Armin, Vodičar Janez ........................................................................... 510

Cohort Fertility Trends in Post-Yugoslav States / Čipin Ivan ............................................................................ 514

Pregled planetarnih in lokalnih demografskih projekcij / Gams Matjaž ............................................................ 518



Pokritost strokovne in znanstvene produkcije v Sloveniji v področju epidemiologije depresije kot dela

demografije zdravja v COBISS.SI / Muhvič Šumandl Romana, Šercar Tvrtko-Matija ................................ 524

Ethical Consideration and Sociological Challenges in the Integration of Artificial Intelligence in Mental Health

Services / Poljak Lukek Saša......................................................................................................................... 529

Novejši razvoj rodnosti po redu rojstva v Sloveniji / Malačič Janez ................................................................. 533

The aggregate loss of GDP resulting from premature mortality: a methodological approach based on GDP

decomposition at occupational level / Redek Tjaša, Došenović Bonča Petra, Farčnik Daša, Istenič Tanja . 539

Mitigating adverse health effects of chronic illnesses on work outcomes: A conceptual framework / Došenović

Bonča Petra, Bavdaž Mojca, Ponikvar Nina, Redek Tjaša ............................................................................ 543

Ko tempo sestavina v stopnji celotne rodnosti izzveni / Sambt Jože ................................................................. 547



Vzgoja in izobraževanje v informacijski družbi / Education in Information Society ............................ 551

PREDGOVOR / FOREWORD ........................................................................................................................... 553

PROGRAMSKI ODBORI / PROGRAMME COMMITTEES ........................................................................... 555

Optimizacija praktičnega pouka digitalne tehnike z avtomatskim testiranjem integriranih vezij / Albreht Jaka

........................................................................................................................................................................ 557

Codeium kot podporni asistent v poučevanju programiranja / Babič Davorin................................................... 560

Uporaba spletnega orodja H5P za formativno spremljanje poučevanja računalniškega mišljenja / Babič Davorin

........................................................................................................................................................................ 566

Utrjevanje snovi preko interaktivne sobe pobega / Bežan Simona .................................................................... 571

Pouk matematike z IKT / Bizjak Blanka ............................................................................................................ 575

Do boljšega razumevanja fizikalnih količin s pomočjo temeljnih znanj računalništva in informatike / Bobnarič

Roman, Keček Vaupotič Lenka ...................................................................................................................... 579

Prepletanje kmetijske stroke, strokovne informatike in statistike ter knjižničnih znanj / Bosnar Darinka ........ 584

Digitalne dimenzije ob Prešernovi klasiki: IKT kot orodje za sodobno obravnavo Krsta pri Savici / Breznik

Andreja ........................................................................................................................................................... 588

Programiranje + fizika → Igra se začne / Candellari Zdenka ............................................................................ 592

Razvijanje digitalnih kompetenc na OŠ Prežihovega Voranca Maribor / Carmona Natalija ............................. 595

Povezanost psihoedukativnega pristopa in ravni pismenosti o anksioznih motnjah / Dernovšček Hafner Nataša,

Šprah Lilijana ................................................................................................................................................. 600

Kaj lahko postorimo za svojo varnost na osebnem računalniku? / Divjak Saša ................................................. 605

Digitalna pismenost v osnovnošolskem izobraževanju / Dončev Aleksandar, Jereb Eva .................................. 606

Učenje španščine z jezikovno aplikacijo Duolingo / Dvoraček Ana .................................................................. 610

Vloga umetne inteligence pri ustvarjanju zgodb v animiranih filmih: »Vpliv na kognitivne procese in

izobraževalni sistem« / Guštin Mojca ........................................................................................................... 614

Matematika, logika in Mathematica / Hafner Izidor .......................................................................................... 620

Reklama za branje knjig / Jelen Polona .............................................................................................................. 622

Uporaba video kamere pri poučevanju športne vzgoje / Jurgele Anže .............................................................. 625

Učenje preko ustvarjanja: uporaba IKT v projektu Knjiga podajanka / Kirn Urša ............................................ 628

Spopadanje z izzivi pri razvijanju digitalnih veščin pri učencih z učnimi težavami v redni osnovni šoli / Kokec

Barbara ........................................................................................................................................................... 633

Knjižnično informacijska znanja pri pouku Slovenščine v 7. razredu / Kokelj Martina .................................... 636

Matematika – vir motivacije za programiranje / Kopinč Gregor ....................................................................... 639

S projektnim delom in umetno inteligenco pri pouku tehnike / Koščak Veronika ............................................. 643

Knjižnično informacijsko znanje in IKT – papir in oblak v šolski knjižnici Gimnazije Ptuj / Kosi Jelka ......... 647

Digitalne kompetence učitelja pri delu z učenci s posebnimi potrebami / Kovač Klaudija ............................... 651

Preverjanje predznanja z digitalno tehnologijo / Koželj Ksenija, Smrekar Erika .............................................. 654

Izdelava interaktivnih tematskih listov s spletnim orodjem Canva / Kuhar Marjeta .......................................... 657

Do rešitve avtentičnih problemov z matematiko in informatiko / Lasič Nastja, Kramar Andreja ..................... 661

Uporaba digitalnih kompetenc v vzgojno izobraževalnem zavodu / Letonja Stanislava ................................... 665

Učna analitika med visokošolskimi učitelji: preliminarni rezultati / Marolt Marjeta, Žnidaršič Anja, Jakšić

Danijela, Slavuj Vanja .................................................................................................................................... 667

Ustvarjanje in vrednotenje labirintov / Mejač Anja, Nančovska Šerbec Irena ................................................... 672

Šolska knjižnica v koraku s časom / Mlakar Tina .............................................................................................. 678



Odkrivanje naravoslovnih konceptov skozi igro: Kombinacija tradicionalnih in digitalnih metod v vrtcu /

Mlinar Renata ................................................................................................................................................. 683

Stolp v Pisi z vidika likovne umetnosti, matematike in informatike / Močnik Alenka ...................................... 688

Ali bo umetna inteligenca spremenila raziskovanje v osnovni šoli / Mrkela Vesna, Voglar Rafaela ................ 692

Poučevanje algoritmov in programiranja z Arduino platformo: Praktični primer iz srednješolskega okolja /

Muha Simon, Konečnik Simon ...................................................................................................................... 694

Uporaba umetne inteligence v osnovni šoli / Pajnik Tina, Strgar Sonja ............................................................ 698

Uporaba digitalne analize pri opisu enostavnih nihal – Interdisciplinarni primer povezovanja fizike, matematike

in računalništva / Pestotnik Stres Luna ......................................................................................................... 703

Umetna inteligenca v izobraževalnem procesu – uporaba virtualnih laboratorijev pri kemiji / Povšič Miha .... 708

Prednosti IKT tehnologije pri pouku športne vzgoje / Rajher Polona ................................................................ 711

Informacijska tehnologija pri inkluzivnem pouku tujega jezika / Rajhman Tina............................................... 715

Pomanjkanje kritičnega mišljenja je nevarnejše od umetne intelligence / Rajkovic Vladislav .......................... 718

Uporaba izobraževalnih aplikacij za prepoznavo travniških rastlin / Rebec Suzana ......................................... 719

Uporaba informacijsko komunikacijske tehnologije pri poučevanju slepih in slabovidnih otrok, mladostnikov ter

izzivi, ki nastajajo pri tem / Ribič Marko ...................................................................................................... 723

Izdelava videoposnetkov za lažje utrjevanje slovničnih struktur v tujem jeziku / Rolih Eva ............................ 726

Izpopolnjevanje jezikovnih in digitalnih veščin s projektnim delom: Ustvarjanje razrednega tabloida s platformo

Canva / Rolih Eva ......................................................................................................................................... 730

Uporaba aplikacije Stanford Cable TV News Analyzer v visokem šolstvu / Romih Dejan, Možina Feliks, Fir

Nejc, Dajčman Silvo, Baruca Arne ................................................................................................................ 733

Razvijanje socialno-emocionalnih veščin z uporabo animiranega filma v predšolskem obdobju / Sterle Luka 737

Psihologija barv na tečaju grafičnega oblikovanja / Strniša Iva ......................................................................... 741

S kvizom spoznajmo živali na travniku / Šebenik Tina ..................................................................................... 746

E-branje in e-knjige: posebnosti in izzivi v šolski praksi / Šebjanič Oražem Maja ........................................... 749

Umetnointeligenčna pismenost in šolske knjižnice / Škrlj Gregor ..................................................................... 754

E-igrifikacija kot metoda dela z učenci s posebnimi potrebami v osnovni šoli / Šuštaršič Urška ..................... 757

Malo po nemško z IKT / Tori Lea ...................................................................................................................... 762

Understanding Factors Affecting Artificial Intelligence Literacy among Higher Education Students / Torre

Cambón Silvia, Kožuh Ines ............................................................................................................................ 767

Pogovorno programerstvo: pot do boljšega razumevanja izotopov / Trost Marina ........................................... 771

Z umetno inteligenco nad domače branje / Traven Eva ..................................................................................... 775

Uporaba orodja OneNote pri delu z učencem z več motnjami: študija primera / Valentinčič Lara ................... 778

Uporaba spletnega okolja MS Teams za izboljšanje učnih dosežkov mladinskega pevskega zbora / Vavpotič

Mitja ............................................................................................................................................................... 781

Analiza projektnega dela za pridobitev praktičnih izkušenj in znanj študentov v delovnem okolju / Werber

Borut ............................................................................................................................................................... 784

Nekaj izkušenj pri uporabi ChatGPT pri poučevanju študentov na FUDŠ – smer Mediji in novinarstvo /

Zakrajšek Srečo, Zakrajšek Miha ................................................................................................................... 791



Digitalna vključenost v informacijski družbi / Digital Inclusion in Information Society ..................... 795

PREDGOVOR / FOREWORD ........................................................................................................................... 797

PROGRAMSKI ODBORI / PROGRAMME COMMITTEES ........................................................................... 799

Exploring Digital Media Literacy: A Case Study of Elderly Deaf Users in Slovenia / Horvat Laura, Čakš Peter,

Lovrenčič Držanič Irena, Kožuh Ines ............................................................................................................. 801

Innovative Technological Solutions for Cities as a Response to the Challenges of Multimodal Mobility for All

Citizens / Demšar Jani, Lovrić Marina, Rener Roman .................................................................................. 805

The role of communication skills in the development of digital literacy competencies using the "DigInGreen"

model / Forte Mateja, Obrecht Matevž, Slomšek Šlamberger Bojana .......................................................... 809

The Moore’s Law for Education and the Need for Inclusion / Balas Marius, Balas Valentina, Rad Dana, Rad

Gavril, Correia Sergio .................................................................................................................................... 812

Multimedia Based Sign Language Dictionaries: How are Potential Users Involved? / Juvan Žana, Pečnik

Klemen ........................................................................................................................................................... 816



(In)Accessibility of Slovenian E-commerce the Year Before the European Accessibility Act / Cerovec Bogdan

........................................................................................................................................................................ 820

Digital Inclusion of Children with Special Needs in Extracurricular Sports Activities / Maver Nuša, Debevc

Matjaž, Košir Katja ........................................................................................................................................ 824

A micro-Learning Units Package for Improving Inclusive Digital Education in HEI / Correia Sergio, Cunha

Ana, Pušnik Maja, Šumak Boštjan ................................................................................................................. 828

Segmentation of students with special needs at the University of Ljubljana / Košir Andrej, Košir Janja, Zaletelj

Janez ............................................................................................................................................................... 832

Introducing a Solution_ The Self-Evaluation Tool in Practice / Leandro Isabel, Marmol Marta ...................... 836

Analysis of the Self-Evaluation Tool: A Pilot Study* / Can Tuncer, Simsek Irfan, Göksu Hüseyin................. 839

Definition of a Framework for Self-Evaluation Tool: Optimizing Evaluation Practices for Enhanced

Performance in HEI / Pušnik Maja, Šumak Boštjan, Kous Katja, Gallo Florian, Dolza Enrico ................... 842

Good practices in creating an inclusive environment in High Education Institutions / Ivanuša Kline Darja,

Carlos Neves José, Sousa Carla, Infante Moro Neves Alfonso, Doratiotto Silvia, Simsek Irfan, Šumak

Boštjan, Pušnik Maja ...................................................................................................................................... 846

Technological Speech Adaptation for Professional Work as a Lawyer and Law Professor: A Case of a Patient

with Multiple Sclerosis / Ahlin Doljak Sara ................................................................................................. 850



Konferenca o zdravi dolgoživosti / Conference on Healthy Longevity ................................................... 855

PREDGOVOR / FOREWORD ........................................................................................................................... 857

PROGRAMSKI ODBORI / PROGRAMME COMMITTEES ........................................................................... 859

Vital Longevity – Challenges and Opportunities for Future Society / Javornik Branka .................................... 861

The 8 Pillars of Health – What We Can Do Ourselves and the Opportunities in Medicine – Gordana's Method /

Kalan Živčec Gordana .................................................................................................................................... 863

The Power of Royal Jelly in Healthy Aging / Kopinč Rok ................................................................................ 864

Advocacy in Action: How Patient Advocacy, Public Policy, and Science Communication Drive Scientific

Progress and Benefit Society / King Melissa ................................................................................................ 865

The Right to Health and Longevity: An Evolving Human Right? / Tietz - Latza Alexander ............................ 866

Should Christians Be Opposed to Technologically Achieved Life Extension? Why Kierkegaard Might Make a

Bad Transhumanist / Buben Adam ............................................................................................................... 867



Legende računalništva in informatike / Legends of Computing and Informatics ................................. 869

PREDGOVOR / FOREWORD ........................................................................................................................... 871

PROGRAMSKI ODBORI / PROGRAMME COMMITTEES ........................................................................... 873

Achievements of the Candidate Marjan Mernik / Mernik Marjan ..................................................................... 875

Contributions of Igor Bernik in the field of informatics / Bernik Igor ............................................................... 878

Dosežki kandidata Marka Bajca / Bajec Marko ................................................................................................. 880

Kratek življenjepis (Borut Robič) / Robič Borut ................................................................................................ 883

Marjan Krisper Career Path and Achivements / Krisper Marjan ....................................................................... 885

Dosežki kandidata Devida Palčiča / Palčič Devid ........................................................................................... 888

"Moja pot skozi 44 let dela na področju računalništva in informatike" / Ojsteršek Milan ................................. 891



ChatGPT v zdravstvu / ChatMED in Medicine ....................................................................................... 895

PREDGOVOR / FOREWORD ........................................................................................................................... 897

PROGRAMSKI ODBORI / PROGRAMME COMMITTEES ........................................................................... 899

Automatic Reviewing of Conference Papers in Healthcare and Other Sciences Using ChatGPT / Gams Matjaž,

Bolješić Luka, Korlat Alex, Kocuvan Primož ................................................................................................ 901

Leveraging Federated Learning for Secure Transfer and Deployment of ML Models in Healthcare / Dodevski

Zlate, Pavleska Tanja, Trajkovik Vladimir .................................................................................................... 905

Use and Limitations of ChatGPT in Mental Health Disorders / Petrović Jovana, Binić Iva, Vacev Maša, Lukić

Stevo ............................................................................................................................................................... 911

Testing ChatGPT’s Performance on Medical Diagnostic Tasks / Perko Alexander, Wotawa Franz ................. 914

HomeDOCtor App: Integrating Medical Knowledge into GPT for Personal Health Counseling / Zadobovšek

Matic, Kocuvan Primož, Gams Matjaž .......................................................................................................... 920



Standards for Use of LLM in Medical Diagnosis / Svetozarević Mihailo, Janković Isidora, Janković Sonja,

Lukić Stevo .................................................................................................................................................... 926

Using Combinatorial Testing for Prompt Engineering of LLMs in Medicine / Perko Alexander, Nica Iulia,

Wotawa Franz ................................................................................................................................................ 930

Meeting Cultural and Linguistic Demands to Accommodate Fine-Tuned LLMs to Local Medical Customs and

Patient Communication / Petrovska Dojchinovska Gordana, Simjanoska Misheva Monika, Mishev Kostadin

........................................................................................................................................................................ 936



Digitalna preobrazba zdravstvene nege - SIZN 2024 / Digital Transformation of Nursing - SIZN 2024

............................................................................................................................................................... 941

PREDGOVOR / FOREWORD ........................................................................................................................... 943

PROGRAMSKI ODBORI / PROGRAMME COMMITTEES ........................................................................... 945

Ali lahko etična odobritev in soglasje pacienta sledita pogovornim posrednikom na področju duševnega

zdravja? / Can Ethical Approval and Patient Consent Keep Pace with Conversational Agents in Mental

Health? / Cilar Budler Leona, Stiglic Gregor ................................................................................................ 947

Uporaba umetne inteligence v procesu zdravstvene nege: Analiza vpliva navodil na kakovost rezultatov / The

use of artificial intelligence in the nursing process: An analysis of the impact of instructions on quality of

outcomes / Hrovat Bukovšek Andreja, Ljubič Andreja ................................................................................ 948

Povezanost dimenzij informacijske varnostne kulture z namero nepooblaščenega dostopa do zdravstvenih

podatkov / The Relationship Between the Dimensions of Information Security Culture and the Intention of

Unauthorised Access to Health Data / Mikuletič Samanta, Žvanut Boštjan ................................................. 949

Odkrivanje znanja v zdravstvenih informacijskih sistemih s pomočjo rudarjenja procesov / Knowledge

discovery in healthcare information systems with process mining / Polančič Gregor .................................. 950

Ocena digitalnih kompetenc medicinskih sester v klinični praksi: opisna raziskava / Assessing digital

competencies of nurses in a clinical practice: a descriptive study / Grmšek Svetlin Anton, Peršolja Melita955

Prevod in validacija vprašalnika SHAIP za ocenjevanje stališč zaposlenih v zdravstveni negi o umetni

inteligenci / Translation and validation of the SHAIP questionnaire for assessment of nursing staff's

perceptions of artificial intelligence / Praprotnik Deja, Žvanut Boštjan ....................................................... 956

Pridobivanje mikrodokazil z uporabo simulacij v zdravstveni negi / Obtaining micro-credentials through

simulation in nursing / Renko Jakob, Karnjuš Igor, Pucer Patrik ................................................................. 957

Uporaba navidezne in obogatene resničnosti v zdravstveni negi / Using Virtual and Augmented Reality in

Nursing / Peterlin Mojca, Ficzko Jelena, Dolničar Danica ........................................................................... 962

Implementacija digitalnih rešitev v specialističnih ambulantah in bolniških oddelkih Splošne bolnišnice Jesenice

ter njihov vpliv na kakovostno zdravstveno obravnavo pacientov / Implementation of digital solutions in

specialist clinics and inpatient wards of the Jesenice General Hospital and their impact on the quality of

patient care / Justin Anton, Bijol Alenka ...................................................................................................... 966



Indeks avtorjev / Author index ................................................................................................................. 971





Zbornik 27. mednarodne multikonference



INFORMACIJSKA DRUŽBA – IS 2024

Zvezek A



Proceedings of the 27th International Multiconference



INFORMATION SOCIETY – IS 2024

Volume A



Slovenska konferenca o umetni inteligenci



Slovenian Conference on Artificial Intelligence



Uredniki / Editors



Mitja Luštrek, Matjaž Gams, Rok Piltaver



http://is.ijs.si



10. –11. oktober 2024 / 10–11 October 2024

Ljubljana, Slovenia





PREDGOVOR




Umetna inteligenca doživlja neverjeten in pospešen razvoj, ko se po tričetrt stoletja, ko je

Alan Mathison Turing postavil temelje računalništva in umetne inteligence, končno približuje

ne le človeški inteligenci, temveč tudi drugim ključnim človeškim lastnostim, kot sta

ustvarjalnost, čustvena inteligenca in zavest. Na številnih področjih umetna inteligenca že

presega zmogljivosti večine ljudi in celo strokovnjakov. Veliki jezikovni modeli dosegajo

tovrstne rezultate tudi pri dosti manj strukturiranih problemih, kot je bilo predstavljivo pred

nekaj leti, npr. pri strokovnih izpitih ter besedilnih nalogah iz matematike in programiranja.



Generativna umetna inteligenca že zdaj spreminja svet. Postala je nepogrešljivo orodje v

poslovnem svetu, raziskavah in vsakdanjem življenju, saj omogoča pisanje besedil,

ustvarjanje kode, generiranje slik in reševanje kompleksnih problemov. Možno je celo, da

smo priča začetkom singularnosti – prelomnega trenutka, ko bo umetna inteligenca presegla

človeško inteligenco in omogočila revolucijo na področju produktivnosti in inovacij, čeprav

bo treba na sodbo o tem še počakati. Optimizem glede prihodnosti je utemeljen: če se bo

razvoj nadaljeval s trenutnim tempom, si lahko predstavljamo svet, kjer bo umetna inteligenca

povsem preoblikovala gospodarstvo, znanost in način življenja, pri čemer bo omogočila višjo

kakovost življenja za vse.



Čeprav nekateri umetno inteligenco vidijo kot grožnjo, njen trenutni razmah resnejših težav še

ni prinesel. Nadejamo se, da bo zadosten del raziskav usmerjen v varnost umetne inteligence,

da bo tako ostalo. Z morebitnimi škodljivimi učinki umetne inteligence se spopadajo tudi

regulatorji, za katere upamo, da bodo uspešno krmarili med tem ciljem in pretiranim

zaviranjem razvoja.



Dostopnost velikih jezikovnih modelov, kot so GPT-ji, pomeni, da so naloge, ki zahtevajo

razumevanje in generiranje naravnega jezika, lažje kot kadar koli prej. Mnogi raziskovalci

verjamejo, da bo prihodnost programiranja prešla iz tradicionalnih jezikov, kot je Python, na

velike jezikovne modele, kjer bo umetna inteligenca generirala kodo in rešitve po meri.

Čeprav je razvoj teh modelov zahtevna naloga, ki presega zmožnosti večine organizacij, se

ljudje navajamo na uporabo tega fenomenalnega orodja. Pričakujemo, da bo umetna

inteligenca postala učinkovit in zanesljiv partner človeštva.



Že letos vidimo, da so konference v sklopu Informacijske družbe posvečene prav velikim

jezikovnim modelom. V okviru Slovenske konference o umetni inteligenci organiziramo

formalno debato dijakov – izkušenih debaterjev, ki se udeležujejo mednarodnih tekmovanj – o

tem, kako bo umetna inteligenca oblikovala prihodnost in zakaj bi to lahko bila najboljša

prihodnost doslej.



Matjaž Gams

Mitja Luštrek

Rok Piltaver

predsedniki Slovenske konference o umetni inteligenci





FOREWORD




Artificial intelligence is experiencing incredible and accelerated development. After three-

quarters of a century since Alan Mathison Turing laid the foundations of computing and

artificial intelligence, it is finally approaching not only human intelligence but also other key

human traits such as creativity, emotional intelligence and consciousness. In many areas,

artificial intelligence already surpasses the capabilities of most people and even experts. Large

language models are achieving such results even in much less structured problems than was

imaginable a few years ago, such as professional exams, and mathematics and programming

tasks described in free text.



Generative artificial intelligence is already transforming the world. It has become an

indispensable tool in the business world, research, and everyday life, enabling text writing,

code generation, image creation, and solving complex problems. It is even possible that we

are witnessing the beginnings of the singularity—the pivotal moment when artificial

intelligence will surpass human intelligence and enable a revolution in productivity and

innovation, although time will show whether this is actually the case. Optimism about the

future is well-founded: if development continues at its current pace, we can imagine a world

where artificial intelligence completely transforms the economy, science, and way of life,

leading to a higher quality of life for all.



Although some see artificial intelligence as a threat, its current rapid progress has not yet led

to serious problems. We hope that a sufficient part of the research will be directed towards AI

safety so that this remains the case. Regulators are also addressing the potential harmful

effects of artificial intelligence, and we hope they will successfully navigate between this goal

and excessive hindering of development.



The accessibility of large language models, such as GPTs, means that tasks requiring the

understanding and generation of natural language are easier than ever before. Many

researchers believe that the future of programming will shift from traditional languages, like

Python, to large language models, where artificial intelligence will generate custom code and

solutions. Although developing these models is a challenging task beyond the capabilities of

most organizations, people are getting accustomed to using this phenomenal tool. We expect

artificial intelligence to become an effective and reliable partner for humanity.



Already this year, we are seeing conferences within the framework of the Information Society

dedicated to large language models. As part of the Slovenian Conference on Artificial

Intelligence, we are organizing a formal debate for high school students—experienced

debaters who participate in international competitions—on how artificial intelligence will

shape the future and why this might be the best future yet.



Matjaž Gams

Mitja Luštrek

Rok Piltaver

Slovenian Conference on Artificial Intelligence chairs



PROGRAMSKI ODBOR / PROGRAMME COMMITTEE



Mitja Luštrek



Matjaž Gams



Rok Piltaver



Cene Bavec



Marko Bohanec



Marko Bonač



Ivan Bratko



Bojan Cestnik



Aleš Dobnikar



Erik Dovgan



Bogdan Filipič



Borka Jerman Blažič



Marjan Krisper



Marjan Mernik



Biljana Mileva Boshkoska



Vladislav Rajkovič



Niko Schlamberger



Tomaž Seljak



Peter Stanovnik



Damjan Strnad



Miha Štajdohar



Vasja Vehovar





PandaChat-RAG:




Towards the Benchmark for Slovenian RAG Applications



Taja Kuzman Urban Rupnik

Tanja Pavleska Primož Cigoj

{taja,tanja}@pc7.io {urban,primoz}@pc7.io

PC7, d.o.o. PC7, d.o.o.

Ljubljana, Slovenia Ljubljana, Slovenia

Jožef Stefan Institute

Ljubljana, Slovenia



Abstract sources, which facilitates the evaluation of the system’s accu-



with external knowledge through document retrieval. Due to Retrieval-augmented generation (RAG) is a recent method for racy [2]. These advantages have spurred quick adoption of RAG 1 systems across various applications. For instance, PandaChat enriching the large language models’ text generation abilities leverages RAG to provide explainable responses with high accu- racy in Slovenian and other languages, integrated in customer its high usefulness for various applications, it already powers service bots and platforms that allow LLM-based retrieval of multiple products. However, despite the widespread adoption, information from texts. there is a notable lack of evaluation benchmarks for RAG systems, Although RAG benchmarking is a relatively recent endeavor, particularly for less-resourced languages. This paper introduces some initial frameworks have already emerged [3, 5, 7]. However, the PandaChat-RAG – the first Slovenian RAG benchmark estab- these benchmarks are only limited to English and Chinese, leav- lished on a newly developed test dataset. The test dataset is based ing a gap in the evaluation of RAG systems for other languages. on the semi-automatic extraction of authentic questions and an- To address this gap, we make the following contributions: swers from a genre-annotated web corpus. The methodology for

the test dataset construction can be efficiently applied to any of • We present the first benchmark for RAG systems for the

the comparable corpora in numerous European languages. The Slovenian language. The benchmark is based on the newly

test dataset is used to assess the RAG system’s performance in re- 2 developed PandaChat-RAG-sl test dataset, which com-

trieving relevant sources essential for providing accurate answers prises authentic questions, answers and source texts.

to the given questions. The evaluation involves comparing the • We introduce a methodology for an efficient semi-automated

performance of eight open- and closed-source embedding models, development of RAG test datasets that is easily replica-

and investigating how the retrieval performance is influenced ble for the languages included in the MaCoCu [1] and

by factors such as the document chunk size and the number of CLASSLA-web corpora collections [10], which include

retrieved sources. These findings contribute to establishing the all South Slavic languages, Albanian, Catalan, Greek, Ice-

guidelines for optimal RAG system configurations not only for landic, Maltese, Ukrainian and Turkish.

Slovenian, but also for other languages. • As the first step of RAG evaluation, we evaluate the re-

triever’s performance in terms of its ability to provide

Keywords relevant sources crucial to retrieve accurate answers to

retrieval-augmented generation, RAG, embedding models, large the posed questions. The evaluation encompasses compar-

language models, LLMs, benchmark, Slovenian ison of performance of several open- and closed-source

embedding models. Furthermore, we provide insights on

the impact of the document chunk size and the number

1 Introduction of retrieved sources, to identify optimal configurations

The advent of large language models (LLMs) has introduced sig- of the indexing and retrieval components for robust and



(NLP). Although LLMs have shown impressive capabilities in gen- The paper is organized as follows: in Section 2, we provide an erating coherent text, they are prone to hallucinations [7, 16], i.e., nificant advancements in the field of natural language processing accurate retrieval.



static and potentially outdated corpora [9]. Retrieval-augmented dataset (Section 3.1) and the RAG system architecture (Section generation (RAG) is a method devised to address these challenges 3.2), which is evaluated in Section 4. Finally, in Section 5, we by augmenting LLMs with external information retrieved from a conclude the paper with a discussion of the main findings and providing false information. Furthermore, they are dependent on of RAG systems; Section 3 introduces the PandaChat-RAG-sl introduction to the previous research concerning the evaluation



provided document collection. Connecting LLMs with a relevant suggestions for future work. database improves the factual accuracy and temporal relevance



of the generated responses. Moreover, RAG contributes to the 2 Related Work explainability of the generated answers by providing verifiable Despite the recent introduction of the RAG architecture, several

Permission to make digital or hard copies of all or part of this work for personal benchmarking initiatives have already emerged [3, 5, 7, 15]. How-

or classroom use is granted without fee provided that copies are not made or ever, since the RAG systems can be applied to various end tasks,



distributed for profit or commercial advantage and that copies bear this notice and the benchmarks focus on different aspects of these systems. Inter the full citation on the first page. Copyrights for third-party components of this

work must be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia 1 https://pandachat.ai/

© 2024 Copyright held by the owner/author(s). 2The PandaChat-RAG benchmark and its test dataset are openly available at https:

https://doi.org/10.70314/is.2024.scai.538 //github.com/TajaKuzman/pandachat-rag-benchmark.





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Kuzman et al.




alia, current benchmarks assess their performance in text citation Table 1: Statistics for the PandaChat-RAG-sl dataset.

[7], text continuation, question-answering with support of exter-

nal knowledge, hallucination modification, and multi-document Number



on the task of Attributable Question Answering [2]. This task Unique texts 160 Words (questions) 1,184 involves providing a question as input to the system, which Words (texts w/o questions) 83,467 then generates both an answer and an attribution, indicating the summarization [12]. The closest to our work is the evaluation of the RAG systems Instances 206

source text on which the answer is based. The advantage of this Total words (questions + texts) 84,651

task over the closed-book question-answering task is that it also

measures the system’s capability to provide the correct source.

The majority of RAG benchmarks assess RAG systems in Eng- inspection of the extracted texts should there be a need to prepare

lish [3, 5, 7, 15] or Chinese [5, 12]. Consequently, the general- a larger dataset.

izability of their findings to other languages remains uncertain. Table 1 provides the statistical overview for the PandaChat-

Furthermore, a limitation of many benchmarks is their reliance RAG-sl dataset. The dataset consists of 206 instances, that is,

on synthetic data generated by LLMs [5, 12, 15]. To avoid poten- triplets of a question, an answer and a source text, derived from

tial biases introduced by LLMs and to better represent the com- 160 texts. The total size of the dataset is 84,651 words, encom-

plexity and diversity of real-world language use, a more reliable passing both the questions and the texts containing the answers.

evaluation would be based on non-synthetic test datasets. De-

spite focusing on a different task, recent research [6] has shown 3.2 RAG System



that resource-efficient development of non-synthetic and non- The RAG pipeline encompasses three main components: index- machine translated question-answering datasets is feasible by ing, retrieval, and text generation. During the indexing phase, the leveraging the availability of general web corpora and genre user-provided text collection is transformed into a database of classifiers. numerical vectors (embeddings) to facilitate document retrieval

by the retriever. This process involves segmenting the documents

into fixed-length chunks, which are then converted into embed-

3 Methodology dings using large language models. The choice of the embedding



The PandaChat-RAG-sl dataset comprises questions, answers, 3.1 model and the chunk size are critical factors that can signifi- PandaChat-RAG-sl Dataset cantly impact the retrieval performance of the model. Selecting an appropriate embedding model is essential to ensure that the and the corresponding source texts that encompass the answers. textual information is converted into a meaningful numerical It was created through a semi-automated process involving the representation for effective retrieval. Moreover, the chunk size, in extraction of texts from the Slovenian web corpus CLASSLA- terms of the number of tokens, plays a crucial role in determining web.sl 1.0 [11], followed by a manual extraction of high-quality the informativeness of the embeddings. Incorrect chunk sizes instances. Since the texts were automatically extracted from a may lead to numerical vectors that lack important information general text collection, the dataset encompasses a diverse range necessary for connecting the question to the corresponding text of topics that were not predefined or decided upon. chunk, thereby compromising retrieval accuracy [12]. The CLASSLA-web.sl 1.0 corpus is a collection of texts, col- When presented with a question, the retrieval component uses lected from the web in 2021 and 2022 [10]. It was chosen due the semantic search (also known as dense retrieval) to retrieve to its numerous advantages: 1) it has high-quality content, with the most relevant text chunks. The search is based on determin- the majority of texts meeting the criteria for publishable quality ing the smallest cosine distance between the chunk vectors and [17]; 2) it is one of the largest and most up-to-date collections the question vector. Lastly, during the text generation phase, of Slovenian texts, comprising approximately 4 million texts; 3) the retriever provides the large language model (LLM) with a the texts are enriched with genre labels, facilitating genre-based selection of top retrieved sources. The LLM is prompted to pro- text selection; and 4) it is developed in the same manner as 6 vide a human-like answer to the provided question based on the other CLASSLA-web corpora [10] and 7 additional MaCoCu web retrieved text sources. The selection of an appropriate number corpora in various European languages [1]. This enables easy of top retrieved sources is crucial in this phase: including more expansion of the benchmark to other languages, including all than just one retrieved source may enhance retrieval accuracy South Slavic languages and various European languages, such as and address situations where the first retrieved source fails to Albanian, Catalan, Greek, Icelandic, Ukrainian and Turkish. encompass all relevant information, especially in the case when The development of the PandaChat-RAG-sl dataset involves more texts cover the same subject matter. However, increasing the following steps: 1) the genre-based selection of texts from the the number of sources also leads to a longer prompt provided CLASSLA-web.sl corpus; 2) the extraction of texts that comprise to the LLM, potentially increasing the costs of using the RAG paragraphs ending with a question (80,215 texts); 3) the extraction system. of questions and answers (paragraphs, following the question); In this study, we assess the indexing and retrieval compo- 4) a manual review process to identify high-quality instances. In nents, focusing on the impact of different embedding models, the genre-based selection phase, we extract texts labeled with chunk sizes, and the number of retrieved sources on retrieval genres that are most likely to contain objective questions and performance. answers, that is, Information/Explanation , Instruction and Legal .

In its present iteration, the dataset consists of 206 instances Embedding Models. The evaluation includes a range of mul-

derived from the first 1,800 extracted texts. It is important to tilingual open-source and closed-source models. The selection

note that this effort can easily be continued with further manual of open-source models is based on the Massive Text Embedding





PandaChat-RAG Benchmark Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia




Benchmark (MTEB) Leaderboard3 [13]. Specifically, we choose



medium-sized multilingual models with up to 600 million pa-

rameters that have demonstrated strong performance on Polish

and Russian – Slavic languages that are linguistically related to

Slovenian. The models used in the evaluation are:

• Closed-source embedding models provided by the OpenAI:

an older model text-embedding-ada-002 (OpenAI-Ada) [8],

and two recently published models: text-embedding-3-

small (OpenAI-3-small), and text-embedding-3-large (OpenAI-

3-large) [14].

• Open-source embedding models, available on the Hugging

Face repository: BGE-M3 model [4], base-sized mGTE

model (mGTE-base) [19], and small (mE5-small), base

(mE5-base) and large sizes (mE5-large) of the Multilin-



gual E5 model [18]. Figure 1: The impact of the chunk size on the retrieval

Chunk size. The impact of the chunk size on retrieval per- performance.

formance is assessed by varying chunk sizes of 128, 256, 512,



and 1024 tokens, with a default chunk overlap of 20 tokens. In 4.2 Number of Retrieved Sources these experiments, the performance is evaluated based on the Figure 2 shows the performance of the RAG systems when in-first retrieved source. creasing the number of retrieved sources. The results demon-

Number of retrieved sources. Previous work indicates that in- strate that increasing the number of retrieved sources initially

creasing the number of retrieved sources improves the retrieval improves the performance, however, after a certain threshold,

accuracy [12]. In this study, we examine the retrieval accuracy the performance levels off.

of embedding models, with a chunk size set to 128 tokens, when Increasing the number of retrieved sources results in larger

the models retrieve 1 to 5 sources. In this scenario, if any of the inputs to the LLM in the text generation component, incurring

multiple retrieved sources matches the correct source, the output higher costs. Using more than two retrieved sources does not

is evaluated as being correct. significantly improve results in most systems. What is more, with

The retrieval capabilities of the RAG system are evaluated the top two retrieved sources, certain embedding models, namely,

on the task of Attributed Question-Answering. The evaluation BGE-M3 and mE5-large, already reach perfect accuracy. Thus,

is based on accuracy, measured as the percentage of questions our findings indicate that using more than the top two retrieved

correctly matched with the relevant source. sources is unnecessary.

The experiments are performed using the LlamaIndex library4.

The chunk size is defined using the SentenceSplitter method in



the indexing phase. Number of retrieved sources (similarity top

k ), the embedding model and the prompt for the LLM model are

specified as parameters of the chat engine. The closed-source

embedding models are used via the OpenAI API, while the ex-

periments with the open-source models are conducted on a GPU

machine.



4 Experiments and Results

In this section, we present the results of the experiments examin-

ing the impact of the chunk size, the number of retrieved sources,

and the selection of the embedding model on the retrieval per-

formance of the RAG system.



4.1 Chunk Size

Figure 1 shows the impact of the chunk size on the retrieval Figure 2: Impact of the number of retrieved sources on the

performance of the RAG systems that are based on different em- retrieval performance.

bedding models. The findings suggest that, with the exception of

the OpenAI-Ada model, all systems demonstrate the best perfor-

mance when the text chunk size is set to 128 tokens. Increasing 4.3 Embedding Models

the chunk size hinders the retrieval performance, which is con- We provide the final comparison of the performance of systems

sistent with previous research [12]. These results confirm that that use different embedding models. We use the parameters

smaller chunk sizes enable the embedding models to capture finer that have shown to provide the best results in the previous ex-

details that are essential for retrieving the most relevant text for periments: the chunk size of 128 tokens and top two retrieved

the given question. sources. As shown in Table 2, the retrieval systems that use the

open-source BGE-M3 and mE5-large embedding models achieve

3https://huggingface.co/spaces/mteb/leaderboard the perfect retrieval score. They are closely followed by the closed-

4https://www.llamaindex.ai/ source OpenAI-3-small and the mE5-base models which achieve





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Kuzman et al.




Table 2: Performance comparison between the open-source Translation . European Association for Machine Translation, Ghent, Belgium,

and closed-source embedding models. (June 2022), 303–304. https://aclanthology.org/2022.eamt-1.41.

[2] Bernd Bohnet et al. 2022. Attributed question answering: Evaluation and

modeling for attributed large language models. arXiv preprint arXiv:2212.08037.



embedding model speed (s) retrieval accuracy [3] Shuyang Cao and Lu Wang. 2024. Verifiable Generation with Subsentence- Level Fine-Grained Citations. arXiv preprint arXiv:2406.06125 .



OpenAI-3-small mE5-large BGE-M3 [4] Jianlv Chen, Shitao Xiao, Peitian Zhang, Kun Luo, Defu Lian, and Zheng 0.58 100 Liu. 2024. BGE M3-Embedding: Multi-Lingual, Multi-Functionality, Multi- 0.58 100 Granularity Text Embeddings Through Self-Knowledge Distillation. (2024). 0.69 99.51 arXiv: 2402.03216 [cs.CL] . [5] Jiawei Chen, Hongyu Lin, Xianpei Han, and Le Sun. 2024. Benchmarking mE5-base 0.29 99.51 large language models in retrieval-augmented generation. In Proceedings OpenAI-3-large 1.19 99.03 of the AAAI Conference on Artificial Intelligence number 16. Vol. 38, 17754– mGTE-base 0.31 99.03 17762. [6] Anni Eskelinen, Amanda Myntti, Erik Henriksson, Sampo Pyysalo, and OpenAI-Ada 0.63 98.54 Veronika Laippala. 2024. Building Question-Answer Data Using Web Regis- mE5-small 0.15 98.54 ter Identification. In Proceedings of the 2024 Joint International Conference

on Computational Linguistics, Language Resources and Evaluation (LREC-

COLING 2024). Nicoletta Calzolari, Min-Yen Kan, Veronique Hoste, Alessan-

accuracy of 99.5%. While having slightly lower scores, all other re- dro Lenci, Sakriani Sakti, and Nianwen Xue, editors. ELRA and ICCL, Torino,

Italia, (May 2024), 2595–2611. https://aclanthology.org/2024.lrec-main.234.



98.5% and 99% in accuracy. Language Models to Generate Text with Citations. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing , 6465–6488. Additionally, Table 2 provides the inference speed of the mod- trieval systems still achieve high performance, ranging between [7] Tianyu Gao, Howard Yen, Jiatong Yu, and Danqi Chen. 2023. Enabling Large

[8] Ryan Greene, Ted Sanders, Lilian Weng, and Arvind Neelakantan. 2022.



els measured in seconds per instance. If inference speed is a New and improved embedding model. https://openai.com/index/new-and-i mproved-embedding-model/. [Accessed 26-08-2024]. (2022). priority, the mE5-base model emerges as the optimal selection, as [9] Angeliki Lazaridou et al. 2021. Mind the gap: Assessing temporal generaliza- it yields high retrieval accuracy of 99.51% and is two times faster tion in neural language models. Advances in Neural Information Processing , 34, 29348–29363. than the two best performing models. In cases where users are re- Systems [10] Nikola Ljubešić and Taja Kuzman. 2024. CLASSLA-web: Comparable Web stricted to closed-source models due to the unavailability of GPU Corpora of South Slavic Languages Enriched with Linguistic and Genre resources, the OpenAI-3-small model stands out as the most suit- Annotation. In Proceedings of the 2024 Joint International Conference on Com-



able option. Its inference speed is comparable to the OpenAI-Ada putational Linguistics, Language Resources and Evaluation (LREC-COLING 2024) , 3271–3282. model, while it achieves a superior retrieval accuracy. [11] Nikola Ljubešić, Peter Rupnik, and Taja Kuzman. 2024. Slovenian web corpus

CLASSLA-web.sl 1.0. In Slovenian language resource repository CLARIN.SI.



5 http://hdl.handle.net/11356/1882. Conclusion and Future Work [12] Yuanjie Lyu et al. 2024. CRUD-RAG: A comprehensive Chinese benchmark



In this paper, a novel test dataset was introduced to assess the for retrieval-augmented generation of large language models.arXiv preprint arXiv:2401.17043 . performance of the RAG system on Slovenian language. A gen-[13] Niklas Muennighoff, Nouamane Tazi, Loic Magne, and Nils Reimers. 2023. eral methodology for efficient creating of non-synthetic RAG MTEB: Massive Text Embedding Benchmark. In Proceedings of the 17th



test datasets was established that can be extended to other lan- Conference of the European Chapter of the Association for Computational Linguistics , 2014–2037. guages. We evaluated the retrieval accuracy of the RAG system, [14] OpenAI. 2024. New embedding models and API updates. https://openai.co examining the impact of the embedding models, the document m/index/new-embedding-models-and-api-updates/. [Accessed 26-08-2024]. (2024). chunk size, and the number of retrieved sources. The assess- [15] Jon Saad-Falcon, Omar Khattab, Christopher Potts, and Matei Zaharia. 2024. ment of embedding models encompassed eight open-source and ARES: An Automated Evaluation Framework for Retrieval-Augmented Gen-



closed-source LLM models. It revealed that open-source models, eration Systems. In Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language specifically, BGE-M3 and mE5-large, reached perfect retrieval Technologies (Volume 1: Long Papers) , 338–354. accuracy, demonstrating their suitability for RAG applications on [16] Kurt Shuster, Spencer Poff, Moya Chen, Douwe Kiela, and Jason Weston.



Slovenian texts. Furthermore, the evaluation of the optimal chunk 2021. Retrieval Augmentation Reduces Hallucination in Conversation. In Findings of the Association for Computational Linguistics: EMNLP 2021 , 3784– size and the number of retrieved sources showed that smaller 3803. chunk sizes yielded superior results. In contrast, increasing the [17] Rik van Noord, Taja Kuzman, Peter Rupnik, Nikola Ljubešić, Miquel Esplà- Gomis, Gema Ramírez-Sánchez, and Antonio Toral. 2024. Do Language Mod- number of retrieved sources enhanced results up to a certain els Care about Text Quality? Evaluating Web-Crawled Corpora across 11 threshold, beyond which the model performance plateaued. Cer- Languages. In Proceedings of the 2024 Joint International Conference on Com-



tain models already achieved perfect accuracy when evaluated putational Linguistics, Language Resources and Evaluation (LREC-COLING 2024) , 5221–5234. based on the top two retrieved sources. [18] Liang Wang, Nan Yang, Xiaolong Huang, Linjun Yang, Rangan Majumder, While the novel test dataset can be used to evaluate all the and Furu Wei. 2024. Multilingual E5 Text Embeddings: A Technical Report.



components of the RAG system, in this paper, we focused on arXiv preprint arXiv:2402.05672. [19] Xin Zhang et al. 2024. mGTE: Generalized Long-Context Text Representation the evaluation of the indexing and retrieval components. In our and Reranking Models for Multilingual Text Retrieval. (2024). https://arxiv future work, we will extend the evaluations to the text generation .org/abs/2407.19669 arXiv: 2407.19669 [cs.CL] .

component with regard to fluency, correctness, and usefulness

of the generated answers. Furthermore, we plan to expand the

benchmark to encompass a wider range of languages. The plans

include extending the dataset and evaluation to South Slavic

languages and other European languages that are covered by

comparable MaCoCu [1] and CLASSLA-web [10] corpora.



References

[1] Marta Bañón et al. 2022. MaCoCu: Massive collection and curation of mono-

lingual and bilingual data: Focus on under-resourced languages. In Proceed-

ings of the 23rd Annual Conference of the European Association for Machine



Choosing Features for Stress Prediction with Machine Learning



Katja Bengeri Junoš Lukan∗ Mitja Luštrek∗

University of Ljubljana Jožef Stefan Institute Jožef Stefan Institute

Ljubljana, Slovenia Department of Intelligent Systems Department of Intelligent Systems

kb96968@student.uni-lj.si Ljubljana, Slovenia Ljubljana, Slovenia

junos.lukan@ijs.si mitja.lustrek@ijs.si



Abstract and Slovenia (26 participants). They answered questionnaires

Feature selection is a crucial step in building effective machine named Ecological Momentary Assessments (EMAs) roughly ev-

learning models, as it directly impacts model accuracy and in- ery 90 minutes, with smartphone sensor and usage data continu-

terpretability. Driven by the aim of improving stress prediction ously collected by an Android application [7], while also wearing

models, this article evaluates multiple approaches for identify- an Empatica E4 wristband recording physiological data. In 15

ing the most relevant features. The study explores filter-based days of their participation, each participant responded to more

methods that assess feature importance through correlation anal- than 96 EMA sessions, on average, which resulted in around 2200

ysis, alongside wrapper methods that iteratively optimize feature labels.

subsets. Additionally, techniques such as Boruta are analysed for

their effectiveness in identifying all important features, while 3 Target and feature extraction

strategies for handling highly correlated variables are also con- To fully leverage the potential of the data, we computed a com-

sidered. By conducting a comprehensive analysis of these ap- prehensive set of features. While some sensors only reported

proaches, we assess the role of feature selection in developing relatively rare events, such as phone calls, others had a high

stress prediction models. sampling frequency, such blood volume pulse which sampled



Keywords 90 min data at 32 Hz. On the other hand, labels were only available every . Therefore, we preprocessed the data in several steps.

Feature selection, Correlation matrix, Balanced accuracy score

3.1 Target variable

1 Introduction While participants responded to various questionnaires, for this

Machine learning models are increasingly being applied to predict study, we selected their responses to Stress Appraisal Measure-

stress, which is critical in various domains such as healthcare, ment [9] as the target variable. It was used to report stress levels

workplace management, and wearable technology. However, one on a scale from 0 to 4, so using it as is the prediction task can be

of the major challenges in developing reliable predictive models approached as a regression problem.

is identifying the most relevant features from extensive datasets, However, many stress detection studies tend towards a dis-

comprising physiological and behavioural information. crete approach, treating stress predominantly as a classification

Feature selection plays a key role in addressing this challenge. task, often only working with a binary target variable. To con-

By selecting only the most informative features, we can reduce vert this into a classification problem, we discretized the target

noise, prevent overfitting, and enhance model accuracy. As we variable into two distinct categories: “no stress”, which included

showed in previous work [8], even simple feature selection tech- all responses with a value of 0, while all others were coded as

niques can increase the 𝐹1 score of predictive models. This paper “stress”. With that, we ensured a balanced distribution of the

builds upon this finding and explores several feature selection target variable values. techniques, ranging from simple correlation-based methods to



more sophisticated wrapper approaches. 3.2 Features

The aim of this work is to assess how feature selection can en-

hance stress prediction models. By comparing different methods, 3.2.1 Data preprocessing. In our work, features were calculated

we aim to identify the optimal strategies for feature selection in on 30-minute intervals preceding each questionnaire session.

stress prediction which would lead to more reliable and more From the wide variety of smartphone data and physiological

easily interpretable machine learning models. measures, a total of 352 features were extracted and grouped into

22 categories, listed in Table 1. Using physiological data from



2 Data collection Empatica wristband, we first calculated specialized physiological features on smaller windows (from 4 s to 120 s , depending on the The data used in this work comes from the STRAW project [1], sensor; see [4] for more details), which were then aggregated results of which have been previously presented at Information over 30 min windows by calculating simple statistical features: Society [6, 8]. The dataset includes the data of 56 participants, mean, median, standard deviation, minimum, and maximum. All recruited from academic institutions in Belgium ( 29 participants) of the categorical features were converted into a set of binary

∗Also with Jožef Stefan International Postgraduate School. features using the one hot encoding technique and the missing

values were replaced with the mode.

Permission to make digital or hard copies of all or part of this work for personal First, some preliminary data cleaning was performed by ex-

or classroom use is granted without fee provided that copies are not made or cluding one of the feature in pairs exhibiting a correlation coeffi-

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this cient of |𝑟| ≥ 0.95. Despite this, some of the remaining features

work must be honored. For all other uses, contact the owner/author(s). still exhibited quite strong correlations as shown in Fig. 1. An

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia interesting observation used in the later stages of feature selec-

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.scai.991 tion was that high correlation, |𝑟 | ≥ 0.8, was mostly observed





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Katja Bengeri, Junoš Lukan, and Mitja Luštrek




Table 1: Feature categories with the number of features on each class. When adjusted for random chance, it is calculated

included in each category in parentheses as

Balanced accuracy 𝑇 𝑃 𝑇 𝑁 = + − 1

1. Empatica electrodermal activity (99) , 12. Phone screen events (7)

𝑇 𝑃 𝐹 𝑁 𝑇 𝑁 𝐹 𝑃 + +

2. Empatica inter-beat interval (50) 13. Phone light (6) in the binary case, where is the number of true positives, 3. Empatica temperature (33) 𝑇 𝑃 14. Phone battery (5)

4. Empatica accelerometer (23) 15. Phone speech (4) 𝑇 𝑁 is the number of true negatives, 𝐹 𝑁 is the number of false

5. Empatica data yield (1) 16. Phone interactions (2) negatives and 𝐹 𝑃 is the number of false positives. This definition

6. Phone applications foreground (47) 17. Phone messages (2) is equivalent to Youden’s J [11], which assigns a 0 to a random 7. Phone location (18) 18. Phone data yield (1)

8. Phone Bluetooth connections (18) 19. Baseline psychological features (7) classifier (indeed, a dummy classifier achieved a score of 0.0208



10. Phone activity recognition (7) 9. Phone calls (10) 20. Language (2) in our case), while a perfect classifier would achieve a score of 1. 21. Gender (2) To evaluate the stress detection models described in the fol-11. Phone Wi-Fi connections (7) 22. Age (1)

lowing sections, we considered several ways of data partitioning.

Since the variations in the results depending on the data split

between features of the same category. As an example, corre- were significant, in order to achieve more consistent accuracy,

lations between features related to phone application use are we employed shuffled 5-fold cross-validation.

shown in Fig. 2. We also considered a leave-one-subject-out cross-validation

technique. However, this method yielded poor results: using all

Empatica accelerometer available features, balanced accuracy was 0.05, while with the

5-fold cross validation it was 0.45. This suggested that the partici-

1.00 pants were quite different from each other, making it challenging



Empatica electrodermal activity to generalize predictions for a subject the model had not encoun- 0.75

tered.

0.50



Empatica inter beat interval 0.25 4.2 Baseline model

0.00 Our initial approach for building a prediction model was to use

Empatica temperature

Limesurvey all available features. This served as a baseline, which we aimed 0.25

Phone activity recognition to improve through feature selection.

Phone applications foreground 0.50 We evaluated various predictive models, as shown in Table 2,

Phone bluetooth Phone battery 0.75 all as implemented in scikit-learn [10]. Among these, the

Phone calls 1.00 Random Forest model yielded the best performance. Phone light



Phone location In this work, we aimed to find the best model for predicting Phone screen Phone speech Phone wifi stress and improve it using the optimal feature subset. Conse- e een ound quently, we used the Random Forest as the benchmark for com- ometer egr ecognition Phone wifi mal activity Limesurvey Phone callsPhone light paring feature selection algorithms. Phone battery Phone location Phone scr Phone speech oder Phone bluetooth



Empatica acceler Empatica temperatur Table 2: Performance of different models for the classifica-Phone activity r Empatica inter beat interval Phone applications for tion problem. The mean over five folds, its standard error, Empatica electr and the maximum are shown.

Figure 1: Correlation matrix of the initial feature set. Only

feature categories with more than two features are labelled. Model Mean Max SEM

Logistic Regression 0.077 0.151 0.025

Support Vector Machines 0.090 0.158 0.022

1.00 . 0 Gaussian Naive Bayes061 0.122 0.020



0.75 Stochastic Gradient Descent 0.027 0.054 0.007

Random Forest 0.475 0.558 0.026

0.50 XGBoost 0 441 0 473 0 013

. . .

0.25

0.00

0.25 In Table 2, SEM represents the Standard Error of the Mean.

0.50 It measures how far the sample mean of the data is likely to be

0.75 from the true population mean.



4.3 Correlation-Based Feature Reduction

Figure 2: Correlation matrix of the feature set in the Phone We began the feature selection process by eliminating highly

applications foreground category. correlated features. For each highly correlated pair, we removed

the feature with the lower rank when sorted by mutual informa-

tion, setting the correlation threshold at |𝑟 | ≥ 0.8 to maintain



4 Prediction models a manageable number of features. This reduction left us with approximately 180 features out of the original 352 for model 4.1 Model performance and validation training and evaluation.

To evaluate the performance of the models we used balanced While selecting the optimal set of features for stress prediction,

accuracy score which is defined as the average of recall obtained we aimed to retain all 22 different categories from Table 1, as





Choosing Features for Stress Prediction with Machine Learning Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia




Empatica accelerometer of features selected varied across folds, ranging from 50 to 93

features.

Empatica electrodermal activity 1.0



0.8 4.6 Sequential Forward Selection

Empatica inter beat interval 0.6 Another feature selection method we employed was Sequential

0.4 Feature Selector (SFS), a wrapper-based technique [2]. SFS and

Empatica temperature

0.2 RFECV differ in their approaches. SFS constructs models for each

Limesurvey

Phone activity recognition 0.0 feature subset at every step, while RFECV builds a single model

Phone applications foreground 0.2 and evaluates feature importance scores. Consequently, SFS is

Phone battery more computationally expensive, as it must evaluate numerous 0.4

Phone bluetooth feature combinations before identifying the optimal subset.

Phone calls 0.6 In the absence of specified parameters for number of fea-Phone light



Phone messages Phone location tures to select (n_features_to_select) and tolerance (tol), the Phone screen Phone speech method defaults to selecting half of the available features. The Phone wifi default configuration was used in our analysis, leading the SFS e een ound ometer egr to select the top 50 features. mal activity Limesurvey ecognition Phone calls Phone light Phone wifi Phone battery Phone scr Phone location Phone speech oder Phone bluetooth Phone messages 4.7 Boruta method Empatica temperatur Empatica acceler Phone activity r Empatica inter beat interval Phone applications for The final feature selection technique we employed was the Boruta Empatica electr

method [5]. With the assistance of “shadow features”, which are

Figure 3: Correlation matrix of the feature set after original features that have been randomly shuffled, the method

correlation-based feature reduction. Only feature cate- identifies a subset of features that are relevant to the classification

gories with more than two features are labelled. task at hand. In our case, shadow features were introduced into

the feature subset obtained after the preprocessing step.

The updated dataset was trained using the Random Forest

model for 100 iterations. In each iteration, all original features

each could provide unique information. Comparing Figs. 1 and 3,



we were left with about half the number of features which were feature were marked as relevant. ranked higher in importance than the highest-ranked shadow still moderately correlated.

Ultimately, a binomial distribution is used to evaluate which



4.4 Feature Selection using the mutual features have enough confidence to be kept in the final selection.

The number of features selected varied across folds, ranging from

information scoring function 47 to 55 features.

Before applying more complex feature selection algorithms, it

was necessary to reduce computational complexity by further 5 Results

reducing our set of 180 features obtained through correlation- In Table 3, the final scores for a Random Forest model built on

based reduction. Therefore, we used the SelectKBest method various feature subsets, as derived from the methods described

and the mutual information scoring function to retain the top 100 above, are presented. The data was split using shuffled 5-fold

features. This resulted in features derived from 19 to 20 categories, cross-validation, to ensure that the results were not overly de-

as categories language, gender, and, in some cases, Empatica pendent on a data split.

accelerometer were not deemed important for predicting stress.

Going forward, we will refer to the elimination of features Table 3: Adjusted balanced accuracy scores of a Random

within highly correlated pairs and the selection of the top 100 Forest model, trained on the different feature sets. Last

features using the mutual information scoring function as the column represents a number of features selected.

preprocessing step.



4.5 Recursive Feature Elimination with All available features Feature set Mean Max SEM N

Cross-Validation (RFECV) 0.464 0.498 0.011 352

Correlation-based reduction 0.483 0.507 0.007 ∼180

One of the previously mentioned complex feature selection meth- Correlation-based, 100 best 0.486 0.498 0.006 100

ods we employed was Recursive Feature Elimination with Cross- Preprocessing, RFECV 0.471 0.511 0.012 50 to 93

Validation (RFECV) [3]. The feature set we got after the prepro- Preprocessing, SFS 0.483 0.520 0.017 50

cessing step was passed to the RFECV algorithm for thorough Preprocessing, Boruta 0.481 0.545 0.020 47 to 55

evaluation. RFECV only 0.465 0.494 0.020 16 to 89

RFECV operates by initially fitting a model to the dataset SFS only 0.426 0.468 0.017 30

and evaluating its performance through cross-validation. After Boruta only 0.456 0.509 0.015 ∼75

the initial fit, RFECV ranks feature importance and iteratively

removes the least important features based on the models feature From Table 3, we can see that the most significant improve-

importances attributes, which in the case of Random Forest are ment in accuracy came after removing the highly correlated

impurity-based feature importances. This process continues until features, with the average adjusted balanced accuracy score ris-

there is no significant improvement in the model’s performance. ing from 0.46 to 0.48. Best mean accuracy was achieved after the

To ensure a reasonable duration for the feature selection process, preprocessing step, with only a minor improvement from 0.483

we set the cross-validation in RFECV to 3 folds. The number to 0.486.





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Katja Bengeri, Junoš Lukan, and Mitja Luštrek




After eliminating highly correlated features, wrapper methods wrapper methods alone were unable to effectively perform correlation-

did not significantly improve the accuracy on average (rows 3 based feature reduction. We can therefore conclude that simply

to 6 in Table 3). The Boruta method performed best among the relying on feature selection methods, however sophisticated, is

three, with the highest overall maximum accuracy in a single fold. not as effective as also considering relationships between fea-

These results led us to investigate whether the wrapper feature tures.

selection method alone could manage correlated features without It should be noted that the improvements in balanced accuracy

their prior removal and to evaluate the impact of the correlation are low in all cases. This indicates that results cannot be easily

threshold. generalized and correlation-based feature selection should not

We employed the RFECV, SFS, and Boruta method on the be seen as sufficient in general. Instead, we can speculate that

entire feature set of 352 features without applying the prepro- no single feature selection method is the best one and that sev-

cessing step. For SFS, only 30 features were selected due to its eral should be considered. We should also note that the Pearson

computational complexity. As shown in the last three rows of correlation coefficient that we used in this work only considers

Table 3, none of the methods alone were able to improve the linear relationships between features. Other methods can select

result achieved with correlation removal. Highly correlated fea- features even if they are related in a different way.

tures were left in the final feature set: for example, we identified

three pairs of features with a correlation coefficient exceeding References

|𝑟 | ≥ 0.8 using SFS alone. Poor results could be attributed either [1] Larissa Bolliger, Junoš Lukan, Mitja Luštrek, Dirk De Bacquer, and Els

to the importance of the correlation removal step or to the feature Clays. 2020. Protocol of the STRess at Work (STRAW) project: how to

subset being too small in the case of the SFS. disentangle day-to-day occupational stress among academics based on EMA,

physiological data, and smartphone sensor and usage data. International

Journal of Environmental Research and Public Health, 17, 23, (Nov. 2020),

5.1 Selecting the best correlation threshold 8835. doi: 10.3390/ijerph17238835.

[2] Francesc J. Ferri, Pavel Pudil, Mohamad Hatef, and Josef V. Kittler. 1994.



As previously mentioned, the biggest improvement in score came Comparative study of techniques for large-scale feature selection. Machine Intelligence and Pattern Recognition , 16, 403–413. doi: 10.1016/b978-0-444-81 from removing the feature inside the highly correlated pair. There- 892-8.50040-7.

fore, we have also experimented with different correlation cut-off [3] Isabelle Guyon, Jason Weston, Stephen Barnhill, and Vladimir Vapnik. 2002.

values to determine the best threshold. Gene selection for cancer classification using Support Vector Machines.

The highest score was achieved with a correlation threshold Machine Learning, 46, 1/3, 389–422. doi: 10.1023/a:1012487302797.

[4] Vito Janko, Matjaž Boštic, Junoš Lukan, and Gašper Slapničar. 2021. Library

of |𝑟| ≥ 0.75 (Table 4). Considering the impact of cross-validation for feature calculation in the context-recognition domain. In Proceedings of

splits and the relatively minor variance in scores, it appears that the 24th International Multiconference Information Society – IS 2021. Slovenian

Conference on Artificial Intelligence (Ljubljana, Slovenia, Oct. 4–8, 2021).

our initial threshold of |𝑟 | ≥ 0.8 was also quite effective. Vol. A, 23–26.

[5] Miron B. Kursa and Witold R. Rudnicki. 2010. Feature selection with the

Boruta package. Journal of Statistical Software, 36, 11, 1–13. doi: 10.18637/js

Table 4: Adjusted balanced accuracy scores of a Random s.v036.i11.

Forest model trained on a feature subset excluding features [6] Junoš Lukan, Larissa Bolliger, Els Clays, Primož Šiško, and Mitja Luštrek.

above the correlation threshold. The number of features 2022. Assessing sources of variability of hierarchical data in a repeated-

measures diary study of stress. In Proceedings of the 25th International Multi-

left after correlation-based feature selection differed over conference Information Society – IS 2022. Pervasive Health and Smart Sensing



validation folds and its range is shown in the final column. (Ljubljana, Slovenia, Oct. 10–14, 2022). Vol. A, 31–34. [7] Junoš Lukan, Marko Katrašnik, Larissa Bolliger, Els Clays, and Mitja Luštrek.

2020. STRAW application for collecting context data and ecological momen-

Threshold tary assessment. In Proceedings of the 23rd International Multiconference Mean Max SEM N

Information Society – IS 2020. Slovenian Conference on Artificial Intelli-

0.55 0.462 0.506 0.018 28 to gence (Ljubljana, Slovenia, Oct. 5–9, 2020). Vol. A, 63–67.

33 [8] Marcel Franse Martinšek, Junoš Lukan, Larissa Bolliger, Els Clays, Primož

0.60 0.467 0.493 0.009 39 to 41 Šiško, and Mitja Luštrek. 2023. Social interaction prediction from smart-



0.65 0.474 0.498 0. phone sensor data. In Proceedings of the 26th International Multiconference 008 47 to 50 Information Society – IS 2023 . Slovenian Conference on Artificial Intelligence 0 . 70 0 . 460 0 . 501 0 . 017 61 to 65 (Ljubljana, Slovenia, Oct. 9–13, 2023). Vol. A, 11–14. 0.75 0.498 0.526 0.012 74 to 80 [9] Edward J. Peacock and Paul T. P. Wong. 1990. The stress appraisal measure

0. (SAM). A multidimensional approach to cognitive appraisal. Stress Medicine, 80 0 . 470 0 . 543 0 . 022 101 to 107 6, 3, (July 1990), 227–236. doi: 10.1002/smi.2460060308.

[10] F. Pedregosa et al. 2011. Scikit-learn: machine learning in Python. Journal

of Machine Learning Research, 12, 2825–2830.

[11] Charles Sanders Peirce. 1884. The numerical measure of the success of

6 Conclusions predictions. Science, ns-4, 93, 453–454. doi: 10.1126/science.ns-4.93.453.b.

This paper examined different feature selection algorithms to

find the most effective subset for stress prediction. The model

using the feature subset after correlation removal achieved the

highest adjusted balanced accuracy score of 0.483.

Alternative feature selection approaches, including the wrap-

per methods SFS and RFECV, as well as the Boruta method, ap-

plied to the preprocessed feature subset, did not lead to further

optimization of the feature subset in terms of model performance.

Additionally, applying these methods to the entire set of features

did not achieve accuracy levels as high as those obtained after

the correlation-based reduction. In the case of SFS, this may be

attributed to its selection of only 30 features.

Therefore, our results underscore the critical role of the correlation-

based reduction step. In contrast, when this step was omitted



Predictive Modeling of Football Results in the WWIN League of



Bosnia and Herzegovina



Ervin Vladić Dželila Mehanović Elma Avdić

International Burch University International Burch University International Burch University

Sarajevo, Bosnia and Herzegovina Sarajevo, Bosnia and Herzegovina Sarajevo, Bosnia and Herzegovina

ervin.vladic@stu.ibu.edu.ba dzelila.mehanovic@ibu.edu.ba elma.avdic@ibu.edu.ba



Abstract a place in the UEFA Conference League. Since the founding of the

WWIN League of Bosnia and Herzegovina, team with the highest

Predictive modeling in football has emerged as a valuable tool for

number of titles was HŠK Zrinjski from Mostar who emerged

enhancing decision-making in sports management. This study

as the winner eight times, followed by Sarajevo which won four

applies machine learning techniques to predict football match

times, Zeljeznicar and Borac both won three times, Siroki Brijeg

outcomes in the WWIN League of Bosnia and Herzegovina. The

won two times and Leotar and Modrica both won once [12].

aim is to evaluate the effectiveness of various models, including

Depending on which entity association they belong to, the teams

Support Vector Machines (SVM), Logistic Regression, Random

that occupy the last two places in the league at the end of the

Forest, Gradient Boosting, and k-Nearest Neighbors (kNN), in

season are relegated to the league below, with two teams from

accurately predicting match results based on key features such

the First League of the Federation of BiH and the First League of

as shots on target, possession percentage, and home/away status.

the RS being promoted in their stead. To elevate football in our

By (1) gathering and analyzing match data from three seasons, (2)

country to the highest level, we must support in-depth analyses

comparing the performance of machine learning models, and (3)

of matches and the factors influencing their outcomes. This will

drawing conclusions on key performance factors, we demonstrate

enable coaches to fine-tune strategies for future games, help

that SVM achieves the highest accuracy at 83%, outperforming

commentators provide more insightful commentary, and allow

other models. These insights contribute to football management,

fans to develop a deeper understanding and get more pleasure

allowing for data-driven strategic planning and performance

from the match.

optimization. Future research will integrate additional factors

The study aims to evaluate the performance of various ML

such as player injuries and weather conditions to improve the

models, including Support Vector Machines (SVM), Logistic Re-

predictive models further.

gression, Random Forest, Gradient Boosting, and k-Nearest Neigh-



Keywords bors (kNN), in predicting match results. By examining key fea-

tures such as shots on target, possession percentage, and home/away

Football match prediction, machine learning, WWIN league, Sup-

status, we conduct an analysis based on match data from three

port Vector Machines, Random Forest

seasons of the WWIN League, encompassing 400 matches and

key performance metrics.

1 Introduction The remainder of the paper is structured as follows: Section

Accurate predictions of match outcomes can inform a wide range II provides an overview of related work in football ML-based

of decisions, from tactical adjustments to player acquisitions, and prediction. Section III describes the methodology, including the

can improve engagement for fans and stakeholders. While pre-

dataset and models used. Section IV presents the results and

dictive modeling has been extensively applied to top-tier football analysis of models performance, with a discussion on the practical

leagues like the English Premier League, there is limited research

implications of the findings for football management. Finally,

on regional leagues such as the WWIN League of Bosnia and Section V concludes the paper and outlines directions for future

Herzegovina. The specificity of the country that is Bosnia and

research.

Herzegovina and the WWIN League, which has not been re-

searched in the sphere of sports research, provides context for

this step. 2 Literature Review

The WWIN League of Bosnia and Herzegovina was established The prediction of the results of football matches has been recently

in the year 2000 and the same year the WWIN was formed by studied extensively because of its relation to betting and decision-

the merging of three leagues, it became a league covering the making in sports. Studies examining the employment of ML

entire territory of Bosnia and Herzegovina. Originally, the league methods are primarily focused on large European leagues, where

consisted of 16 clubs, and, from the 2016-2017 season, the league extensive and highly detailed data is available. The application of

contains 12 clubs which makes the level of the league higher these techniques to regional football leagues, such as the WWIN

[25]. The winner is the team that has the most points by the League of B&H, remains underexplored.

completion of thirty-three rounds; this position will grant a team Rodrigues and Pinto [15] used a variety of ML methods, in-

a place in the UEFA Champions League qualifications [10]; the cluding Naive Bayes, K-nearest neighbors, Random Forest, and

remaining two teams and the winner of the cup will compete for SVM, to predict the match outcomes based on previous match

data and player attributes. Their studies revealed excellent re-

Permission to make digital or hard copies of all or part of this work for personal

sults in terms of soccer betting profit margins, with the Random

or classroom use is granted without fee provided that copies are not made or

Forest approach obtaining a success rate of 65.26% and a profit

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this margin of 26.74%. Rahman [13] dedicated his work to employ-

work must be honored. For all other uses, contact the owner /author(s).

ing deep learning frameworks especially Deep Neural Networks

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

(DNNs) for football match outcome prediction, particularly dur-

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.scai.1642 ing FIFA World Cup 2018. The study classified match outcomes





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Vladić et al.





with 63.3% accuracy with DNN architectures with LSTM or GRU

cells. Baboota and Kaur [3] used machine learning approaches

to predict English Premier League match results. The models

compared included Support Vector Machines, Random Forest;

and Gradient Boosting. From their study , they ascertained that

Gradient Boosting outperformed other models in accuracy and

overall predictiveness. Authors in [16] used machine learning

techniques, notably SVM and Random Forest Classifier, to predict

English Premier League (EPL) football match results. They got

54.3% accuracy with SVM and 49.8% with Random Forest after

evaluating data from 2013/2014 to 2018/2019 seasons. Another

study [8] employed a few machine learning algorithms to pre-

dict matches of the English Premier League season 2017-2018.

Models including Linear Regression, SVM, Logistic Regression,

Random Forest, and Multinomial Naïve Bayes classier show that

the K-nearest neighbors give the best accurate predictions.

In summary, while existing studies have demonstrated the

effectiveness of machine learning in football matches prediction,

there remains a gap in the application of these techniques to

regional leagues like the WWIN League, due to the availability Figure 1: Workflow diagram

and quality of data. The characteristics of these leagues, such



as smaller datasets and potentially different factors influencing Table 1: Class Distribution match outcomes, require a tailored approach. In lesser-known

football leagues models might perform differently due to varia-

Match Type Count

tions in competitive structures and gameplay strategies, as well.

The study of Munđar and Šimić [11] in which they developed a Home Win 301

simulation model using the Poisson distribution to predict the Away Win 142

seasonal rankings of teams in the Croatian First Football League, Draw 151

highlighted the predictive power of statistical models and demon-

strated the significance of home advantage in determining match

The table sums up a type of match result in terms of its fre-

outcomes, which is also an important factor in the WWIN League.

quency in the dataset.

In the recorded 594 matches, 301 ended in home team victories,

142 in away team victories and 151 were tied. The following pie

chart describes the percentage distribution of the match outcome.

3 Materials and Methods Curiously, home wins are in the majority, comprising 50.7% of all

In this section, we describe the study conducted, detailing the



data collection and feature selection processes, the machine learn-

ing models applied, the evaluation metrics used to assess model

performance, and the approach taken to analyze key features

influencing match outcomes. As a result of providing numerous

procedures that are declared in this section, we represent the

graphical illustration of our methodology. The steps involved

in predicting the outcomes of the WWIN League of Bosnia and

Herzegovina, including data collection, preprocessing, model

development, and algorithm evaluation.



3.1 Dataset

The authors created the dataset for this study by consolidating in-

formation from rezultati.com [14], 1XBET [1], and Sofascore [24]. Figure 2: Class distribution of the dataset

The unique dataset represents the seasons 2021/2022, 2022/2023,

matches. However, away victories contribute to approximately

and 2023/2024 of WWIN League of Bosnia and Herzegovina. The

23.9% of all recorded match results, while 25.4% contribute to

platforms provide a wide range of football match data so it is

draw results. The Fig.2 depicts the frequency of each of the match

easy to find important information for examination. The dataset

outcomes.

includes key match facts as date, day of the week, time, home

team, away team, final as well as half-time goals scored in the

game, referee details, shots taken at goal as well as corner kicks 3.2 Machine Learning Prediction

resulting from these attempts on target, bookings made during In football, the concept of machine learning prediction entails de-

play by both teams and other relevant performance indicators. veloping models to forecast match outcomes based on the teams’



Predictive Modeling of WWIN League Football Results Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



and players’ histories and other attributes [5]. These models Gradient Boosting combines multiple weak learners (typically

employ such methods as regression analysis, classification, and decision trees) to create a stronger predictive model, improving

neural nets to determine the results given the data fed as the accuracy by focusing on correcting errors from previous models

input. [6].

k-Nearest Neighbors (kNN) is an instance-based learning method

3.2.1 Models initialisation, preprocessing, training and testing.

that classifies data by identifying the majority label among the

While implementing Logistic Regression, we have set the max_iter

k closest points. Though simple, it can be computationally ex-

=1000 and random_state = 42. Again, with the same classifier,

pensive as it requires storing all training data and performing

the kernel argument was assigned a linear value while the ran-

real-time comparisons [7].

dom_state was set to 42 to keep the results predictable. Gaussian

Naive Bayes was employed with no modification of its settings Last but not the least, the trained mod- 3.2.3 Evaluation Metrics.

because of the model’s simplistic nature. For Random Forest, we els are assessed by metrics such as accuracy of the models [19],

used the default parameters since the algorithm is capable of precision of the models [21], the recall of the models [22], and

changing the setting on its own based on the complexity of given F1-score value of the models [20]. This evaluation enables one

data. We initiated the Gradient Boosting with the default param- to analyze how well each of the models is likely to perform in

eters so that the gradients could easily learn and an ensemble terms of match outcome prediction.

could be formed. Last but not least, we left all the parameters of

k-Nearest Neighbors (kNN) for default value because the algo- 4 Results and Discussion rithm can find the optimal number of neighbors appropriate for

In this study, we employed six different classifiers to predict

the distribution of the data.

football match outcomes and conducted a comparative analysis

Following that, we proceed with the process of dividing this

of their performance. The effectiveness of each classifier was

gathered data into two sets: the training and the testing ones. We

evaluated based on its accuracy, providing a clear comparison of

split the data into training, where 70% of the data was allocated

their predictive capabilities across the dataset.

and the testing data where 30% was allocated.



Subsequently, the phase of model preprocessing is created for 4.1 Model Performance

which it is essential to filter data effectively to ensure proper

Among the classifiers employed, SVM predicted the most accurate

model training. In the case of feature transformation, we used

results at 83% This model performed almost well, with balanced

scikit-learn’s ColumnTransformer [17] to empower the numeric

precision and recall across all three classes (A, D, and H), show-

features normalization via the StandardScaler [23] while trans-

ing that it can predict match outcomes. In comparison, Random

forming the categorical variables into the binary format by the

Forest achieved a lower accuracy of 65%, with especially evident

use of the OneHotEncoder [18]. This technique pays a lot of

deficits in precision and recall for class ’D’. Logistic Regression

attention to ensuring that feature types are standard as well

performed worse than Support Vector Machines, with accuracy of

as harmonious. This method ensures consistency by creating a

77%. Despite its simplicity and computational efficiency, Gaussian

pipeline where preprocessing processes and the model are joined

Naive Bayes had the lowest accuracy of any classifier tested, at

in the same line of work. This means that there is always uni-

39%. This model struggled to predict class ’D’, with low accuracy

formity in the training and the testing of the model, hence a

and recall scores. Random Forest, an established ensemble learn-

manageable variability. Assuming the pipeline has been defined

ing approach, performed not so good, with an accuracy of 54%.

and is ready to proceed, we proceed to the next step of model

This model has generally balanced accuracy and recall across all

training.

classes, making it an acceptable alternative for predicting match

3.2.2 Models in Detail. In this study, many supervised learn- results. Gradient Boosting, another ensemble learning technique, ing classifier techniques that have proven valuable in the sports has a little higher accuracy than Random Forest at 64%. While

area for predictive purposes are employed. Logistic Regression Gradient Boosting is recognized for its ability to manage compli-

is a statistical technique that predicts the probability of a binary cated connections, it produced poorer recall ratings, especially

classification, using a sigmoid function to map outputs to a [0,1] for class ’D’. Lastly, k-Nearest Neighbors (kNN) resulted in 51%

probability space. Coefficients indicate the strength and direction accuracy, showing that the classifier was relatively poor, they

of relationships between variables, with positive values increas- had relatively fair precision and recall with all the classes.

ing the likelihood of an event and negative values decreasing it For making the match predictions, we employed the following

[9]. classification models – Logistic Regression, Support Vector Ma-

Random Forest extends the bagging method by generating chine, Gaussian Naive Bayes, Random Forest, Gradient Boost and

multiple decision trees using randomly selected data samples. k-Nearest Neighbors. We obtained the results varying from 39%

Each tree operates independently, and the final prediction is the to 83%, in which Support Vector Machines were the most effective.

average result across all trees, reducing overfitting and improving Our findings are partially consistent with prior research because

accuracy in classification tasks [4]. classifiers like Support Vector Machines, Logistic Regression,

SVM aims to find the best hyperplane to separate data points and Random Forest have manifested robustness in predicting

by class, maximizing the margin between them. It handles non- the match outcome across datasets. Nevertheless, the results are

linear boundaries by transforming the input data into a higher- not in conformity with some emerging works, particularly con-

dimensional space [2]. cerning the efficacy of Gaussian Naive Bayes, which performed

Naïve Bayes, based on Bayes’ theorem, assumes feature inde- poorly in our study in contrast to other research results. It should

pendence, making it fast and easy to implement, especially in be noted that results may vary significantly between different

applications like spam detection and text classification. Despite studies depending on the quality, the quantity, and the nature of

the simplicity of this assumption, it performs well in practice the data that had been used for creating the models of Gaussian

[26]. Naive Bayes.





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Vladić et al.




Model Accuracy Precision Recall F1 score establish a strong foundation for data-driven decision-making in

football management. Future work should incorporate additional

Logistic Regression 77% 75% 74% 74%

factors such as player injuries and weather conditions to enhance

SVM 83% 86% 83% 84%

predictive accuracy.

Gaussian NB 39% 47% 42% 36%



Random Forest 54% 43% 46% 43% References

Gradient Boosting 64% 64% 59% 60%

[1] 1XBET. 2007–2024. 1xbet. Retrieved May 26, 2024, from https://1xlite- 57954

kNN 51% 49% 49% 49%

2.top/en?tag=s_245231m_5435c_. (2007–2024).

Table 2: Model Performances [2] Mariette Awad and Rahul Khanna. 2015. Support vector machines for classi-

fication. In . Rahul Khanna, editor. Apress, 39–66. Efficient Learning Machines

doi: 10.1007/978- 1- 4302- 5990- 9_3.

[3] Rahul Baboota and Harleen Kaur. 2019. Predictive analysis and modeling

football results using a machine learning approach for the english premier

league. , 35, 2, 741–755. doi: 10.1016/j.ijf International Journal of Forecasting

The Table 2 shows how accurate various machine learning

orecast.2018.01.003.

models are in predicting WWIN League of Bosnia and Herzegov- [4] Leo Breiman. 2001. Random forests. Machine Learning, 45, 5–32. doi: 10.102

3/A:1010933404324.

ina match outcomes.

[5] Rory P. Bunker and Fadi Thabtah. 2019. A machine learning framework for



4.1.1 sport result prediction. , 15, 1, 27–33. Applied Computing and Informatics Key factors influencing match outcomes.

While this study

[6] Stefanos Fafalios, Pavlos Charonyktakis, and Ioannis Tsamardinos. 2020.

does not perform formal feature analysis, the observed perfor- . Gnosis Data Analysis PC, (Apr. 2020). Gradient Boosting Trees

[7] Gongde Guo, Hui Wang, David Bell, Yaxin Bi, and Kieran Greer. 2003. Knn

mance trends allow us to draw conclusions about the key factors

model-based approach in classification. In On The Move to Meaningful Inter-

influencing match outcomes. In line with prior research, home net Systems 2003: CoopIS, DOA, and ODBASE: OTM Confederated International

advantage emerged as a critical factor, with teams winning at . Springer Berlin Heidelberg, Conferences, CoopIS, DOA, and ODBASE 2003

Catania, Sicily, Italy, (Nov. 2003), 986–996.

home in over 50% of cases (Table 1) which reinforces the psy-

[8] Ishan Jawade, Rushikesh Jadhav, Mark Joseph Vaz, and Vaishnavi Yamgekar.

chological and tactical advantages that come with playing on 2021. Predicting football match results using machine learning. International

familiar ground. , 8, 7, 177. https://w Research Journal of Engineering and Technology (IRJET)

ww.irjet.net.

Offensive metrics, particularly shots on target, also revealed Logistic Regression [9] Daniel Jurafsky and James H. Martin. 2023. . Stanford

themselves as strong predictors of success. Teams that generated University, 5. https://web.stanf ord.edu/~juraf sky/slp3/5.pdf .

[10] Haris Kruskic. 2019. Uefa champions league explained: how the tournament

more attempts on goal were significantly more likely to win, rein-

works. Bleacher Report. Retrieved from https://bleacherreport.com/articles

forcing the widely accepted view that aggressive, forward-driven

/2819840- uef a- champions- league- explained- how- the- tournament- works.

play translates directly into better results. This trend mirrors ob- (2019).

[11] Dušan Munđar and Diana Šimić. 2016. Roatian first football league: teams’

servations from other football leagues, where offensive intensity

performance in the championship. roatian Review of Economic, Business and

is often directly correlated with victory. , 2, 1, 15–23. https://hrcak.srce.hr/f ile/245359. Social Statistics 2

[12] Prva Liga BiH. 2022. Osvajači trofeja. Retrieved from https://plbih.ba/osvaja

4.1.2 Limitations and future work. Despite the promising results, ci- trof eja/. (2022).

[13] Ashiqur Rahman. 2020. A deep learning framework for football match

this study has several limitations. First, the dataset used does

prediction. , 2, 2, 165. doi: 10.1007/s42452- 019- 1821- 5. SN Applied Sciences

not account for external factors such as player injuries, weather [14] 2006–2024. Rezultati. Retrieved May 26, 2024, from https://www.rezultati.c

conditions, or team morale, all of which can influence match om/. (2006–2024).

[15] Fátima Rodrigues and Ângelo Pinto. 2022. Prediction of football match

outcomes. Future research should incorporate these factors to im-

results with machine learning. , 204, 463–470. doi: Procedia Computer Science

prove the accuracy of predictions. Second, while SVM performed 10.1016/j.procs.2022.08.057.

[16] Sayed Muhammad Yonus Saiedy, Muhammad Aslam HemmatQachmas, and

well in this context, more advanced models such as deep learn-

Dr. Amanullah Faqiri. 2020. Predicting epl football matches results using

ing could potentially offer even better predictive performance, International Journal of Engineering Applied

machine learning algorithms.

particularly when dealing with larger datasets. , 5, 3, 83–91. http://www.ijeast.com. Sciences and Technology

[17] scikit-learn. 2024. Columntransformer. Retrieved from https://scikit- learn.o

Future work will explore the integration of additional domain-

rg/stable/modules/generated/sklearn.compose.ColumnTransf ormer.html.

specific features, such as player statistics, team form, and envi- (2024).

ronmental conditions, to further refine the predictive models. [18] scikit-learn. 2024. Onehotencoder. Retrieved from https://scikit- learn.or

g/stable/modules/generated/sklearn.preprocessing.OneHotEncoder.html.

We will also experiment with more complex algorithms, such as

(2024).

neural networks, to capture the intricate relationships between [19] scikit-learn. 2024. Sklearn.metrics.accuracy_score. Retrieved from https://sc

ikit- learn.org/stable/modules/generated/sklearn.metrics.accuracy_score.h

features that may be missed by traditional machine learning

tml. (2024).

models.

[20] scikit-learn. 2024. Sklearn.metrics.f1_score. Retrieved from https://scikit- le

arn.org/stable/modules/generated/sklearn.metrics.f 1_score.html. (2024).



5 Conclusion [21] scikit-learn. 2024. Sklearn.metrics.precision_score. Retrieved from https://s cikit- learn.org/stable/modules/generated/sklearn.metrics.precision_score

This study demonstrates that machine learning, particularly SVM, .html. (2024).

[22] scikit-learn. 2024. Sklearn.metrics.recall_score. Retrieved from https://sciki

effectively predicts football match outcomes in the WWIN League

t- learn.org/stable/modules/generated/sklearn.metrics.recall_score.html.

of Bosnia and Herzegovina. Support Vector Machine has been (2024).

[23] scikit-learn developers. 2024. Sklearn.preprocessing.standardscaler. https:

found to be the highest accurate classifier with 83% of accuracy

//scikit- learn.org/stable/modules/generated/sklearn.preprocessing.Standa

rate on match result prediction. SVM has moderate accuracy and

rdScaler.html. (2024).

recall with all three outcome classes: Home Win, Away Win, and [24] Sofascore. 2024. Sofascore. Retrieved May 26, 2024, from https://www.sof as

core.com/. (2024).

Draw, indicating football prediction applicability. However, it

[25] SportMonks. 2022. Premier league api bosnia. Retrieved from https://www

has also revealed that other classifiers’ performances are varying .sportmonks.com/f ootball- api/premier- league- api- bosnia/. (2022).

with Logistic Regression producing 77% of accuracy and Gauss- [26] Geoffrey I. Webb. 2016. Naïve bayes. In Encyclopedia of Machine Learning

and Data Mining. Claude Sammut and Geoffrey I. Webb, editors. (Jan. 2016),

ian Naïve Bayes a poor 39% accuracy. Both Random Forest and

1–2. doi: 10.1007/978- 1- 4899- 7502- 7_581- 1.

Gradient Boosting, which are ensemble learning algorithms, have

similar levels of accuracy; 54% and 64% respectively. While fur-

ther refinement of the models is needed, the current findings





Sarcasm Detection in a Less-Resourced Language




Lazar Ðoković Marko Robnik-Šikonja

lazardjokoviclaki02@gmail.com marko.robnik@f ri.uni- lj.si

University of Ljubljana, Faculty of Computer and University of Ljubljana, Faculty of Computer and

Information Science Information Science

Ljubljana, Slovenia Ljubljana, Slovenia



Abstract user annotation via distant supervision through hashtags, such

as , , , etc. This method is popular since 1) #sarcasm #sarcastic #not

The sarcasm detection task in natural language processing tries

only the author of a post can determine whether it is sarcastic or

to classify whether an utterance is sarcastic or not. It is related

not, and 2) it allows large-scale dataset creation. However, this

to sentiment analysis since it often inverts surface sentiment. Be-

method introduces large amounts of noise due to lack of context,

cause sarcastic sentences are highly dependent on context, and

user errors, and common misuse on social media platforms. The

they are often accompanied by various non-verbal cues, the task

sarcasm detection datasets created through manual annotation

is challenging. Most of related work focuses on high-resourced

tend to be of higher quality but are typically much smaller. These

languages like English. To build a sarcasm detection dataset for

problems are further compounded for non-English datasets, both

a less-resourced language, such as Slovenian, we leverage two

manually labeled and automatically collected. Further, as sarcasm

modern techniques: a machine translation specific medium-size

strongly relies on its context, using classical machine translation

transformer model, and a very large generative language model.

(MT) from English often produces inadequate results. This makes

We explore the viability of translated datasets and how the size of

sarcasm detection in less-resourced languages, like Slovenian, an

a pretrained transformer affects its ability to detect sarcasm. We

even bigger challenge. Therefore, developing reliable sarcasm

train ensembles of detection models and evaluate models’ perfor-

detection models is of crucial importance for robust sentiment

mance. The results show that larger models generally outperform

analysis in these languages.

smaller ones and that ensembling can slightly improve sarcasm

We develop a methodology for sarcasm detection in less-resour-

detection performance. Our best ensemble approach achieves an

ced languages and test it on the Slovenian language. We address

F 1-score of 0.765 which is close to annotators’ agreement in the

the problem of missing datasets by comparing state-of-the-art

source language.

machine translation with large generative models. We explore



Keywords the viability of such datasets and how the number of parameters

affects a model’s ability to detect sarcasm. We construct various

natural language processing, large language models, sarcasm

ensembles of large pretrained language models and evaluate their

detection, neural machine translation, BERT model, GPT model,

performance.

LLaMa model, ensembles

The rest of this work is organized as follows. In Section 2, we



1 discuss the proposed approach for detecting sarcasm in a less- Introduction

resourced language such as Slovenian. We present the creation

Sentiment analysis is a popular task in natural language process-

of a dataset, details of the training methodology and deployed

ing (NLP), concerned with the extraction of underlying attitudes

ensemble techniques. We lay out our experimental results and

and opinions, usually categorized as “positive”, “negative”, and

their interpretations in Sections 2.3 and 4. In Section 5, we provide

“neutral”. Detection of sentiment is challenging if the utterances

conclusions and directions for future work.

are ironic or sarcastic. is a form of verbal irony that trans-Sarcasm

forms the surface polarity of an apparently positive or negative 2 Sarcasm Detection Dataset

utterance/statement into its opposite [6]. Sarcasm is frequent

Existing attempts at automatic sarcasm detection have resulted

in our day-to-day communication, especially on social media

in the creation of datasets in a small number of languages with

[5]. This poses a significant problem for sentiment analysis tools

differing sizes and quality. It is unclear if models trained on these

since sarcasm polarity switches create ambiguity in meaning.

datasets would generalize well to unseen languages [1]. Since

Sarcasm is highly dependent on its context. For example, the

sentence “ ” could be interpreted as sarcastic,

I just love hot weather no dataset exists for Slovenian, we leverage recent advances

in machine translation and large language models (LLMs) to

depending on the situation, e.g., during summer heat waves.

create a dataset for supervised sarcasm detection. We thus apply

Historical developments of sarcasm detection are surveyed by

a translate-train approach when fine-tuning our models.

Joshi et al. [3], while recent developments are covered by Moores

The prevalence of research done on sarcasm in English means

and Mago [5]. The problem of automatic sarcasm detection in

that English datasets are usually larger and of higher quality than

text is most commonly formulated as a classification task. Unfor-

their counterparts in other languages. Further, as the translation

tunately, sarcasm detection is affected by the lack of large-scale,

from (and to) English is usually of better quality, we consider

noise-free datasets. Existing datasets are mostly harvested from

only English datasets.

microblogging platforms such as Twitter and Reddit, relying on

Preliminary tests showed poor quality and poor translation

Permission to make digital or hard copies of all or part of this work for personal 1 ability of Sarcasm on Reddit dataset, and News Headlines

or classroom use is granted without fee provided that copies are not made or 2 Dataset For Sarcasm Detection

. Hence, we chose the recent

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this dataset from the shared task. We SemEval-2022

iSarcasmEval3

work must be honored. For all other uses, contact the owner /author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia 1

www.kaggle.com/datasets/danofer/sarcasm

2

© 2024 Copyright held by the owner/author(s). www.kaggle.com/datasets/rmisra/news-headlines-dataset-for-sarcasm-detection

3

https://doi.org/10.70314/is.2024.scai.4212 github.com/iabufarha/iSarcasmEval





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Ðoković and Robnik-Šikonja




believe that relatively low performance scores obtained in this You will be provided with a sarcastic/non-sarcastic sentence in

shared task could be improved with the use of larger LLMs. English, and your task is to translate it into the Slovenian language.

It should keep the original meaning. Examples:

2.1 iSarcasmEval Dataset • love getting assignments at 6:25pm on a Friday!! //

iSarcasmEval is a dataset of both English and Arabic sarcastic obožujem, ko mi v petek ob 18:25 pošljejo naloge!!

and non-sarcastic short-form tweets obtained from Twitter. We • I still can’t believe England won the World Cup //

use only the English part, which is pre-split into a train and test Še vedno ne morem verjeti, da je Anglija zmagala na sve-

set. Both sets are unbalanced, the former having 867 sarcastic and tovnem prvenstvu



2601 non-sarcastic examples, while the latter has 200 sarcastic • taking spanish at ut was not my best decision // and 1200 non-sarcastic examples. The authors of the shared task



claim that both distant supervision and manual annotation of Učenje španščine na UT ni bila moja najboljša odločitev ,

datasets produce noisy labels in terms of both false positives We manually assessed the outputs of both transformers in order to

determine the best translations for fine-tuning detection models.

and false negatives [1]. Thus, they ask Twitter users to directly

provide one sarcastic and three non-sarcastic tweets they have

posted in the past. These responses are then filtered to ensure 2.3 Translation Results

their quality. The produced dataset is not entirely clean since it During translation, the T5 model sometimes had trouble with

contains links, emojis, and capitalized text. We chose to leave all examples that had multiple newline characters in a row. It oc-

of these potential features in the text, as they commonly occur casionally dropped parts of texts it didn’t understand (mostly

in online conversations and could be indicative of sarcasm. slang and various types of informal text styles). This shows that

Let us mention, that an ensemble approach with 15 trans- a 10B parameter model is not large enough to robustly translate

former models and transfer from three external sarcasm datasets all features of a language such as English into a less-resourced

proved to be the most accurate modeling technique for English language such as Slovenian.

[9] achieving an F1-score of 0.605. On the other hand, the GPT model performed surprisingly

well in most instances and it seemed to have a more nuanced

2.2 Translating iSarcasmEval understanding of phrases used in online speech. It consistently

translated entire texts, keeping the original structure and mean-

Our preliminary testing using smaller BERT-like classifiers showed

ing. Consequently, we used GPT’s translations when training

that the models learned the distribution of the data and defaulted

to the majority classifier (1200 1400 0.857 test accuracy). To / =

sarcasm detection models. The translations can be seen in our

6

repository .

try to dissuade this, we merged the train and test sets, kept all

the sarcastic instances, and randomly sampled an equal number

of non-sarcastic examples. This left us with a balanced dataset of 3 Model Training

2134 samples. We tested the performance of a wide range of LLMs of different

To enable task specific instructions that would preserve sar- sizes. Their overview is contained in Table 1.

casm, we skipped classical machine translation tools, and tried

two alternative translation approaches: Table 1: Summary of used sarcasm detection models.

• using a medium-sized T5 model trained specifically for

Model Parameters

neural machine translation,

• SloBERTa 110M leveraging a significantly larger model via OpenAI’s API.

BERT-BASE-MULTILINGUAL-CASED 179M

The T5 model uses both the encoder and decoder stacks of

XLM-RoBERTa-BASE 279M

the Transformer architecture and is trained within a text-to-

XLM-RoBERTa-LARGE 561M

text framework. We chose Google’s 32-layer T5 model called

4 META-Llama-3.1-8B-INSTRUCT

8.03B

MADLAD400-10B-MT , which has 10.7 billion parameters and is

META-Llama-3.1-70B-INSTRUCT 70.6B

pretrained on the [4] dataset with 250 billion tokens MADLAD-400

META-Llama-3.1-405B-INSTRUCT 406B

covering 450 languages. Fine-tuning for machine translation was

GPT-3.5-TURBO-0125 ?

done on a combination of parallel data sources in 157 languages,

GPT-4o-2024-05-13 ?

including Slovenian.

As a generative model, we chose OpenAI’s decoder-based

5 3.1 Encoder Models Under 1B Parameters

GPT-4o-2024-05-13 . Its true size is not known to the public,

but it’s speculated that it is significantly smaller than , since The four smallest models are encoder-based models that embed GPT-4

input text and use a classification head to assign it a class. They

it is much faster and more efficient. OpenAI claims that it has the

required additional fine-tuning to perform sarcasm detection. For

best performance across non-English languages of any of their

models, thus making it suitable for our task. these models, we conducted hyperparameter optimization.

When generative decoder-based models, it is nec-

prompting We chose the [7, 8] model in order to check whether SloBERTa

essary to craft clear and specific instructions to achieve the best using a monolingual Slovenian model impacts sarcasm detection

performance. We also wanted to compare BERT and RoBERTa-

results. We used few-shot learning [2], and randomly sampled

three training instances, manually translated them, and included like models, so we used their multilingual variants and fine-tuned

them on Slovenian data.

them in the following prompt where the double forward slash

was used as a delimiter between the query and the expected The models were trained for a maximum of five epochs with

response. the use of early stopping, where the training was halted if the

validation loss didn’t improve after two epochs.

4

huggingface.co/google/madlad400-10b-mt

5 6

platform.openai.com/docs/models/gpt-4o github.com/GalaxyGHz/Diploma





Sarcasm Detection Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia




3.2 Llama 3.1 Models When 𝑛 is set to zero, this approach is equivalent to hard

voting, and in the case of 𝑛 being equal to the predictor count, it

Since the teams that competed in the 2022 shared task on sarcasm

is equivalent to soft voting. We report both results. Additionally,

mostly used BERT and RoBERTa models, we extend the testing to

we compare the results of voting using all trained models with

include significantly larger models. We chose Meta’s open-source

the results obtained by using only the models with large weights

Llama family of models, more specifically, their newest Llama 3.1

in our regularized logistic regression ensemble.

variants. These come in three different sizes, which was perfect

for studying the effects of parameter counts on sarcasm detection.

We decided to use the “ ” version of all three models 4 Sarcasm Detection Results

instruct

since these were fine-tuned to be better at following instructions. Table 2 summarizes all our results. It is roughly sorted by model

When prompting LLama and GPT generative models, the fol- size, smaller models being on top and larger ones being on bot-

lowing few-shot classification prompt was given, with two pos- tom. The tag indicates that a model was not fine-tuned, (NFT)

itive and two negative examples randomly sampled from our while the tag means that a model was trained with LoRA. (LoRA)

dataset. Results are rounded to three decimal places.

You will be provided with text in the Slovenian language, and your

task is to classify whether it is sarcastic or not. Use ONLY token 0 Table 2: Summary of performance results for all tested

(not sarcastic) or 1 (sarcastic) as in the examples: models. The best scores are in bold.



• Spanje? Kaj je to... Še nikoli nisem slišal za to? 1

Model Accuracy F-score



1

• Lepo je biti primerjan z zidom 1 SloBERTa 0.621 0.632

• To sploh nima smisla. Nehaj kopati. 0 BERT-BASE-MULTILINGUAL-CASED 0.499 0.666

• XLM-RoBERTa-BASE Dne 12. 10. 21 ob 10:30 je bil nivo reke 0,37 m. 0 . 0.578 0.579

XLM-RoBERTa-LARGE 0.550 0.597

We used full versions of the 8B and 70B parameter models, Llama-3.1-8B-INSTRUCT (NFT) 0.560 0.676

while the 405B parameter model was loaded in 16-bit precision 0.569 Llama-3.1-8B-INSTRUCT (LoRA) 0.682 mode. To minimize the use of resources and costs, we employed 0.660 0.724 Llama-3.1-70B-INSTRUCT (NFT)

LoRA parameter-efficient fine-tuning. We provided the models 0.637 Llama-3.1-70B-INSTRUCT (4-bit-LoRA) 0.717 with training and validation sets and trained them for a maximum Llama-3.1-405B-INSTRUCT (16-bit-NFT) 0.686 0.751

GPT-3.5-TURBO-0125 (NFT) 0.564 0.679

of 10 epochs. No hyperparameter optimization was conducted

GPT-3.5-TURBO-0125 0.749 0.760

in this case due to time constraints. We used the validation loss

GPT-4o-2024-05-13 (NFT) 0.686 0.746

to choose the best model, and we used the same early stopping L2-LOGISTIC-REGRESSION 0.759 0.765 technique as with the smaller models. 0.707 L2-LOGISTIC-REGRESSION-NON-COMMERCIAL 0.749

HARD-VOTING-ALL 0.670 0.738

3.3 SOFT-VOTING-ALL 0.658 0.732 GPT 3 and 4 Models

HARD-VOTING-BEST-5 0.686 0.749

We also tested two models offered on the OpenAI platform, SOFT-VOTING-BEST-5 0.686 0.749

GPT-4o-2024-05-13 and GPT-3.5-TURBO-0125. We first used

them in few-shot mode and classified all our examples without Individual Model Performance

any fine-tuning. When fine-tuning, the platform’s tier system Out of all of the used models, only BERT-BASE-MULTILINGUAL-

limited us to only the smaller model. We failed to learn any pattern in our data and defaulted to GPT-3.5-TURBO-0125-CASED fine-tuned the model for a maximum of three epochs. In the end, the dummy classifier.

we used the model with the lowest validation loss to classify the sometimes predicts the correct token GPT-3.5-TURBO-0125

examples in the test set. but then continues to generate additional text, such as 11 and

1\n1. This happens with a small quantity of examples in our



3.4 Sarcasm Detection Ensembles testing set. We decided to truncate these responses and only kept

the first token as the answer.

When constructing ensemble models, we tried two techniques:

The Llama models sometimes refused to generate tokens zero

stacking and voting. In both cases, we used the predicted proba-

or one. We decided to drop these examples altogether. We report

bility of the sarcastic class from each model as input features.

the test accuracy and trained the ensemble models without them.



3.4.1 Stacking With Regularized Logistic Regression. Smaller encoder models performed poorly when compared to Our first

some of the larger models. Only the model manages SloBERTa

ensemble used stacking approach, and logistic regression with

to achieve an accuracy above 0.6. Despite being the smallest of

Ridge regularization as the meta-level classifier. This choice was

the four small models we tested, performed the best. SloBERTa

motivated primarily by the need for feature selection, as we

This suggests that the three larger multilingual encoder models

wanted to identify the most important model predictions and

may lack sufficient understanding of Slovenian. It also highlights

determine which models would be assigned a lower weight. The

that model size alone does not necessarily correlate with better

best models were then used for voting.

performance when it comes to sarcasm detection.

3.4.2 Standard and Mixed Voting. The second ensembling method The Llama models fared better, achieving accuracies of up to

was voting. We tried cut-off-based mixed voting inspired by [9]. 0.686 with the 405B model being comparable to GPT-4o in perfor-

Formally, we used hard voting when the absolute difference be- mance. They still fell short of the fine-tuned GPT-3.5-TURBO-0125 tween the number of sarcastic and non-sarcastic predictions was model, which managed to successfully classify about three-quarters

greater than 𝑛, and we used soft voting otherwise. We optimized of our examples with a F1-score of 0.76.

the value of 𝑛 based on the ensembles performance on our vali- Some models had significantly higher F 1-scores and lower dation set. accuracies. We show the confusion matrix of one of the models





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Ðoković and Robnik-Šikonja




Table 3: Confusion Matrix for non-fine-tuned Slovenian sarcasm detection power, but we also note that a pos-

Llama-3.1-405B-INSTRUCT model. sible alternative could be local fine-tuning of the Llama-3.1-8B-

-INSTRUCT model. Our testing shows that using aggressive quan-

Predicted \ Actual Positive Negative tization combined with LoRA results in significant performance

degradation.

Positive 202 123 We also constructed ensemble models based on voting and



Negative stacking methods. Observations showed that voting didn’t result

11 91

in any performance improvements. On the other hand, stacking

that exhibited the largest difference in Table 3. These models with the use of a regularized logistic regression managed to

seem to have a tendency to incorrectly classify non-sarcastic text improve on the performance of its base models.

as sarcastic, leading to a high rate of false positives. Additional work needs to be done in dataset construction.

Our testing also showed that loading the Sarcastic examples could be extended with context or labels of the Llama-3.1-70B-

-INSTRUCT model in 4-bit mode and fine-tuning it with LoRA types of sarcasm they represent. This might help guide models

does not produce satisfactory results, and it is thus better to towards better understanding of sarcasm. Future work could

conduct full fine-tuning with the smaller Llama model or to use also explore incorporating heterogeneous models into ensembles

one of OpenAI’s models via their fine-tuning API. or the creation of Mixture of Experts (MoE) ensembles, whose

GPT-3.5-TURBO-0125 performed the best among individual baseline models would focus on different aspects of sarcasm.

models, so if costs associated with the use of OpenAI’s API are

acceptable, we recommend its use for sarcasm detection in Slove- Acknowledgements nian. This shows that very large models can effectively identify

This research was supported by the Slovenian Research and In-

sarcasm. We believe that with better parameter tuning, Llama 8B

novation Agency (ARIS) core research programme P6-0411 and

could be one of the best (and most economical) options for sar-

projects J7-3159, CRP V5-2297, L2-50070, and PoVeJMo.

casm detection in Slovenian, provided that the user has sufficient

hardware resources. References

Ensemble Model Performance [1] Ibrahim Abu Farha, Silviu Vlad Oprea, Steven Wilson, and Walid Magdy. 2022.

We observed that the regularized logistic regression mostly re- SemEval-2022 task 6: iSarcasmEval, intended sarcasm detection in English

and Arabic. In Proceedings of the 16th International Workshop on Semantic

lied on the best-performing models. Its focus on the best model Evaluation (SemEval-2022)

, 802–814. doi: 10.18653/v1/2022.semeval- 1.111.

( ), however, suggests that there is signifi- [2] Tom Brown et al. 2020. Language models are few-shot learners. In Advances GPT-3.5-TURBO-0125 cant overlap between the various model predictions. in Neural Information Processing Systems

. Vol. 33, 1877–1901. https://proceedi

ngs.neurips.cc/paper_f iles/paper/2020/f ile/1457c0d6bf cb4967418bf b8ac142

We decided to discard when BERT-BASE-MULTILINGUAL-CASED

f 64a- Paper.pdf .

constructing our voting ensembles since its dummy classification [3] Aditya Joshi, Pushpak Bhattacharyya, and Mark J. Carman. 2017. Automatic

sarcasm detection: a survey. , 50, 5, Article 73, 22 pages. ACM Comput. Surv.

didn’t contribute to overall model performance. Both of these

doi: 10.1145/3124420.

two voting classifiers had an odd number of predictors, so there [4] Sneha Kudugunta, Isaac Caswell, Biao Zhang, Xavier Garcia, Derrick Xin,

was no need for a tie-breaker mechanism. Aditya Kusupati, Romi Stella, Ankur Bapna, and Orhan Firat. [n. d.] Madlad-

400: a multilingual and document-level large audited dataset. In Proceedings

Voting proved to be ineffective in our setups, even scoring

of the 37th International Conference on Neural Information Processing Systems

lower than some of its base models. hard voting generally out- Article 2940, 13 pages.

[5] Bleau Moores and Vijay Mago. 2022. A survey on automated sarcasm detec-

performed soft voting. We also note that there was no benefit

tion on twitter. . doi: 10.48550/arXiv.2202.02516. arXiv preprint

in using mixed voting, at least for the sets of predictors that we

[6] Smaranda Muresan, Roberto Gonzalez-Ibanez, Debanjan Ghosh, and Nina

obtained as hard voting always had a higher accuracy. This was Wacholder. 2016. Identification of nonliteral language in social media: a

case study on sarcasm. Journal of the Association for Information Science and

true for both the classifiers that used all and only five of the base

Technology, 67, 11, 2725–2737. doi: 10.1002/asi.23624.

models. [7] Matej Ulčar and Marko Robnik Šikonja. 2021. Sloberta: slovene monolingual

Regularized logistic regression managed to improve on the large pretrained masked language model. In Proceedings of Data Mining and

Data Warehousing, SiKDD, 17–20. http://library.ijs.si/Stacks/Proceedings/Inf

scores of individual models, raising accuracy by one percent, thus

ormationSociety/2021/IS2021_Volume_C.pdf .

achieving the best performance out of all of the tested approaches. [8] Matej Ulčar and Marko Robnik Šikonja. 2021. Slovenian RoBERTa contextual This shows that there is still performance to be gained from embeddings model: SloBERTa 2.0 CLARIN.SI data & tools . . Nasl. z nasl. zaslona.

Fakulteta za računalništvo in informatiko. http://hdl.handle.net/11356/1397.

ensembles; however, it is still necessary to use commercial models

[9] Mengfei Yuan, Zhou Mengyuan, Lianxin Jiang, Yang Mo, and Xiaofeng Shi.

for top performance. 2022. Stce at SemEval-2022 task 6: sarcasm detection in English tweets. In Pro-

ceedings of the 16th International Workshop on Semantic Evaluation (SemEval-

5 2022), 820–826. doi: 10.18653/v1/2022.semeval- 1.113. Conclusion

In this work, we presented the task of sarcasm detection in the

less-resourced Slovenian language. Our code and results are freely

7

available .

We tackled the missing dataset problem by using two LLMs

to perform neural translation of an English dataset into Slove-

nian. The translations generated by out-GPT-4o-2024-05-13

paced those generated by a large T5 model specifically trained

for neural machine translation in terms of quality.

We used this data to train a plethora of Transformer-based

models in various settings. We found that fine-tuning GPT-3.5-

-TURBO-0125 via OpenAI’s API results in the highest individual



7

github.com/GalaxyGHz/Diploma



Speech-to-Service: Using LLMs to Facilitate Recording of



Services in Healthcare



Maj Smerkol Rok Susič Mariša Ratajec

maj.smerkol@ijs.si rs36117@student.uni- lj.si mr97744@student.uni- lj.si

Jožef Stefan Institute University of Ljubljana, Faculty of University of Ljubljana, Faculty of

Ljubljana, Slovenia Mathematics and Physics Electrical Engineering

Ljubljana, Slovenia Ljubljana, Slovenia



Helena Halbwachs Anton Gradišek

h.halbwachs@senecura.si anton.gradisek@ijs.si

SeneCura Kliniken- und Jožef Stefan Institute

Heimebetriedsgesellschaft m.b.H. Ljubljana, Slovenia

Vienna, Austria



Abstract by significantly lowering the number of clicks required in the

UI. The system is built using open-source or publicly accessible

Digital tracking of services is one of the main administrative bur-

components, particularly a speech-to-text system that transcribes

dens of the healthcare staff. Here, we present a proof-of-concept

the recorded conversation, and a large language model (LLM)

study of a so-called speech-to-service (S2S) system that is aimed

that leverages its natural language processing capabilities. The

at facilitating recording of activities, extracting information from

recommender system shows possible required tasks, serving as a

the conversation between a healthcare provider and recipient.

reminder and to suggest tasks that are expected soon, which may

The system comprises of a speech recorder, a diarization compo-

lower the number of visits per patient. These recommended tasks

nent, an LLM to interpret the conversation, and a recommenda-

are then suggested to the healthcare worker, who can review and

tion system integrated in a smart tablet that records completed

confirm them using the LLM-assisted interface. LLMs, such as

activities and suggests possible other activities that may have

ChatGPT and Llama, have seen a surge in popularity in a wide

still be required. We tested the system on 350 conversations and

variety of topics since their popularization in particular with the

obtained 95% accuracy, 97% precision and 94% recall.

unveiling of ChatGPT3 in the autumn of 2022.



Keywords Several LLM based systems have been proposed recently, in-

cluding administrative task automation [6], decision making pro-

healthcare, LLM, speech recognition, recommendation system

cess [10], improving existing automatic speech recognition (ASR)



1 Introduction systems [1], and providing patients with needed information [9].

A recent study [11] concludes that utilising ASR to ease some

Healthcare workers, including nurses, technicians, and care per- administrative tasks leads to faster, more efficient work and even

sonnel form the backbone of the health system as they care for increase workers’ moods. patients and tend to their needs. However, with the standard-



ization and systematization of the healthcare professions and 2 System Architecture services often becomes a large bureaucratic burden, as health-

care workers have to record all the activities and services they This paper describes two early prototype systems, both aiming

provide to the patients. This process is of course needed as it to alleviate the workload of healthcare workers by easing the

provides traceability and ensures that all the required activities task of documenting care actions performed. These are the ASR

were taken care of, but the problem is that the interfaces designed system that logs care actions based on captured dialogue between

for activity logging are often not user-friendly and require the the healthcare worker and the patient, and a recommender sys-

users to choose the activities from a extensive lists of drop-down tem that predicts the required services at a specific time. This

menus. In total, this amounts to substantial time required only for recommender system relies on the historical data, appropriate

tedious administrative tasks, time that would be more beneficially for long-term patient care facilities.

spent otherwise. Both systems are limited in scope and only target the most

With the aim to alleviate the administrative burden of activity common healthcare services in the dataset for detection or pre-

logging, we explored the possibilities of novel technologies to as- diction respectively, which can still greatly easy the workload

sist the healthcare staff in their logging tasks. We developed and for medical workers, since the top 10 most common tasks out of

tested a proof-of-concept system that records the conversation around 200 care action types represent around 80% of all services

between the healthcare worker and a patient, identifies the activ- performed.

ities, and allows the healthcare worker to batch-confirm them on The recommender system allows the care workers to anticipate

a dedicated smart tablet. Batch-confirmation saves a lot of time tasks in advance and server as a reminder. This aims to lower the

number of patient visits, which also alleviates the workload.

Permission to make digital or hard copies of all or part of this work for personal

or classroom use is granted without fee provided that copies are not made or



distributed for profit or commercial advantage and that copies bear this notice and 2.1 Speech-to-Service ASR the full citation on the first page. Copyrights for third-party components of this

work must be honored. For all other uses, contact the owner /author(s).

The ASR system consists of a speech diarization model, capable

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

of segmenting the recorded speech based on who is currently

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.scai.4550 speaking, a speech transcription model that transcribes the audio





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Smerkol et al.




to text, and a LLM fine-tuned to extract specific information from hundred conversations that way, and manually checked for mis-

the text. Figure 1 shows the architecture of the prototype system. takes in the model output. Many conversations were removed

due to selected actions not being mentioned or other reasons. Fi-



nally, the resulting dataset contains 350 conversations and JSON

formatted lists of tasks.

For the prediction of services required during a visit, we have

acquired a log of all services performed in one long-term patient

care facility over a period of 6 months, with the next version

expanding to data from six facilities. The tasks in dataset include

measurements (body temperature, heart rate, blood pressure, ...),

medical tasks (monitoring medicine intake, performing exam-

inations, turning the patient in bed) and care tasks (breakfast,

lunch, cleaning). There are over 200 different tasks mentioned.



Figure 1: Overview of the ASR system. The dataset includes limited patient information—patient ID, care

type, and a detailed chronological history of services received.

Care types (CareType I, CareType II, CareType III/A, CareType

1

We employ speaker-diarization pretrained model [4] for di-

III/B, CareType IIII) represent an estimate of how much assistance

arization, pyannote/speaker-diarization-3.1 pretrained model [7]

a person requires. Legal restrictions on accessing sensitive health

2

for transcription, and fine-tuned Llama3 model [2] for informa-

data prevented us from obtaining more detailed patient records,

tion extraction via generating JSON formatted output.

so we developed prediction models based on these limited data



2.2 Recommender System points, balancing accuracy with regulatory constraints. The data preprocessing involved determining each patient’s The recommender system prototype is based on machine-learning

presence in the facility by identifying the timestamps of their

prediction of events that are expected to occur in a certain time

first and last recorded service. Patients with a stay of less than

window for a specific patient with addition of tasks that com-

four months were excluded from the analysis to ensure sufficient

monly follow predicted tasks. Due to the sensitive nature of the

data for reliable predictions.

data, we base our predictions only on the time window, patient



ID and care type. Thus we consider multi-output binary clas- 4 Methods sifiers that do not require large amounts of data for training.

This section describes the methodology used to develop the ASR

Additional tasks are added to the list based on a Markov chain

system and the recommender system.

model that commonly follow, e.g. the task ’clean table’ follows



the task ’lunch’. 4.1 Clustering

The feature vector includes the time of day, day of week, week

of month and month of year as numbers, allowing for capture The primary goal of the clustering process was to group patients

of periodic events with different periods. Due to lack of patient with similar patterns in terms of the type and frequency of ser-

data, we opted for personalized models, trained for each patient vices they received, allowing us to predict relevant services more

separately. We believe that results can be further improved by effectively for each cluster (since it was not clear, even among

adding more patient-related attributes. The model training used experts, whether care type and actual care provided were corre-

five month period of data collected, with cross-validation, and lated).

the accuracy was evaluated on the data collected during the sixth The clusters, as shown in Figure 2, demonstrate that patients

month. Due to patients’ medical states changing over time, some within the same care type tend to receive similar services. Some

data drift is expected, which is reflected in our results. deviations, where multiple classifications appear within a cluster,

are likely due to temporary conditions we could not fully exclude



3 Dataset (for instance, an individual categorized under "Care Type II" may

temporarily receive services typical of "Care Type III/A" (e.g.

To fine-tune the information extraction model based on Llama3,

due to a broken arm), while their care type classification remains

we have created a dataset of conversations in text form and ap-

unchanged). Despite this, the care types differentiate well enough

propriate outputs for each of them, as the task on hand is very

across clusters, leading us to use "CareType" as one of the key

specific and we did not find any existing appropriate dataset. We

attribute for further service predictions.

automated the process and manually removed any bad exam-

In the clustering process, we excluded CareType IIII because

ples. A real dataset, ideally recorded in the target environment,

this group is characterized by highly diverse healthcare needs

is needed for for final implementation - LLM generated datasets

due to specific diseases, and experts advised us to omit it for this

used for training LLMs are only appropriate in preliminary stud-

part of the analysis.

ies.

3



To generate the dataset, we prepared a BERT LLM via prompt- 4.2 Recommender System

ing [5]. A training sample was generating by first randomly select-

ing 2 of the 10 target actions, and programmatically generating To recommend the required services, we constructed the train-

the target output JSON. The BERT model was then tasked with ing dataset using a detailed log of care actions performed over

generating a conversation, in which these two tasks are men- a 6-month period. For each patient, the data was divided into

tioned as done during the conversation. We generated several consecutive 4-hour time windows. In each window, we examined

whether specific care actions were performed, marking them as

1

pyannote/speaker-diarization-3.1

2 "positive" if they occurred within that time frame. This granu-

meta/meta-llama-3-8b

3google-bert/bert-base-multilingual-cased lar approach allowed us to capture the temporal dynamics of

Speech-to-Service Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia





added. Thus, the transcribed text resembles a play with unknown

characters speaking.

The information extraction model is Llama3 [2], and fine-

tuned utilising a LoRA few-shot fine tuning. Our approach was

to fine-tune the model for the task of extracting information about

specific care actions and generate the output in a JSON format,

providing structured data directly. A small training dataset was

prepared as described in the section 3.



5 Results and Discussion



5.1 LLM Based Infromation Retrieval Model

The Llama3 based information extraction model is evaluated

using a 5-fold cross validation, achieving , 95% accuracy 97%

precision, and 94% recall. For evaluation the model’s JSON-

Figure 2: Clustering of patients closely aligns with pre- formatted strings were deserialized to objects and tested against

existing care type assignments, ranging from minimal known correct objects to be able to interpret the results as multi-

personal assistance (CareType I) to moderate assistance label binary classification.

(CareType II), and full or intensive personal assistance The LLM infromation extraction model sometimes generates

(CareTypes III/A and III/B) for those with more severe care invalid JSON after fine-tuning, most commonly due to duplicated

needs. keys or getting stuck in a loop, generating same elements until

maximum output size is generated. The generated strings are

therefore post-processed to fix these mistakes via simple string

manipulation, however this indicates that experiments with dif-

service delivery, ensuring that for each time window, we had a

ferent output formats or avoiding generating the answers should

clear record of the services provided. As a result, we generated

be performed.

over 1000 labeled examples per patient, with each example rep-

The whole ASR pipeline including diarization and transcrip-

resenting a specific time window and its associated care actions.

tion has not yet been evaluated and falls within the scope of

This enabled the model to learn patterns in service requirements

future work.

throughout the day and week.



To identify the best predictive model, we evaluated various 5.2 Recommender System

classification algorithms, including Random Forest, Decision Tree,

Tables 1 and 3 present the classification results. Table 1 reports

K-Neighbors, Support Vector Classifier (SVC), Gradient Boost-

the average performance across all patients, including standard

ing, and Naive Bayes. Each model was trained using a multi-

deviations for the different models, while Table 3 shows classifica-

output classification approach, with features including the fre-

tion accuracy by care type, with averages and standard deviations

quency of the top services provided and the relevant time at-

across all patients within each care type, based on the model with

tributes. To ensure robust model evaluation, we implemented

the best results, which in this case is K-Neighbors (KNN).

5-fold cross-validation and subsequently tested the models on

Results are reported in two ways, tables 1 and 3 show accuracy

the sixth month’s data to assess their predictive performance.

considering all target attributes, only considering a prediction



4.3 Speech Recognition and Information correct when all targets are predicted correctly. The table 2 show

average of accuracies for each target attribute.

Extraction

Due to limited availability of training data, only the information- Table 1: Cross-validation and test accuracy (mean ± stan-

extraction model based on Llama3 was fine tuned using few-shot dard deviation) across all patients for various classification

LoRA (low-rank adaptor) supervised training. The diarization models.

and transcription models are used unchanged.



The diarization model used is speaker-diarization [4]. Initial Model CV Accuracy Test Accuracy

experiments with few-shot LoRA fine tuning [3] did not improve

RandomForest 0.71 ± 0.14 0.66 ± 0.16

the performance, hinting at the need for a larger training dataset.

DecisionTree 0.65 ± 0.16 0.66 ± 0.16

The model’s performance is satisfactory at the task of segmenta-

tion, but less so at the task of identifying which segments belong KNeighbors 0.73 ± 0.13 0.71 ± 0.16

to which speaker, especially for longer conversations. For a two- SVC 0.63 ± 0.12 0.63 ± 0.14 speaker situation, the model seems to assume the speakers take

GradientBoosting 0.68 ± 0.12 0.66 ± 0.15

turns speaking, causing mistakes when a single speaker pauses

NaiveBayes 0.57 ± 0.17 0.55 ± 0.20

before continuing to speak.

The transcription model used is whisper [8]. The model tran-

scribes each segment separately. As mentioned above, the speak- The K-Neighbors (KNN) classifier outperformed other models,

ers are not robustly recognised, and we cannot reliably assign a achieving an average CV accuracy of 73%, a test accuracy of 71%,

speaker to each line of text. Still, labeling each line of text even and R2 score of 0.44. This made it the most effective model for

with an ambiguous label improves the downstream task of infor- predicting service plans. Random Forest also performed reason-

mation extraction. The transcribed lines of text are concatenated, ably well, achieving a test accuracy of 66%, though it did not

and at the start of each utterance a label marking it as such is surpass KNN in overall performance.





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Smerkol et al.




Table 2: Majority Class Percentage and Task-wise Test Ac- failure at each step. The information retrieval model itself is not

curacy (mean ± standard deviation) across all patients for inefficient considering computational time and memory required,

various classification models. but diarization and transcription steps are. The required service

prediction should also be further improved. Using current dataset

Model Majority Class Task-wise an alternative approach that may improve performance is using

Percentage Accuracy sequence modelling or event prediction approaches. Finally, the

two models could work in tandem - predicting the required ac-

RandomForest 0.72 ± 0.19 0.89 ± 0.10

tions and using that information in the ASR pipeline could be

DecisionTree 0.72 ± 0.19 0.89 ± 0.11

beneficial.

KNeighbors Based on the proof-of-concept study, we conclude the sug- 0.72 ± 0.19 0.91 ± 0.10

gested approach is in principle feasible and can be beneficial

SVC 0.65 ± 0.16 0.88 ± 0.10

to healthcare providers. However, in view of regulations, spe-

GradientBoosting 0.65 ± 0.16 0.89 ± 0.09

cial caution has to be paid during the implementation of any

NaiveBayes 0.72 ± 0.19 0.85 ± 0.15

sort of such system in a real-world setting. Recording and di-

arizing conversations between healthcare staff and the patients

Table 3: Classification performance of K-Neighbors (KNN) is likely to include highly personal data, which falls under the

by CareType, showing cross-validation and test accuracy EU relevant legislation, specifically the GDPR (General Data Pro-

(mean ± standard deviation), averaged across all patients 4 tection Regulation ) and the EU AI Act (Artificial Intelligence

within each care type. Act (Regulation (EU) 2024/1689) 5 ). Furthermore, indiscriminately

recording conversations and feeding them into an LLM will likely



CareType CV Accuracy Test Accuracy be considered as "high risk" in view of the AI Act. This means

that implementing such services will require extensive screening,

CareType I 0.79 ± 0.12 0.76 ± 0.16

documentation, clear division of ownership and access roles, and

CareType II 0.79 ± 0.11 0.78 ± 0.13 other compliance with legal requirements.



CareType III/A 0.68 ± 0.13 0.66 ± 0.15 Acknowledgements

CareType III/B 0.70 ± 0.14 0.68 ± 0.17

We thank the healthcare provider organization for the dataset

CareType IIII 0.68 ± 0.10 0.67 ± 0.12

and for insightful discussions.



References

Since all predictive accuracy values exceed the 70% majority

[1] Ayo Adedeji, Sarita Joshi, and Brendan Doohan. 2024. The sound of health-

class baseline, this is an excellent result. In multi-label classifica- care: improving medical transcription asr accuracy with large language

models. . arXiv preprint arXiv:2402.07658

tion, where multiple services are predicted simultaneously, it’s

[2] AI@Meta. 2024. Llama 3 model card. https://github.com/meta- llama/llama3

important to not only focus on overall accuracy but also on the

/blob/main/MODEL_CARD.md.

accuracy of each individual task. By achieving 90% accuracy on [3] Shamil Ayupov and Nadezhda Chirkova. 2022. Parameter-efficient finetun-

ing of transformers for source code. , abs/2212.05901. https://api.sema ArXiv

the most common tasks, the model ensures that key services are

nticscholar.org/CorpusID:254564456.

reliably predicted. [4] Hervé Bredin and Antoine Laurent. 2021. End-to-end speaker segmentation

The lower test accuracy compared to cross-validation can be for overlap-aware resegmentation. In Proc. Interspeech 2021. Brno, Czech

Republic, (Aug. 2021).

explained by temporal changes in patient conditions, as the test

[5] Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2018.

set only included the last month of data. As patient care needs BERT: pre-training of deep bidirectional transformers for language under-

shift over time, predicting long-term patterns is more challenging standing. , abs/1810.04805. http : / / arxiv . org / abs / 1810 . 04805 arXiv: CoRR

1810.04805.

than shorter-term cross-validation, where care remains more

[6] Senay A. Gebreab, Khaled Salah, Raja Jayaraman, Muhammad Habib ur

stable. Rehman, and Samer Ellaham. 2024. Llm-based framework for administrative

task automation in healthcare. In 2024 12th International Symposium on

The test accuracy also reflected noticeable differences across

Digital Forensics and Security (ISDFS), 1–7. doi: 10.1109/ISDFS60797.2024.10

care types. CareType I and CareType II showed higher accuracy

527275.

rates, while more complex types, such as CareType III/A, III/B, [7] Alec Radford, Jong Wook Kim, Tao Xu, Greg Brockman, Christine McLeavey,

and Ilya Sutskever. 2022. Robust speech recognition via large-scale weak

and IIII, exhibited a drop in accuracy of around 10%. This is likely

supervision. (2022). doi: 10.48550/ARXIV.2212.04356.

due to the more diverse and unpredictable care needs in these [8] Alec Radford, Jong Wook Kim, Tao Xu, Greg Brockman, Christine McLeavey,

and Ilya Sutskever. 2022. Robust speech recognition via large-scale weak

groups, making service prediction more challenging.

supervision. (2022). doi: 10.48550/ARXIV.2212.04356.

This approach, particularly with the strong performance of

[9] Prakasam S, N. Balakrishnan, Kirthickram T R, Ajith Jerom B, and Deepak S.

our K-Neighbors (KNN) model, demonstrated the potential of 2023. Design and development of ai-powered healthcare whatsapp chatbot.

machine learning to enhance personalized planning in healthcare. 2023 2nd International Conference on Vision Towards Emerging Trends in

Communication and Networking Technologies (ViTECoN), 1–6. https://api.se

In future work, including additional patient-specific features manticscholar.org/CorpusID:259280109. beyond time-based data, such as health-related attributes, could [10] Raja Vavekanand, Pinja Karttunen, Yue Xu, Stephanie Milani, and Huao Li.

2024. Large language models in healthcare decision support: a review.

further improve accuracy, particularly for the more complex care

[11] Markus Vogel, Wolfgang Kaisers, Ralf Wassmuth, and Ertan Mayatepek.

types. 2015. Analysis of documentation speed using web-based medical speech

recognition technology: randomized controlled trial. Journal of medical



6 Conclusions Internet research, 17, 11, e247.

This is early work and further improvements are underway. The

whole ASR pipeline needs to be evaluated and we expect no-

ticeably worse performance comparing to only the information 4

https://gdpr-info.eu/

extraction model due to larger complexity and possibility for 5https://artificialintelligenceact.eu/the-act/





Performance Comparison of Axle Weight Prediction




Algorithms on Time-Series Data



Žiga Kolar David Susič Martin Konečnik

Jožef Stefan Institute Jožef Stefan Institute Cestel Cestni Inženiring d.o.o

Jamova cesta 39 Jamova cesta 39 Špruha 32

Ljubljana, Slovenia Ljubljana, Slovenia Trzin, Slovenia

ziga.kolar@ijs.si david.susic@ijs.si martin.konecnik@cestel.si



Domen Prestor Tomo Pejanovič Nosaka Bajko Kulauzović

Cestel Cestni Inženiring d.o.o Cestel Cestni Inženiring d.o.o Cestel Cestni Inženiring d.o.o

Špruha 32 Špruha 32 Špruha 32

Trzin, Slovenia Trzin, Slovenia Trzin, Slovenia

domen.prestor@cestel.si tomo.pejanovic@cestel.si bajko@cestel.si



Jan Kalin Matjaž Skobir Matjaž Gams

Zavod za gradbeništvo Slovenije Cestel Cestni Inženiring d.o.o Jožef Stefan Institute

Dimičeva ulica 12 Špruha 32 Jamova cesta 39

Ljubljana, Slovenia Trzin, Slovenia Ljubljana, Slovenia

jan.kalin@zag.si matjaz.skobir@cestel.si matjaz.gams@ijs.si



Abstract including road maintenance planning, traffic management, and

Accurate vehicle axle weight estimation is essential for the main- the prevention of overloading, which can lead to premature road

tenance and safety of transportation infrastructure. This study wear and increased accident risks [8]. Traditional methods for

evaluates and compares the performance of various algorithms axle weight measurement often rely on static scales or weigh-

for axle weight prediction using time-series data. The algorithms in-motion (WIM) systems. While these methods provide direct

assessed include traditional machine learning models (e.g., ran- measurements, they are susceptible to limitations such as high

dom forest) and advanced deep learning techniques (e.g., con- installation and maintenance costs, potential measurement inac-

volutional neural networks). The evaluation utilized datasets curacies due to environmental factors, and the need for frequent

comprising time-series data from 10 sensors positioned on a sin- calibration.

gle lane of a bridge, with the goal of predicting each vehicle’s axle In recent years, the advent of advanced computational tech-

weights based on the signals from these sensors. Each algorithm’s niques has opened new avenues for improving axle weight predic-

performance was measured against the OIML R-134 recommen- tion. Machine learning (ML) and deep learning (DL) algorithms, in

dation, where a prediction was classified as accurate if the error particular, offer promising alternatives by leveraging time-series

was within ±4 percent for two-axle vehicles and ±8 percent for data to model complex, non-linear relationships inherent in ve-

vehicles with more than two axles. Tests were conducted on sev- hicular weight patterns. These methods can enhance prediction

eral bridges, with this paper presenting detailed results from the accuracy, handle large volumes of data, and adapt to varying con-

Lopata bridge. Findings indicate that deep learning models, par- ditions, making them suitable for real-world applications where

ticularly convolutional neural networks, significantly outperform traditional methods may fall short.

traditional methods in terms of accuracy and their ability to adapt This study systematically evaluates and compares the per-

to complex patterns in time-series data. This study provides a formance of various axle weight prediction algorithms using

valuable reference for researchers and practitioners aiming to time-series data. We focus on a diverse set of algorithms, includ-

enhance axle weight prediction systems, thereby contributing to ing machine learning models like random forests (RF) [6] and

more effective infrastructure management and safety monitoring. advanced deep learning techniques such as convolutional neural

networks (CNN) [4].

Keywords The objective of this research is to explore the potential of



time-series data, axle weight, machine learning, neural network combining traditional WIM systems with advanced ML and DL models to enhance axle weight predictions. By comparing the



Accurate axle weight prediction plays a pivotal role in the mainte-1 performance of different methodologies, including the SIWIM Introduction traditional model, random forest (IJS RF), a hybrid approach (AVERAGE(IJS, SIWIM traditional)), and a CNN-based model, this nance and safety of transportation infrastructure [7]. The precise study aims to identify the most effective strategies for accurate estimation of axle weights is essential for various applications, and reliable axle weight estimation. Additionally, it examines the

Permission to make digital or hard copies of all or part of this work for personal impact of synthetic data generation on the performance of these

or classroom use is granted without fee provided that copies are not made or models, providing a comprehensive evaluation of their practical

distributed for profit or commercial advantage and that copies bear this notice and applicability in real-world scenarios.

the full citation on the first page. Copyrights for third-party components of this The study aimed to predict the axle weights of vehicles using

work must be honored. For all other uses, contact the owner/author(s). ten input signals from sensors placed under the Lopata bridge.

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.scai.4752





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Kolar et al.




Each predictive algorithm’s performance was evaluated accord- Bosso et al. [1] proposed a method using weigh-in-motion

ing to the OIML R-134 recommendation, which is deemed accu- (WIM) data and regression trees to identify patterns in over-

rate if the error margin for predicting the axle weight is within loaded truck weights and travel. The analysis reveals that truck

±4% for vehicles with two axles and within ±8% for vehicles with type is the key predictor of overloading, while time of day is

more than two axles. crucial for axle overloading, with most incidents occurring late

The dataset comprised 1478 samples, i.e. passing of a vehicle, at night or early morning. These insights can enhance enforce-

each containing 10 signals per vehicle. For each sample, a static ment strategies and inform pavement management and design,

weight for each axle was assigned as the target value. Static optimizing infrastructure longevity and safety.

weight refers to the weight measured by a scale when the vehicle He et al. [2] introduced a new method that uses only the

is stationary. flexural strain signals from weighing sensors to identify axle

This paper is structured as follows: Section 2 reviews several spacing and weights, reducing installation costs and expanding

state-of-the-art approaches. Section 3 details the preprocessing BWIM applications. The method’s accuracy is validated through

steps necessary before applying machine learning methods. In numerical simulations and laboratory experiments with a scaled

Section 4, algorithms used for predicting axle weights are pre- vehicle-bridge interaction model, showing promising results for

sented. Section 5 presents the final results of the axle weight accurate axle spacing and weight identification.

predictions. Finally, Section 6 summarizes the findings and pro-

poses ideas for future research. 3 Data Preprocessing

Before applying various algorithms to the dataset, several pre-



The prediction of axle weights using time-series data has often signals from each sample, padding was performed to standardize them to the length of the longest signal. Samples with a gross 2 processing steps were necessary. Due to the differing lengths of Related Work been studied in recent years, resulting in a substantial body of weight below 5 kN were excluded from both the training and related work. Below, several state-of-the-art (SOTA) approaches test datasets. Each signal was cropped by removing data to the are described. left of the leftmost peak value minus 100 and to the right of the Zhou et al. [10] differentiated between high-speed and low- rightmost peak value plus 100. The peak values were calculated speed weigh-in-motion (WIM) systems and analyzed the char- acteristics of axle weight signals. They proposed a nonlinear in advance. To address the limited availability of data required for deep curve-fitting algorithm, detailing its implementation. Numerical learning, which typically necessitates tens of thousands of sam-simulations and field experiments assessed the method’s perfor-ples for effective training, synthetic data generation was em-mance, demonstrating its effectiveness with maximum weighing ployed. The original dataset comprised 1,478 samples (from Janu-errors for the front axle, rear axle, and gross weights recorded ary 2022 to December 2023) i.e. passing of a vehicle, each contain-at 4.01%, 5.24%, and 3.92%, respectively, at speeds of 15 km/h or ing 10 signals per vehicle. An additional 20,000 synthetic samples lower. were generated using a specific algorithm. This algorithm oper-Wu et al. [8] introduced a modified encoder-decoder architec-ates by iterating 20,000 times, during each of which a random ture with a signal-reconstruction layer to identify vehicle proper-training sample and a random strain factor were selected. The ties (velocity, wheelbase, axle weight) using the bridge’s dynamic strain factor is a random value ranging between 0.5 and 0.99. The response. This unsupervised encoder-decoder method extracts selected signal from the training sample was then scaled by the higher features from the original data. A numerical bridge model chosen strain factor. This scaling process effectively models the based on vehicle-bridge coupling vibration theory demonstrated feature that doubling the amplitude of the signal corresponds to



the method’s applicability. Results indicated that the proposed ap- doubling its weight. proach accurately predicts traffic loads without additional sensors A crucial aspect of data preprocessing involved the normal-or vehicle weight labels, achieving better stability and reliability ization of sensor signals to ensure uniformity across the dataset. even with significant data pollution. Each signal was normalized to have a mean of zero and a stan-Xu et al. [9] applied wavelet transform for denoising and re-dard deviation of one, which helps in improving the convergence constructing the WIM signal, and used a back propagation (BP) of machine learning algorithms by ensuring that each feature neural network optimized by the brain storm optimization (BSO) contributes equally to the learning process. algorithm to process the WIM signal. Comparing the predictive The selection of training and test data was conducted using a abilities of BP neural networks optimized by different algorithms, rolling window approach [3]. Specifically, for each testing month, they found the BSO-BP WIM model to exhibit fast convergence the training data comprised all available data up to but not includ-and high accuracy, with a maximum gross weight relative error ing the testing month. For instance, if May 2023 was designated of 1.41% and a maximum axle weight relative error of 6.69%. as the testing month, the training dataset consisted of data from Kim et al. [5] developed signal analysis algorithms using artifi-January 2022 through April 2023. This process was systematically cial neural networks (ANN) for Bridge Weigh-in-Motion (B-WIM) repeated for each testing month from March 2022 to December systems. Their procedure involved extracting information on ve-2023. hicle weight, speed, and axle count from time-domain strain

data. ANNs were selected for their effectiveness in incorporating

dynamic effects and bridge-vehicle interactions. Vehicle exper- 4 Methodology

iments with various load cases were conducted on two bridge Four methods were identified as applicable for predicting vehicle

types: a simply supported pre-stressed concrete girder bridge and axle weights. The first method, known as SIWIM traditional [11],

a cable-stayed bridge. High-speed and low-speed WIM systems calculated the number of axles, axle lengths, and axle weights by

were used to cross-check and validate the algorithms’ perfor- utilizing influence lines to model the signal and determine the

mance. correct output. For validation purposes, each predicted output





Performance Comparison of Axle Weight Prediction Algorithms Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia





The detailed architecture of the CNN is shown in Figure 1. 2D

Convolutional layers (Conv2D) were used instead of 1D Convolu-

tional layers due to the input data consisting of 10 sensor signals.

The number of filters and kernel size are specified within the

parentheses of each Conv2D layer, while the pooling size is de-

fined in each 2D MaxPooling layer parentheses (MaxPooling2D).

The last Dense layer has 100 neurons. To mitigate overfitting, a

Dropout layer was added after the final Dense layer. Additionally,

Batch Normalization was applied after each 2D Convolutional

layer to further reduce the risk of overfitting.

Although Long Short-Term Memory (LSTM) and Gated Re-

current Unit (GRU) neural networks could be used for this task,

a Convolutional Neural Network (CNN) was chosen instead be-

cause of its strengths in capturing spatial hierarchies and local

patterns within the data. CNNs are highly effective at extracting

local features and detecting patterns, while LSTM and GRU are

better suited for handling temporal dependencies, which are not

that relevant to this specific task.



5 Results





Figure 1: Architecture of CNN for predicting axle weights.



was stored in a separate file alongside the signal data, enabling

direct comparison with the actual values. Figure 2: Accuracies of all algorithms for each testing

The second method used the random forest [6] (named IJS month.

RF) for predicting vehicle axle weights. The model relied on ac-

curately identifying the positions of peaks to function correctly.

Peak values were determined using the 𝑓 𝑖𝑛𝑑_𝑝𝑒𝑎𝑘𝑠 method from The results of each method described in Section 4 are illus-

the SciPy library, which identifies peaks based on a specified trated in Figure 2. Among the methods evaluated, SIWIM tradi-

minimum height. Once the peaks were identified, the algorithm tional exhibited the poorest performance, with fluctuating trends

extracted values within a ±5 range of each peak. These extracted observed throughout the entire two-year period. The CNN be-

values were then used as input features for the random forest gan to outperform the other three approaches after December

model. Additionally, the random forest model incorporated tem- 2022. Conversely, the AVERAGE(IJS, SIWIM traditional) method

perature, axle distances and gross weight as input features. Ran- showed superior performance during the initial testing months

dom forest algorithms are not inherently suited for time series from March 2022 to June 2022.

data; however, they perform effectively with numerical data such The performance of the CNN improved with an increasing

as temperature, axle distance, and gross weight. Therefore, this amount of data, whereas the IJS RF and AVERAGE(IJS, SIWIM

algorithm was chosen for analyzing this type of input data. traditional) methods were more effective during the initial phase

The third method integrated the first two approaches by aver- when less training data was available. However, the improvement

aging the outputs from the SIWIM traditional and IJS RF models in CNN’s accuracy was not linear. This non-linear trend can be

(named AVERAGE(IJS, SIWIM traditional)). This approach is mo- attributed to the random initialization of the CNN’s weights

tivated by the concept that combining multiple models can often before each training session, occasionally leading to suboptimal

yield more accurate results than relying on a single model alone convergence.

[12]. An additional analysis was conducted to compare the perfor-

The final method employed a convolutional neural network mance of the models under varying environmental conditions,

(CNN) to predict axle weights. The CNN utilized synthetic data, such as temperature fluctuations and differing traffic patterns.

as detailed in section 3, during the training phase. This method This analysis revealed that the CNN model maintained its accu-

processed all 10 signals as input to calculate the axle weights. racy more consistently across different conditions, indicating its





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Kolar et al.




robustness and adaptability. Furthermore, the inclusion of syn- be better tailored to capture specific nuances in the time-series

thetic data in training the CNN model contributed to its superior data.

performance, as it allowed the model to learn from a more di- After developing individual models for each sensor, the next

verse set of examples. Future research should focus on expanding step would be to combine the predictions from these models into

the range of synthetic data and exploring additional ensemble a single final prediction. This can be achieved using an ensemble

techniques to further enhance prediction accuracy. method, such as a random forest classifier. The random forest

Despite achieving high accuracy with the CNN model, with classifier would take the ten individual predictions (one from

the highest accuracy reaching 0.94, this most accurate method each sensor model) as input features and produce a consolidated

still falls short of meeting the OIML R-134 recommendation by final axle weight prediction.

4.4%. Furthermore, the results show that more static data could This method not only holds the potential to improve the ac-

be needed for the learning phase. Having 1000 static samples curacy and robustness of the axle weight predictions but also

which were augmented might not be sufficient to reach the OIML provides a scalable framework that can be adapted to different

R-134 recommendation. datasets and sensor configurations. Future work should explore

In summary, the results indicate that while traditional meth- the implementation of this approach, including the optimization

ods such as IJS RF and AVERAGE(IJS, SIWIM traditional) perform of individual sensor models and the integration of their predic-

well with limited data, convolutional neural networks (CNNs) tions through an ensemble method.

demonstrate superior performance as more data becomes avail- By advancing the CNN model in this manner, it is anticipated

able, despite some variability in their convergence. In addition, that the performance gap relative to the OIML R-134 recommen-

a sufficient number of training examples is needed to approach dation could be further reduced, bringing the predictions closer

the desired OIML R-134 recommendation. to the required accuracy levels with a smaller amount of data

and enhancing the overall efficacy of the axle weight prediction



6 system. Conclusion and Discussion



algorithms evaluated encompassed traditional machine learning acknowledge the funding from the Slovenian Research and Inno- vation Agency (ARIS), Grant (PR-10495) and Basic core funding models, such as random forests, and advanced deep learning P2-0209. The author(s) made use of chatGPT to assist with this techniques, notably convolutional neural networks. article. ChatGPT was commonly employed as a tool for enhanc- The major findings reveal that CNNs achieved significantly ing the language of the initial draft, without altering the length better results in predicting axle weights during the latter months prediction algorithms using time-series data collected from 10 This study received funding from company Cestel. The authors sensors positioned on the Lopata bridge was conducted. The In this study, a performance comparison of various axle weight Acknowledgements



of the experiment. The CNNs’ ability to adapt to and learn from of the text. ChatGPT 4 was accessed/obtained from chatgpt.com and used with modification in July 2024. complex patterns within the time series data was a key factor in



4.4%. their superior performance. Despite achieving high accuracy with References the CNN model, reaching a peak accuracy of 0.94, this method [1] Mariana Bosso, Kamilla L Vasconcelos, Linda Lee Ho, and Liedi LB Bernucci. still falls short of meeting the OIML R-134 recommendation by 2020. Use of regression trees to predict overweight trucks from historical weigh-in-motion data. Journal of Traffic and Transportation Engineering , 7, 6, 843–859. (English Edition) Overall, there are three implications of this study. First, it [2] Wei He, Tianyang Ling, Eugene J OBrien, and Lu Deng. 2019. Virtual axle demonstrates the potential of deep learning techniques to en-method for bridge weigh-in-motion systems requiring no axle detector.



hance the accuracy of axle weight predictions where sufficient Journal of Bridge Engineering , 24, 9, 04019086. [3] Hamed Kalhori, Mehrisadat Makki Alamdari, Xinqun Zhu, Bijan Samali, data is available, thereby facilitating more reliable infrastructure and Samir Mustapha. 2017. Non-intrusive schemes for speed and axle iden-management. Second, for smaller datasets, it is more effective tification in bridge-weigh-in-motion systems. Measurement Science and , 28, 2, 025102. to use classical machine learning systems in combination with Technology [4] Teja Kattenborn, Jens Leitloff, Felix Schiefer, and Stefan Hinz. 2021. Review methods like SIWIM traditional. Third, it provides a valuable on convolutional neural networks (cnn) in vegetation remote sensing. ISPRS

benchmark for researchers and practitioners, guiding the de- journal of photogrammetry and remote sensing, 173, 24–49.

velopment and implementation of more effective axle weight [5] Sungkon Kim, Jungwhee Lee, Min-Seok Park, and Byung-Wan Jo. 2009.

Vehicle signal analysis using artificial neural networks for a bridge weigh-

prediction systems. in-motion system. Sensors, 9, 10, 7943–7956.



possible: To achieve the OIML R-134 recommendation, two options are [6] Steven J Rigatti. 2017. Random forest. Journal of Insurance Medicine, 47, 1, 31–39. [7] Mohhammad Sujon and Fei Dai. 2021. Application of weigh-in-motion

technologies for pavement and bridge response monitoring: state-of-the-art

• Just add more data - if the trend continues, adding another review. Automation in Construction, 130, 103844.



• half a year of measurements would enable achieving the [8] Yuhan Wu, Lu Deng, and Wei He. 2020. Bwimnet: a novel method for iden- tifying moving vehicles utilizing a modified encoder-decoder architecture. standard. Another option would be to apply measurements Sensors , 20, 24, 7170. on a bridge with more traffic. [9] Suan Xu, Xing Chen, Yaqiong Fu, Hongwei Xu, and Kaixing Hong. 2022. Research on weigh-in-motion algorithm of vehicles based on bso-bp. Sensors , Improve the methods by incorporating advanced ensemble 22, 6, 2109. techniques. [10] ZF Zhou, P Cai, and RX Chen. 2007. Estimating the axle weight of vehicle

in motion based on nonlinear curve-fitting. IET science, measurement &



ment involves modeling each sensor individually. This approach [11] A Žnidarič, J Kalin, M Kreslin, M Mavrič, et al. 2016. Recent advances in bridge wim technology. In Proc. 7th International Conference on WIM . entails building a separate CNN model for each of the ten sen- To introduce the ensemble approaches, one potential improve- technology, 1, 4, 185–190.

[12] Hui Zou and Yuhong Yang. 2004. Combining time series models for fore-

sors, allowing for more specialized and potentially more accurate casting. International journal of Forecasting, 20, 1, 69–84.

predictions from each sensor’s data. By focusing on the unique

characteristics and data patterns of each sensor, the models can



Comparison of Feature- and Embedding-based Approaches for



Audio and Visual Emotion Classification



Sebastijan Trojer Zoja Anžur

st5804@student.uni- lj.si zoja.anzur@ijs.si

Jožef Stefan Institute Jožef Stefan Institute

Faculty of Computer and Information Science Ljubljana, Slovenia

Ljubljana, Slovenia



Mitja Luštrek Gašper Slapničar

mitja.lustrek@ijs.si gasper.slapnicar@ijs.si

Jožef Stefan Institute Jožef Stefan Institute

Jožef Stefan International Postgraduate School Ljubljana, Slovenia

Ljubljana, Slovenia



Abstract nature, lacking explainability and interpretability of the internally

derived features [9]. Furthermore, while some research suggests

This paper presents a comparative analysis of feature- and embe-

superior performance of embeddings compared to traditional

dding-based approaches for audio-visual emotion classification.

features [20], this is not universally agreed upon [8], especially

We compared the performance of traditional handcrafted fea-

when taking into account potentially much higher computational

tures, using MediaPipe for visual features and Mel-frequency

complexity of deriving embeddings with deep artificial neural

cepstral coefficients (MFCCs) for audio features, against neural

networks (ANNs).

network (NN)-based embeddings obtained from pretrained mod-

Our research question is thus, whether it is better to compute

els suitable for emotion recognition (ER). The study employs

embeddings using SOTA pretrained DL models instead of using

separate uni-modal datasets for audio and visual modalities to

hand-crafted features, as ANN embeddings promise to increase

rigorously assess the performance of each feature set on each

detection accuracy at the cost of interpretability and computa-

modality. Results demonstrate that in the case of visual data NN-

tional complexity. In this work we compared the performance of

based embeddings significantly outperform handcrafted features

hand-crafted features and embeddings obtained with pretrained

in terms of accuracy and F1 score when training a traditional

SOTA models for the down-stream task of emotion recognition.

classifier. However, for audio data the performance is similar

We independently compared ER performance of audio and video

on all feature sets. Handcrafted features, such as facial blend-

modality, using established benchmark datasets for each. Hand-

shapes, computed from MediaPipe keypoints and MFCCs, re-

crafted features were chosen based on literature and embeddings

main relevant in resource-constrained settings due to their lower

were computed with task-suitable pretrained models available

computational demands. This research provides insights into

in existing Python libraries. Both were formatted in a way that

the trade-offs between traditional feature extraction methods

allowed us to then train a set of traditional ML models, listed in

and modern deep learning techniques, offering guidance for the

Section 3.3, for ER, using hand-crafted features, embeddings, or

development of future emotion classification systems.

a union of both as inputs.



Keywords

emotion recognition, embeddings, hand-crafted features 2 Related Work

Performance comparison of hand-crafted features and learned

1 Introduction embeddings has been discussed in depth in computer vision do-

main. Schonberger et al. [15] demonstrated that hand-crafted

Automated emotion recognition (ER) often focuses on two modali-

ties – video and audio. This is akin to human emotion recognition, features (e.g., SIFT) still perform on par or better than learned

embeddings in image reconstruction. They warned of high vari-

as we heavily rely on audio-visual characteristics, such as facial

expressions and audio cues, to deduce emotional state [7]. Both ance across datasets when using learned embeddings as features.

Similarly, Antipov et al. [2] reported similar performance of hand-

audio and video are relatively simple to obtain using sensors,

crafted features (e.g., HOG) and learned embeddings when classi-

as such sensors are unobtrusive and easily available (e.g., web-

cameras) and can be used to train machine learning (ML) models fying pedestrian gender from images using small datasets. They

also highlighted superior generalization performance of embed-

for emotion recognition.

In the past decade deep-learning (DL) approaches achieved dings across (unseen) datasets. In emotion recognition from audio,

Papakostas et al. [13] compared using hand-crafted MFCC-based

state-of-the-art (SOTA) results in many domains, including emo-

tion recognition [16]. However, despite the superior performance features with embeddings from a custom convolutional neural

network (CNN) trained on spectrograms. The latter slightly out-

of such models, many doubts have been cast on their black-box

performed hand-crafted features by 1% on average in terms of

Permission to make digital or hard copies of all or part of this work for personal

F1 score, again showing similar performance. Ye et al. [21] re-

or classroom use is granted without fee provided that copies are not made or

cently showed that using a union of both hand-crafted features

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this and learned embeddings achieves superior performance in user

work must be honored. For all other uses, contact the owner /author(s).

identification, compared to using each input individually.

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

There is moderate (but not universal) agreement in recent

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.scai.6883 literature that performance between hand-crafted features and



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Sebastijan Trojer, Zoja Anžur, Mitja Luštrek, and Gašper Slapničar



learned embeddings is similar, however, most work comparing the other hand, we chose SOTA pretrained models trained for

their performance is limited to a single modality or task. We related tasks. We extracted embeddings at a model-specific point

compared performance between two different modalities for the before the learning layers, and formatted them using principal

task of ER and investigated potential performance improvements component analysis (PCA) in order to reduce their dimensionality

of feature-level fusion (hand-crafted + embeddings). while maintaining the relevant information.



3 Methodology in ER from audio [10], as they give a good approximation of the 3.2.1 Audio Features. MFCCs are historically well-established

Our task consisted of two parts – hand-crafted features and em- human auditory system’s response. For each audio clip, we com-

beddings computation, and ER model training for classification. puted a common set of statistical aggregate features (averages,

Both were done on (separate) audio and visual modality and will standard deviations) for MFCCs, Root Mean Square (RMS) en-

be described per-modality in the following sections. ergy (volume), Zero-Crossing Rate, Spectral Bandwidth, Spectral

Contrast, and Spectral Roll-off, using the python library. librosa

3.1 Datasets For embeddings we decided to investigate models pretrained

on similar audio tasks (e.g., emotion recognition) and use them

As mentioned previously, the ER task is most-often audio-visual,

so we decided to use an audio and a visual dataset to indepen- to the point where embeddings are available, which typically

means the upper part of the ANN architecture, responsible for

dently evaluate the performance of different feature sets. While

many datasets exist that contain both modalities, they often have computation of embeddings representing the features. Three pre-

trained models were investigated in our evaluation, all based

a problem of imprecise coarse emotion labelling [18], as labels

on the wav2vec2 architecture, which is a self-supervised model

are video-based, while emotions can be exhibited and changed in

for learning speech representations proposed by Facebook AI

much shorter windows. Splitting video into frames yields a large

Research (FAIR) [3]. Full wav2vec2 pretraining framework com-

number of (different) instances with the same label, so we wanted

prises a latent feature encoder, a context network using the trans-

a dataset with individual image labels. As our focus was on com-

former architecture, a quantization module and contrastive loss

paring the performance of hand-crafted and embedding-based

features, we chose two well-established benchmark datasets ded- (pre-training objective). For our purposes the feature encoder

is important, which is a 7-layer 1D CNN reducing the dimen-

icated to audio and visual emotion classification. These datasets

contain short audio clips and individual images with precise sionality of audio inputs into a sequence of feature vectors. The

initial model version was pretrained on the LibriSpeech dataset,

short-term and per-frame labels, circumventing the mentioned

another version was fine-tuned on IEMOCAP dataset specifically

per-video labelling problem.

for ER, and finally a large general cross-lingual model (XLSR)

3.1.1 Audio Dataset. For evaluation on audio data we decided was trained on millions of hours of unlabeled audio data in 53 to use the crowd-sourced emotional multimodal actors dataset

(later extended) languages [5]. These three variants were used

(CREMA-D) [4]. It contains short clips of 91 actors between the

to extract their corresponding embeddings. Since the input data

ages of 20 and 74 coming from a variety of races and ethnicities,

from CREMA-D is of inconsistent shape (varying by < 1 sec), we

who exhibited six different emotions (Anger, Disgust, Fear, Happy,

had to employ an additional adaptive average pooling layer to en-

Neutral, Sad). Each actor produced about 80 clips (small vari- sure consistently shaped outputs. We designed this pooling layer

ation), saying specific sentences exhibiting different emotions.

so that we lost minimal information (short segment length for

The distribution of labels was balanced, each class representing

pooling) and the outputs were then flattened. PCA was employed

approx. 16% of the data. The intended emotions were verified

to subsequently reduce them to 10 dimensions. The number of di-

with 2,443 crowd-sourced human raters as baseline. These raters

mensions was chosen arbitrarily and could be changed, however,

predicted emotions based on audio only, video only, or both,

we believe that 10 dimensions offer a good balance between re-

achieving 40.9%, 58.2% and 63.6% recognition of intended (acted)

tained information and computational (and spatial) requirements.

emotion respectively.

Moreover, this number of PCA components is on the same order



3.1.2 of magnitude as the number of hand-crafted features, making Visual Dataset.

For visual data we chose the extended

them more comparable.

Cohn-Kanade dataset (CK+) [11], which a staple dataset in ER



evaluation from facial expressions. It contains images of 118 3.2.2 Visual Features. For visual features, we focused on the

adults, aged between 18 and 50, again of different ethnicities. Par-

movement of specific facial keypoints, such as the corners of

ticipants were instructed to perform a series of 23 facial displays,

the mouth and eyebrows, which form the basis of the Facial

relating to one of seven emotions (Anger, Contempt, Disgust, Fear,

Action Coding System (FACS) – a taxonomy that categorizes

Happy, Sad, Surprise). The distribution of classes in CK+ is not human facial expressions based on muscle movements [6]. We

balanced – is the majority class at 25% and the Surprise Contempt

employed the MediaPipe (MP) framework [12] to extract values

minority class at 6%, with others in between. This distribution

representing the activation of various facial blendshapes, which

also changes between subjects. CK+ images were reshaped to

correspond approximately to the regions defined in FACS. In this

48x48 pixels, put in grayscale format and cropped using frontal

paper, we classify MediaPipe features as “handcrafted” because,

face Haar cascade classifier [1] as part of preprocessing. The

despite being neural network-based, they quantify predefined

emotion labels were validated by experts via facial activation

facial areas with human-interpretable metrics. This contrasts

unit rules (e.g., = Activation unit 12 must be present = Lip Happy

with CNN-based embeddings, which capture patterns without

corner puller active).

direct interpretability.

For comparison, we used embeddings from two pretrained

3.2 Feature Computation models: FaceNet [17] and EfficientNet [19] from the HSEmotion

For selection of hand-crafted features we relied on literature library [14]. FaceNet architecture is based on GoogleNet, which

and previous work in ER for each modality. For embeddings on is a variant of deep CNN, and is trained using triplet loss. It





Comparison of Feature- and Embedding-based Approaches Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia




was optimized for facial recognition, verification, and clustering. (1) Hand-crafted statistical features relating to MFCCs

EfficientNet comprises several inverted bottleneck convolutional (2) 10-component PCA of wav2vec2 embeddings from a model

residual blocks. It achieved SOTA results on the AffectNet ER trained on LibriSpeech

dataset, while being relatively light-weight. Again, PCA was used (3) 10-component PCA of wav2vec2 embeddings from a model

to reduce the embeddings to 10 dimensions. trained on IEMOCAP

(4) 10-component PCA of wav2vec2 embeddings from a cross-

3.2.3 Computational and Spatial Requirements. In order to have

lingual XLSR model

a clear overview of the trade-off between computational and

(5) Union of hand-crafted and best-performing embeddings

spatial requirements of each feature computation method, and

(from above)

their classification performance discussed in the next section, we

These were compared in experiments as described in Section 3.3,

first report the average times to compute and disk sizes of the

using a set of four ML models. Results of best-performing model

output (per one instance) for each method in Table 1.

for each set in terms of accuracy and F1 are given in Table 2.

Table 1: Average time and disk space needed for feature Fused data was acquired by concatenating the feature sets.

computation using each method.

Table 2: Best performing models for each feature set and

Modality Feature method Avg. Time Avg. Space corresponding accuracy and F1 scores for audio data. Note

MFCC stats that embeddings were represented with 10 components 19 ms < 1 kB

wav2vec2 LibriSpeech 99 ms 194 kB obtained from PCA.

Audio

wav2vec2 XLSR 274 ms 258 kB

wav2vec2 IEMOCAP 101 ms 5 kB Feature set Best model Accuracy F1 score

MediaPipe 10 ms Majority 0.17 0.00 0.05 0.00 < 1 kB N/A ± ±

FaceNet 29 ms 2 kB MFCC stats RF 0.46 0.08 0.43 0.09 ± ±

Video

EfficientNet 5 kB 2 ms ± ±

wav2vec2 LibriSpeech SVM 0.47 0.08 0.45 0.09

When interpreting the results in Table 1, it must also be con- ± ±

wav2vec2 XLSR SVM 0.30 0.05 0.27 0.05

sidered that DL-based methods require additional computational ± ±

wav2vec2 IEMOCAP SVM 0.47 0.08 0.42 0.09

time when doing PCA on top of the raw embeddings. best 0.52±0.09 0.50±0.10 MFCC + wav2vec2 SVM



3.3 4.2 Image Emotion Classification Emotion Classification

To stay consistent with the audio experiments we performed the

Data splitting is a crucial step in evaluation of ML models, as

same LOSO experiments described in Section 3.3. We compared

it must be done in a way to avoid overfitting and provide a ro-

model performances using the following features as inputs:

bust evaluation of generalization capabilities of a model. The

aim of this research was primarily not to evaluate the absolute (1) MediaPipe blendshapes

performance of ER, but rather compare the performance when (2) 10-component PCA of FaceNet embeddings

using hand-crafted vs. embedding features. Therefore it was cru- (3) 10-component PCA of EfficientNet embeddings

cial to consistently ensure that the same data splits and models (4) Union of MP and FaceNet embeddings

were used in each experiment, for each of the compared inputs. (5) Union of MP and EfficientNet embeddings

We decided for the most robust leave-one-subject-out (LOSO)

Accuracy and F1 scores for the best performing models for

evaluation, always using default model hyperparameters. Such

each set of features are again reported in Table 3

experimental setup minimized overfitting and also gave a good

overview of generalization performance of emotion classifiers.

Table 3: Best-performing models for each feature set and



4 corresponding accuracy and F1 scores for visual data. Note Experiments and Results

that embeddings were represented with 10 components

The outputs of the previous step were used as inputs (features) obtained from PCA. to train a traditional ML model for emotion classification. We

evaluated several options: taking the 10 PCA components of em-

Feature set Best model Accuracy F1 score

beddings obtained from each pretrained model as inputs, taking

only hand-crafted features as inputs, and taking union of both N/A Majority 0.25±0.00 0.40±0.00

MediaPipe RF 0.62 0.28 0.51 0.29 ± ±

as input. Each of these cases was evaluated for audio and visual

FaceNet SVM 0.45 0.30 0.36 0.30 ± ±

modality separately, using the LOSO experimental setup. Several

EfficientNet RF 0.93±0.16 0.90±0.20

popular ML models were compared (with default hyperparame-

Mediapipe + FaceNet XGB 0.70 0.28 0.60 0.29 ± ±

ters), including k-nearest Neighbours (kNN), Random Forest (RF),

Mediapipe + EfficientNet XGB 0.93±0.17 0.90±0.21

Support Vector Machines (SVM) with linear kernel, and eXtreme

Gradient Boosting (XGB). We monitored classification accuracy

and macro F1 score as metrics of the model performance. All 4.3 Discussion

results were compared with baseline majority classifier and are From Tables 2 and 3 we can observe that for audio the best

reported as averages across all iterations of LOSO, where majority

performance is achieved when using union of hand-crafted and

was always taken from the train data (all except left-out). embedding features, while for visual ER the performance of only

embeddings or union is nearly identical. The improvement of

4.1 Audio Emotion Classification feature union is thus generally small, as for visual data we get

As mentioned in Section 3 we investigated the following options the same result as using only the best embeddings (1% difference

as feature inputs: in standard deviation), while for audio data the improvement in





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Sebastijan Trojer, Zoja Anžur, Mitja Luštrek, and Gašper Slapničar




both metrics is about 5% compared to individual feature sets. All used data was simulated/acted, so interpretation of these results

results substantially outperform the baseline majority classifiers. must take that into account. Numbers are expected to decrease

For audio data we can see that the best embedding set (wav2vec2 on a more realistic dataset, as emotions in everyday life are quite

LibriSpeech) performs nearly the same as hand-crafted features subtle [18]. It would thus make sense to run similar experiments

(MFCC stats), which is in agreement with some literature [13]. It on more realistic data as well, although such data is more scarce.

is surprising that LibriSpeech embeddings slightly outperform

IEMOCAP ones, since the latter were trained specifically for Acknowledgements emotion recognition, while the former were not. The subpar

This work was supported by bilateral Weave project, funded by

performance of XLSR is expected, since it is a more general cross-

the Slovenian Agency of Research and Innovation (ARIS) under

lingual unsupervised model, while investigated data is spoken in

grant agreement N1-0319 and by the Swiss National Science

English. For visual data on the other hand the best embeddings

Foundation (SNSF) under grant agreement 214991.

(EfficientNet) substantially outperform hand-crafted facial ex-

pression features (MediaPipe) and those obtained from FaceNet. References This is expected, as EfficientNet was trained specifically for emo- [1] Shahad Salh Ali, Jamila Harbi Al’ Ameri, and Thekra Abbas. 2022. Face tion recognition, while FaceNet was trained for face recognition. detection using Haar cascade algorithm. In 2022 Fifth College of Science

International Conference of Recent Trends in Information Technology (CSCTIT),

In terms of ML models, we consistently observed best perfor-

198–201. doi: 10.1109/csctit56299.2022.10145680.

mance of SVM for ER from audio data, while for video data the [2] Grigory Antipov, Sid-Ahmed Berrani, Natacha Ruchaud, et al. 2015. Learned

vs. hand-crafted features for pedestrian gender recognition. In Proceedings

best model is not as homogeneous. Importantly, performance of

of the 23rd ACM international conference on Multimedia, 1263–1266.

different models (RF, SVM and XGB) was often within 1%.

[3] Alexei Baevski, Yuhao Zhou, Abdelrahman Mohamed, et al. 2020. Wav2vec

Another important observation is the relative stability of re- 2.0: a framework for self-supervised learning of speech representations.

sults across subjects when classifying from audio, with standard Advances in neural information processing systems, 33, 12449–12460.

[4] Houwei Cao, David G Cooper, Michael K Keutmann, et al. 2014. Crema-d:

deviations around 8%. The same was not observed in the eval- crowd-sourced emotional multimodal actors dataset. IEEE transactions on uation from visual data, with much higher standard deviations, , 5, 4, 377–390. affective computing

[5] Alexis Conneau, Alexei Baevski, et al. 2020. Unsupervised cross-lingual rep-

indicating lower stability and greater variation between subjects.

resentation learning for speech recognition. . arXiv preprint arXiv:2006.13979

To address our initial research question, we observed simi- [6] Paul Ekman and Erika L Rosenberg. 1997. What the face reveals: Basic and

lar performance of hand-crafted features and embeddings from applied studies of spontaneous expression using the Facial Action Coding

System (FACS). Oxford University Press, USA.

SOTA DL models for audio-based ER, with union of both achiev-

[7] Monica Gori, Lucia Schiatti, and Maria B. Amadeo. 2021. Masking emotions:

ing the best results. The image-based visual ER achieved much face masks impair how we read emotions. , 12, 669432. Frontiers in Psychology

[8] Léo Grinsztajn, Edouard Oyallon, and Gaël Varoquaux. 2022. Why do tree-

better performance with learned embeddings as inputs, while the

based models still outperform deep learning on typical tabular data? Ad-

union of features showed no improvement. However, the cost , 35, 507–520. vances in neural information processing systems of hand-crafted features and embeddings in terms of computa- [9] Xuhong Li, Haoyi Xiong, Xingjian Li, et al. 2022. Interpretable deep learning:

interpretation, interpretability, trustworthiness, and beyond. Knowledge and

tional power required to compute, and spatial requirements to Information Systems, 64, 12, 3197–3234.

save, is not the same. While hand-crafted features are usually [10] MS Likitha, Sri Raksha R Gupta, K Hasitha, et al. 2017. Speech based human

emotion recognition using MFCC. In 2017 international conference on wireless

computed quickly and represented with a few numbers, as re-

communications, signal processing and networking (WiSPNET), 2257–2260.

ported in Table 1, the embeddings require loading a (commonly

[11] Patrick Lucey, Jeffrey F Cohn, Takeo Kanade, et al. 2010. The extended

large) pretrained ANN, which performs a large number of matrix cohn-kanade dataset (ck+): a complete dataset for action unit and emotion-

specified expression. In 2010 ieee computer society conference on computer

multiplications, resulting in high-dimensional embeddings (e.g.,

vision and pattern recognition-workshops. IEEE, 94–101.

64 512). This in turn requires additional dimensionality reduc- [12] Camillo Lugaresi, Jiuqiang Tang, Hadon Nash, et al. 2019. Mediapipe: a ×

tion, such as PCA employed in this work. Our results indicate framework for building perception pipelines. (2019). https://arxiv.org/abs/1

906.08172.

that for image-based visual ER, the additional cost is worthwhile,

[13] Michalis Papakostas, Evaggelos Spyrou, Theodoros Giannakopoulos, et al.

due to large improvements in performance, while audio-based 2017. Deep visual attributes vs. hand-crafted audio features on multidomain

speech emotion recognition. , 5, 2, 26. Computation

ER achieved much smaller improvement, making the use of em-

[14] Andrey Savchenko. 2023. Facial expression recognition with adaptive frame

beddings from pretrained models less attractive. Proceedings of the 40th Inter-

rate based on multiple testing correction. In

Finally, hand-crafted features mostly offer direct interpretabil- (Proceedings of Machine national Conference on Machine Learning (ICML)

Learning Research). Vol. 202. Pmlr, (July 2023), 30119–30129. https://procee

ity (e.g., audio loudness), while embeddings are commonly black-

dings.mlr.press/v202/savchenko23a.html.

box in nature, lacking explainability without suitable mechanisms [15] Johannes L Schonberger, Hans Hardmeier, Torsten Sattler, et al. 2017. Com-on top. The clear meaning of hand-crafted features can be helpful parative evaluation of hand-crafted and learned local features. In Proceedings

of the IEEE conference on computer vision and pattern recognition, 1482–1491.

when training traditional ML models, where feature importance

[16] Liam Schoneveld, Alice Othmani, and Hazem Abdelkawy. 2021. Leveraging

can be compared and subsequently interpreted. recent advances in deep learning for audio-visual emotion recognition.

Pattern Recognition Letters, 146, 1–7.

[17] Florian Schroff, Dmitry Kalenichenko, and James Philbin. 2015. Facenet: a



5 unified embedding for face recognition and clustering. In 2015 IEEE Confer- Conclusion

ence on Computer Vision and Pattern Recognition (CVPR). IEEE, (June 2015).

doi: 10.1109/cvpr.2015.7298682.

In summary we compared using hand-crafted features, embed-

[18] Gašper Slapničar, Zoja Anžur, Sebastijan Trojer, et al. 2024. Contact-free

dings of pretrained SOTA models, or union of both, as inputs for emotion recognition for monitoring of well-being: early prospects and future

ER models using audio and visual data. We found that embedding- ideas. In Intelligent Environments 2024: Combined Proceedings of Workshops

and Demos & Videos Session. IOS Press, 58–67.

based approach is substantially superior with visual data, out-

[19] Mingxing Tan and Quoc V. Le. 2019. Efficientnet: rethinking model scaling

weighing the computational cost – the latter is in fact the lowest for convolutional neural networks. . http://arxiv.org/abs/1905.11946. CoRR

[20] Pawan Kumar Verma, Prateek Agrawal, Ivone Amorim, et al. 2021. Welfake:

when using EfficientNet. For audio data, the improvement was

word embedding over linguistic features for fake news detection. IEEE

only seen in union of inputs, and was relatively low. Transactions on Computational Social Systems, 8, 4, 881–893.

As future work it would be worthwhile to compare merged [21] Cuicui Ye, Jing Yang, and Yan Mao. 2024. Fdhfui: fusing deep representa-

tion and hand-crafted features for user identification. IEEE Transactions on

audio-visual features and embeddings in a single ER problem on

Consumer Electronics.

the same dataset having both modalities. Furthermore, currently





Multi-modal Data Collection and Preliminary Statistical




Analysis for Cognitive Load Assessment



Ana Krstevska Sebastjan Kramar Hristijan Gjoreski

Department of Intelligent Systems Department of Intelligent Systems Faculty of Electrical Engineering and

Jožef Stefan Institute Jožef Stefan Institute Information Technologies

Ljubljana, Slovenia Ljubljana, Slovenia Skopje, Macedonia

ana.krstevska2001@gmail.com sebastjan.kramar@ijs.si hristijang@feit.ukim.edu.mk



Martin Gjoreski Junoš Lukan Sebastijan Trojer

Università della Svizzera italiana (USI) Department of Intelligent Systems Department of Intelligent Systems

Lugano, Switzerland Jožef Stefan Institute Jožef Stefan Institute

martin.gjoreski@usi.ch Jožef Stefan International Postgraduate Ljubljana, Slovenia

School st5804@student.uni-lj.si

Ljubljana, Slovenia

junos.lukan@ijs.si



Mitja Luštrek Gašper Slapničar

Department of Intelligent Systems Department of Intelligent Systems

Jožef Stefan Institute Jožef Stefan Institute

Jožef Stefan International Postgraduate School Ljubljana, Slovenia

Ljubljana, Slovenia gasper.slapnicar@ijs.si

mitja.lustrek@ijs.si



Abstract 1 Introduction

To mitigate distractions during complex tasks, ubiquitous Human attention is a critical resource that is increasingly targeted

computing devices should adapt to the user's cognitive load. by mobile applications, online services, and other forms of digital

However, accurately assessing cognitive load remains a significant engagement. In an era of constant connectivity, capturing and

challenge. This study aims to present sophisticated, multi-modal retaining user attention has become a primary objective for many

data collection, which can enable accurate estimation of cognitive technologies. However, as users engage in cognitively demanding

load using wearable and contact-free devices. A total of 25 tasks, distractions can lead to performance degradation and

participants participated in six cognitive load-inducing tasks, each increased stress. Therefore, to minimize interruptions and maintain

presented at two levels of difficulty. Simultaneously, physiological productivity, ubiquitous computing systems must become capable

and behavioral data were collected from a multi-modal sensory of recognizing and adapting to the user’s cognitive load in real time.

setup, including: Empatica E4 wristband, Emteq OCOsense

Cognitive load, defined as the mental effort required to process

glasses, an eye tracker, a thermal camera, a depth camera and an

information and perform tasks, triggers a series of physiological

RGB video camera. Additionally, participants provided subjective

responses in the human body. These responses are largely governed

measures of cognitive load by completing standardized NASA

by the activation of the sympathetic nervous system. When

Task Load Index (NASA TLX) and Instantaneous Self-Assessment

cognitive load increases, measurable changes can be observed in

(ISA) questionnaires following each cognitive task. Preliminary



performance metrics, and the perceived difficulty of tasks, as rate variability, etc. Furthermore, changes are also reflected in reported in the completed questionnaires. facial expressions, posture, and other behavioural patterns. Keywords This study seeks to offer a unique multi-modal dataset with a rich statistical analyses were conducted on participant demographics, movements, electrodermal activity (EDA), respiration rate, heart physiological markers, including blood pressure, brain activity, eye



cognitive load inference, wearable sensors, contact-free set of wearable and unobtrusive sensors to capture the subtle unobtrusive sensors changes that occur with the gradual activation of the sympathetic

Permission to make digital or hard copies of part or all of this work for personal or nervous system. Rather than solely focusing on maximizing data

classroom use is granted without fee provided that copies are not made or distributed accuracy through the use of numerous devices, this approach also

for profit or commercial advantage and that copies bear this notice and the full citation

on the first page. Copyrights for third-party components of this work must be honored. aims to identify the minimum set of sensors required to achieve

For all other uses, contact the owner/author(s). reliable cognitive load assessment. To that end, rich multi-modal

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia data was collected from a myriad of sensors, including wearables

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.scai.6961



35

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Ana Krstevska et al.





(OCOsense glasses and Empatica E4 wristband) and contact-free

unobtrusive sensors such as an advanced eye tracker, a thermal

camera, a depth camera, and an RGB video camera. To the best of

our knowledge, no prior dataset exists containing such rich multi-

modal data obtained with such an elaborate sensory setup.



2 Related Work

The challenge of cognitive load estimation has been extensively

studied across various fields. Significant emphasis has been placed

on reducing cognitive load in dynamic environments, such as

aviation [1]. Recent research has increasingly focused on



transitioning from direct measurements, such as

electroencephalography (EEG), to indirect methods of cognitive

load assessment. For instance, ocular metrics, including pupil

diameter and blink rate, have been shown to accurately estimate

cognitive load [2, 3, 4]. Additionally, facial temperature variations

have been widely correlated with cognitive workload, providing Figure 1: Experimental setup

another non-invasive means of assessment [5, 6]. Novak et al.

demonstrated that biometric indicators, such as galvanic skin Calibration data for the OCOsense glasses was then recorded by

response and skin temperature, can signal increased cognitive load; having participants replicate four facial expressions — smiling,

however, these measures are insufficient to distinguish between frowning, brow raising, and eye squeezing — three times each.

varying levels of cognitive load [7]. Wang et al. demonstrated that Calibration for the eye tracker followed, during which participants

visual cues—including face pose, eye gaze, eye blinking, and yawn tracked a moving dot with their eyes. This calibration aimed to

frequency—can serve as indicators of cognitive load [8]. optimize participant's seating position for accurate eye-tracking.

This research aims to address the complexities of cognitive load The experiment's main phase involved participants completing

estimation by integrating a wide range of psychophysiological cognitive load-inducing tasks that tested three aspects of cognition:

signals, offering a more comprehensive approach to this task. attention, memory, and visual perception. For each cognitive

domain, two widely recognized tasks were presented, each with



cognitive load-inducing tasks. The study was conducted in a quiet, that were not expected to induce cognitive load, such as meditation with open eyes, listening to music to relieve stress and passive temperature-controlled room, with participants tested individually. viewing of aesthetically pleasing images. These tasks provided At the beginning of each session, participants were seated in a baseline data for periods of minimal cognitive load. comfortable chair in front of a 24” monitor and given instructions In summary, each recording session included six cognitive load- about the experiment and their expected role. The Empatica E4 inducing tasks (with two levels of difficulty) and three relaxation wristband was then fitted to the participant's non-dominant hand, The objective of our data collection was to capture participants' the differentiation of cognitive load levels. Following each category of cognitive tasks, participants engaged in relaxation tasks cognitive load under varying levels of difficulty imposed by 3 Experimental Setup two levels of difficulty (easy and difficult). This design allowed for

tasks, totaling 15 tasks. After each task, participants completed the

and the OCOsense glasses for emotion recognition were equipped

NASA Task Load Index (NASA TLX) questionnaire, a validated

in line with product instructions.

instrument for assessing cognitive load across six dimensions:

Data collection was further enriched through the use of

mental demand, physical demand, temporal demand, performance,



D455 (providing depth data at 30 fps), a Logitech BRIO stream 4k to 100. In this study, the unweighted version of the NASA TLX, known as the Raw NASA TLX, was used. Additionally, webcam at 10 fps with HDR and noise-canceling microphones and participants completed a single-item Instantaneous Self- a FLIR Lepton 3 thermal camera delivering a full 160x120 pixel tracker (60 frames per second), an Intel RealSense Depth Camera effort, and frustration [9]. Each question was rated on a scale of 0 unobtrusive sensing technologies, including a Tobii Spark eye



thermal resolution with 8 fps. We used this set of devices to Assessment (ISA) of workload, providing a subjective measure of the cognitive load induced by the task [10]. These questionnaires continuously monitor participants throughout the recording served as subjective assessments of cognitive load and as reference session. The experimental setup can be observed in Figure 1. points for the difficulty of each task [11].



4 Data Collection Protocol The tasks were implemented using PsychoPy, an open-source

software package commonly used in neuroscience and

Prior to the experiment, participants completed a brief sleep experimental psychology research [12]. For attention-related tasks,

questionnaire to gather information about their sleep patterns (e.g., participants completed the N-back and Stroop tests. In the N-back

hours slept the night before and usual sleep duration) and rated their task, participants were presented with a sequence of letters and

levels of fatigue and focus on a scale of 1 to 10. asked to determine whether the current letter matched the one



36

Multi-modal Data Collection and Preliminary Statistical Analysis Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

for Cognitive Load Assessment



presented N trials earlier (with task difficulty increasing as N for the N-back tasks. Notably, the N-back tasks were always

increased) [13]. Participants completed both a 2-back and a 3-back presented first to participants, suggesting that they may have

task. In the Stroop test, participants identified whether the word required additional time to adjust to the testing environment and

matched the color in which it was written, with the easier version fully engage with the task.

involving two colors (red and blue) and the more difficult version Next, an inferential statistical analysis was performed on the

incorporating five colors [14]. relationship between task scores and various variables of the sleep

Memory-related tasks included a memory game and a question- pattern. To investigate the potential influence of tiredness on

answering task based on a previously shown image. In the memory performance, responses from the sleep patterns questionnaire were

game, participants recalled as many words as possible from a set, analyzed. A non-parametric Kruskal-Wallis test was performed to

with the easier version comprising seven words and the more determine whether there was a statistically significant difference in

difficult version consisting of 15 words. In the question-answering overall scores across different levels of tiredness (low, medium,

task, participants focused on an image and then answered questions and high). The resulting p-value (0.91) indicated no significant

about it (e.g., remembering the number of particular objects in the difference in performance between these groups. Thus, tiredness

image), with the hard version using an image with greater detail. levels did not show a statistically significant impact on

The visual perception tasks included a "spot the difference" task performance within a 95 % confidence interval.

and a pursuit test. In the "spot the difference" task, participants were Similarly, the effect of focus level (low vs. high) on overall

presented with two images and were asked to identify as many performance was examined using a non-parametric Mann-Whitney

differences as possible within a one-minute time frame. The test. The p-value was 0.12, indicating no statistically significant

difficulty of this task varied, with the more challenging version difference in performance between low and high focus groups at

involving an image that contained greater detail compared to the the 5 % significance level.

simpler, easier version. The pursuit test required participants to Furthermore, the relationship between hours of sleep the night

visually track irregularly curved, overlapping lines. As with the before the experiment and participant performance was examined

"spot the difference" task, the pursuit test was administered at two using Spearman’s correlation. The p-value was 0.42, indicating no

levels of difficulty. The more difficult version featured a more statistically significant correlation between overall performance

intricate image, with longer and more tangled lines, as opposed to scores and hours of sleep the night before the experiment.

the less complex image used in the easier version of the task. The potential influence of participants' highest education level

on overall performance was also investigated. To assess this, a non-

5 Statistics parametric Kruskal-Wallis test was conducted. The results (p-value

In this section, we present some descriptive demographic and task- of 0.33) indicated no statistically significant difference in

related statistics for the participants involved in the experiment. performance scores across different education levels among the

The average age of participants was 29.28 years, with a standard participants.

deviation of 8.31. In terms of educational background, the majority Overall, the small sample size may have constrained the ability

of participants (44 %) had obtained a Bachelor's degree (BSc), to detect significant effects. The limited variability in the sample's

followed by those with a Master's degree (MSc), 28 %. A smaller educational background and other factors likely contributed to the

portion had completed only high school (16 %) or had earned a PhD lack of observed differences, emphasizing the need for a larger,

(12 %). more diverse sample to better understand the impact of these Additionally, 60 % of the participants were male.

We then looked at the descriptive statistics derived from the variables on cognitive load performance.

performance of the participants in each task. These indicate that

participants performed consistently well on tasks such as the 2-back



task, both easy and difficult versions of the Stroop test, the easy

memory task (where participants recalled an average of 5 out of 7

words), the easy version of the "spot the difference" task (with an

average detection rate of approximately 90 % of all the

differences), and both versions of the pursuit test. Notably,

participants performed slightly better on the difficult version of the

Stroop test, likely due to their increased familiarity with the task.

However, performance was lower on tasks such as the 3-back

test (which most participants perceived as highly or extremely

difficult), the difficult memory task (with an average recall rate of

39 %), and both the easy and difficult question-answering tasks.

The difficult version of the "spot the difference" task also showed

lower performance, with participants detecting only 25 % of the

differences on average. Consistent performance among subjects Figure 2: Reported perceived difficulty per cognitive task (with low standard deviation) was observed across all tasks except



37



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Ana Krstevska et al.



As shown in Figure 2, participants consistently perceived the background, and other demographic factors is desirable to enhance

difficulty of the two N-back tasks and the difficult version of the the generalizability of the findings.

"spot the difference" task as somewhat high or high. This suggests In future work, the collected data will be processed and utilized

a general consensus regarding the difficulty of these tasks. In to train machine learning models aimed at estimating cognitive

contrast, the NASA TLX-based perceived difficulty of remaining load. Ground truth for the machine learning models can be derived

tasks, exhibited significant variability among participants. from various sources, including perceived task difficulty reported

To assess differences in performance across task difficulties and through the standardized questionnaires, the designed difficulty

evaluate the potential for differentiating cognitive load using level of the tasks or the participants' performance on the tasks.

machine learning models, we conducted additional inferential These machine learning models will leverage sophisticated ML

statistical analyses. The Wilcoxon signed-rank test was used to techniques to effectively integrate and analyze multi-modal data,

compare participant performance on the easier and more difficult aiming to enhance the accuracy of cognitive load predictions. We

versions of each cognitive task. also plan to further expand the dataset with another phase of data

Statistically significant differences in performance were found collection, offering a rich dataset both in terms of modalities, as

between the two difficulty levels for all tasks. For the N-back, "spot well as in terms of participants. The collected dataset will serve as

the difference", and pursuit tasks, participants performed a stepping stone towards robust multi-modal cognitive load

significantly better on the easier versions, indicating that increased assessment, allowing for creation and benchmarking of ML models

difficulty negatively impacted performance. Conversely, for the and will be made available to general public after the collection is

Stroop, memory, and question-answering tasks, participants finalized.



performed better on the more difficult versions. Acknowledgements

The statistical analysis conducted in this study provides initial

evidence supporting the validity of the data collection protocol, This work was supported by the Jožef Stefan Institute and

particularly with respect to the selection of tasks and task difficulty Università della Svizzera italiana (funded by SNSF through the

levels. The tasks chosen for this experiment varied significantly in project XAI-PAC (PZ00P2_216405)).



differences in performance between the easier and more difficult [1] Jonathan Mead, Mark Middendorf, Christina Gruenwald, Chelsey Credlebaugh, versions of each task. These results indicate that cognitive load and terms of their cognitive demands, as reflected by the substantial References

and Scott Galster. 2017. Investigating Facial Electromyography as an Indicator of

performance are task-specific, and the significant differences Cognitive Workload. In 19th International Symposium on Aviation Psychology, 377–

observed support the feasibility of using machine learning models 382.

[2] Muneeb Imtiaz Ahmad, Ingo Keller, David A. Robb, and Katrin S. Lohan. 2020.

to differentiate between varying levels of cognitive load. A framework to estimate cognitive load using physiological data. Personal and

Ubiquitous Computing, 27, 2027–2041.



6 Conclusion and Future Work [3] Tobias Appel, Christian Scharinger, Peter Gerjets, and Enkelejda Kasneci. 2018. Cross-subject workload classification using pupil-related measures. In Proceedings of

This study employs a novel approach to data collection for the 2018 ACM Symposium on Eye Tracking Research & Applications, 4, 1–8.

[4] Tobias Appel, Natalia Sevcenko, Franz Wortha, Katerina Tsarava, Korbinian

cognitive load inference by combining psychophysiological data Moeller, Manuel Ninaus, Enkelejda Kasneci, and Peter Gerjets. 2019. Predicting



obtained from multi-modal sensory setup, including wearable and Cognitive Load in an Emergency Simulation Based on Behavioral and Physiological Measures. In Proceedings of the 21st ACM International Conference on Multimodal unobtrusive contact-free sensors. The decision to utilize a diverse Interaction (ICMI) , 154–163.

set of devices was motivated by the hypothesis that integrating data [5] Fangqing Zhengren, George Chernyshov, Dingding Zheng, and Kai Kunze. 2019.

from multiple sources could provide a more accurate assessment of Cognitive load assessment from facial temperature using smart eyewear. In

Proceedings of the 2019 ACM International Joint Conference on Pervasive and

cognitive load, while also aiming to identify the minimal sensor Ubiquitous Computing and Proceedings of the 2019 ACM International Symposium



configuration required to achieve reliable results. This is on Wearable Computers, 657–660. [6] Yomna Abdelrahman, Eduardo Velloso, Tillman Dingler, Albrecht Schmidt, and particularly relevant in dynamic and high-stakes environments, Frank Vetere. 2017. Cognitive Heat: Exploring the Usage of Thermal Imaging to

such as driving, where accurate cognitive load assessment could Unobtrusively Estimate Cognitive Load. Proceedings of the ACM on Interactive,

have life-saving implications. To the best of our knowledge, no Mobile, Wearable and Ubiquitous Technologies, 33, 1–20.

[7] Klemen Novak, Kristina Stojmenova, Grega Jakus, and Jaka Sodnik. 2017.

prior research has incorporated such a comprehensive and Assessment of cognitive load through biometric monitoring. In 7th International

multifaceted setup for cognitive load evaluation. Conference on Information Society and Technology (ICIST).

[8] Zixuan Wang, Jinyun Yan, and Hamid Aghajan. 2012. A framework of personal

The statistical analyses conducted thus far offer promising assistant for computer users by analyzing video stream. In Proceedings of the 4th

validation for the data collection protocol. The selection of tasks Workshop on Eye Gaze in Intelligent Human Machine Interaction, 1–3.

and task difficulty levels proved effective in eliciting a range of [9] Sandra G. Hart, and Lowell E. Staveland. 1988. Development of NASA-TLX (Task

Load Index): Results of Empirical and Theoretical Research. In Advances in

cognitive load levels, as evidenced by the significant performance Psychology, 52, 139-183

differences between task difficulties. [10] Andrew J. Tattersall, and Penelope S. Foord. 2007. An experimental evaluation

of instantaneous self-assessment as a measure of workload. Ergonomics, 39, 740-748.

To further enhance the validity of the data collection protocol, [11] Thomas Kosch, Jakob Karolus, Johannes Zagermann, Harald Reiterer, Albrecht



several changes could be implemented in potential subsequent Schmidt, and Paweł W. Woźniak. 2023. A Survey on Measuring Cognitive Workload in Human-Computer Interaction. ACM Computing Surveys , 55, 1–39. collections. Refining task difficulty levels could offer more [12] Jonathan Peirce, Rebecca Hirst, and Michael MacAskill. 2022. Building

granularity in cognitive load differentiation, ensuring a clearer Experiments in PsychoPy. Sage Publications

distinction between varying levels of cognitive load. Furthermore, [13] Michael J. Kane, and Andrew Conway. 2016. The invention of n-back: An

extremely brief history. The Winnower

increasing the diversity of participants in terms of age, educational [14] John Ridley Stroop. 1992. Studies of interference in serial verbal reactions.

Journal of Experimental Psychology, 121, 15-23



38





Predicting Health-Related Absenteeism with Machine Learning:




A Case Study



Aleksander Piciga Matjaž Kukar

ap7377@student.uni- lj.si matjaz.kukar@f ri.uni- lj.si

Faculty of Computer and Information Science, Faculty of Computer and Information Science,

University of Ljubljana University of Ljubljana

Ljubljana, Slovenia Ljubljana, Slovenia



Abstract 0.10

Health-related absenteeism, or sick leave, is a complex issue with

significant financial and operational implications for businesses. 0.08

We present a machine learning approach to predict employee

0.06

absenteeism in a Slovenian company. The study involved pre-

teeism

processing and augmenting the dataset by incorporating domain

0.04

knowledge, and evaluating various machine learning models. Absen

Gradient Boosted Regression Trees emerged as the most effective 0.02

model, significantly outperforming the baseline model which

merely predicted the previous year’s absenteeism rate. Key at- 0.00

2014 2015 2016 2017 2018 2019 2020 2021 2022

tributes influencing absenteeism were identified, notably includ- Year

ing current absenteeism, performance evaluations, and various



job type and location-related features. Results highlight the po- Figure 1: The increase in absenteeism rate in Slovenia be-

tential of machine learning in proactively managing absenteeism tween 2014 and 2022 [8]. We can observe a steady increase

and offer recommendations for future research, such as modeling throughout the years. absenteeism as a time series and incorporating additional data

sources. We also show that the current data is not detailed and

granular enough to further improve the results. and augmenting the company’s employee data by incorporating

domain knowledge, and evaluating various machine learning

Keywords models. The findings highlight key attributes influencing ab-

absenteeism, data analysis, data augmentation, machine learning senteeism and offer recommendations for future research and

interventions.

1 Introduction The significance of our work extends beyond Company X,

offering a blueprint for organizations tackling absenteeism. By

Absenteeism — temporary absence from work due to health

showcasing machine learning’s efficacy in predicting absenteeism

reasons — is awidespread issue. In Slovenia, it has been on the rise

and revealing its drivers, we contribute to the broader field and

since 2014 (Figure 1), with an average of 56,128 individuals absent

pave the way for data-driven interventions promoting a healthier,

daily in 2022, representing approximately 5.91% of the workforce

more productive workforce. This aligns with the growing trend

[8]. This carries substantial financial burdens: direct costs like sick

of using AI and ML to address complex organizational challenges.

pay and indirect costs from overstaffing, reduced productivity

Insights from such analyses can aid in strategic workforce plan-

and service quality [2]. The complexity of absenteeism, rooted

ning, optimize resource allocation, and ultimately contribute to

in personal and organizational factors, makes it challenging to

a more sustainable and resilient organization.

predict and manage effectively [10].

In section 2 we detail the data and preprocessing, section 3

Recent years have witnessed a growing interest in leverag-

outlines the methodology, section 4 presents the results, and

ing artificial intelligence (AI) and machine learning (ML) to ad-

section 5 discusses the findings and concludes the study.

dress the absenteeism challenge [5]. Various machine learning

techniques, including neural networks, decision trees, random

forests, and gradient boosting, have been employed to predict ab- 2 Materials

senteeism and identify its underlying causes [3, 9]. These studies The data used in our work spanned six years, from 2017 to 2022,

have demonstrated the potential of machine learning in providing and initially comprised 13,798 instances (aggregated employee

valuable insights for proactive absenteeism management. records) with up to 49 attributes each. They include demographic

This paper presents a case study conducted in collaboration details, work-related factors, performance evaluations and the

1

with a Slovenian IT company aiming to improve absenteeism current year’s absenteeism rate for each employee, but no partic-

prediction and management. The study includes preprocessing ulars about sick leave and other personal data.

1

The initial dataset required substantial preprocessing to pre-

The company asked to remain anonymous, so it is referred to as Company X.

pare it for analysis and machine learning [6]. The data cleaning

Permission to make digital or hard copies of all or part of this work for personal

process involved addressing inconsistencies in attribute values,

or classroom use is granted without fee provided that copies are not made or

such as removing extraneous spaces and converting text to low-

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this ercase for uniformity. A significant challenge in the dataset was

work must be honored. For all other uses, contact the owner /author(s).

the presence of missing values, denoted by ’/’. Their meaning and

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

handling were discussed with a company representative to de-

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.scai.7260 termine their origins and ensure appropriate treatment. In some





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Piciga et al.




cases, missing values were imputed based on the average values discern statistically significant differences across groups defined

of similar instances. For example, missing values in ’Kilometers by categorical attributes.

to work’ were attributed to errors in data entry and were imputed

using the average value for employees living in the same location 3.2 Data augmentation/Feature Engineering and working at the same place. On the other hand, missing values

The original dataset underwent a series of transformations to

in performance evaluations were due to employee’s absence on

enhance its suitability for machine learning. This included data

evaluation days.

cleaning, handling missing values, and the creation of new at-

The target variable — health-related absenteeism rate in the

tributes based on domain knowledge and insights from the EDA.

following year — is a continuous variable ranging from 0 to 1. It

New attributes were engineered based on domain knowledge

signifies the proportion of workdays an employee is absent due

and statistical analysis. These included indicators for elevated

to health reasons compared to the total number of workdays. The

absenteeism, receipt of bonuses or awards, high and low perfor-

distribution of this target variable is heavily skewed to the right,

mance evaluations, and absenteeism rates within the employee’s

with most values clustered near zero, indicating that the majority

team and job type. External factors, such as average absenteeism

of employees have low absenteeism rates. However, there exist

rates in the employee’s residential and work locations, were also

some outliers with extremely high absenteeism rates (Figure 2).

incorporated. The feature engineering process was iterative, in-

volving close collaboration with domain experts to ensure the

0.0125 derived attributes were meaningful and captured relevant aspects 0.2750

Median of employee behavior and organizational dynamics.

95th percentile



yees 10 3 3.3 Machine Learning Models

emplo Several well-known machine learning models were employed

of 2 10 for absenteeism prediction, including Decision Trees, Linear Re-

er

b

gression with L1 regularization, K-Nearest Neighbors (KNN),

Num 1 10 Support Vector Regression (SVR), Gradient Boosted Regression

Trees (GBRT), and Random Forest. Hyperparameter optimization



0.0 was conducted by using Optuna toolkit [1] to optimize model 0.2 0.4 0.6 0.8 1.0

Absenteeism in the following year performance.



Figure 2: Log-distribution of the target variable. Most work- 3.4 Model Evaluation and Selection

ers have very little absence, causing a right-tailed distribu- Model evaluation was performed using Mean Absolute Error

tion with an “outlier” spike on the right. (MAE), Root Mean Squared Error (RMSE), and coefficient of de-

2

termination (R ). The models were trained on past years’ data

The skewed distribution of the target variable has implica- and tested on the subsequent year, with the training set size in-

tions for the statistical analysis and machine learning modeling. creasing each year. The MAE provided an intuitive measure of

Therefore, non-parametric statistical tests, such as the Spear- the average prediction error, while the RMSE penalized larger er-

man’s rank correlation and Kruskal-Wallis test, were employed rors more severely. The R2 quantified the proportion of variance

in EDA and data preprocessing. Additionally, the presence of in the target variable explained by the model. The models were

outliers necessitates careful consideration during model building also compared against a baseline model that simply predicted

and evaluation. the previous year’s absenteeism, to gauge the added value of

The final dataset, comprising 10,347 instances and 42 attributes, the machine learning approach. A baseline model predicting the

serves as the foundation for the subsequent machine learning, previous year’s absenteeism rate was used for comparison.

where various models are trained to predict absenteeism rates.

3.5 Model Interpretation

3 Methods SHAP (SHapley Additive exPlanations) values [4, 7] were cal-

The research methodology encompassed a multi-faceted approach, culated to interpret model predictions and assess attribute im-

integrating exploratory data analysis, feature engineering, and portance. SHAP values provide insights into the contribution of

the application of diverse machine learning models. The ultimate each attribute to the model’s output, aiding in understanding

goal was to establish a robust predictive framework for health- the factors driving absenteeism. SHAP values provide a unified

related absenteeism, while also ensuring model interpretability framework for interpreting any machine learning model, quanti-

to observe actionable insights. fying the contribution of each feature to the model’s prediction

for a given instance. By analyzing the SHAP values, it was possi-

3.1 Exploratory Data Analysis (EDA) ble to identify the most influential attributes and their directional

impact on absenteeism.

The initial phase involved a thorough EDA to understand the

underlying data distribution, identify potential outliers, and un-

cover preliminary relationships between attributes and the target 3.6 Data Splitting

variable (absenteeism in the following year). Given the skewed To ensure robust model evaluation and mitigate the risk of over-

nature of the target variable, visualizations like histograms and fitting, the dataset was split into training and testing sets in a

box plots were complemented by non-parametric statistical tests. prequential manner (year after year). The models were trained

The Spearman’s rank correlation coefficient was employed to as- on the training set and their performance was assessed on the

sess monotonic relationships between continuous attributes and unseen testing set for the subsequent year. This comprehensive

the target variable, while the Kruskal-Wallis test was utilized to methodological framework enabled a systematic exploration of





Predicting Health-Related Absenteeism Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia




the factors influencing health-related absenteeism and the devel- the latest year’s data. These results were statistically significantly

opment of a predictive model to proactively manage this critical better than the baseline model, demonstrating the effectiveness of

issue. GBRT in capturing the complex patterns underlying absenteeism.

Figure 4 reveals a general trend of MAE improvement for

4 Results most models in later years, surpassing the baseline model in

The primary objective of our work was to develop machine learn- the final year. This suggests that the models benefit from the

increasing amount of training data available in later years. RMSE

ing models capable of predicting health-related absenteeism in

2

and R charts (not shown) exhibit almost identical properties.

the subsequent year. The models were evaluated using three key

It is clear that ML models profit tremendously from increasing

metrics: Mean Absolute Error (MAE), Root Mean Squared Error

2 amounts of data, as can be expected.

(RMSE), and the coefficient of determination (R ). The baseline

Given the observed performance gains in later years with

model, which simply predicted the previous year’s absenteeism,

larger training sets, we explored the impact of incorporating data

served as a benchmark for comparison (Table 1).

from previous years. Figure 4 showcases the change in MAE for

Table 1: Model performance averaged year-over-year. the final year when models were trained on data from the past

year and the past three years, respectively.

Model RMSE MAE R2

Random Forest 0.107 0.052 0.344

0.050 Random Forest

GBRT 0.108 0.051 0.333 GBRT

Linear Regression

Linear Regression 0.049 Regression Decision Tree 0.108 0.051 0.331

Baseline Model

Regression Decision Tree 0.112 0.051 0.281 0.048

KNN 0.121 0.057 0.173

MAE 0.047

SVR 0.117 0.075 0.215

0.046

Baseline Model 0.121 0.051 0.156

0.045



As we can see, all machine learning models outperform the

Dataset Year Years

baseline model in terms of RMSE and R . This indicates their Three Past 2 Base With

Past

superior ability to explain the variance in the target variable With

Dataset

(absenteeism in the following year). While the MAE remains rel-

atively consistent across models, the improvement in RMSE and

2 Figure 4: Impact of additional attributes from past years

R suggests that the models are particularly effective in handling

larger deviations in absenteeism predictions. on MAE.

To establish the statistical significance of the model improve-

ments, we conducted a paired T-test comparing the predictions The GBRT model exhibited notable improvement with the

of each model against the baseline model. All the selected models inclusion of additional data, achieving an MAE of 0.044, RMSE

demonstrated statistically significant improvements (p < 0.05) of 0.093, and R2 of 0.36. This underscores the value of historical

2

in RMSE and R ; this ensures that their superior performance is data in enhancing the predictive capabilities of machine learning

statistically substantiated and not due to chance. models for absenteeism and suggests that including even more

historical data per employee would be beneficial.

4.1 Performance Trends and Impact of

Additional Data per Employee 4.2 Interpretability and Additional Insights

Analysis of SHAP values yielded the following key attributes

To gain deeper insights into model behavior, we examined their

performance trends over the years. Figure 3 illustrates the evolu- influencing absenteeism:

tion of MAE for each model. Current absenteeism rate •

• Performance evaluations

• With respect to the employee’s job type and location:

Random Forest

0.056 GBRT –

Linear Regression

Absenteeism rate

0.054 Baseline Model – Regression Decision Tree

Proportion of employees with elevated absenteeism

0.052 – Proportion of employees without bonuses

Our findings suggest that absenteeism is influenced by a combina-

MAE 0.050

tion of individual factors (current absenteeism, performance eval-

0.048 uations) and organizational factors ( job type, location, bonuses).

Additionally, a rather simple EDA visualisation of functional

0.046

grouping of employees was quite surprising (Figure 5). Its inter-

2018 2019 2020 2021 pretation can be quite speculative, possibly related to increased

Year

job satisfaction or engagement in certain groups. Another, some-

what surprising finding from EDA is that the COVID-19 pandemic

Figure 3: MAE trend over time with additional training did not significantly influence absenteeism rates in 2020, but it

data from past years. may have in 2021 (Figure 6).

Finally, t-SNE visualization of the full dataset shows that em-

Among the evaluated models, GBRT exhibited the best perfor- ployees cannot easily be separated in clusters with similar ab-

2

mance, achieving an MAE of 0.045, RMSE of 0.10, and R of 0.40 on senteeism (Figure 7). We can identify some distinct subgroups





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Piciga et al.




The findings align with existing literature highlighting the

0.20

multifactorial nature of absenteeism. The strong influence of

0.15 current absenteeism on future absenteeism emphasizes its pre-

dictive power, suggesting that past behavior can be a significant

variable 0.10

indicator of future trends. The negative correlation between per-

arget formance evaluations and absenteeism suggests that employees

T 0.05

with higher evaluations tend to be less absent, potentially due to

0.00 increased job satisfaction or engagement. The impact of denied

ort bonuses on absenteeism points to the potential role of financial

Commercial Tec incentives and recognition in influencing employee attendance. Supp hnology

Work field

The limitations of our work include the relatively short time

span and the potential influence of unmeasured external factors.

Figure 5: Target variable according to functional partition- Future research could address these limitations by: modeling

ing within the company. absenteeism as a time series to capture its dynamic nature, incor-

porating additional data sources such as employee surveys, par-

ticipation in wellness programs, and (within legal limits) health

0.150 and personal circumstances data analyzing absenteeism at a finer

0.125 granularity (e.g., monthly or daily), exploring the inclusion of

0.100 employee health records and workplace environmental factors in

variable predictive models, and conducting longitudinal studies to track

0.075

absenteeism patterns over extended periods.

arget 0.050

T While quantitative improvements of ML model predictions

0.025 are not overwhelming, the gained insights can enable targeted

0.000 interventions to reduce absenteeism and promote a healthier

2017 2018 2019 2020 2021 workforce. By leveraging ML and data-driven insights, organi-

Year zations can proactively manage absenteeism, thus improving

productivity, financial stability, and employee well-being.

Figure 6: Target variable by year. Note the sharp increase

in 2021, possibly attributable to the COVID-19 pandemic. Acknowledgements

The authors sincerely thank to Company X for providing the data,

domain expertise and several fruitful discussions.. The authors

(like the cluster of red dots on the left), however most data points

acknowledge the financial support from the Slovenian Research

are quite intermingled. This suggests that with our current set of

Agency (research core funding No. P2-209).

attributes, we shouldn’t anticipate a significant improvement in

predictive performance.

References

[1] T. Akiba, S. Sano, T. Yanase, T. Ohta, and M. Koyama. 2019. Optuna: a

next-generation hyperparameter optimization framework. In Proceedings

t 50 of the 25th ACM SIGKDD International Conference on Knowledge Discovery

40 . Association for Computing Machinery, 2623–2631. isbn: & Data Mining

onen 30 9781450362016. doi: 10.1145/3292500.3330701.

20

[2] M. Bregant, E. Boštjančič, J. Buzeti, M. Ceglar Ključevšek, A. Hiršl, M. Klun,



comp 10 T. Kozjek, N. Tomaževič, and J. Stare. 2012. Izboljševanje delovnega okolja z 0 inovativnimi rešitami . Združenje delodajalcev Slovenije. − 10 t-SNE [3] B. Hu. 2021. The application of machine learning in predicting absenteeism − 20 2021 2nd International Conference on Computing and Data Science at work. In − 30 (CDS) , 270–276. doi: 10.1109/CDS52072.2021.00054. − 40 Second [4] Y. Meng, N. Yang, Z. Qian, and G. Zhang. 2021. What makes an online − 50

− review more helpful: an interpretation framework using xgboost and shap 60

−50 −40 −30 −20 −10 0 10 20 30 40 50 60 values. Journal of Theoretical and Applied Electronic Commerce Research, 16,

First t-SNE component 3, 466–490. doi: 10.3390/jtaer16030029.

[5] I. H. Montano, G. Marques, S. G. Alonso, M. López Coronado, and I. de la

Torre Díez. 2020. Predicting absenteeism and temporary disability using



Figure 7: Data visualized in 2D space with t-SNE projection. machine learning: a systematic review and analysis. Journal of Medical Systems , 44, 9, (Aug. 2020), 162. doi: 10.1007/s10916- 020- 01626- 2. Red dots represent examples with absenteeism in the next [6] A. Piciga. 2024. Napovedovanje zdravstvenega absentizma s strojnim učenjem. year above 0.25. Blue shades depict examples with absen- Bachelor’s Thesis. Univerza v Ljubljani, Fakulteta za računalništvo in infor-

teeism between 0 (light blue) and 0.25 (dark blue). matiko. https://repozitorij.uni- lj.si/IzpisGradiva.php?lang=slv&id=160413.

[7] E. Štrumbelj and I. Kononenko. 2014. Explaining prediction models and in-

dividual predictions with feature contributions. Knowledge and Information

Systems, 41, 3, (Dec. 2014), 647–665. doi: 10.1007/s10115- 013- 0679- x.

5 Discussion and Conclusion [8] M. Zaletel, D. Vardič, and M. Hladnik. 2024. Zdravstveni statistični letopis Slovenije 2022. (2024). Retrieved June 5, 2024 from https://nijz.si/publikacije

Our work successfully demonstrates the application of machine /zdravstveni- statisticni- letopis- 2022/.

[9] W. Zaman, S. Zaidi, A. I. Abdullah, and B. Touhid. 2019. Predicting absen-

learning to predict health-related absenteeism. The GBRT model’s Proceedings of the 3rd Interna-

teeism at work using tree-based learners. In

superior performance highlights its ability to capture complex . Association for tional Conference on Machine Learning and Soft Computing

Computing Machinery, 7–11. doi: 10.1145/3310986.3310994.

data relationships, outperforming simpler models and the base-

[10] S. Zupanc. 2011. . Absentizem, kolegialnost in obremenjenost posameznikov

line. Also, identifying key attributes influencing absenteeism,

Bachelor’s Thesis. Univerza v Ljubljani. http://www.cek.ef .uni- lj.si/UPES/z

such as current absenteeism, denied bonuses, work type and lo- upanc1175.pdf .

cation, and performance evaluations, provides valuable insights.





Puzzle Generation for Ultimate Tic-Tac-Toe




Maj Zirkelbach Aleksander Sadikov

mz5153@student.uni- lj.si aleksander.sadikov@f ri.uni- lj.si

University of Ljubljana, Faculty of Computer and University of Ljubljana, Faculty of Computer and

Information Science Information Science

Ljubljana, Slovenia Ljubljana, Slovenia



Abstract our application, which is designed to enhance players’ tactical

and strategic thinking.

Ultimate Tic-Tac-Toe is an interesting and popular variant of

In Section 2 we present the related work, and in Section 3 we

Tic-Tac-Toe that lacks available resources for improving game-

detail the technical aspects of the developed application. In Sec-

play skills. In this paper, we present a semi-automatic system for

tion 4 we present the implemented agents and their approximate

generating puzzles as a part of a larger tutorial application aimed

strength. In Section 5 we provide a description of different types

at teaching Ultimate Tic-Tac-Toe. The puzzles are designed to en-

of puzzles and the methodology for their construction. In Section

hance players’ tactical and strategic understanding by presenting

6 we present the evaluation and discuss the results in Section 7.

game scenarios where they must identify correct continuations.

Finally, in Section 8 we present the conclusions and give possible

To ensure the quality of generated puzzles we tested the appli-

extensions and enhancements for future work.

cation with a group of volunteers. The results have shown that

the number of solved puzzles positively impacted users’ ability

to reach higher strength levels but had less of an effect on lower



levels. 2 Related Work



Keywords There are many implementations of the Ultimate Tic-Tac-Toe

available online, mostly appearing as mobile games aimed pri-

Ultimate Tic-Tac-Toe, puzzle generation, minimax algorithm, tu-

marily at entertainment and lacking advanced playing agents

tor application

[12] [9] [10], as well as web and desktop applications developed

to create the strongest possible programs [15] [7] [13]. An exam-

1 Introduction ple of the latter is an agent based on the ideas of the AlphaZero

For centuries, people have enjoyed playing board games like program [14], currently considered one of the strongest players of

chess and Go. Over time, these games have led to the develop- this game [13]. During the development of this agent, significant

ment of extensive theory and the accumulation of knowledge, strategies were discovered, which were also useful in developing

helping players navigate their complexity. Today, advanced arti- our application. Some researchers have attempted to solve the

ficial intelligence (AI) programs such as AlphaZero [14] surpass game theoretically, but the spatial complexity proved too great

even the strongest human players, offering new insights into to allow for a complete solution [5].

strategies. However, many lesser-known games have yet to be It is important to differentiate between the various versions of

thoroughly explored, despite their popularity. One such game is Ultimate Tic-Tac-Toe. One variant allows the game to continue

Ultimate Tic-Tac-Toe, an advanced version of the classic Tic-Tac- playing on already-won local boards, which drastically changes

Toe. This game is played on a 3x3 grid of local Tic-Tac-Toe boards, the game’s dynamics. In this variant, researchers have demon-

creating a global board (Figure 1a). The goal is to win three local strated an optimal strategy for the starting player, who can win

boards in a row, while players must make their moves within in 43 moves [1]. Further research has focused on enabling a more

specific local boards determined by their opponent’s previous balanced game by introducing random opening moves, which

move. For example, if a player moves in the top-left corner of a reduces the predictability of forced wins [4]. Despite these inter-

local board, the next player must play on the top-left local board. esting findings, research on these variants is not so relevant for

If the designated board is full or decided, the player can choose us, as it does not contribute to the understanding of the main

any other available board. Despite its apparent simplicity, the game.

game has enough spatial complexity that it cannot currently be While there is a lack of educational material specific to our

solved using brute-force methods. game, much can be learned from related fields, such as chess,

While there are several online implementations of the game, which has been extensively researched. The paper by Gobet and

most focus on building strong AI agents; however, There is a Jansen [8] describes a scientific approach to learning chess, which

noticeable lack of resources aimed at teaching and helping players includes methods to improve memory, perception, and problem-

understand the deeper strategies of the game, which could make solving skills in players. In this context, it focuses on the acquisi-

the learning curve more manageable for new and aspiring players. tion and organization of knowledge, including both explicit and

Thus, we have created an application that addresses the lack of implicit learning of tactics and strategies. This approach facil-

learning tools available for Ultimate Tic-Tac-Toe. This article itates a deep understanding of games and the development of

places particular emphasis on the puzzle generation aspect of more effective learning methods.

Chess also offers highly sophisticated practical solutions from

Permission to make digital or hard copies of all or part of this work for personal chess.com

which we can learn a great deal. Platforms such as [2]

or classroom use is granted without fee provided that copies are not made or

and [11] offer extensive resources and tools for lichess.org

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this learning chess, especially in the areas of tactics and openings.

work must be honored. For all other uses, contact the owner /author(s).

These platforms allow players to learn through interactive lessons,

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

solving puzzles, and studying various openings, which contribute

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.scai.7299 to a deeper understanding of the game and improve playing skills.



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Zirkelbach et al.



This approach has proven extremely effective in helping players The chess rating system is used to measure the playing strength

master complex strategic and tactical concepts in chess. of chess players. The most commonly used system is the Elo rat-

On the mentioned platforms, the methods for learning tac- ing [6], which predicts the likelihood of one player winning

tics are designed to allow players to solve problems based on against another based on their ratings:

concrete game situations, which improves pattern recognition

1

and decision-making abilities in real games. Similarly, learning 𝐸𝐴 , − =

𝑅 𝑅

𝐵 𝐴

openings involves demonstrating optimal opening moves and +

1 10 400

their continuations, helping players develop effective strategies

where 𝐸 represents the expected score for player A, 𝑅 is the

𝐴 𝐴

at the beginning of the game. rating of player A, and 𝑅 is the rating of player B. 𝐵

We have applied similar methods in our Ultimate Tic-TacToe



application. For example, adapting approaches for learning tactics Table 1: Table of approximate agent strengths. Each agent

can help users improve their recognition and solving of complex played 100 games (50 as X and 50 as O) against the agent

situations in the game while learning openings helps to under- one level lower. The results column shows the number

stand key opening moves and their impact on the further course of points each agent earned with each symbol, as well

of the game. By incorporating these methods into our application, as the total score. A win awarded 1 point, while a draw

we ensured more effective learning processes and improved the awarded 0.5 points. The last line shows the results of the

overall gaming experience. strongest freely available agent against level 9. It had the

same amount of time to think, and they played 30 games.



3 Application Details Result Agent Estimated Rating

In addition to puzzle-solving, the app offers a comprehensive X O Combined

learning experience through various other features. It includes Confused Chimp - 1 - - - 1

AI opponents of different difficulty levels, game analysis, and Goofy Goblin - 2 49 49 98 620

exploration of effective opening strategies, allowing players to Casual Carl - 3 41.5 35.5 77 835

refine their understanding in all phases of the game. The user in- Average Joe - 4 37 25 62 926

terface ensures smooth navigation between these modes, making Hustling Hugo - 5 39.5 34.5 74 1114

the app a versatile tool for both playing and learning Ultimate Witty Walter - 6 43 30 73 1293 Tic-Tac-Toe. By integrating these elements, the app serves as a

Thinking Tiffany - 7 35 24 59 1361

resource for players at all levels, helping them to deepen their

Brainy Bob - 8 42,.5 26.5 69 1506

understanding and improve their skills.

Bossman - 9 36.5 22.5 59 1574

To reach a broader audience, the application was developed

UT T T AI 14.5 12.5 27 1948

for both Android and Windows, the dominant operating systems

in the market [15]. It uses Flutter components to deliver a respon-

sive and user-friendly interface. Local data storage is utilized

for user settings, progress, and puzzle data, ensuring efficient 5 Puzzle Description and Methodology

performance and data management. In this section, we describe different types of puzzles and the

We employed modern technologies and mobile development methodology employed to generate them for our game.

practices, including state management patterns, to create an eas-

ily expandable app for future updates and enhancements. The 5.1 Puzzles entire project is hosted on GitHub, though it is not open-source.

The puzzles in the application are divided into tactical and strate-

Test versions of the app for Android and Windows are avail-

gic, with each type of puzzle covering different aspects of the

able on Google Drive: https://drive.google.com/drive/f olders/1Sn

game and helping players improve specific skills.

O_mN_ZVa2wXd0OGI07kLiYKQTDHuEe?usp=drive_link, while

Tactical puzzles are useful for understanding tactical ideas

the Android production version is accessible on Google Play Store:

and are particularly applicable in the endgame and middlegame

https://play.google.com/store/apps/details?id=com.uttt_tutor.

phases. They focus on specific situations that require precise

and thoughtful moves, helping players develop the ability to

think quickly and effectively. In total, we generated 1,263 tactical

4 AI Agents and Rating System puzzles, distributed across five levels. The number of puzzles for

each level is shown in Table 2.

Playing against intelligent agents allows users to refine their

Unlike tactical puzzles, strategic puzzles aim to understand

skills by competing against various virtual opponents. The appli-

the position and long-term plans. They are instrumental in the

cation includes nine different agents, each varying in difficulty

opening and middlegame, where it is crucial to recognize strategic

and gameplay strategies. These agents are designed using Mini-

ideas and develop plans that provide an advantage as the game

max and Monte Carlo Tree Search [3] algorithms, which provide

continues. There are 50 strategic puzzles available, currently

different levels of complexity and depth in move analysis. The

arranged in one level, with the possibility of expansion in the

agents and their approximate strengths are shown in Table 1.

future.

To better understand the quality of the agents and evaluate

user progress, we need to establish a system for measuring their

strength. Since Ultimate Tic-Tac-Toe is not widely popular, there 5.2 Tactical Puzzle Generation

is no established system for rating player abilities. Therefore, we To generate tactical puzzles, we developed a specialized minimax

decided to use the chess rating system as an approximation for agent that builds a tree of all possible moves leading to victory

our agents. from the solver’s perspective. A key step in this process is the





Puzzle Generation for Ultimate Tic-Tac-Toe Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia




Table 2: Number of tactical puzzles on each level. strategic understanding of the position, which our agents, using

relatively simple heuristics, are incapable of. Therefore, we re-

sorted to the most powerful freely available agent [13], which is

Level Puzzle Depth Quantity

based on the ideas of the AlphaZero program.

1 1 273

Thus, we generated the strategic puzzles manually. We searched

2 3 493

for interesting and instructive positions that arose in games be-

3 5 231

tween the aforementioned agent and our stronger programs. We

4 7 176

focused on moments when there was a significant deviation in

5 9 90

the position evaluation between the two agents. When the agent

with better strategic understanding detected an important change,

we saved the given position, studied it more closely, and based

selection of tree branches to retain only relevant and correct so-

on our understanding of the game, formulated a solution. The

lutions. It is essential to preserve all of the winner’s possibilities

most common examples of such situations involved sacrificing

while limiting the loser’s responses to those that make finding a

the edge board to gain control over the central board. A basic

solution as difficult as possible. Therefore, we select the continu-

example of this can be seen in Figure 2.

ation that allows the longest possible game for the loser while

leading to the fewest continuations for the winner.

From the tree, we extract all the correct solutions for the given

position. For a high-quality puzzle, it must not have too many

solutions. The criterion we set is that the number of solutions

must be less than the depth of the puzzle. We also decided to

discard all puzzles that have multiple correct continuations for

the first move. This way, we avoid trivial puzzles that would

be too simple. An example of a level 3 tactical puzzle with its

generated solution tree is shown in Figure 1.





(a) User interface of the most powerful freely available agent. For



the given position, it ran 1000 simulations and assessed the move



F2 as the best with an 82% probability. It evaluates the position

with a value of +16.85, which means it assigns approximately 58.4%

win probability to player X (a value of 0 means a draw, 100 a win,

and -100 a loss).





(a) Level 3 tactical puzzle. (b) Solution tree.



Figure 1: An example of tactical puzzle and its generated

solution tree.



The generation of tactical puzzles for different difficulty levels (b) Minimax agent with a search depth of 12. It marks the move

was automated by conducting matches between agents of equal F2 as the worst, as it does not recognize the long-term advantage.

strength, with the search depth of both agents corresponding



to the depth of the puzzle we wanted to find. We chose this Figure 2: Different interpretations of the same position,

approach to ensure that the resulting positions were interesting based on which we built the strategic puzzle. and balanced, as otherwise, the stronger side would usually have

an overly obvious advantage at the start of the puzzle which

would make it boring to solve.

6 Evaluation and Results

5.3 Strategic Puzzle Generation We conducted a quality analysis of the application with 14 vol-

Automating the creation of strategic puzzles is impossible without unteers. Their task was to use the app for an extended period

a program that could interpret the given position and simultane- to improve their knowledge of the game. We were interested

ously provide a human-understandable explanation. Additionally, in determining whether using the app had a positive impact on

generating strategic puzzles requires an agent with an advanced the development of their Ultimate Tic-Tac-Toe playing skills and





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Zirkelbach et al.




whether progress was dependent on motivation or the time spent understanding, making progress slower and more challenging.

learning. This reflects the diminishing returns on improvement as you

To assess individual progress, participants played against the climb the rating ladder.

agent at the start of testing to determine their initial skill level. It must also be mentioned that users were free to use any

The application then tracked the highest level each user defeated, tools within the app during testing and solving more puzzles did

providing an estimate of their improvement over time. This not correlate with longer app usage. For a clearer assessment of

progress, in relation to the number of puzzles solved, is illus- puzzle significance, a controlled test focusing solely on puzzle-

trated in Figure 3. For a more concrete interpretation of obtained solving would be more appropriate.

level strengths, refer to Section 4.

8 Conclusion

500 In this work, we presented methods for generating puzzles for

the game of Ultimate Tic-Tac-Toe. To evaluate the quality of

400 these puzzles, we tracked how the number of solved puzzles im-

pacted individual user progress. Our results indicate a correlation

300 between the number of puzzles solved and the ability to reach

stronger AI levels.

However, the evaluation could be refined by focusing exclu-

sively on the puzzle-solving component, isolating it from other

60

functionalities of the application. Additionally, the automation

of tactical puzzle generation could be expanded to cover the mid-

50

dlegame phase, rather than being limited to endgame scenarios.

Another area of improvement is providing clearer assessments of

40 puzzle difficulty. This could be achieved by implementing a rating

Puzzles

d system that ranks puzzles based on completion rates, offering a

e

more accurate measure of challenge for each puzzle.

30

Solv



of Acknowledgements

er

20

The author would like to thank the family and friends who par-

ticipated in testing the application.

Numb

10

References

[1] Guillaume Bertholon, Rémi Géraud-Stewart, Axel Kugelmann, Théo Lenoir,

0

and David Naccache. 2020. At most 43 moves, at least 29: optimal strategies

and bounds for ultimate tic-tac-toe. (2020). doi: 10.48550/ARXIV.2006.02353.

0 [2] Chess.com. 2024. Chess.com. (June 2024). https://www.chess.com/. 1 2 3 4 5 6 7 8 9

[3] Rémi Coulom. 2006. Efficient selectivity and backup operators in monte-

AI Level Beaten In: Proceedings Computers and Games 2006

carlo tree search. In . Springer-

Verlag.

[4] Justin Diamond. 2022. A practical method for preventing forced wins in

Figure 3: Progress in relation to the number of solved puz- ultimate tic-tac-toe. (2022). doi: 10.48550/ARXIV.2207.06239.

zles. Each arrow represents a human tester and indicates [5] Nelson Elhage. 2020. Solving ultimate tic tac toe. (July 2020). https://minim

the change in the achieved level from the beginning to the ax.dev/docs/ultimate/.

[6] Arpad E Elo. 1978. . Arco Pub. The Rating of Chessplayers, Past and Present

end of the application’s use. [7] Ofek Gila. 2019. Ultimate tic tac toe. (2019). https://www.theof ekf oundatio

n.org/games/UltimateTicTacToe/.

[8] Fernand Gobet and Peter J Jansen. 2006. Training in chess: a scientific



7 approach. . Education and chess Discussion

[9] Henryk. 2023. Ultimate tic tac toe. (Dec. 2023). https://play.google.com/stor

e/apps/details?id=com.henrykvdb.sttt.

The results in Figure 3 indicate that solving more puzzles im-

[10] HPStudios. 2024. Ultimate tic tac toe. (June 2024). https://play.google.com/s

pacted users’ ability to reach higher levels, but had less of an tore/apps/details?id=com.MertTaylan.UltimateTicTacToe.

effect on lower levels. This is likely due to the fact that begin- [11] Lichess. 2024. Lichess. (June 2024). https://lichess.org/.

[12] Levi Moore. 2020. Ultimate tic tac toe. (Nov. 2020). https://play.google.com

ners can improve relatively quickly by simply playing the game,

/store/apps/details?id=com.ZeroRare.UltimateTicTacToe.

whereas advanced players require more effort to progress (eg. it [13] Arkadiusz Nowaczynski. 2021. Ar-nowaczynski/utttai: alphazero-like ai

solution for playing ultimate tic-tac-toe in the browser. (Dec. 2021). https:

is a lot easier to gain 100 rating points when you are rated 500 as

//github.com/ar- nowaczynski/utttai.

compared to when you are rated 1500).

[14] David Silver et al. 2018. A general reinforcement learning algorithm that

The reason for this is that at lower ratings, there is generally masters chess, shogi, and go through self-play. , 362, 6419, 1140–1144. Science

eprint: https : / / www . science . org / doi / pdf / 10 . 1126 / science . aar6404. doi:

more room for rapid improvement because the skill gap between

10.1126/science.aar6404.

players tends to be more pronounced, and beginners can quickly [15] Michael Xing. 2019. Ultimate tic-tac-toe. (Oct. 2019). https://www.michaelx

benefit from fundamental knowledge and tactical awareness. As ing.com/UltimateTTT/v3/.

a result, achieving a higher rating initially is easier as players

can fix obvious mistakes and exploit weaker opponents’ errors.

However, as players reach higher levels, the competition be-

comes tougher, and the differences in skill become more nu-

anced. Players at this level are more consistent and less likely

to make blunders, so improving further requires mastering ad-

vanced strategies, pattern recognition, and deeper positional





Ethical Consideration and Sociological Challenges in the Integration of Artificial




Intelligence in Mental Health Services



Saša Poljak Lukek

sasa.poljaklukek@teof.uni-lj.si

University of Ljubljana, Faculty of Theology

Ljubljana, Slovenia



Abstract 1 treatment options. Chatbots provide therapy via

This article explores the transformative potential natural language processing [5], while digital

of artificial intelligence (AI) in the field of mental platforms support online mostly cognitive

health, with a particular focus on ethical behavioral therapeutic interventions [6]. Avatar

considerations and social challenges. As AI therapy uses AI to help patients manage

tools become increasingly sophisticated, their conditions like dementia, autism spectrum

ability to support mental health interventions disorder, and schizophrenia [7].

presents both opportunities and challenges. We

discuss the importance of a human-centered 1.2 The Prospect of artificial

approach to AI development and the need for intelligence in mental health services

comprehensive ethical guidelines to ensure The future orientation underlines the importance

patient safety and well-being. In addition, this of digital health in overcoming challenges such

paper explores key social trends such as the as limited access to services, especially in

evolving dynamics of modern families, aging underserved regions, and outlines measures to

population, migration and considers how AI can ensure equitable access to digital health

be integrated into these contexts to improve solutions across the European region [8]. The

mental health care. use of AI in mental health services raises

questions about the role of non-human

Keywords: interventions, transparency in the use of

Artificial Intelligence, Mental Health, Human- algorithms and the long-term impact on the

Centered Approach, Ethics, Modern Family understanding of illness and the human

Dynamics, Aging Populations, Migration condition [9]. There are also concerns about

potential bias, gaps in ethical and legal



1 frameworks, and the possibility of misuse Introduction



health care has shown some positive effects on Accessibility and personalization of services. AI offers new mechanisms to reach those who the treatment of mental health problems [1], might not otherwise be served. AI-supported including early detection [2,3], providing tools can improve the early detection and feedback and personalized treatment plans [4], diagnosis of mental disorders [12]. AI chatbots and developing of novel diagnose tools [2]. have shown promise in increasing referrals to AI in mental health services is implemented health services However, there are at least two potentially positive effects of the use of AI in healthcare: Research on the application of AI in mental 1.1 [10,11]. Artificial intelligence in mental



through models like chatbots, digital platforms, mental health services, especially for minority groups who are blocked from accessing and avatar therapy, enhancing accessibility and traditional care [13]. These technologies can

1 provide initial assessments, psychoeducation

This Publication is a Part of the Research Program The Intersection and even treatment, expanding access to mental

Insights, financed by the University of Ljubljana. health support [12]. AI-driven virtual assistants of Virtue, Experience, and Digital Culture: Ethical and Theological

and wearable devices enable continuous



monitoring and personalized care, which could 2.2 Aging Populations

improve patient outcomes [11,14]. AI offers promising solutions for supporting an

The integration of artificial intelligence into aging population, particularly in addressing

mental health services represents a promising cognitive decline and mental health challenges.

avenue for the development of personalized AI applications can monitor vital signs, health

treatment plans through the sophisticated indicators, and cognition, as well as provide

analysis of large datasets, enabling the support for daily activities [20]. With an

identification of optimal therapeutic strategies increasing number of elderly individuals, AI can

tailored to specific client profiles [15,16]. This support mental health care by providing

data-driven methodology enables the dynamic companionship through intelligent animal-like

adaptation of therapy to the evolving needs of robots (e.g., Paro, Harp seal) and assisting in

the client. monitoring and managing conditions like

dementia [21,22]. AI can also help in tracking

2 Overcoming Sociological cognitive health and providing timely

Challenges through the Integration of interventions to maintain mental well-being in

Artificial Intelligence in Mental Health older adults. These technologies have the

Services potential to enhance independent living and

quality of life for older adults and their families.



2.1 Modern Family Dynamics

Modern family trends show that family 2.3 Migration



structures and attitudes have changed Migrants often face mental health challenges due to displacement, cultural adjustment and significantly in recent decades [17]. There is a language barriers. AI can help migrants access growing acceptance of different family forms, mental health services by providing culturally including unmarried cohabitation, same-sex and linguistically relevant resources and relationships and joint custody arrangements support. Chatbots and AI-driven platforms can [18]. These changes reflect an expansion of bridge gaps in care by providing immediate help developmental idealism and increasing support and continuity of care across different regions for individual freedom in family choice [17]. [23]. On the other hand, there is a growing need for Recent research highlights the increasing role of mental health services for families [19]. As the digitalization and artificial intelligence (AI) in most vulnerable members of the family - the migration and mobility systems, especially in the children - are usually also at risk, quick and context of the COVID-19 pandemic [24]. While effective action in family mental health is of these technologies offer opportunities for great importance. Many families are struggling improving human rights and supporting with various psychological problems. Together international development, they also bring with the changing family structure, this means a challenges that require careful consideration of great burden for every family member. In design, development and implementation addition, access to psychologists, psychiatrics aspects. The integration of AI into migration and therapists is limited, leading to an acute processes requires a focus on human rights at shortage of mental health professionals all stages that goes beyond technical feasibility worldwide. and companies' claims of inclusivity [24]. The accessibility of services is probably the

strongest argument for the integration of AI in

healthcare [12]. AI-powered conversational 3 Ethical Consideration in the

agents can improve the accessibility of mental Integration of Artificial Intelligence in

health services by being available online at all Mental Health Services

times and in underserved areas, being scalable, One of the main caveats to the use of AI in

reliable, fatigue-free, and providing consistent mental health is the introduction of new ethical

support, being culturally sensitive to adapt, and standards to ensure user safety. The approach to

helping with education and symptom integrating AI into services should therefore be

management. human-centered [25]. Any innovation should

therefore focus on people in their most



vulnerable position. It is important to assess all bias, especially among marginalized groups, the

risks with sufficient accuracy and avoid misuse risks associated with data privacy and security,

of AI as much as possible. The most important and the challenges posed by the lack of

areas for ethical consideration when integrating transparency of AI models.

AI into mental health services should be privacy,



bias, transparency, security. 4 Conclusion

Data privacy and security are critical in digital We propose to define AI as a new ethical entity

healthcare and require robust measures to in the field of mental health [30]. AI represents a

protect sensitive information and prevent novel artifact that changes interactions,

unauthorized access. Protecting privacy rights concepts, epistemic fields and normative

and ensuring informed consent are critical to requirements. This change requires a

maintaining trust and ethical standards in the redefinition of the role of AI, which lies on a

use of personal health data [11]. Combining spectrum between a tool and an agent. This shift

multiple data streams increases the risk of underscores the need for new ethical standards

unauthorized use, which exacerbates privacy and guidelines that recognize the unique status

issues. Ensuring informed consent and of AI as a distinct and influential actor in the field

maintaining transparency, especially in of mental health.

emergency operations, are critical to addressing The integration of AI into services can, on the

these ethical concerns and protecting the rights one hand, provide more efficient and faster

of participants [26]. solutions to some of the sociological challenges

The use of AI in mental health treatment raises of today's society, but on the other hand,

ethical concerns about bias, particularly among requires a precise and correct definition of the

marginalized populations who are already limits within which these models can be used.

discriminated against and lack access to mental These efforts aim to bridge the gap between

health care. It is uncertain whether AI-assisted technology and human-centered care and

psychotherapy can effectively address cultural ensure that AI complements, rather than

differences and close treatment gaps in diverse replaces, the therapeutic benefits of human

populations [27]. In addition, populations that interaction.

are traditionally marginalized in fields such as

psychology and psychiatry are most vulnerable

to algorithmic biases in AI and machine learning Literature

[27,28]. These biases limit the ability of AI to

provide culturally and linguistically appropriate [1] Sandhya Bhatt. 2024. Digital Mental Health: Role of

Artificial Intelligence in Psychotherapy. Annals of

mental health resources, exacerbating existing Neurosciences, 0,0, 1-11.

inequalities. The persistence of such biases in AI doi:10.1177/09727531231221612



systems not only risks increasing health [2] Sijia Zhou, Jingping Zhao and Lulu Zhang. 2022. Application of Artificial Intelligence on Psychological inequalities, but also exacerbates existing social Interventions and Diagnosis: An Overview. Frontiers in inequalities and raises critical ethical Psychiatry , 13(March), 1 – 7.

considerations [9]. https://doi.org/10.3389/fpsyt.2022.811665



The future of artificial intelligence in clinical [3] Klaudia Kister, Jakub Laskowski, Agata Makarewicz and Jakub Tarkowski. 2023. Application of artificial settings is affected by a significant ethical intelligence tools in diagnosis and treatmentof mental dilemma concerning the trade-off between the disorders. Current Problems of Psychiatry , 24, 1 – 18.

performance and interpretability of machine https://doi.org/10.12923/2353-8627/2023-0001



learning models [29]. The lack of transparency [4] Rachel L. Horn and John R. Weisz. 2020. Can Artificial Intelligence Improve Psychotherapy Research and in AI models makes it difficult to detect and Practice? Administration and Policy in Mental Health and correct biases. This underscores the need for Mental Health Services Research , 47, 5, 852 – 855.

greater transparency to ensure ethical and fair https://doi.org/10.1007/s10488-020-01056-9



clinical decision-making. [5] Kerstin Denecke, Alaa Abd-alrazaq and Mowafa Househ. 2021. Artificial Intelligence for Chatbots in Mental In summary, the integration of AI into mental Health: Opportunities and Challenges. In: Househ, M., health services requires the establishment of Borycki, E., Kushniruk, A. (eds) Multiple Perspectives on

strict ethical standards to protect the safety and Artificial Intelligence in Healthcare. 115– 128.

privacy of users. A human-centered approach is https://doi.org/10.1007/978-3-030-67303-1_10

essential, with a focus on dealing with potential

[6] Elias Aboujaoude, Lina Gega, Michelle B. Parish and Century of Family Attitude Trends in the United States.

Donald M. Hilty. 2020. Editorial: Digital Interventions in Sociology of Development, 9, 1, 1–32.

Mental Health: Current Status and Future Directions. https://doi.org/10.1525/sod.2022.0003

Front. Psychiatry 11, 111. doi: 10.3389/fpsyt.2020.00111 [19] WHO. 2022. World mental health report: transforming

[7] Kay T. Pham, Amir Nabizadeh & Salih Selek. 2022. mental health for all. (June 2022) Retrieved August 20,

Artificial Intelligence and Chatbots in Psychiatry. 2024 from

Psychiatric Quarterly, 93, 1, 249–253. https://www.who.int/publications/i/item/978924004933

https://doi.org/10.1007/s11126-022-09973-8 8.

[8] WHO. 2022. Regional digital health action plan for the [20] Sara J.Czaja and Marco Ceruso. 2022. The Promise of

WHO European Region 2023–2030 (RC72). (July 2022). Artificial Intelligence in Supporting an Aging Population.

Retrieved August 20, 2024 from Journal of Cognitive Engineering and Decision Making,

https://www.who.int/europe/publications/i/item/EUR- 16, 4, 182– 193.

RC72-5 https://doi.org/10.1177/15553434221129914

[9] Amelia Fiske, Peter Henningsen and Alena Buyx. 2019. [21] Maria R. Lima. 2024. Home Integration of Conversational

Your Robot Therapist Will See You Now: Ethical Robots to Enhance Ageing and Dementia Care.

Implications of Embodied Artificial Intelligence in ACM/IEEE International Conference on Human-Robot

Psychiatry, Psychology, and Psychotherapy. Journal of Interaction, 115– 117.

Medical Internet Research, 21, 5, e13216. https://doi.org/10.1145/3610978.3638378

https://doi.org/10.2196/13216 [22] Wendy Moyle. 2019. The promise of technology in the

[10] Elizabeth C. Stade, Shannon Wiltsey Stirman, Lyle future of dementia care. Nature Reviews Neurology, 15,

Ungar, Cody L. Boland,H. Andrew Schwartz, David B. 6, 353–359. https://doi.org/10.1038/s41582-019-0188-y

Yaden, Joao Sedoc, Robert J. DeRubeis, Robb Willer and [23] Zahra Abtahi, Miriam Potocky, Zarin Eizadyar, Shanna L.

Johannes C. Eichstaedt. 2024. Large Language Models Burke, Nicole M. Fava. 2022. Digital Interventions for the

Could Change the Future of Behavioral Healthcare: A Mental Health and Well-Being of International Migrants:

Proposal for Responsible Development and Evaluation. A Systematic Review. Research on Social Work Practice,

Mental Health Res 3, 12. 33, 5, 518-529. Doi:10.1177/10497315221118854

https://doi.org/10.1038/s44184-024-00056-z [24] Marie McAuliffe, Jenna Blower and Ana Beduschi. 2021.

[11] David B. Olawade, Ojima Z. Wada, Aderonke Odetayo, Digitalization and artificial intelligence in migration and

Aanuoluwapo Clement David-olawade, Fiyinfoluwa mobility: Transnational implications of the covid-19

Asaolu and Judith Eberhardt. 2024. Enhancing mental pandemic. Societies, 11, 4, 135.

health with Artificial Intelligence: Current trends and https://doi.org/10.3390/soc11040135

future prospects. Journal of Medicine, Surgery, and [25] Luke Balcombe and Diego de Leo. 2022. Human-

Public Health, 3, 100099. Computer Interaction in Digital Mental Health.

https://doi.org/10.1016/j.glmedi.2024.100099 Informatics, 9,1, 14.

[12] Koki Shimada. 2023. The Role of Artificial Intelligence in https://doi.org/10.3390/informatics9010014

Mental Health: A Review. Science Insights 43, 5, 1119- [26] Nicholas C. Jacobson and Matthew D. Nemesure. 2021.

1127. doi:10.15354/si.23.re820 Using Artificial Intelligence to Predict Change in

[13] Max Rollwage, Johanna Habicht, Keno Juechems, Ben Depression and Anxiety Symptoms in a Digital

Carrington, Sruthi Viswanathan, Mona Stylianou, Tobias Intervention: Evidence from a Transdiagnostic

U. Hauser and Ross Harper. 2023. Using Conversational Randomized Controlled Trial. Psychiatry Research, 295,

AI to Facilitate Mental Health Assessments and Improve 113618.

Clinical Efficiency Within Psychotherapy Services: Real- https://doi.org/10.1016/J.PSYCHRES.2020.113618

World Observational Study. JMIR AI, 2, e44358. [27] Bennett Knox, Pierce Christoffersen, Kalista Leggitt, Zeia

https://doi.org/10.2196/44358 Woodruff and Matthew H. Haber. 2023. Justice,

[14] David D. Luxton. 2020. Ethical implications of Vulnerable Populations, and the Use of Conversational

conversational agents in global public health. Bulletin of AI in Psychotherapy. American Journal of Bioethics, 23,

the World Health Organization, 98, 4, 285–287. 5, 48–50.

https://doi.org/10.2471/BLT.19.237636 https://doi.org/10.1080/15265161.2023.2191040

[15] Leonard.Bickman. 2020. Improving Mental Health [28] Zoha Khawaja and Jean C. Bélisle-Pipon. 2023. Your

Services: A 50-Year Journey from Randomized robot therapist is not your therapist: understanding the

Experiments to Artificial Intelligence and Precision role of AI-powered mental health chatbots. Frontiers in

Mental Health. Adm Policy Ment Health, 47, 795–843. Digital Health, 5, 1278186. doi:

https://doi.org/10.1007/s10488-020-01065-8 10.3389/fdgth.2023.1278186

[16] Silvan Hornstein, Valerie Forman-Hoffman, Albert [29] Danilo Bzdok and Andreas Meyer-Lindenberg,. 2018.

Nazander, Kristian Ranta and Kevin Hilbert. 2021. Machine Learning for Precision Psychiatry: Opportunities

Predicting therapy outcome in a digital mental health and Challenges. Biological Psychiatry: Cognitive

intervention for depression and anxiety: A machine Neuroscience and Neuroimaging, 3, 3, 223–230.

learning approach. DIGITAL HEALTH, 7, 1-11. https://doi.org/10.1016/J.BPSC.2017.11.007

doi:10.1177/20552076211060659 [30] Jana Sedlakova and Manuel Trachsel. 2023.

[17] Josef Ehmer. 2021. A historical perspective on family Conversational Artificial Intelligence in Psychotherapy: A

change in Europe. In Norbert F. Schneider and Michaela New Therapeutic Tool or Agent? American Journal of

Kreyenfeld (eds). Research Handbook on the Sociology Bioethics, 23, 5, 4–13.

of the Family, 143–161. https://doi.org/10.1080/15265161.2022.2048739

https://doi.org/10.4337/9781788975544.00018

[18] Keera Allendorf, Linda Young-Demarco and Arland

Thornton. 2023. Developmental Idealism and a Half-





Optimization Problem Inspector: A Tool for Analysis of




Industrial Optimization Problems and Their Solutions



Tea Tušar

Jordan N. Cork

Andrejaana Andova

Bogdan Filipič

Jožef Stefan Institute and Jožef Stefan International Postgraduate School

Ljubljana, Slovenia

{tea.tusar, jordan.cork, andrejaana.andova, bogdan.f ilipic}@ijs.si



Abstract OPI is a web application, implemented by a Python library called

This paper presents the Optimization Problem Inspector (OPI) optimization-problem-inspector included in the PyPi Python

1

package index . It is highly interactive and requires no program-

tool for assisting researchers and practitioners in analyzing indus-

ming knowledge to be used.

trial optimization problems and their solutions. OPI is a highly

Freely available contemporary software tools for multiobjective

interactive web application requiring no programming knowl-

optimization, such as DESDEO [12], jMetal [7] (and jMetalPy [2]),

edge to be used. It helps the users to better understand their

the MOEA Framework [8], ParadisEO-MOEO [10], platEMO [17],

problem by: 1) comparing the landscape features of the analyzed

pygmo [3], pymoo [4], and Scilab [15], provide the implementa-

problem with those of some well-understood reference problems,

tion of various optimization algorithms and test problems. While

and 2) visualizing the values of solution variables, objectives, con-

the majority of them include some visualization of solutions,

straints and any other user-specified solution parameters. The

the plots are mostly focused on showing algorithm results for

features of OPI are presented using a bi-objective pressure vessel

the purpose of comparing algorithm performance and not to

design problem as an example.

increase problem understanding. In addition, none of these tools



Keywords compute additional problem features as OPI does. Therefore, OPI

brings a unique perspective to optimization problem analysis and

optimization, black-box problems, sampling, problem characteri-

understanding.

zation, visualization

Next, Section 2 presents the real-world problem that will be

used to showcase the features of OPI in Section 3. The paper

1 Introduction

concludes with some remarks in Section 4.

Industrial optimization problems often require simulations to



evaluate solutions. For example, in electrical motor design [18, 2 Real-World Use Case

19], assessing the efficiency and electromagnetic performance of a

Our chosen real-world problem is a version of the well-known

proposed design is done by running a simulator that analyzes the

pressure vessel design problem, first proposed more than 30

motor magnetic field and flux distribution. Such evaluations are

years ago [16]. In this work, we adapt the formulation from [5]

black boxes to the user and the optimization algorithm alike, i.e.,

to handle the pressure vessel volume as a constraint, as well as

the underlying functions cannot be explicitly expressed, which

an objective. We also remove one unnecessary constraint and

makes the problem hard to understand and solve.

use the original boundary constraints for the first two variables.

The established way to gain a better understanding of indus-

A pressure vessel is a tank, designed to store compressed

trial problems is through the analysis of their solutions. Depend-

gasses or liquids. It consists of a cylindrical middle part capped

ing on the problem at hand, this can be a challenging task, as

at both ends by hemispherical heads. The pressure vessel has

industrial problems often have a large number of variables, mul-

four design variables (see Figure 1): the shell thickness, 𝑥 1 𝑇s, =

tiple objectives and constraints [20].

the head thickness, 𝑥 2 𝑇 , the inner radius, 𝑥 𝑅 h = 3 =, and the

The Optimization Problem Inspector (OPI) presented in this

length of the cylindrical section of the vessel, 𝑥 4 𝐿. The two =

paper is a tool conceived to ease this task for both problem experts

thickness variables are integer multiples of 0.0625 inches, which

and optimization algorithm developers. OPI provides two ways

correspond to the available thicknesses of rolled steel plates,

to further the understanding of an optimization problem:

while the length and the radius are continuous. The problem

(1) It computes a set of landscape features of the analyzed

has three constraints, two on the search variables and one on

problem and compares them to those of well-understood

1

reference problems. https://pypi.org/project/optimization- problem- inspector/

(2) It provides visualizations of solutions through the values

of their variables, objectives, constraints and any other

𝑥 𝐿 =

user-specified solution parameters. 𝑥2 = 4 =

𝑇 𝑥 𝑇

h 1 s



Permission to make digital or hard copies of all or part of this work for personal 𝑥 = 3 𝑅

or classroom use is granted without fee provided that copies are not made or

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this

work must be honored. For all other uses, contact the owner /author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.scai.8265 Figure 1: Pressure vessel design variables.





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Tea Tušar, Jordan N. Cork, Andrejaana Andova, and Bogdan Filipič




the volume. Its two objectives are to minimize the total costs, If needed, the sample can be generated by the tool itself, based

including the costs of the material, forming and welding, and to on the variable information provided in the problem specification

maximize the volume. The problem is formally defined as follows: step. However, this is not a required step in using OPI. A user

that already has a set of (evaluated) solutions to work with can

min 𝑓1 0.6224𝑧1𝑥3𝑥4 1.7781𝑧2𝑥 3.1661𝑧 𝑥4 x ) = + +

( 2 2

3 1 skip it and input the data directly (see Section 3.3).

+ 2 19.84𝑧 𝑥 Sample generation requires one to choose the number of de- 3

1

sired samples, set to a default of 100, and the sample generation

4

max 𝑓2 ( 2 3 x ) = 𝜋 𝑥 𝑥 4 3 +

𝜋 𝑥 method. Three sample generation methods are supported: ran- 3

3

dom, Sobol and Latin Hypercube, with random sampling being

subject to 𝑔1 ( x) = 0.0193𝑥3 − 𝑧1 ≤ 0 the default. The user may alter the settings of these sampling

𝑔2 0.00954𝑥3 𝑧2 0 methods, such as the random generator seed. Selecting the but-( x ) = − ≤

𝑔3 𝑓2 1 296 000 x ) = ( x ) ≥

( ton to generate and download the sample will download it in a

csv-formatted file.

𝑥 1 18, . . . , 32 ∈ {}

In the pressure vessel use case, OPI warns the user that not

𝑥 2 10, . . . , 32 all sample generation methods are appropriate. In fact, the Sobol ∈ { }

𝑥 3, 𝑥4 10 ∈ [, 200] sampler and the Latin Hypercube Sampler are not compatible

with non-continuous parameters. If used nevertheless, they may

where 𝑧1 = 0.0625𝑥1 produce unexpected results.

𝑧2 0.0625𝑥2 =

3.3 Data

3 Optimization Problem Inspector Features In OPI, the data is essentially a set of evaluated solutions, where

OPI is a web application, organized into five functional sections each solution must contain a value for all objectives, constraints

and a help section, providing guidance to the user. OPI expects the and other parameters included in the problem specification. The

user to provide the problem specification and its data—evaluated

evaluation is conducted externally to the tool.

problem solutions. Then, it generates and visualizes comparisons The data needs to be uploaded in a file in csv format. If any to artificial reference problems and visualizes the provided data.

parameters from the problem specification are missing from the

Next, we will describe the main features of OPI through its data, the tool will display a warning message. Any data parame-

five content sections: problem specification, sample generation, ters that are not included in the problem specification, are ignored

data, comparison to reference problems, and data visualization. without raising any warnings. When correctly input, the user

will be able to view the data they have input, inspecting it in

3.1 Problem Specification tabular format.

Inputting the data completes the setup stage of the process.

In the first OPI section, the user can provide the specification

The user may then begin generating visualisations to assist them

of the industrial problem to be studied. The tool needs this in-

in understanding their problem.

formation to properly generate the samples, described in the

Section 3.2, and setup the visualisations.

The problem specification must be given in the file for-yaml 3.4 Comparison to Reference Problems

mat and needs to contain some basic information about problem

The first visualization mechanism provided by OPI visually com-

parameters (variables, objectives, constraints) to be included in

pares the problem to a set of artificial reference problems with

the analysis. OPI can handle one or more objectives and zero or

known properties. This is conducted by displaying the landscape

more constraints. In addition to variables, objectives and con-

features of the user-defined problem alongside the same features

straints, the user can specify any number of other parameters

of each of the reference problems in a parallel coordinates plot.

that they want analyzed and visualized, for example, the name

The plot is interactive—the user can highlight some of the prob-

of the algorithm that found a solution or the time required to

lems by brushing along one of the parallel axes. In addition, the

evaluate a solution.

feature values can be viewed in a table and downloaded to a file

For each of the parameters, the user needs to specify its name csv

in format.

and its grouping (whether it is a variable, objective, constraint or

The reference problems can be set by the user, however, con-

something else). For variables, their type (continuous, integer or

fined within the collection labelled here as GBBOB, i.e., gener-

categorical) and the upper and lower bounds (for non-categorical

alised BBOB, where BBOB stands for the well-known suite of 24

types) are also required. An example file, specifying a con-yaml

Black-Box Optimization Benchmarking problems with diverse

strained multiobjective problem with several variables, is already

properties [9]. OPI provides a generator of GBBOB problems that

provided within the tool to guide the user.

match the analyzed problem in terms of the number of variables

For the pressure vessel design problem, we can input four

and objectives and the presence or absence of constraints. For

variables (first two are integer and last two are continuous), two

objectives and (optionally) the constraint, any single-objective

objectives and three constraints. Alternatively, we can decide to

BBOB problem instance can be used. The user can specify the de-

skip the individual constraints and only use the total constraint yaml

sired GBBOB problems in the format. OPI already contains

violation instead.

five GBBOB problems to start.

A problem can be characterized by a large number of features,

3.2 Sample Generation most hard to interpret by a human. In OPI, we included the fol-

In OPI, a sample is a set of x-values, corresponding to the variables lowing problem landscape features that are understandable to an

set in the problem specification section. In other words, a sample expert user [1, 11, 13, 14]: , , , , CorrObj MinCV FR constr_obj_corr

is a set of non-evaluated solutions. , , and a set of neighborhood features. H_MAX UPO_N PO_N CorrObj

Optimization Problem Inspector Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia





Figure 2: The initial part of the parallel coordinate plot visualizing feature values for the analyzed problem and the chosen

set of artificial test problems.



is a feature that shows the correlation between the objectives. well as the color map, can also be specified by the user. Both vi-

MinCV represents the minimum constraint violation among all so- sualizations support interaction and can be downloaded in html

lutions in the population. represents the proportion of feasible or format. FR png

solutions in the population. presents the max-constr_obj_corr

imum correlation between the constraints and all the problem 3.5.1 Scatter Plot Matrix. 2

The scatter plot matrix consists of 𝑛

objectives. is the maximum information content among

H_MAX plots for 𝑛 chosen problem parameters as it contains 2-D scatter

all objectives. is the proportion of unconstrained non-

UPO_N plots for all possible parameter pairs. In OPI, the user can apply

dominated solutions, while is the proportion of the con-

PO_N brushing and linking to select the desired solutions in one or

more of the scatter plots. These are then highlighted in all scatter

strained non-dominated solutions. The neighbourhood features

plots in the matrix.

denoted by are a collection of features neighbourhood_feats

Figure 3 shows such a scatter plot for our pressure vessel

explaining the neighborhood of solutions, e.g., how many neigh-

problem. This visualization includes data from two sources. The

bors of a solution dominate the solution, how many neighbors

first comes from a random sampling of the search space (shown

are dominated by the solution, how many are incomparable to

in light blue) and the second from running the NSGA-II algo-

the solution, how close the neighboring solutions are, etc. OPI

rithm [6] on this problem for 2 10 function evaluations to

· 6

offers a total of 16 features, but the user can choose which to

achieve a good approximation of the Pareto front (shown in

compute and visualize.

black). The two sources are set apart by a custom parameter that

Figure 2 shows the initial part of the parallel coordinates plot

is then used for coloring the solutions. Some solutions from Fig-

(as the entire plot would not fit the paper) for the pressure vessel

problem. In the comparison, we use the default five GBBOB ure 3 are highlighted – see the rectangle in the 𝑥 3 (, 𝑥1) scatter

plot (third from the left in the top row).

reference problems as well as a custom created one. We notice

These plots clearly show the linear relationship of the near-

that the pressure vessel problem is most similar to the custom

optimal solutions between 𝑥 1 and 𝑥2 as well as 𝑥1 and 𝑥3. When

GBBOB problem with the first objective equal to the step ellipsoid

only 𝑓1 and 𝑓2 are chosen, it is distinctively visible that the Pareto

function 𝑓7, the second to the multimodal peaks function 𝑓22, and

set approximation is piece-wise linear and disconnected.

the linear constraint 𝑓5. This similarity might be due to our mixed-

integer problem containing plateaus in the continuous landscape 3.5.2 Parallel Coordinates Plot. The parallel coordinates plot

space in which the features are computed, which is similar to the

shows all chosen parameters as parallel coordinates and solu-

step ellipsoid function, and having linear constraints.

tions as lines in the plot. Similarly as with the scatter plot matrix,

interaction via brushing and linking is supported to select solu-

3.5 Data Visualization tions that fit the desired values.

In the data visualization section of the web application, the sup-

plied data can be visualized using either a scatter plot matrix or 4 Conclusions

a parallel plot. In both cases, the user can choose which prob- This work presented the features of Optimization Problem Inspec-

lem parameters to visualize among all those listed in problem tor – a web application to support problem experts and algorithm

specification. Additionally, a simple data filtering that limits any designers in gaining a better understanding of industrial optimiza-

variable between the desired minimum and maximum values is tion problems. The tool provides comparisons to well-understood

also supported and can be manipulated via the OPI interface in reference problems and interactive and highly-customizable vi-

yaml format. The parameter used for coloring the solutions, as sualizations, which can be exported in html and png formats.





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Tea Tušar, Jordan N. Cork, Andrejaana Andova, and Bogdan Filipič





Figure 3: Random (light blue) and near-optimal (black) solutions of the pressure vessel design problem visualized in OPI

with a scatter plot matrix containing variables 𝑥1 to 𝑥4 and objectives 𝑓1 and 𝑓2.



Samples can be exported and solutions imported using the stan- [7] Juan José Durillo and Antonio J. Nebro. 2011. jMetal: A Java framework

dard format, which makes the data exchange between OPI

csv for multi-objective optimization. , 42, 10, Advanced Engineering Software

760–771. doi: 10.1016/J.ADVENGSOFT.2011.05.014.

and various optimization software easy to do. OPI functionality

[8] David Hadka. 2024. MOEA Framework: A free and open source Java frame-

is made to be simple and at the same time flexible. Therefore, it work for multiobjective optimization. https://github.com/MOEAFramewor

k/MOEAFramework. Computer software, version 4.4. (2024).

is utilisable by non-experts and experts, alike, providing a wide

[9] Nikolaus Hansen, Steffen Finck, Raymond Ros, and Anne Auger. 2009. Real-

range of angles from which to view the problems. Parameter Black-Box Optimization Benchmarking 2009: Noiseless Functions

Definitions. Research Report RR-6829. INRIA. https://hal.inria.f r/inria- 0036



Acknowledgements 2633v2. [10] Arnaud Liefooghe, Laetitia Jourdan, and El-Ghazali Talbi. 2011. A software

The authors acknowledge the financial support from the Slove- framework based on a conceptual unified model for evolutionary multiob-

jective optimization: ParadisEO-MOEO. European Journal of Operational

nian Research and Innovation Agency (research core funding Research

, 209, 2, 104–112. doi: 10.1016/J.EJOR.2010.07.023.

No. P2-0209 “Artificial Intelligence and Intelligent Systems”, and [11] K. M. Malan, J. F. Oberholzer, and A. P. Engelbrecht. 2015. Characterising

constrained continuous optimisation problems. In Proceedings of the 2015

project No. N2-0254 “Constrained Multiobjective Optimization

IEEE Congress on Evolutionary Computation (CEC 2015), 1351–1358. doi:

Based on Problem Landscape Analysis”) and from the Jožef Ste-

10.1109/CEC.2015.7257045.

fan Innovation Fund (project “A Tool for Analysis of Industrial [12] Giovanni Misitano, Bhupinder Singh Saini, Bekir Afsar, Babooshka Shavazi-

pour, and Kaisa Miettinen. 2021. DESDEO: The modular and open source

Optimization Problems and Their Solutions”). This publication is

framework for interactive multiobjective optimization. , 9, 148277– IEEE Access

also based upon work from COST Action “Randomised Optimi- 148295. doi: 10.1109/ACCESS.2021.3123825.

sation Algorithms Research Network” (ROAR-NET), CA22137, [13] Mario A. Muñoz, Michael Kirley, and Saman K. Halgamuge. 2015. Ex-

ploratory landscape analysis of continuous space optimization problems

supported by COST (European Cooperation in Science and Tech- IEEE Transactions on Evolutionary Computation

using information content. ,

nology). We are grateful to Jernej Zupančič for implementing the 19, 1, 74–87. doi: 10.1109/TEVC.2014.2302006.

[14] Cyril Picard and Jürg Schiffmann. 2021. Realistic constrained multiobjec-

core functionalities of the Optimization Problem Inspector.

tive optimization benchmark problems from design. IEEE Transactions on

Evolutionary Computation, 25, 2, 234–246. doi: 10.1109/TEVC.2020.3020046.

References [15] Philippe Roux and Perrine Mathieu. 2016. Scilab: I. Fundamentals. In Scilab,

[1] Hanan Alsouly, Michael Kirley, and Mario Andrés Muñoz. 2023. An instance from theory to practice. D-Booker Editions.

space analysis of constrained multiobjective optimization problems.

IEEE [16] E. Sandgren. 1990. Nonlinear integer and discrete programming in mechan-

Transactions on Evolutionary Computation ical design optimization. , 112, 2, 223–229. Journal of Mechanical Design

, 27, 5, 1427–1439. doi: 10.1109

[17] Ye Tian, Ran Cheng, Xingyi Zhang, and Yaochu Jin. 2017. PlatEMO: A MAT-

/TEVC.2022.3208595.

LAB platform for evolutionary multi-objective optimization. IEEE Computa-

[2] Antonio Benítez-Hidalgo, Antonio J. Nebro, José García-Nieto, Izaskun

Oregi, and Javier Del Ser. 2019. jMetalPy: A Python framework for multi- tional Intelligence Magazine

, 12, 4, 73–87. doi: 10.1109/MCI.2017.2742868.

objective optimization with metaheuristics.

Swarm and Evolutionary Com- [18] Tea Tušar, Peter Korošec, and Bogdan Filipič. 2023. A multi-step evaluation

putation process in electric motor design. In Slovenian Conference on Artificial In-

, 51, 100598. doi: 10.1016/J.SWEVO.2019.100598.

[3] Francesco Biscani and Dario Izzo. 2020. A parallel global multiobjective telligence, Proceedings of the 26th International Multiconference Information

framework for optimization: pagmo. , 5, 53, Society (IS 2023) Journal of Open Source Software

. Vol. A. Jožef Stefan Institute, Ljubljana, Slovenia, 48–51.

[19] Tea Tušar, Peter Korošec, Gregor Papa, Bogdan Filipič, and Jurij Šilc. 2007.

2338. doi: 10.21105/joss.02338.

A comparative study of stochastic optimization methods in electric motor

[4] Julian Blank and Kalyanmoy Deb. 2020. Pymoo: Multi-objective optimization

in Python. , 8, 89497–89509. doi: 10.1109/ACCESS.2020.2990567.

IEEE Access design. , 27, 2, 101–111. doi: 10.1007/S10489- 006- 0022- 2. Applied Intelligence

[20] Koen van der Blom, Timo M. Deist, Vanessa Volz, Mariapia Marchi, Yusuke

[5] Carlos A. Coello Coello. 2002. Theoretical and numerical constraint-handling

Nojima, Boris Naujoks, Akira Oyama, and Tea Tušar. 2023. Identifying

techniques used with evolutionary algorithms: A survey of the state of the

art. , 191, 11, 1245–

Computer Methods in Applied Mechanics and Engineering properties of real-world optimisation problems through a questionnaire. In

1287. doi: https://doi.org/10.1016/S0045- 7825(01)00323- 1. Many-Criteria Optimization and Decision Analysis: State-of-the-Art, Present

[6] Kalyanmoy Deb, Amrit Pratap, Sameer Agarwal, and T. Meyarivan. 2002. A Challenges, and Future Perspectives. Natural Computing Series. Dimo Brock-

fast and elitist multiobjective genetic algorithm: NSGA-II.

IEEE Transactions hoff, Michael Emmerich, Boris Naujoks, and Robin C. Purshouse, editors.

on Evolutionary Computation Springer, 59–80. doi: 10.1007/978- 3- 031- 25263- 1_3.

, 6, 2, 182–197. doi: 10.1109/4235.996017.

Multi-Agent System for Autonomous Table Football: A Winning



Strategy



∗ ∗

Marcel Založnik Kristjan Šoln

Jožef Stefan Institute Faculty of Electrical Engineering, University of Ljubljana

Ljubljana, Slovenia Ljubljana, Slovenia

marcel.zaloznik@gmail.com ks4835@student.uni- lj.si





Abstract

This paper presents a multi-agent system (MAS) for autonomous

table football, developed for the FuzbAI competition at the Uni-

versity of Ljubljana. Our system consists of four independent

agents, each dynamically assigned specific roles—Goalkeeper,

Defender, Midfielder, and Attacker—based on real-time game

analysis. This role-based architecture enabled seamless coordi-

nation between offensive and defensive strategies, allowing our

team to secure first place. We describe the simulation framework

used, the processing of sensor data, and the control strategies

that allowed the agents to execute precise actions in a dynamic

environment. The results highlight the effectiveness of adaptive,

role-based decision-making, demonstrating the potential of MAS

in real-time, competitive settings.



Keywords

multi-agent system, autonomous table football, role-based strat- Figure 1: Table setup for the FuzbAI autonomous football

competition.

egy, real-time decision making, AI in robotics



1 Introduction selecting roles that dictated their actions during gameplay. This

The FuzbAI competition, held as part of the “Dnevi Avtomatike”

strategic approach enabled our team to outperform competitors

event at the Faculty of Electrical Engineering, University of Ljubl-

and ultimately secure first place in the competition.

jana, is a premier contest for students specializing in automation

This paper delves into the development and implementation

and artificial intelligence [11]. This event challenges participants

of our multi-agent system. We will explore the architectural

to develop intelligent autonomous agents capable of playing table choices, the role-based decision-making strategies employed by

football without human intervention. The competition not only

each agent, and the overall system’s performance in the context

serves as a platform for demonstrating technical skills but also of the FuzbAI competition. fosters innovation in the application of AI and machine learning



techniques in real-time environments. Figure 1 illustrates the 2 Competition Setup and System Description table setup used in the competition.

The FuzbAI competition required all participants to develop pro-

The FuzbAI competition is structured in a way that teams

grams capable of playing table football autonomously. To facil-

must design and implement a fully autonomous system capable

itate this, the competition provided a standardized simulation

of effectively competing against other AI-driven systems. Each

environment and a set of initial tools that every team used as

match is a test of the participants’ ability to integrate advanced

the foundation for their development. This section describes the

algorithms and robotics, simulating the dynamics of a real foot-

simulation framework, the types of data available from the sys-

ball game on a miniature scale. The competitive format includes

tem, and the means by which agents could interact with both the

both qualification rounds and knockout stages, ensuring that

simulated and real game environments.

only the most capable and innovative solutions advance to the



final stages. 2.1 Simulation Framework

Our entry into the FuzbAI competition focused on the develop-

Participants were provided with a Python-based simulation frame-

ment of a multi-agent system (MAS), where each of our four rods

work designed to emulate a real table football match, as shown in

functioned as an independent agent. These agents were designed

figure 2. This simulator accurately replicated the physics of the

to collaborate through a streamlined decision-making process,

game, including the movement of the ball and rods, and managed

∗Both authors contributed equally to this research. the interactions between the environment and the agents control-

ling the rods. The framework included fundamental functionali-

Permission to make digital or hard copies of all or part of this work for personal

ties such as ball tracking, rod positioning, and interaction rules,

or classroom use is granted without fee provided that copies are not made or

allowing all teams to concentrate on AI development without

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this needing to construct the simulation infrastructure themselves.

work must be honored. For all other uses, contact the owner /author(s).

One of the key features of the competition setup was that the

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

interaction protocols for the simulator and the physical table

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.scai.8341 were identical. The same signals and commands used to control





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Založnik and Šoln





3 Related Work

Research on multi-agent systems (MAS) and their application

in robotic football has been extensively explored. This section

reviews some contributions that have informed the development

of autonomous systems for table football and real football.

Moos et al. (2024) [5] developed an automated football table

as a research platform for reinforcement learning, highlighting

the challenges of transferring learned behaviors from simulation

to real-world environments and the need for robust algorithms

to handle uncertainties. While reinforcement learning is a com-

mon approach in such studies, we did not achieve satisfactory

results with it. Therefore, we decided to use multi-agent systems

instead. Klančar et al. (2002) [4] investigated cooperative con-

trol in robot football (real football) using multi-agent systems,

Figure 2: Simulator interface. focusing on behavior-based control and dynamic role assignment

among robots to optimize performance. Their approach empha-

sized effective communication for coordination in multi-agent

settings. This work particularly inspired our approach to multi-

the actuators in the simulator were also used for the real table

agent systems, where we focused on behavior-based control and

without any modification. This feature ensured that teams could

dynamic role assignment. Ribeiro et al. (2024) [6] proposed a

seamlessly transition their algorithms from the simulated envi-

probability-based strategy (PBS) for robotic football (real foot-

ronment to the physical table setup, which was used in the final

ball), utilizing real-time data for centralized decision-making

rounds of the competition. As a result, the simulation provided

without relying heavily on pre-defined plays. Their approach

a consistent testing ground that mirrored the actual physical

demonstrated flexibility across different environments. Smit et

setup, enabling teams to develop and refine their strategies under

al. (2023) [8] explored scaling multi-agent reinforcement learn-

uniform conditions.

ing (MARL) to a full 11v11 simulated football environment (real

football), focusing on computational efficiency and the use of

2.2 Sensor Data attention mechanisms to enhance scalability in large-scale multi-Both the simulation environment and the real table provided

agent settings. Song et al. (2024) [9] conducted an empirical study

each team with data from two cameras, one placed on each side on the Google Research Football platform (real football), intro-

of the table. Each camera captured different views of the game,

ducing a population-based MARL training pipeline to quickly

and teams had to decide how to combine the information from develop competitive AI players, highlighting the importance of

both cameras. The data provided by each camera included: scalable training frameworks. Scott et al. (2022) [7] examined



• end-to-end learning in RoboCup simulations (real football), op- Ball position: The coordinates of the ball on the 2D plane

timizing both low-level skills and high-level strategies through

of the table.

• competitive self-play, providing a comprehensive approach to Ball speed: Velocity of the ball.

• multi-agent training in competitive environments.

Ball size: Area of the ball in the captured image (in pixels).

• Rod positions: Calibrated position of all rods (in the inter-

val 0, 1 ). [ ] 4 MAS Approach to Autonomous Table

• Rod angles: Calibrated angle of all rods (in the interval Football Control

[−32, +32]). In this section, we describe the the methodology of our approach.

This camera data was streamed continuously, requiring teams We describe agent architecture, different agent roles and outline

to process and merge the inputs from both cameras to accurately the actions they can take. Then, we discuss the conditions and

interpret the game’s state. The accuracy and frequency of the priorities for role assignment during the game and evaluate the

behavior of the system as a whole.

data were sufficient to enable real-time decision-making by the

autonomous agents, whether interacting with the simulator or

the physical table. 4.1 Agent Architecture

There exist several agent architectures, commonly used in MAS.

2.3 Actuator Outputs Approaches, such as [4, 10, 12, 13], use role-based approach for

interaction between agents and with the environment. In role-

To interact with the environment, each agent could send com-

based approach, based on the concepts from role theory [1], the

mands to the actuators that controlled the rods. The system

agents are assigned roles which affect their behavior. While the

allowed for two primary types of commands:

overall long-term goal of the system is typically predefined and

• Translatory movement: Moving the rod left or right across

does not change, e.g. win a table football match, the current role

the table.

of an individual agent defines agent’s short-term goals, which

• Rotational movement: Rotating the rod to control the angle

influences agent behavior, their decision-making process, and

at which the players struck the ball.

how they interact with the rest of the system. Furthermore, sepa-

Precise and timely commands were crucial for effective game ration of agent functionality into independent roles can provide

control, as they enabled the agents to optimally position their fig- simplification and decoupling of individual tasks, leading to a

ures, strike the ball accurately, and execute defensive or offensive more modular system, which can simplify and improve the ex-

strategies effectively. tensibility of the implementation [3].



Autonomous Table Football: Winning Strategy Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



There exist several approaches to role and behavior implemen- opponent scoring even if the actuators fail to respond fast enough

tation in MAS, such as merging different roles, role models and to block this style of attack. Here, communication between the

class members [2, 3, 4]. In our implementation, we simplify the two agents is performed implicitly, as each agent perceives the

architecture by allowing an agent to occupy only a single role at roles of other agents as a part of the overall environmental state.

a time, and defining the roles in a way that allows reassigning be- is an agent tasked with blocking opponent attacks Defender

tween iterations of the algorithm without regard to the previous by intercepting the ball when it is in the opponent’s possession or

role or state of the agent. moving towards the goal. This role utilizes a single action, follow

Each role defines a set of possible actions an agent can take. similar to the Goalkeeper’s action. Whenever the Defender follow

The agents decide which action to take based on their priority role is active, the agent tracks the position and velocity of the ball,

and the current environment. More complex roles can be im- trying to match either its current coordinate or the estimated

plemented in a stateful manner, meaning the decision on which intersection with the trajectory of the ball. The agent identifies

action to take is dependent on previous actions as well. An agent the figure closest to the intersection and attempts to move the

can only be assigned a single role at a time, but can switch be- rod using minimal amount of movement. This approach allows

tween roles throughout iterations regardless if the particular goal for faster adjustments during the game, improving defensive

is fulfilled, when appropriate conditions arise. Additionally, every efficiency.

agent must have a role assigned at all times. is a an agent role with the primary task of raising Midfielder

An action is a discrete, autonomous task that an agent can the figures to allow passing the ball from behind the current

take on by making appropriate decisions and acting onto the agent. This role, although simple, is essential in order to avoid

environment, e.g. by sending commands to the actuators. This accidentally breaking a friendly attack by an Attacker agent

advances the agent toward the goal imposed by the current role. behind the current rod.

An agent can only execute a single action at a time. Additionally, is an agent with the task of kicking the ball towards Attacker

every agent must be actively executing an action at all times. the opponent goal in an attempt to score a point. Unlike other

These concepts were implemented using an Object Oriented roles, the Attacker role is implemented in a stateful manner.

approach, as suggested by the authors of the competition. In our Actions can only happen in a specified order, when the corre-

implementation, each agent repeatedly executes a fast processing sponding conditions are met. The role implements , follow kick

routine. Every iteration, the environment data is updated and role and actions. prevent back-kick

selection for the agent is performed. Then, as the agent decides on Whenever the agent is assigned this role, the action is follow

a role for that iteration, the appropriate role processing function executed first. During the action, the agent slightly raises follow is called, executing individual actions. the figures in order to prepare for a kick. The figure closest to

the ball is selected and rod offset is adjusted in order to align

the figure with the ball. Whenever the agent determines that the

4.2 Role Description alignment with the ball is sufficient, the agent moves onto the

A typical table football setup consists of four rods per player, each next state, the action. Here, the rod is rotated in order to kick with a number of mounted figures. In this implementation, each strike the ball. During this state, it is still necessary to track the

rod is considered an agent, resulting in a system with four agents position of the ball, as the ball can move significantly within a

for which we define the following roles, typically associated with few iterations of the algorithm. As the rod completes the forward

table football games. rotation, the agent monitors the position of the ball and assesses

Goalkeeper is the final line of defense, primarily responsible if the figure successfully hit the ball. In that case, the next action

for intercepting the ball before it reaches the goal. Typically the is set back to , and the agent is usually assigned a new role follow left-most rod, which is nearest to the goal and has a single figure, according to the environment. However, if the figure missed the

the goalkeeper follows the ball position using two possible ac- ball during the kick, the agent moves onto the prevent back-kick

tions: and . The action simply tries action. This final action is meant to prevent an accidental kick follow misaligned follow follow to align the figure on the rod with the current ball position. How- in the opposite of the intended direction. The rod is translated

ever, if the velocity of the ball exceeds a predefined threshold, sideways and slowly rotated into a neutral position, in order to

the agent instead attempts to estimate the ball trajectory based circumvent the ball. While executing this action, role switching

on its velocity vector. This estimation is simplified by assuming for the current agent is disabled as well.

that the ball maintains a straight-line path. The figure is there- During execution, the agent aligns the rod position with the

fore positioned at the intersection of the rod and the estimated ball; however, a perfectly aligned figure results in a straight shot,

trajectory in an attempt to intercept a fast-moving ball. which is easily defended by maintaining alignment with the ball.

The action is an augmented variant of the A more effective strategy involves kicking at an angle to aim misaligned follow

former action, designed to increase the overall defense surface of for the goal or create a rebound off the wall, which is harder

the defending agents. A common scenario in table football occurs to defend. This role achieves this by slightly misaligning the

when an attacker attempts to bypass the defenses by slightly figure before and during the kick. The agent computes the angle

pushing the ball parallel to the rod and striking it immediately between the ball’s current position and the selected target, with

after. Even though a human player might react fast enough to the figure’s required misalignment set proportionally to this angle

block such an attack, actuator response times are often insuffi- and adjusted by a tunable parameter for fine-tuning. This attack

cient. A defense strategy against such attacks is is to misalign the strategy significantly increases the performance of the Attacker

goalkeeper and defender figures, increasing the defense surface. role.

Here, this is implemented by the action, and is misaligned follow

activated whenever the ball is relatively slow, in the possession

of the opponent and another agent in front of the Goalkeeper is

currently in a Defender role. This decreases the chances of the





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Založnik and Šoln




4.3 Role Assignment success rate. Additionally, even though there are no explicit,

intentional passes between agents, the strategy of simply passing

Individual roles are assigned to agents according to defined as-

the ball as far forward as possible is enough for a successful

signment conditions and rules. Some approaches use an objective

gameplay.

function in order to select a role, often taking role priority into

The system overall is sensitive to changes in parameters and

account [4]. In this approach, we instead define a simple set of

requires precise tuning. The simulator, although effective, does

conditions which, along with role priority, decide on the most

not perfectly simulate the physical table, and additional parame-

appropriate role for a particular agent based on the current state

ter tuning is required when transitioning from the simulator to

of the environment.

real-world application.

If in a particular instant, more roles fulfill the assignment con-

ditions for a particular agent, the role with higher priority is

selected. In this implementation, the highest priority belongs 5 Conclusion

to the Attacker role, followed by the Goalkeeper, Defender and This paper presented a multi-agent system (MAS) for autonomous

finally the Midfielder with the lowest priority. This ordering is table football, developed for the FuzbAI competition. Our role-

based on the strictness of assignment conditions for each role, based design allowed each rod to act as an independent agent,

and the importance of that particular role. For example, the At- dynamically adapting to the game state. This approach enabled

tacker role has the strictest selection conditions among all roles, effective coordination between offense and defense, contributing

and therefore is assigned the highest priority, while the Mid- to our system’s first-place win.

fielder role has a very broad assignment condition and is not as The results demonstrate the effectiveness of a modular, adap-

important compared to an Attack agent. tive architecture in dynamic environments, highlighting the im-

We define the role selection conditions as follows. The At- portance of robust decision-making and quick role-switching.

tacker role is selected whenever the ball speed drops below a Future work could include machine learning to predict opponent

specified threshold, and the ball is within kicking clearance of behavior and optimize strategies, as well as expanding the system

the rod. The Goalkeeper role is selected if that particular agent to more complex environments. Overall, our MAS showed strong

belongs to the left-most rod, closest to the player’s goal. The performance in a competitive setting, offering valuable insights

Defender role is selected whenever the ball is in front of the rod. for future developments in autonomous systems.

Lastly, the Midfielder role is selected whenever the ball is behind

the rod, as the role’s only task is to raise the figures to allow the References

ball to pass forward. [1] Bruce J Biddle. 1986. Recent developments in role theory. Annual review of

sociology, 12, 1, 67–92.

This set of conditions combined with the defined role priority,

[2] G. Cabri, L. Ferrari, and L. Leonardi. 2004. Agent role-based collaboration and

allow the agents to switch between roles effectively and covers coordination: a survey about existing approaches. In 2004 IEEE International the main functionality required to play the game. Role priority Conference on Systems, Man and Cybernetics (IEEE Cat. No.04CH37583)

. Vol. 6.

IEEE. isbn: 1062-922X. doi: 10.1109/ICSMC.2004.1401064.

ensures that the agent works toward a correct goal based on

[3] E.A. Kendall. 1999. Role modelling for agent system analysis, design, and

the circumstances. For example, any rod, even the Goalkeeper, implementation. In Proceedings - 1st International Symposium on Agent should attempt to kick the ball if it is close and slow enough, Systems and Applications and 3rd International Symposium on Mobile Agents,

ASA/MA 1999. IEEE, 204–218. isbn: 0769503403. doi: 10.1109/ASAMA.1999

while only the left-most rod should attempt to be the goalkeeper.

.805405.

[4] Gregor Klančar, Marko Lepetič, Boštjan Potočnik, Rihard Karba, and Drago

Matko. Cooperative control of mobile agents in soccer game. Faculty of

4.4 Behavior of the System as a Whole Electrical Engineering, University of Ljubljana, Slovenia, (2002).

[5] Janosch Moos, Cedric Derstroff, Niklas Schröder, and Debora Clever. 2024.

The system’s primary offensive strategy is for the Attacker agents Learning to play foosball: system and baselines. In Cornell University Li-

brary, arXiv.org. doi: 10.48550/arxiv.2407.16606.

to advance the ball as far forward as possible, ultimately aiming

[6] António Fernando Alcântara Ribeiro, Ana Carolina Coelho Lopes, Tiago

for the goal, while Midfielder agents ensure that they do not

Alcântara Ribeiro, Nino Sancho Sampaio Martins Pereira, Gil Teixeira Lopes,

obstruct forward passes. During opponent attacks, the systems and António Fernando Macedo Ribeiro. 2024. Probability-based strategy

for a football multi-agent autonomous robot system. , 13, 1. doi: Robotics

primary defensive strategy is for the Defender and Goalkeeper

10.3390/robotics13010005.

roles to intercept the ball. In certain situations, they collaborate [7] Atom Scott, Keisuke Fujii, and Masaki Onishi. 2022. How does ai play to expand the defense surface, compensating for the limitations football? an analysis of rl and real-world football strategies. In International

Conference on Agents and Artificial Intelligence. Vol. 1. Elsevier Scopus, 42–52.

posed by actuator response times. Once the opponent’s attack

isbn: 2184-3589. doi: 10.5220/0010844300003116.

ends, agents detect the change in the environment and the roles [8] Andries Smit, Herman A. Engelbrecht, Willie Brink, and Arnu Pretorius.

2023. Scaling multi-agent reinforcement learning to full 11 versus 11 sim-

are reassigned to shift the game towards offensive play.

ulated robotic football. , 37, 1. Autonomous agents and multi-agent systems

The system’s game strategy can be adjusted by modifying

doi: 10.1007/s10458- 023- 09603- y.

parameters such as role priority, assignment rules, or individual [9] Yan Song, He Jiang, Zheng Tian, Haifeng Zhang, Yingping Zhang, Jiangcheng

Zhu, Zonghong Dai, Weinan Zhang, and Jun Wang. 2024. An empirical study

actions. For instance, a more defensive strategy can be achieved

on google research football multi-agent scenarios. International journal of

by tightening the conditions for assigning the Attacker role. , 21, 3, 549–570. doi: 10.1007/s11633- 023- 1426- 8. automation and computing

Overall, the implemented algorithm performs well, with the [10] Manuela Veloso, Peter Stone, and Kwun Han. 1998. The cmunited-97 robotic

soccer team: perception and multiagent control. In Proceedings of the second

combination of discrete roles resulting in a competent gameplay. international conference on Autonomous agents, 78–85.

However, delays and noise present in measurements, and delays [11] Laboratorij za avtomatiko in kibernetiko. 2024. Dnevi avtomatike. Accessed:

2024-08-21. (2024). https://dnevi- avtomatike.si/?page_id=270.

due to actuator response times, sometimes cause the system to

[12] Franco Zambonelli, Nicholas R. Jennings, and Michael Wooldridge. 2003.

miss, e.g. during attacks. The action of the prevent back-kick

Developing multiagent systems: the gaia methodology. ACM transactions on

Attacker role proves essential in such situations, performing , 12, 3, 317–370. ObjectType-Article-2. software engineering and methodology

doi: 10.1145/958961.958963.

careful repositioning. Another surprisingly successful strategy is

[13] Xiaoqin Zhang, Haiping Xu, and Bhavesh Shrestha. 2007. An integrated

aiming at the goal or the wall during the action. Even if the role-based approach for modeling, designing and implementing multi-agent attack

ball does not follow the intended trajectory due to measurement systems. , 13, 2, 45–60. doi: 10.100 Journal of the Brazilian Computer Society

7/bf 03192409.

noise and system delays, it still considerably increases the attack



Towards a Decision Support System for Project Planning:



Multi-Criteria Evaluation of Past Projects Success



Miha Hafner Marko Bohanec

Elea iC d.o.o. Jožef Stefan Institute

Department for Tunnels and Geotechnics, and Department of Knowledge Technologies

Jožef Stefan International Postgraduate School Ljubljana, Slovenia

Ljubljana, Slovenia marko.bohanec@ijs.si

miha.hafner@elea.si



Abstract • Objectives that support project goals include concrete and



in which project assets, timelines, budgets, milestones, goals [7]. subcontractors, etc., are determined before the new project starts. • Scope and requirements concerning project boundaries, In this paper, we address infrastructure design projects in the e.g., the need for experts, potential subcontractors, technical Project planning typically refers to the project management step milestones, and other steps and strategies to achieve the measurable project characteristics such as deliverables,

context of a specific company (Elea iC) and explore the idea of

equipment and other requirements to finish the project.

using data about past-finished projects to help project managers

• Constraints and limitations concerning project deadlines,



this context is the ability to evaluate/assess the success of costs, etc. [8]. and project leaders in project planning. A crucial requirement in

Besides that, each project should finish with the client’s and

finished/new projects. This paper proposes a solution using a

stakeholders’ satisfaction [8].

multi-criteria model to evaluate finished projects. This way, we

To achieve the above for the new project, project planning is

add project success information to the finished projects database,

vital at the beginning of each new project [6], [8]. It is the project

which we shall use in the decision support system being designed

management and project leaders' task to recognize and include

to extract knowledge for the new project plan.

all these in the project plan so that the work and processes lead



Keywords to successful project completion.

This study aims to support this process in the context of

Project success evaluation, multi-criteria model, decision support Elea iC company, an interdisciplinary provider of engineering

systems, data analysis, data mining, project management, project services and projects in Slovenia [5]. We wanted to include the

leading tools. knowledge obtained from past–finished projects in the project

planning process for the new projects. The company collected

this data from 2001. The assumptions are as follows:

1 Introduction 1. The finished projects in the database offer valuable

Infrastructure, such as tunnels, bridges, schools, houses, sewage information for the new project planning phase.

systems, roads, etc., and its design discipline play a vital role in 2. The project workflows established in the company and

society. Thus, infrastructure design must have properly and requirements remained similar over the years.

thoroughly defined requirements, objectives, scope and The main challenge related to this question is the new project

constraints concerning many expert fields such as civil success assessment and its consideration in light of the available

engineering, architecture, geology, geotechnics, environmental finished project data [7]. Unfortunately, the actual finished

engineering, urban planning, and other expert fields [1], [2], [3]. projects database does not contain much information about the

The term design is connected to the process that ends with finished projects' success. To bridge this, we had to construct a

technical documentation, technical approvals, models, and other project success evaluation model, evaluate finished projects in

deliverables prepared at the end of the design process. Each such the database and add this information to the database. The

process is referred to as the project [4]. The projects are expected expected result of those steps is a database suitable for applying

to have clearly defined: data-analysis and knowledge extraction methods, such as

• Goals defining the project's desired result, e.g., a building hierarchical clustering and machine learning [20].

permit for a bridge, static analysis of a retaining wall, This paper describes the finished project success evaluation

architectural design for a subway station, geotechnical component (hereafter called FPSE), which is part of the future

exploration for a tunnel, etc. [4]. decision support system (DSS, [12], [13], [14]) for project

planning (hereafter called E-DSS). First, we present the general

Permission to make digital or hard copies of part or all of this work for personal or architecture of the E-DSS, explaining the role and integration of

classroom use is granted without fee provided that copies are not made or distributed FPSE in its context. In section 3, we present the database of

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must finished projects and its preparation for supporting the

be honored. For all other uses, contact the owner/author(s). configuration of new projects. The evaluation model used and

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s). the experimental evaluation of FPSE are presented in sections 4

https://doi.org/10.70314/is.2024.scai.8463 and 5, respectively. Section 6 concludes the paper.





2 Figure 1 also shows the element E-DSS administrator used E-DSS Architecture


to upgrade FPSE periodically by upgrading the database of the

E-DSS is a DSS under construction to support the project

finished projects or making changes in EM according to the

management and project leaders in the Elea iC company

system's operational requirements and expected results.

(hereafter called “the user”) in configuring the new project plan

parameters when a new project starts.

The user is expected to define the E-DSS input as shown in

Figure 1: the new project objectives, requirements, desires, and

expectations. Practically, this means that the user collects all the

available new project data by:

• Extracting the new project data from the new project

assignment and contracts containing relevant information

for the project planning.

• Checking the company and potential subcontractors' state of

the resources and assets needed to complete the new project.

Examples of those data include projected monetary value, project

scope and goals, project start and finish date, the expert fields

needed for project completion, etc.

The E-DSS output (Figure 1) consists of the new project plan Figure 2: FPD+S development workflow

configuration together with the corresponding success scores This paper focuses on the development of FPSE. The

(+S). The configuration comprises the data such as the number workflow is shown in Figure 2, consisting of the following steps:

of employees involved, the number of subcontractors, work Step A. Finished projects database preparation (FPD), as

distribution, work duration, the number of pauses, etc. Project described in section 3.

success scores are assessed assuming this configuration settings. Step B. Project success evaluation model (EM), as described

in section 4.



Step C. Finished projects database with EM success scores



(FPD+S): The result of the FPSE is the upgraded database

of the finished projects with the finished projects' success

scores (FPD+S).



3 Data Description

E-DSS is a data-driven system that operates on data from past-

finished projects. This data was collected in Elea iC company

from the year 2001 to 2023. At the beginning of the data

collection, the number of the observed variables was relatively

small, but it has grown substantially over the years. At the time

Figure 1: E-DSS architecture of this study, the database contained data on 4704 finished

Accordingly, E-DSS is composed of the following projects, described by 39 numeric variables; 6 of them were

components (Figure 1): date/time/year variables, and 2 categorical variables.

• NPPE (New Project Parameters Extraction) is the Data preparation (Step A, Figure 2) was carried out as

component that extracts the potential new project follows:

configuration parameters and corresponding data to support 1. Data cleaning: replacing “Nan” and deleting erroneous data;

the decision-making. NPPE is currently under construction 2. Outlier’s removal using the Interquartile range approach

and is aimed to operate interactively with the user and [18];

support: searching for similar projects in FPD+S according 3. Data imputation: replacing the missing values using a

to a predefined range, searching the projects by desired descriptive statistic (e.g. mean, median, or most frequent)

success score, project segmentation, and project group along each column or using a constant value [19]. We

identification—unsupervised descriptive analytics and employed the mean strategy.

parameter prediction by supervised machine learning 4. Sensitive data and information removal. For this reason, all

methods. The component NPPE+S inside NPPE evaluates numeric data was scaled to a range between 0 and 1.

the success of the potential new project's configuration We ended up with the database FPD containing data on 3132

parameters obtained. The evaluation is made by EM, which finished projects described by 27 numeric variables. The

is part of FPSE. variables describe the main project management characteristics,

• such as project financial results, workload distribution, number FPSE (Finished Project Success Evaluation) consists of:

o of employees, subcontractors, etc. Table 1 shows the list of all EM (Project success Evaluation Model) for evaluating

the new project configuration (described in section 4). variables together with their basic statistics.

o This way, the finished project database (FPD) was prepared FPD+S is the database of finished projects with project

success evaluations (section 3). for the FPSE component. FPD is the main resource for

Exploratory Data Analysis for observing the data and its





properties, such as variable correlations, variable information • Pauses Time Share: the ratio between the months the


gain, etc. These operations are invoked interactively by the user employees did not work and the total number of months.

in the context of NPPE and are not discussed further in this paper. • Hour Income: the ratio between project value and the

number of work hours necessary to finish the project.

Table 1: Basic statistics of the variables after data cleaning,

outliers’ removal, and data scaling





Figure 3: Multi-criteria model for the projects’ success



evaluation



Evaluation parameters represent outputs of the model:

• WORK DISTRIBUTION: assesses the characteristics of the

work distribution in the project duration.

• PROJECT PAUSES: assesses the work pauses.

• PROJECT WORKFLOW: combines evaluation parameters

WORK DISTRIBUTION and PROJECT PAUSES

• PROJECT FINANCIAL RESULT: assesses the project's

success from the financial point of view.

• PROJECT SUCCESS SCORE: overall success score,

determined by aggregation of all subordinate parameters.

Aggregation functions map subordinate EM parameters to the

4 Evaluation of Projects’ Success corresponding parent parameters. Employed is the weighted

The project success evaluation model (EM), developed in Step B average function, using weights shown in Figure 3. Currently,

(Figure 2), is aimed at: weights are chosen to make all parameters equally important.

• The evaluation of the projects in FPD resulting in the

FPD+S (Figure 2).

• 5 Experimental Evaluation of FPSE The evaluation of potential new projects suggested through

interaction between the user and NPPE+S (Figure 1). Figure 4 shows an example of evaluating a project from FPD.

Project success evaluation involves multiple criteria that have Input parameters’ values (terminal nodes) were obtained from

to be aggregated into a single evaluation score. Different criteria the data base, while evaluation parameters’ values (green nodes)

might be of different importance and affect the score differently, were calculated by EM. The example project shows good

i.e., with different weights. For this purpose, we chose MAUT workflow score (0.75), but has a poor financial score (0.29), both

(Multi-Attribute Utility Theory) [11], a multi-criteria modelling leading to an average success score (0.52) of the project. Several

approach that facilitates both hierarchical structuring of criteria other projects of different types were evaluated in this way,

and using weights for the aggregation of scores. confirming the appropriateness of EM structure and

Considering the above requirements, available FPD data and conformance with requirements of potential users. In this way,

other multi-criteria approaches to project evaluation ([15], [16], the quality of EM was assessed on a sample of past projects.

[17]), we developed the EM as presented in Figure 3. Further assessment is planned in the next stages while

EM consists of three components [10]: input parameters, configuring new projects, where EM’s results can be confronted

evaluation parameters and aggregation functions. with opinions of project leaders actively involved in the process.

EM already enables evaluation of multiple finished projects.

Input parameters are variables in the leaf nodes of the model:



• Project Work Concentration In Step C (Figure 2), FPD was extended by adding five variables : explains the distribution of the corresponding the five Evaluation parameters of EM. All work on the project. If the value is closer to 0 or 1, the projects in FPD were evaluated by EM, resulting in FPD+S. majority of the work is done at the beginning or at the end Basic statistics of FPD+S is presented by the distribution of of the project, respectively. the variables in Figure 5. The variables marked with red colour

• Time Reserve: explains if the project work ended earlier on the x-axis are E-DSS input parameters, the green uppercase

than defined in the contract. variables are those corresponding to success scores, and the blue

• Number of Pauses: the number of times the work on the variables are potential new project parameters. The distribution

project stopped. of the final project evaluation, PROJECT_SUCCESS_SCORE,





(average = 0.52, min = 0.15, max = 0.94) indicates that it well and created decision trees for prediction of individual output


covers the range of possible outcomes and enables the parameters that may lead to high new project success scores.

discrimination and sorting of projects. Future work will primarily continue by further data analysis

and data mining of FPD+S, attempting to design effective



algorithms for interactive exploration of past projects and

suggesting as good as possible configurations of new projects.

On this basis, we shall make a detailed functional specification

of the NPPE+S component and design/implement the E-DSS.

Despite that E-DSS considered here is tailor-made for the

specific business environment and is bound to the specific data

base, the approach seems general enough to be applied to similar

environments, projects and processes [9]. This work is a

showcase of substantial efforts needed to prepare a corporate

database for decision-support, which is often neglected in the

Figure 4: Example of evaluating a project using EM literature. The main contribution is a combination of data

processing with MAUT-based multi-criteria decision modelling.





References

[1] Fransje L. Hooimeijer, Jeremy D. Bricker, Adam J. Pel, A. D. Brand, Frans

H.M. Van de Ven, Amin Askarinejad. 2022. Multi-and interdisciplinary

design of urban infrastructure development. In Proceedings of the Institution

of Civil Engineers: Urban Design and Planning. Vol.175. TU Delft. 153-168.

[2] Simon Christian Becker, Philip Sander. 2023. Development of a Project

Objective and Requirement System (PORS) for major infrastructure projects

to align the interests of all the stakeholders. In Expanding Underground -

Knowledge and Passion to Make a Positive Impact on the World. CRC Press,

London, UK, 3369-3376. DOI:10.1201/9781003348030-408.

[3] Michel-Alexandre Cardin, Ana Mijic, Jennifer Whyte. 2023. Data-driven

infrastructure systems design for uncertainty, sustainability, and resilience.

In D. M. Fabio Biondini, Life-Cycle of Structures and Infrastructure

Systems. CRC Press, London, UK, 2565 – 2572. DOI:

10.1201/9781003323020-312.

[4] Saša Žagar. 2016. Organizacijski model v projektivnem podjetju Elea iC

d.o.o.. Maribor, B.Sc. Thesis, Retrieved July 12, 2024 from

https://dk.um.si/IzpisGradiva.php?id=58799&lang=eng.

[5] Elea iC webpage. https://www.elea.si/en/.

[6] Jürg Kuster, Eugen Huber, Robert Lippmann, Alphons Schmid, Emil

Schneider, Urs Witschi, Roger Wüst. 2015. Project Management Handbook.



Figure 5: Distribution of the FPD+S features, including EM [7] Springer-Verlag, Berlin Heidelberg, Germany. Anton Hauc. 2007. Projektni management. (2nd. ed.). GV Založba,

project success assessments Ljubljana, Slovenija.

[8] Harvey A. Levine. 2002. Practical Project Management: Tips, Tactics, and

Tools. John Wiley & Sons, Inc., New York, NY.

[9] Nadja Damij, Talib Damij. 2014. Process management. Springer-Verlag,

6 Conclusions Berlin Heidelberg, Germany.

E-DSS is a DSS under construction, aimed at supporting the [10] Marko Bohanec. 2012. Odločanje in modeli. DMFA – založništvo,

Ljubljana, Slovenija.

project management and project leaders' process in the new [11] Salvatore Greco, Matthias Ehrgott, José Rui Figueira. 2016. Multiple Criteria

infrastructure project planning phase. We presented the design Decision Analysis, State of the Art Surveys. Springer, Portsmouth, UK. DOI

10.1007/978-1-4939-3094-4

and development of the FPSE component, consisting of a multi- [12] Maria Rashidi, Maryam Ghodrat, Bijan Samali and Masoud Mohammadi.

criteria project success evaluation model EM and a data base of 2018. Decision Support Systems. In Management of Information Systems.

projects, extended with success evaluation scores FPD+S. IntechOpen, London, UK, 19-38. DOI: 10.5772/intechopen.79390.

[13] Daniel Joseph Power. 2013. Decision Support, Analytics, and Business

EM has been developed using the MAUT approach and has Intelligence. Business Expert Press, New York, NY. DOI

turned out to be “fit-for-purpose”. It employs the data that is 10.4128/9781606496190.

[14] Sofiat Abioye, Lukumon Oyedele, Lukman Akanbi, Anuoluwapo Ajayi,

available in the projects’ database. It meaningfully describes Juan Manuel Davila Delgado, Muhammad Bilal, Olugbenga Akinade, Ashraf

aspects of the project's success and offers practical and functional Ahmed. 2021. Artificial intelligence in the construction industry: A review of present status, opportunities and future challenges. Journal of Building

model for the evaluation of multiple projects in the database. Engineering 44, Elsevier. https://doi.org/10.1016/j.jobe.2021.103299.

FPSE is a key decision-support resource for E-DSS. E-DSS [15] Erwin Berghuis. 2018. Measuring Systems Engineering and Project Success,

will allow the user to interactively search for similar past Master’s Thesis. University of Twente. https://purl.utwente.nl/essays/75088

[16] Ali Beiki Ashkezari, Mahsa Zokaee, Amir Aghsami, Fariborz Jolai, Maziar

projects, to filter them according to the success score and Yazdani. 2022. Selecting an Appropriate Configuration in a Construction

simulate the effects of alternative project configurations, Project Using a Hybrid Multiple Attribute Decision Making and Failure

ultimately proposing the best one. Approaches based on Analysis Methods. Buildings, MDPI, Volume 12, 643. DOI:

https://doi.org/10.3390/buildings9050112.

unsupervised descriptive analytics (clustering) and supervised [17] Urban Pinter, Igor Pšunder. 2013. Evaluating construction project success

machine learning methods for prediction of E-DSS output with use of the M-TOPSIS method. Journal of civil engineering and

management, Volume 19(1), 16–23. doi:10.3846/13923730.2012.734849

parameters are foreseen for this purpose. Actually, we have [18] Interquartile range. Retrieved May 15, 2024 from

already tested hierarchical clustering and decision tree https://en.wikipedia.org/wiki/Interquartile_range

[19] SimpleImputer. Retrieved May 15, 2024 from https://scikit-

classification methods on FPD+S, and first results are learn.org/stable/modules/generated/sklearn.impute.SimpleImputer.html

encouraging. We obtained meaningful clusters of past projects [20] Aggarwal C. Aggarwal. 2015. Data Mining: The Textbook. Springer, New

York, USA.





Minimizing Costs and Risks in Demand Response Optimization:




Insights from Initial Experiments



Mila Nedić Tea Tušar

Faculty of Mathematics and Physics Jožef Stefan Institute and

University of Ljubljana Jožef Stefan International Postgraduate School

Ljubljana, Slovenia Ljubljana, Slovenia

mn38120@student.uni- lj.si tea.tusar@ijs.si



Abstract techniques. However, baselines can be exploited, e.g., when con-

sumers artificially increase consumption before an event to inflate

This paper presents a method for changing the energy use of

their baseline and maximize the awarded rebate.

consumers participating in Demand Response (DR) programs,

1

Through the SEEDS project , we are developing a method-

focusing on peak balancing to improve grid stability. Multiple ob-

ology for providing energy flexibility services to prosumers –

jectives including costs and risks are considered, and a weighted

participants in energy markets capable of both producing and

sum is used to transform them into a single objective. This re-

consuming energy – in order to enhance grid stability. Machine

sults in an optimization problem that can be optimally solved.

learning is used to predict the baseline energy usage of prosumers

To calculate the costs, the load consumption baseline needs to

and their flexibility, while mixed-integer linear programming

be established. Since this is challenging and can be exploited,

(MILP) is used to optimize the operation of prosumers within

we conduct initial experiments to test whether our method to

their flexibility. Our approach will be tested in the Slovenian pilot,

adjust the baseline can be easily manipulated. We explore an

in collaboration with Petrol d.d. and Elektro Celje d.d.

original scenario and three of its variants to examine the effects

Our work integrates prosumer flexibility into DR optimization,

of various parameters on the optimization outcome. Our results

focusing on minimizing costs and risks while limiting energy

indicate that 1) an excessive emphasis on risk results in no energy

fluctuations. While the goal is to eventually use this approach on

change, 2) enforcing a net zero energy change minimizes energy

real-world data from the pilot, this paper reports on some initial

use while still securing the rebate, and 3) without an adjustment

experiments verifying whether the current problem formulation

period, the consumer is less inclined to increase the load just be-

results in solutions with desired properties. In particular, we wish

fore the demand period. In future work, we will reformulate some

to test if our adjusted consumer baseline approach can be easily

objectives to avoid exploitation and better reflect the real-world

exploited.

needs of DR.

Research on prosumer flexibility, optimization techniques,



Keywords and demand response optimization includes a wide range of

approaches [8]. In [3], Balázs et al. quantify residential prosumer

multiobjective optimization, mixed-integer linear programming,

flexibility using engineering models and real-world data. Their

demand response, baseline consumption, electrical grid

work provides valuable insight into prosumer behavior and en-

ergy management. Capone et al. [4] optimize district energy

1 Introduction

systems by balancing costs and carbon emissions with genetic al-

Peaks in energy demand can strain the electrical grid, leading to

gorithms and linear programming, showing significant emission

inefficiencies and potential failures. A widely used strategy for

reductions at a modest cost increase. Magalhães and Antunes [7]

balancing these peaks is Demand Response (DR), in which the

compare thermal load models in demand response strategies

Distribution System Operator (DSO) forecasts future peaks and

using MILP, finding that discrete control formulations improve

requests from consumers to adjust their energy use to reduce

computational efficiency. Thus, our methodology is in line with

them. In the peak time rebate DR program [2], consumers receive

related work while the actual optimization problem (its vari-

a rebate if they reduce their load in the demand period. On the

ables, objectives and constraints) differs from existing ones as it

other hand, if they commit to respond to the demand, but fail to

is adapted to our specific use case.

do so, they can be penalized. It is therefore of utmost importance

This paper is further organized as follows. In Section 2, we

to accurately assess whether and how much a consumer reduced

provide a brief overview of the optimization problem, followed

their load to meet the demand.

by its detailed definition in terms of its variables, constraints and

The load reduction of a consumer is computed as the differ-

objectives. The optimization approach is explained in Section 3,

ence between its baseline (the amount of energy the customer

where we discuss the scalarization technique used to transform

would have consumed without a demand request) and its actual

our multi-objective problem into a single-objective MILP form

use [2]. The importance of establishing a baseline and the various

and the method used to solve it. The experiments and their results

ways of calculating it are presented in [5]. Common methods

are given in Section 4. Finally, conclusions and further work ideas

for calculating baselines include simple historical data averages,

are described in Section 5.

exponential moving averages and short-term load forecasting



2 Optimization Problem

Permission to make digital or hard copies of all or part of this work for personal

or classroom use is granted without fee provided that copies are not made or The problem formulation in this work assumes a peak time rebate

distributed for profit or commercial advantage and that copies bear this notice and

DR program in which the DSO and the consumer have a contract

the full citation on the first page. Copyrights for third-party components of this

work must be honored. For all other uses, contact the owner /author(s). stipulating the following conditions: 1) the consumer can chose

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia whether to respond to a demand request, 2) if the consumer

© 2024 Copyright held by the owner/author(s).

1

https://doi.org/10.70314/is.2024.scai.8587 https://project- seeds.eu/



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Mila Nedić and Tea Tušar



A F

participates in DR, it receives a rebate proportional to the reduced 𝐸 instead of the forecast one 𝐸 , where the adjustment is deter-

𝑡 𝑡

A

load, 3) if the consumer participates in DR but does not reduce the mined by the energy amounts in the adjustment period – the 𝑛

load by at least 75 % of the required amount, it is penalized, 4) the S

intervals before the start of the demand period 𝑡 . More formally,

load reduction is estimated using an adjusted consumer baseline, the adjusted timetable is computed as

which takes into account the forecast consumer energy usage as

𝑡 1

well as its actual consumption before the demand period.   1 ∑︁ F  S −

A 𝐸 𝐸 𝐸 , 𝑛 𝑗   𝑡 − F A −

if > 0;

The optimization task is to set the energy consumption of A 𝑗 =

𝐸 𝑛

all loads of a consumer participating in DR taking into account 𝑗 𝑡 − 𝑡 A S =

𝑛



their flexibility so that consumer costs, risks and energy fluctua-  F  𝐸 , otherwise

𝑡



tions are minimized. This ensures efficient grid operation while

for all intervals 𝑡 𝑡 , . . . , 𝑡 in the demand period. Then, }

maintaining economic feasibility for the consumer.

∈ { S E

R ∈

To formally define our optimization problem, we first intro- 𝑡 the recognized load reduction 𝐸 at demand time interval 𝑡

duce its variables, followed by the constraints and the objective 𝑡 , . . . , 𝑡 is determined as { S E }

functions we aim to optimize. Finally, we provide an overview

of the weighted sum approach, which serves as the scalarization 𝐸 𝐸 𝑡 𝑡 = R A −

𝐸 ,

𝑡

technique to transform all objective values into a single one. R

while the total recognized load reduction 𝐸 is computed as



2.1 Variables E 𝑡

∑︁

A solution is specified by the energy amounts 𝐸 𝐸 𝐸 . for each 𝑡 R

𝑐,𝑖

∈ R R =

consumer load 𝑐 and time interval 𝑖 1, . . . , 𝑛 . They 𝑡 = S ∈ C ∈ { }

𝑡

correspond to the change of consumption from the forecast one.

R

A rebate is awarded if 𝐸 is negative (the consumption has

These are the only variables of this optimization problem.

been reduced). If the total recognized load reduction exceeds the

From these energy amounts and the forecast timetable of en-

T

total demanded energy reduction 𝐸 , the rebate is capped, i.e.,

ergy usage, the resulting energy consumption 𝐸𝑖 in time interval

𝑖 1, . . . , 𝑛 is computed as ( ∈ { } B R T R

R 𝑝 min 𝐸 , 𝐸 , if 𝐸 < 0

∑︁ 𝑓 . =

𝐸 𝑖 𝐸𝑖 = F +

𝐸𝑐,𝑖. 0, otherwise

𝑐 ∈ C

Finally, a penalty is added to the total costs if the demand

2.2 Constraints has not been met, that is, the ratio between the recognized and

D

The energy amounts of a solution need to adhere to two kinds of demanded energy reduction, 𝐸 , in any of the demand time

constraints. The first type are the interval energy constraints: intervals 𝑡 𝑡 , . . . , 𝑡 is lower than 75 %, ∈ { } S E



min max ≤ ≤

𝐸 R 𝐸 𝐸 , 𝑐,𝑖 𝑐,𝑖 𝑐,𝑖 𝐸 

P  P 𝑡 T S E 𝑝   | 𝐸 | , if < 75 % for one or more 𝑡 ∈ { 𝑡 , . . . , 𝑡}

for each consumer load 𝑐 and time interval 𝑖 1, . . . , 𝑛 . 𝐸 ∈ { }

∈ C 𝑓 D . =

 0 

, otherwise

The second are the total energy constraints:



𝑛

∑︁ The second optimization objective 𝑓2 represents risks. In order

𝐸 𝐸𝑐,𝑖 𝑐 ≤ 𝑇 ,min 𝑇 ,max ≤

𝐸 ,

𝑐 to penalize any changes to the timetable when the risks are high,

𝑖 1 =

the objective function is defined as

for each consumer load 𝑐 . ∈ C

𝑛



2.3 𝑓 2 𝑟𝑖 𝐸𝑐,𝑖 , Objective Functions

= ∑︁ ∑︁

𝑖 1 𝑐 =∈ C

The three objectives to be minimized in this scenario are the

where 𝑟 represents the risk at time interval 𝑖.

costs, risks and energy fluctuations. 𝑖

To penalize unnecessary energy fluctuations, the third objec-

The first optimization objective 𝑓1 consists of all costs associ-

ated with the solution and equals tive 𝑓3 averages the consecutive changes in energy amounts for

all consumer loads, i.e.,

𝑓1 𝑓 𝑓 𝑓 , − +

= E R P

𝑛



where 𝑓 represents the energy costs, 𝑓 is the rebate for the 𝑓 E R3 𝐸𝑐,𝑖 𝑐,𝑖 1 𝐸 − − recognized load reduction and P 𝑖 (𝑛 − 1) |C | 𝑓 = 1 ∑︁ ∑︁

.

is the penalty that is charged 2 = 𝑐 ∈ C

in case the recognized load reduction does not meet the require-

ments. 2.4 Weighted Sum Approach

E

The energy costs 𝑓 equal the sum of energy costs over all Since the optimal solutions to this problem appear to reside in

time intervals 𝑖 1, . . . , 𝑛 , the convex region of the objective space, we use a weighted sum ∈ { }

approach to transform all objective values into a single one. The

𝑛

∑︁

𝑓 𝑝 single objective function to be minimized thus equals E =

𝑖 𝑖 𝐸 ,

𝑖 1 =

𝑓 𝑤 1𝑓1 𝑤2 𝑓2 𝑤3𝑓3 = + +

where 𝑝𝑖 is the interval energy price.

The solution gains a rebate it the load is reduced in the demand under the condition 𝑤 1 + 𝑤2 = 1. The weight 𝑤3 can be set

period 𝑡 , . . . , 𝑡 . Note that the recognized load reduction 𝐸 , independently of 𝑤 1 { S E R }

𝑡 2

and 𝑤 and serves as a measure of limiting

𝑡 𝑡 , . . . , 𝑡 , is computed from the adjusted timetable energy the energy fluctuations. } ∈ { S E





Minimizing Costs and Risks in Demand Response Optimization Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia




3 Optimization Approach The three scenario variants differ from the basic as follows.



3.1 The first scenario variant has no demand. In the second and Setting Weights in the Weighted Sum

third scenario variant, the total energy change is set to 0 kWh

To obtain diverse solutions with the weighted sum approach, a

ensuring the reduction in energy consumption in some intervals

good strategy for setting the weights is needed. While we plan to

is matched with its increase in others. Additionally, the third

use a more sophisticated approach for this purpose in future work, =

scenario variant has no adjustment period, i.e, 𝑛 A 0.

these initial experiments were made by choosing equidistant



values of 𝑤 1 from the interval 0, 1 and defining 𝑤2 as 1 𝑤1. [ ] − 4.2 Results and Discussion

−3

In order to limit energy fluctuations, we set 𝑤 3 to 10 . Smaller

We discuss here the results of our original scenario and its three

weights proved insufficient in limiting the fluctuations while

variants. They are depicted also in plots in Figures 1 to 4, which

larger weights interfered with the first two objectives, which are

show with a black line how the consumer load changes from its

more important than the third.

planned timetable. Consumer load flexibility at each time interval



3.2 is shown in gray (there is no flexibility in the first four and last Linearization

four intervals). The demand period is denoted in red and the

Since all of the objective functions specified in Section 2.3 are

adjustment period in blue. In most cases (unless the risk has

either non-linear or contain non-linear parts, specific techniques

a large weight), the consumer reduces the load in the demand

are required to linearize these objectives and ensure the problem

period enough to meet the required demand and earn the entire

fits the MILP form. In particular, it is necessary to linearize the ab-

available rebate while not incurring any penalty. The amount of

solute value of a real variable, the product of a binary variable and

this reduction and the energy change outside of this period differ

a real variable, the minimum of two variables, along with other

for the various scenario variants.

non-linear function conditions. We use standard approaches to

achieve linearization for all these cases [9]. When the risk has a large weight, the 4.2.1 Original Scenario.

load does not change outside of the demand period (see the top

3.3 Tool and Solver plot in Figure 1). However, when the impact of risk is minimal

2 (bottom plot in Figure 1), the load is reduced everywhere except

We use the OR-Tools Python library to implement and solve

during the adjustment period. This strategy artificially increases

the single-objective MILP problem. The library is a comprehen-

the perceived load reduction to maximize the rebate, as dictated

sive tool for solving optimization problems, including linear pro-

by the rebate calculation formula.

gramming, integer programming, and combinatorial optimiza-



tion. Specifically, we use the SCIP (Solving Constraint Integer 4.2.2 Scenario Variant #1: No Demand. If the optimization is

3

Programs) solver [1] integrated within OR-Tools for solving

called without a demand, the result depends on the weighting of

MILP problem instances.

the first two objectives. As long as the impact of risk is significant

To solve a MILP problem using OR-Tools and the integrated

(top plot in Figure 2), the load does not change. Otherwise, the

SCIP solver, the following steps are performed: import the linear

load is reduced to the maximum extent in each interval (bottom

solver wrapper, declare the SCIP solver, define the variables with

plot in Figure 2). This approach minimizes the function 𝑓 , there-

𝐸

their respective bounds, set the constraints and the objective

fore reducing costs. This means that the consumer behavior can

function and lastly, analyze and display the solution.

change when optimized even if no demand is present.



4 Experiments 4.2.3 Scenario Variant #2: Zero Total Energy Change. Due to the

zero energy constraint, the consumer makes adjustments solely

We first conduct experiments using a basic scenario with a single

within the demand and adjustment periods (see Figure 3). During

consumer load. Then, we variate some parameters of this scenario

the adjustment period, the user offsets the consumption from the

to see how they affect the resulting solutions.

demand period, thereby achieving a maximal rebate. To adhere



4.1 Experimental Setup to the requirement of minimizing risks and fluctuations in other

intervals, no additional changes are made, as such actions would

The basic scenario has the following parameters:

increase the objective value.

• Time is represented as 28 15-minute intervals.

• The demand period starts at 𝑖 = 13 and ends at 𝑖 = 16. 4.2.4 Scenario Variant #3: Zero Total Energy Change and No Ad-

• T The total required reduction 𝐸 equals −8 kWh and the justment Period. When the baseline is not adjusted, the load is

required reduction 𝐸 at each interval equals 2 kWh. increased in intervals outside of the demand period, regardless D −

• The adjustment period has a duration of four intervals. whether they occur before or after it. The specific intervals when

• The load change needs to be within [−3 kWh, 3 kWh] for this happens depend on the solver and are random as they lead

each interval 𝑖 5, 6, . . . , 24 and is fixed to 0 kWh for the to the same objective function value. An example of such a case =

remaining intervals. in depicted in Figure 4.

• F

The forecast timetable energy 𝐸 is constant and equals The last two variants additionally confirm that the usage of

𝑖

12 kWh for all time intervals. the adjustment period enables exploitation – the entire rebate

• The total energy constraint is unbounded. can be gained with a smaller load reduction in the demand period

• The risk equals 0 if the load is increased in the adjustment period. . 50 for all time intervals.

• R = =

All prices are constant: 𝑝 𝑖 0.25 EUR, 𝑝 0.50 EUR and

P = 5 Conclusions

𝑝 1.00 EUR.

This paper focuses on demand response optimization and the

2

https://developers.google.com/optimization

3 growing role of prosumers in energy systems. A standard MILP

https://github.com/google/or- tools/blob/stable/ortools/linear_solver/samples/mi

p_var_array.py framework is used to set the consumer load energies within





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Mila Nedić and Tea Tušar




]

for generating a set of diverse solutions representing various

Wh

Adjustment Demand

[k period period Load flexibility

trade-offs between costs and risks.

3

2

1 By creating three scenario variants, we were able to explore

0

change

−1 the effect of some parameters on the optimization outcome. We

−2

−3 observe that:

Energy

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29

Time intervals • Regardless of the variant, the optimal load schedule does

] not deviate from the forecast one if the importance of risk

Wh

Adjustment Demand

k period period Load flexibility 𝑤 is too high, i.e., if the weight 2 is too large. This critical

[

3

2 value of 𝑤 depends on the scenario variant.

2

1

0 •

change −

If the consumer is obliged to a zero sum in load increase

1

−2 and reduction, the optimal solution uses the minimal nec-

−3

Energy essary resources to earn a rebate while avoiding excessive

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29

Time intervals energy changes.

• When the adjustment period is unspecified, the prosumer

Figure 1: Results for the original scenario with 𝑤1 = 0.6 is less likely to increase the load just before the demand

and 𝑤2 = 0.4 (top) and 𝑤1 = 0.8 and 𝑤2 = 0.2 (bottom). period.

Moving forward, we need to refine the objectives. The cur-

] rent method to assess the baseline consumption is susceptible

Wh

[ Load flexibility to exploitation and should be amended. We could calculate the k

3

2 consumer baseline from similar consumers that do not partici-

1

0 pate in DR as suggested in [6]. We will also need to revise the

change

−1

−2 penalty calculation to account for the imminent change of tariffs − 3

Energy in the Slovenian energy market. We additionally plan to improve

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29

Time intervals

the calculation of risks to ensure more robust optimization and

]

real-world applicability. Finally, we intend to develop a better

Wh

[k Load flexibility

3

strategy for setting the weights, targeting values with the most

2

1 significant impact rather than evenly distributing them.

0

change −

1

−2 Acknowledgements

−3

Energy

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29

Time intervals

The SEEDS project is co-funded by the European Union’s Horizon

Europe innovation actions programme under the Grant Agree-



Figure 2: Results for the variant without demand with 𝑤 ment n°101138211. The authors acknowledge the financial sup- 1 =

0.5 𝑤 2 0.5 𝑤1 0.7 𝑤2 0.3 port from the Slovenian Research and Innovation Agency (re-and = (top) and = and = (bottom).

search core funding No. P2-0209). The authors wish to thank

Bernard Ženko, Martin Žnidaršič and Aljaž Osojnik for helpful

]

Adjustment Demand discussions when shaping this work.

Wh

[k

period period Load flexibility

3

2 References

1

0

change

−1 [1] Tobias Achterberg. 2009. SCIP: Solving constraint integer programs. Mathe-− 2

−3 matical Programming Computation, 1, 1–41. doi: 10.1007/s12532- 008- 0001- 1.

Energy [2] 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 AEIC Load Research Committee. 2009. Demand response measurement &

verification: Applications for load research. Tech. rep. AEIC Load Research

Time intervals

Committee.

[3] István Balázs, Attila Fodor, and Attila Magyar. 2021. Quantification of the

Figure 3: Results for the variant with zero total energy flexibility of residential prosumers. Energies, 14, 4860. doi: 10.3390/en141648

change with 𝑤1 = 0.6 and 𝑤2 = 0.4. 60.

[4] Martina Capone, Elisa Guelpa, and Verda Vittorio. 2021. Multi-objective

optimization of district energy systems with demand response. , 227, Energy

120472. doi: 10.1016/j.energy.2021.120472.

]

[5] Antonio Gabaldón, Ana García-Garre, María Carmen Ruiz-Abellón, Antonio

Wh

Demand

[k period Load flexibility Guillamón, Carlos Álvarez-Bel, and Luis Alfredo Fernandez-Jimenez. 2021.

3

Improvement of customer baselines for the evaluation of demand response

2

1 through the use of physically-based load models. Utilities Policy

, 70, 101213.

0

change

−1 doi: 10.1016/j.jup.2021.101213.

−2

− [6] Joe Glass, Stephen Suffian, Adam Scheer, and Carmen Best. 2022. Demand

3

Energy

response advanced measurement methodology: Analysis of open-source

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29

baseline and comparison group methods to enable CAISO demand response

Time intervals

resource performance evaluation. Tech. rep. California Independent System

Operator (CAISO).

Figure 4: Results for the variant with zero total energy [7] Pedro L. Magalhães and Carlos Henggeler Antunes. 2020. Comparison of ther-

change and no adjustment period with mal load models for MILP-based demand response planning. In Sustainable and

𝑤 1 0.6 𝑤2 = =

Energy for Smart Cities. Springer International Publishing, Cham, 110–124.

0.4 [8] Javier Parra-Domínguez, Esteban Sánchez, and Ángel Ordóñez. 2023. The .

prosumer: A systematic review of the new paradigm in energy and sustainable

development. , 15, 13. doi: 10.3390/su151310552. Sustainability

[9] Nace Sever. 2022. Časovno razporejanje terenskih nalog z mešanim celoštevil-

their flexibility so that the costs, risks and energy fluctuations . Bachelor’s Thesis. University of Ljubljana, skim linearnim programiranjem

Faculty of Mathematics and Physics. https://repozitorij.uni- lj.si/IzpisGradiva

are all minimized. Since the objectives are scalarized with the

.php?lang=slv&id=140427.

weighted sum approach, correctly setting their weights is crucial





Predicting Hydrogen Adsorption Energies on Platinum




Nanoparticles and Surfaces with Machine Learning



Lea Gaˇspariˇ c Anton Kokalj Saˇso Dˇ zeroski

lea.gasparic@ijs.si tone.kokalj@ijs.si saso.dzeroski@ijs.si

Joˇ zef Stefan Institute, Joˇ zef Joˇ zef Stefan Institute, Joˇ zef Joˇ zef Stefan Institute, Joˇ zef

Stefan international postgraduate Stefan international postgraduate Stefan international postgraduate

schoole school school

Ljubljana, Slovenia Ljubljana, Slovenia Ljubljana, Slovenia



Abstract The mechanism of HER includes adsorbed hydrogen

The growing interest in hydrogen gas as a fuel drives re- atom (H*) as an intermediate. Consequently, the adsorp-

search into environmentally friendly hydrogen production tion energy of hydrogen is often used as a descriptor of

methods. One viable approach of obtaining hydrogen is the catalytic activity of the material [15, 21]. The most

the electrocatalysis of water, which includes the hydrogen straightforward approach to obtain the adsorption energies

evolution reaction (HER) as one of the half-reactions. In is with density-functional theory (DFT) calculations. How-

the search of highly active catalysts for the HER, machine ever, as the size of the system and the number of different

learning can be effectively utilized to develop models for adsorption sites increase, a full DFT analysis becomes com-

calculating hydrogen adsorption energy, a key descriptor of putationally unfeasible. To address this challenge, machine-

catalytic activity. In this study, we learned models for pre- learning methods can be employed to predict hydrogen

dicting hydrogen adsorption energy on platinum. We used adsorption energies based on DFT results, enabling the

various machine-learning (ML) techniques on two datasets, investigation of more complex systems [10]. For example,

one for extended surfaces and the other for nanoparticles. bimetallic nanoparticles were investigated by J¨ ager et al.

The respective results reveal that ML models for extended [8] and Zhang et al. investigated amorphous systems [20].

surfaces are more accurate than those for nanoparticles, This contribution focuses on the use of machine learning

and that the features describing the local environment are for predicting hydrogen adsorption energies on platinum

the most significant for the predictions. For surfaces, the using electronic and geometric descriptors. Two separate

coordination number is the most relevant feature, while the datasets were constructed, one for surfaces and the other

d-band center is the most important for nanoparticles. The for nanoparticles. By employing supervised learning and

ML models developed in this study lack sufficient accuracy attribute ranking, we built ML models, assessed their accu-

to provide reliable results, highlighting the need for further racy and analyzed whether the two datasets exhibit similar

investigation with additional features or larger datasets. correlations. The idea of the contribution is illustrated in

Figure 1.

Keywords



platinum, hydrogen, DFT calculations, decision trees, fea-

ture ranking



1 Introduction

A lot of scientific and societal interest is devoted to hydro-

gen fuel, which can generate electrical power by producing

water as a byproduct. One environmentally friendly method

of producing hydrogen is through the electrocatalysis of

water, where hydrogen and oxygen gases are formed. This Figure 1: Supervised machine learning and feature

process involves two reactions: oxygen and hydrogen evolu- ranking was performed for hydrogen adsorption

tion reactions. Considerable effort is being directed towards energy on platinum catalysts modeled as surfaces

improving catalysts for both reactions and understanding and nanoparticles.

the fundamental processes involved [21, 13]. In this contri-

bution, we will focus on the hydrogen evolution reaction

(HER), for which platinum is known to be a highly ac- 2 Materials and Methods



tive catalyst due to its near-optimal hydrogen adsorption 2.1 DFT Calculations and Datasets

free energy [15, 21]. However, the high cost of platinum

motivates ongoing research of alternative materials. We utilized DFT calculations to calculate hydrogen ad-

sorption energies (a target variable for ML) and electronic

Permission to make digital or hard copies of all or part of this descriptors for ML. We also utilized geometric descriptors.

work for personal or classroom use is granted without fee provided Two datasets were constructed, one for platinum nanopar-

that copies are not made or distributed for profit or commercial

advantage and that copies bear this notice and the full citation on ticles and the other for platinum surfaces.

the first page. Copyrights for third-party components of this work DFT calculations were performed with the Perdew-Burke-

must be honored. For all other uses, contact the owner/author(s). Ernzerhof (PBE) approximation [17], a plane-wave basis

© 2024 Copyright held by the owner/author(s). set, and PAW pseudopotentials [3]. Energy cutoffs were set Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

https://doi.org/10.70314/is.2024.scai.8689 to 50 and 575 Ry for wavefunctions and electron density,

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Gaˇ spariˇ c et al.





respectively. Methfessel-Paxton smearing [12] of 0.02 eV

was employed.

Pt(111), Pt(100), and Pt(110) surface slab models were

constructed with the calculated lattice parameter of bulk

Pt (3.97 ˚ A). The models of Pt(111) and Pt(100) surfaces

consist of 4 atomic layers, with the bottom layer fixed to Figure 2: Models of extended surfaces used to cal-

bulk positions, while Pt(110) has 6 atomic layers with the culate hydrogen adsorption energies. bottom two layers fixed. To achieve a greater variety of

adsorption sites, Pt(111) and Pt(100) were also modeled

with a missing-row defect. All surface models are shown in

to the size of nanoparticles were also utilized, in particular:

Figure 2. Calculations accounted for the dipole correction

the number of all atoms (𝑁all) in the nanoparticle, the

and periodic images of slabs were separated by at least

15 ˚ number of surface atoms (𝑁surf ), the maximal (𝑟max) and A of vacuum. Different sizes of surface supercells were

used, and the k-point grid for (1× minimal (𝑟min) distances from the center of the nanoparticle

1) surface unit cells of

Pt(111), Pt(100), and Pt(110) were 12 × to the surface atoms and the distance from the center of 12× 1, 11× 11×

1,

and 11 the nanoparticle to the adsorption site (𝑟 ads). The datasets × 8 × 1, respectively. For larger supercells, the number

for surfaces and nanoparticles contained 46 and 85 data

of k-points was adapted accordingly.

points, respectively.

Calculations with nanoparticles were performed with

the gamma k-point and Martyna-Tuckerman correction

for isolated systems [11]. Nanoparticles were modeled with 2.2 Machine-Learning Methods

different shapes and sizes, consisting of 3 and up to 116 The prepared datasets were analyzed using the Weka soft-

atoms. Their periodic images were separated by at least ware package [4]. The target value in both datasets is the

15 ˚ A of vacuum. All calculations were preformed with the hydrogen adsorption energy, making this a regression task.

Quantum ESPRESSO package [5]. Supervised machine learning was employed to develop mod-

The hydrogen adsorption energy was calculated as: els for predicting the target value, which were evaluated

by 10-fold cross-validation.



𝐸 1 One of the used methods is linear regression, that com- ads = 𝐸 H * − 𝐸 * − 𝐸 H 2 (1) 2 putes the linear relationship between the target value and

where 𝐸 H* is the calculated energy of optimized adsorp- the descriptors. The relevant descriptors included in the

tion system, 𝐸* is the energy of the standalone platinum equation were selected according to the M5 method [18].

system, and 𝐸H2 is the energy of the hydrogen molecule. This method iteratively removes descriptors with the small-

All performed calculations included only one adsorbed H est effect on the model until the error of the model no

atom per supercell or nanoparticle. longer decreases.

As an electronic descriptor, we used the d-band center, We also used the random forest method [7, 1] with 100

which is considered to be a good indicator of metal reac- trees of unlimited depth. With this method, multiple deci-

tivity [6]. It was obtained through DFT calculations using sion trees were constructed by selecting relevant features

the following equation: from a random subset of int(log 2(𝑚) + 1) features, where

∞ 𝑚 is the total number of features. The final values are the

∫︀ 𝑛 (𝐸)𝐸𝑑𝐸 averages of the predictions from the individual trees.

d

𝜀 −∞ To obtain an explainable ML model, we also built regres- = (2)

d ∞

∫︀ sion trees using the M5’ method [18, 19]. In this method,

𝑛d(𝐸)𝑑𝐸

−∞ trees are built by splitting the training sets according to

where attributes that maximize the standard deviation reduction. 𝐸 is the energy and 𝑛 d is the projected density of

states on d-orbitals of the atoms forming the adsorption After the trees are constructed, they are pruned to avoid

site. overfitting and smoothed to address discontinuities between

For the geometric descriptors, we determined the average the leaves. For our datasets, we used unpruned trees to

coordination number of Pt atoms forming the adsorption prevent the formation of trees that are too small and give

site, as well as the generalized coordination number (GCN) poor predictions. We also restricted tree branching to a

of the adsorption site [2], calculated as: minimum of 6 instances per leaf node for surfaces and 20

for nanoparticles to avoid overfitting the data and to ensure

𝑁𝑖 trees of sufficient size.

∑︁ CN(𝑗)



GCN(𝑖) = (3) We also performed variable importance estimation and 𝑗 CNmax =1 ranking for our selected descriptors with all data points

where 𝑖 denotes an atom or a group of atoms forming the used as a test set. To evaluate the importance of the de-

adsorption site, 𝑁𝑖 is the number of first nearest neighbors scriptors with respect to hydrogen adsorption energy, we

of 𝑖, which are denoted with 𝑗. CN(𝑗) is the coordination employed two methods: ReliefF [9] and correlation [16]. The

number of atom 𝑗 and CNmax is the maximal coordination ReliefF method is more sensitive to feature interactions

of a given site found in the bulk material. and works by calculating the distances between training in-

In addition, the type of adsorption site was used as a stances and identifying the ’nearest hit’ and ’nearest miss’.

descriptor. For extended surfaces, the coverage of H atoms, It then adjusts the weights of the differing descriptors be-

the surface area per H atom and surface type were also used tween the target and nearest instances. The correlation

for learning. For nanoparticles, some descriptors relevant method evaluates the Pearson correlation coefficient [16]

Hydrogen on platinum Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



between the features and the target variable, without ac- The regression-tree models shown in Figure 3 have lower

counting for interactions between features. It gives scores accuracy and, consequently, are less reliable.

ranging from −1 to 1, with 1 being the highest correla- The ML models could be improved by expanding the

tion score, a score of −1 indicates anti-correlation, and 0 dataset or by calculating additional descriptors. For sur-

indicates no correlation. faces, more data can be obtained through calculations on

a wider variety of surface types and by accounting for dif-

3 Results and Discussion ferent surface defects. However, expanding the dataset for



3.1 nanoparticles is limited by their size, since DFT calcula- Machine-Learning Models

tions for larger particles are computationally too demand-

Supervised machine learning was performed using linear ing. Therefore, a larger number of different smaller particles

regression, random forest, and M5’ regression tree. The can be tested instead. Using more sophisticated descriptors

obtained Pearson’s correlation coefficients and root mean such as atom-centered symmetry functions, smooth overlap

squared errors (RMSE) between true and predicted values of atomic positions and many body tensor representation

are shown in Table 1. could also improve the results, but would require different

We can observe that not all ML models provide better sampling of adsorption structures. The use of transfer learn-

RMSE values compared to those calculated with a simple ing from pre-trained models based on chemical structures

arithmetic average, referred to as the default predictor. For could also lead to significant improvements. surfaces, linear regression and random forest perform the



best and yield similar results. The regression tree model 3.2 Feature Ranking performs the worst and has higher RMSE compared to

Feature ranking was performed for both surfaces and nanopar-

the default predictor. For nanoparticles, all methods yield

ticles, with the results presented in Figure 4. The ReliefF

errors close to those of the default predictor and correlation

and correlation importance criteria provide different rank-

coefficients bellow 0.5.

ings of features. For surfaces, the coordination number is

The obtained results indicate that with the selected

identified as the most relevant descriptor, followed by the

descriptors, the hydrogen adsorption energies are more

generalized coordination number. In contrast, for nanopar-

accurately predicted on surfaces, which are simpler as com-

ticles, the d-band center is the most important descriptor.

pared to nanoparticles. Surfaces have high symmetry and

Features describing the size of different nanoparticles show

only a handful of different adsorption sites, while nanopar-

lower relevance for predictions. The most relevant features

ticles have different shapes and sizes, consist of different

in both data sets describe the local environment of the

facets, and each nanoparticle has numerous different ad-

adsorption site, indicating the local nature of adsorption.

sorption sites. This gives a huge variety of adsorption sites

The importance of the d-band center is already well-

that can make the prediction of adsorption energies harder.

documented in the literature [14], as it correlates with

Considering the best models, the obtained adsorption

energies have an error of ± the reactivity of metals. As seen from the graphs, the d- 0.13 eV for surfaces and ±

0.22 eV

band center is not so strongly correlated with the hydrogen

for nanoparticles. Due to the exponential dependence of

binding energy on surfaces. This can be attributed to the

reaction rate and adsorption energy, even a small error in

fact that on a perfectly flat surface, all surface atoms have

adsorption energy hugely affects the reaction rate. Hence,

the same d-band center. In contrast, on nanoparticles, the

the models, particularly for nanoparticles, do not provide

d-band center varies for each adsorption site because the

sufficiently accurate results for any practical use.

atoms are not equivalent. Therefore, the d-band center is

The selected ML models also provide insights into the

expected to be more relevant for nanoparticles. For the

relations between the considered features and the target

ranking based on correlation, the calculated factors for the

variable. The linear regression model for nanoparticles

includes only the d-band center and a factor for the hollow

adsorption site, whereas the equation for surfaces is more



complex. It includes adsorption site, surface type, and both

coordination numbers. This indicates that for nanoparticles,

the d-band center is the most relevant factor, while for

surfaces, geometric factors exhibit greater predictive value.



Table 1: Pearson’s correlation coefficients (CC) and

root mean squared errors (RMSE) in eV units for

all three used ML methods. For comparison, RMSE

of the default predictor is also given.



surfaces Nanoparticles

CC RMSE CC RMSE



linear regression 0.71 0.13 0.38 0.22 Figure 3: Schematic representation the obtained

random forest 0.69 0.13 0.34 0.22 random-tree models for ideal surfaces and nanopar-

M5’ decision tree 0.49 0.19 0.34 0.22 ticles. Nodes are denoted with orange and the resulting

classes are represented with turquoise circles and in-

default predictor / 0.18 / 0.23 clude the number of data points in the class.





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Gaˇ spariˇ c et al.





[3] Andrea Dal Corso. 2014. Pseudopotentials periodic table:

From H to Pu. Comput. Mater. Sci., 95, (Dec. 2014), 337–350.

(files: H.pbe-kjpaw psl.1.0.0.UPF, Pt.pbe-n-kjpaw psl.1.0.0.UPF).

doi: 10.1016/j.commatsci.2014.07.043.

[4] Eibe Frank, Mark A. Hall, and Ian H. Witten. 2016. The

WEKA Workbench. Online Appendix for ”Data Mining:

Practical Machine Learning Tools and Techniques”. Fourth

Edition. Morgan Kaufmann. https://ml.cms.waikato.ac.nz/w

eka/Witten et al 2016 appendix.pdf.

[5] Paolo Giannozzi et al. 2009. Quantum ESPRESSO: a modular

and open-source software project for quantum simulations of

materials. J. Phys: Condens. Matter, 21, 39, 395502. Code

available from http : / / www . quantum - espresso . org/. doi:

10.1088/0953-8984/21/39/395502.

[6] Bjørk Hammer and Jens K. Nørskov. 1995. Electronic factors

determining the reactivity of metal surfaces. Surf. Sci., 343,

3, (Dec. 1995), 211–220. doi: 10.1016/0039-6028(96)80007-0.

[7] Tin Kam Ho. 1995. Random decision forests. In Proceedings

of 3rd international conference on document analysis and

recognition. Vol. 1. IEEE, 278–282.

Figure 4: Variable importance scores calculated by the [8] Marc O. J. J¨ ager, Yashasvi S. Ranawat, Filippo Federici ReliefF and correlation criteria. Importance scores for Canova, Eiaki V. Morooka, and Adam S. Foster. 2020. Efficient

correlation criteria are given as absolute values. machine-learning-aided screening of hydrogen adsorption on

bimetallic nanoclusters. ACS Comb. Sci., 22, 12, (Dec. 2020),

768–781. doi: 10.1021/acscombsci.0c00102.



d-band center are negative. This indicates that a lower 1997. Overcoming the myopia of inductive learning algorithms [9] Igor Kononenko, Edvard ˇ Simec, and Marko Robnik-ˇ Sikonja.

d-band center corresponds to a higher adsorption energy with RELIEFF. Applied Intelligence, 7, 1, (Jan. 1997), 39–55.

and consequently a less reactive site, which is physically doi: 10.1023/A:1008280620621.

[10] Jin Li et al. 2023. Machine learning-assisted low-dimensional

intuitive. electrocatalysts design for hydrogen evolution reaction. Nano-

It is also interesting to note that the surface type de- Micro Lett., 15, 1, (Oct. 2023), 227–27. doi: 10.1007/s40820-

scriptor is not very relevant according to correlation, yet 023-01192-5.

[11] Glenn J. Martyna and Mark E. Tuckerman. 1999. A reciprocal

it becomes the second most important feature when other space based method for treating long range interactions in ab

descriptor are considered. This can be attributed to the initio and force-field-based calculations in clusters. J. Chem.

fact that this descriptor has the same value for all adsorp- Phys., 110, 6, (Feb. 1999), 2810–2821. doi: 10.1063/1.477923.

[12] Michael Methfessel and Anthony Thomas Paxton. 1989. High-

tion sites on the same surface. However, when combined precision sampling for brillouin-zone integration in metals.

with other descriptors, it can give additional information, Phys. Rev. B, 40, 6, (Aug. 1989), 3616–3621. doi: 10.1103

as similar adsorption sites on different surfaces can yield [13] /PhysRevB.40.3616. Bishnupad Mohanty, Piyali Bhanja, and Bikash Kumar Jena.

considerably different adsorption energies. 2022. An overview on advances in design and development

of materials for electrochemical generation of hydrogen and

4 oxygen. Mater. Today Energy, 23, (Jan. 2022), 100902. doi: Conclusion

10.1016/j.mtener.2021.100902.

We applied different ML techniques to predict the adsorp- [14] Anders Nilsson, Lars G. M. Pettersson, Bjørk Hammer, Thomas

Bligaard, Claus Hviid Christensen, and Jens K. Nørskov. 2005.

tion energy of hydrogen on platinum surfaces and nanopar- The electronic structure effect in heterogeneous catalysis.

ticles using simple geometric and electronic descriptors. Catal. Lett., 100, 3, (Apr. 2005), 111–114. doi: 10.1007/s1056

Models for predicting adsorption energy on surfaces per- 2-004-3434-9.

[15] Jens Kehlet Nørskov, Thomas Bligaard, Ashildur Logadottir,

formed better, with the linear regression and random forest John R. Kitchin, Jingguang G. Chen, Stanislav Pandelov, and

methods showing the highest correlation coefficient and Ulrich Stimming. 2005. Trends in the exchange current for

hydrogen evolution. J. Electrochem. Soc., 152, 3, (Jan. 2005),

accuracy. In contrast, predictions for nanoparticles yielded J23. doi: 10.1149/1.1856988. lower correlation coefficients and accuracy similar to the [16] Karl Pearson. 1895. Vii. note on regression and inheritance

one calculated by a default predictor. Therefore, the mod- in the case of two parents. proceedings of the royal society

of London, 58, 347-352, 240–242.

els presented in this contribution do not provide accurate [17] John P. Perdew, Kieron Burke, and Matthias Ernzerhof. 1996.

estimation of hydrogen adsorption energies. Utilizing more Generalized gradient approximation made simple. Phys. Rev.

sophisticated descriptors and larger training data sets could Lett., 77, 18, (Oct. 1996), 3865–3868. doi: 10.1103/PhysRev

Lett.77.3865.

enhance the performance of these models. [18] John R et al. Quinlan. 1992. Learning with continuous classes.

Differences between datasets are also evident in feature In 5th Australian joint conference on artificial intelligence.

ranking. For surfaces, coordination numbers are the most [19] Vol. 92. World Scientific, 343–348. Yong Wang and Ian H Witten. 1997. Inducing model trees

relevant descriptors, while for nanoparticles, the d-band for continuous classes. In Proceedings of the ninth European

center shows the highest relevance. All these relevant de- conference on machine learning number 1. Vol. 9. Citeseer,

128–137.

scriptors are related to the local environment of the adsorp- [20] Jiawei Zhang, Peijun Hu, and Haifeng Wang. 2020. Amor-

tion site, indicating that adsorption is a local phenomenon. phous catalysis: machine learning driven high-throughput

screening of superior active site for hydrogen evolution reac-

tion. J. Phys. Chem. C, 124, 19, (May 2020), 10483–10494.

References doi: 10.1021/acs.jpcc.0c00406.

[1] [21] Jing Zhu, Liangsheng Hu, Pengxiang Zhao, Lawrence Yoon Leo Breiman. 2001. Random forests. Machine Learning , 45,

1, (Oct. 2001), 5–32. Suk Lee, and Kwok-Yin Wong. 2020. Recent advances in elec- doi : 10.1023/A:1010933404324.

[2] trocatalytic hydrogen evolution using nanoparticles. Chem. Federico Calle-Vallejo, Jos´ e I. Mart´ ınez, Juan M. Garc´ ıa-

Lastra, Philippe Sautet, and David Loffreda. 2014. Fast pre- Rev., 120, 2, (Jan. 2020), 851–918. doi: 10.1021/acs.chemrev

diction of adsorption properties for platinum nanocatalysts .9b00248.

with generalized coordination numbers. Angew. Chem. Int.

Ed., 53, 32, (Aug. 2014), 8316–8319. doi: 10.1002/anie.20140

2958.



SmartCHANGE Risk Prediction Tool: Demonstrating Risk



Assessment for Children and Youth



Marko Jordan Nina Reščič Sebastjan Kramar

Jožef Stefan Institute, Jožef Stefan Institute, Jožef Stefan Institute,

Department of Intelligent Systems Department of Intelligent Systems Department of Intelligent Systems

Ljubljana, Slovenia Jožef Stefan International Ljubljana, Slovenia

marko.jordan@ijs.si Postgraduate School sebastjan.kramar@ijs.si

Ljubljana, Slovenia

nina.rescic@ijs.si



Marcel Založnik Mitja Luštrek

Jožef Stefan Institute, Jožef Stefan Institute,

Department of Intelligent Systems Department of Intelligent Systems

Ljubljana, Slovenia Jožef Stefan International

marcel.zaloznik@ijs.si Postgraduate School

Ljubljana, Slovenia

mitja.lustrek@ijs.si



Abstract healthy lifestyle can improve physical, social, and mental well-

being, especially among youth, while mitigating the risks of

Non-communicable diseases (NCDs) have become a significant

NCD-related morbidity and mortality [15], [14], [5].

public health challenge in developed countries, driven by com-

Traditionally, clinical prevention strategies for NCDs have

mon risk factors such as obesity, low physical activity, and un-

been directed at adults, as the risk factors typically become ev-

healthy lifestyle choices. Early childhood and adolescence are

ident in adulthood. However, recent evidence suggests that fo-

crucial for establishing healthy behaviours, and early interven-

cusing interventions on children and adolescents can be a more

tion can play a crucial role in preventing or delaying the onset

effective strategy for reducing NCD risk through behaviour mod-

of NCDs later in life. However, current tools for identifying high-

ification [13]. While NCDs may not appear in childhood or ado-

risk individuals are primarily designed for adults, which results

lescence, early signs can alreadexistnt. Tackling risk factors and

in missed early detection opportunities in younger populations.

promoting healthy habits during these stages can prevent or de-

The SmartCHANGE project (https://smart-change.eu/) seeks to

lay NCDs later in life [12]. Childhood and youth are also crucial

bridge this gap by developing reliable AI tools that assess risk

periods for establishing healthy lifestyle habits. Since risk fac-

factors in children and adolescents as accurately as possible while

tors for NCDs often persist from childhood into adulthood [9],

promoting optimized risk reduction strategies.

early risk assessment and reduction of risk factors can potentially

In developing the risk assessment tool, we addressed the chal-

lower the incidence of NCD. Lastly, NCDs in youth are a signifi-

lenge of merging diverse datasets, predicting missing data to cre-

cant global health challenge, with nearly one in five adolescents

ate longitudinal datasets, implementing existing validated models

worldwide being overweight or obese [1].

for diabetes (QxMD) and cardiovascular disease (SCORE2), and

Identifying high-risk individuals for future health problems

ultimately creating a simple online application to demonstrate

is essential for targeted preventive interventions. Existing tools

the functionality of the developed risk tool.

focus mainly on adults [6], for instance predicting 10-year risk of



Keywords developing cardiovascular disseased [17] or diabetes [8], missing

the opportunity to identify high-risk individuals during child-

risk tool, dataset merge, neural networks, online application

hood and adolescence, a critical period for forming lifestyle habits.

However, recognition of health risks is not a trivial task. For

1 Introduction instance, only 35% of doctors in the UK are aware of the rec-

In developed countries, non-communicable chronic diseases (NCDs) ommendations for physical activity, and only 13% can specify

have emerged as the foremost public health challenge over recent the recommended weekly duration. Moreover, more than 80% of

decades. According to the World Health Organization (WHO), parents of inactive children incorrectly believe that their children

NCDs account for more than 70% of mortality in the European are sufficiently active [4]. Developing risk prediction tools for

region [18]. Common risk factors for NCD include obesity, poor children and youth would significantly improve NCD prevention

physical fitness, and unhealthy lifestyle habits such as inadequate and promote cost-effective strategies.

physical activity, sedentary behaviour, poor nutrition, insufficient This paper presents the development of an initial demo ap-

sleep, smoking, and excessive alcohol consumption. Embracing a plication of a risk assessment tool designed for children and

adolescents in the SmartCHANGE project [3] - merging datasets,

Permission to make digital or hard copies of all or part of this work for personal predicting missing data to build longitudinal datasets, and im-

or classroom use is granted without fee provided that copies are not made or

plementing existing validated models for diabetes (QxMD) and

distributed for profit or commercial advantage and that copies bear this notice and

cardiovascular disease (SCORE2) and finally, the application de-

the full citation on the first page. Copyrights for third-party components of this

work must be honored. For all other uses, contact the owner /author(s). velopment.

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.scai.8844

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Trovato et al.





Table 1: Overview of Selected Datasets



Dataset Name SLOfit LGS YFS AFINA-TE

Country of Origin SI BE FI PT

Age Range 5 - 20 5 - 25 0 - 60 5 - 25

Longitudinal Study Yes Yes Yes No

# of Participants 280,165 17,991 3,596 1,632

# of Measurements 3,121,399 31,127 32,364 1,632

# of Variables 13 80 24 59

% of Missing Values 2.55% 16.25% 39.49% 33.53%



2 (a) Example of the datasets pre-imputation. Methodology



2.1 Datasets

To estimate the risk of non-communicable diseases in children,



ideally, one would need a dataset that tracks risk factors from a

young age (when the prediction is made) to an older age (when

these diseases typically emerge). Such comprehensive longitu-

dinal datasets would allow for accurate predictions of an indi-

vidual’s likelihood of developing a disease later in life based on

their early risk factors. However, such datasets are currently un-

available, so we must rely on a collection of partial and often

heterogeneous datasets.

In our study, we have chosen 16 types of variables that are

used by risk models SCORE2 [17] and QxMD [8]. The datasets

we were using are described in Table 1. The SLOfit program is a

school fitness monitoring initiative in Slovenia [11]. The Leuven (b) Example of the datasets post-imputation.

Growth Study (LGS)[2, 16] is a longitudinal study initiated in



1969 that evaluates physical fitness. The Cardiovascular Risk in Figure 1: The YFS dataset (blue) covers a broad range of vari-

Young Finns Study (YFS)[10], started in the late 1970s, focuses on ables across a wide age span but includes a relatively small

early cardiovascular disease risk factors. The AFINA-TE dataset number of participants. In contrast, the SLOfit dataset

[7] is part of an intervention program in Portugal designed to (green) has many participants but includes fewer variables

enhance physical fitness, activity, and nutritional knowledge over a shorter age span. In the first step, we imputed the

among children and adolescents. missing variables across the datasets (grey).



2.2 Data Imputation Through Datasets



The first step involved imputing missing values within each

dataset (see Figure 1 for representation). To guide this process,

we calculated the coverage for each variable. Initially, we used

only fully observed variables—such as height, weight, and sex—as

features in models to impute missing values for other variables.

The variables were imputed based on their coverage using ma-

chine learning on existing features. After this initial imputation

sweep, we had a complete, though potentially imperfect, dataset.

In the second sweep, we treated all columns as complete, incor-

porating the newly imputed values from the first sweep. This

allowed us to train models with a more comprehensive dataset,

improving the accuracy of the imputation. Figure 2: Longitudinal filling of the datasets.



2.3 Longitudinal Data Imputation

For instance, a vertical jump one standard deviation above the

In the second step, we employed a similar approach but focused

mean in the LGS dataset was considered equivalent to a stand-

on merging the datasets to fill the new merged dataset longitudi-

ing long jump one standard deviation above the mean in the

nally (see Figure 2 for representation). To maximize their overlap,

SLOfit dataset. After matching and standardizing the columns

we treated certain variables as equivalent—such as vertical jump

across datasets, we merged the individual datasets into a single,

from the LGS dataset and standing long jump from the SLOfit

comprehensive dataset and repeated the imputation process.

dataset.

With a merged dataset free of missing values, we built models

Since the raw values of these variables differ, we standardized

to predict attribute values at age 55—the oldest age supported

them by converting them to z-scores, which were calculated as

by our data—using values from age 14. Due to the lack of data

follows:

𝑣 𝑎𝑟 𝑖𝑎𝑏𝑙 𝑒 𝑚𝑒𝑎𝑛 covering the entire age range from 14 to 55, we approached this −

𝑧_𝑠𝑐𝑜𝑟 𝑒 . =

in two stages: predicting from age 14 to 18 and then from 18 to

𝑠𝑡 𝑎𝑛𝑑𝑎𝑟 𝑑_𝑑𝑒 𝑣𝑖𝑎𝑡 𝑖𝑜𝑛

Short title to put in the header Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia





Ind. 18 Ind. 55 Pop.

Height [cm] 3.11 3.47 1.62

Weight [kg] 4.79 13.60 10.58

SBP [mmHg] 1.46 2.39 10.91

Total cholesterol [mmol/L] 0.05 0.10 0.64

HDL [mmol/L] 0.02 0.08 0.21

LDL [mmol/L] 0.05 0.17 0.51

Smoking [1-9] 1.01 1.72 2.26

Table 2: Mean absolute error for individual forecasting to

ages 18 and 55, and for population forecasting.



Figure 3: Population-based approach using z-scores.

with greater accuracy in this approach. In the future, we may

explore combining both methods or select the more accurate one

depending on the variable.

55. The models used were simple neural networks with a single

hidden layer.

This individual forecasting approach required available data 4 Demo Application

To show the general idea of the project, we constructed a demo

for the same person from the start to the end age. However,

application implemented with Python in the Dash framework.

since we had more data available for different people of various

ages, we also explored a population-based approach to forecast In the app, a user can specify the inputs (some inputs, such as

steroid use, were fixed to make the app more concise) to the

the typical evolution of each variable. While this method is less

personalized, it is also less prone to anomalies caused by atypical models, which in turn yielded two plots which showed how

the cardiovascular and diabetes risk evolved from the currently

individuals. In the population-based approach, we again used z-

scores, assuming that each person’s z-score remains constant. For selected age up to an age of an older adult, at age 55. In a different

plot, we also showed how a risk factor chosen changes over time.

example, if someone’s blood pressure is one standard deviation

below the mean at age 14, it is assumed to stay one standard

deviation below the mean at age 55 (see Figure 3). 4.1 Risk Prediction using Demo Application

The developed demo application interface offers a dynamic tool

2.4 Risk Models for visualizing health risks based on various user-input parame-

ters used in risk models (Figure 4). By allowing users to adjust

The SCORE2 and QxMD models were used in the application

these parameters, the dashboard generates real-time projections

to assess cardiovascular disease and type 2 diabetes risk. These

of two key risk metrics: a 10-year cardiovascular risk score and

models were chosen for their validity, robustness and effective-

a 10-year risk of developing diabetes. These risks are shown in

ness in predicting these chronic conditions. By incorporating

two line graphs, illustrating how these conditions’ probability

both, healthcare practitioners can comprehensively evaluate car-

evolves with age. Additionally, the dashboard includes a feature

diometabolic risk factors, aiding in well-informed patient man-

that tracks the progression of a selected health parameter (BMI,

agement and intervention decisions.

systolic blood pressure, total cholesterol, HDL) over time, provid-

The SCORE2 model, developed by the European Society of

ing insight into how this factor might change as the individual

Cardiology, estimates the risk of cardiovascular events over ten

ages. The developed tool intuitively explains how lifestyle and

years. It calculates the risk score by incorporating variables such

physiological factors contribute to long-term health risks, offer-

as age, sex, smoking status, blood pressure, and lipid profile. Ad-

ing valuable insights for clinical decision-making and personal

ditionally, SCORE2 considers regional variations in risk factors,

health management.

providing more accurate predictions tailored to specific popula-

tions [17].

The QxMD Diabetes Risk Calculator, a comprehensive clinical 4.2 Further Development of the Application

decision support tool, is employed to evaluate the risk of devel- The current version of the demo application is developed based on

oping type 2 diabetes mellitus. This model integrates risk factors, the data and models currently available. However, there remains

including age, BMI, family history, physical activity level, and an open question regarding the specific needs and preferences of

dietary habits, to estimate an individual’s diabetes risk [8]. the medical experts who will ultimately use the final application.

To address this, we plan to present the current version to these

3 Evaluation experts and, based on their feedback, refine and enhance the

application in subsequent iterations.

Table 2 presents the cross-validated evaluation results of our

forecasting models. As anticipated, the errors in the first stage

of individual forecasting are shallow due to the relatively short 5 Conclusion

period. The emistakesin the second stage are higher but still con- The SmartCHANGE project represents a significant step toward

sidered acceptable, with the notable exceptions of weight and improving the early detection and prevention of non-communicable

smoking. We hypothesize that the high variability during puberty, diseases (NCDs) in children and youth. While the tool presented

which many adolescents experience around age 14, complicates in this paper is a demo version demonstrating some basic func-

accurate weight forecasting. In population forecasting, the errors tionalities, our future work will focus on developing a more

are generally more significant, which aligns with the less per- comprehensive web application for medical professionals and a

sonalized nature of this method. However, weight is forecasted mobile application for families. We also plan to enhance the tool

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Trovato et al.





Figure 4: The figure is a dashboard interface that allows users to input various health-related parameters and observe the

evolution of associated risks over time.



by replacing the current SCORE2 and QxMD risk models with cardiovascular risk: a report of the american college of cardiology/american

heart association task force on practice guidelines. , 129, 25, (June Circulation

more advanced models—Test2Prevent for diabetes and Healthy

2014), S49–S73.

Heart Score for cardiovascular disease—incorporating features

[7] Noelia González-Gálvez, Jose Carlos Ribeiro, and Jorge Mota. 2022. Car-

related to diet and physical activity. Additionally, the application diorespiratory fitness, obesity and physical activity in schoolchildren: the

effect of mediation. International journal of environmental research and public

will be updated to meet medical experts’ needs based on their

health, 19, 23, 16262–16270. ObjectType-Article-1. doi: 10.3390/ijerph19231

feedback. 6262.

[8] S. J. Griffin, P. S. Little, C. N. Hales, A. L. Kinmonth, and N. J. Wareham.



Acknowledgements 2000. Diabetes risk score: towards earlier detection of type 2 diabetes in general practice. Diabetes/Metabolism Research and Reviews , 16, 3, 164–171.

This work was carried out as a part of the SmartCHANGE project, [9] D. R. Jacobs, J. G. Woo, A. R. Sinaiko, S. R. Daniels, J. Ikonen, and M. Juonala.

2022. Childhood cardiovascular risk factors and adult cardiovascular events.

which received funding from the European Union’s Horizon Eu- New England Journal of Medicine, 386, 19, (May 2022), 1765–1777.

rope research and innovation program under grant agreement [10] Markus Juonala et al. 2008. Cohort profile: the cardiovascular risk in young

finns study. , 37, 6, 1220–1226. International Journal of Epidemiology

No 101080965. The SLOfit dataset for wasvided by the University

[11] Gregor Jurak et al. 2020. Slofit surveillance system of somatic and motor

of Ljubljana (courtesy of Gregor Jurak et al.), the LGS dataset was

development of children and adolescents: upgrading the slovenian sports

provided by KU Leuven (courtesy of Martine ThomThomashe educational chart. , 56, 1, Acta Universitatis Carolinae. Kinanthropologica

28–40. doi: 10.14712/23366052.2020.4.

AFINA-TE dataset was provided by the University of Porto (cour-

[12] H. C. Jr McGill, C. A. McMahan, E. E. Herderick, G. T. Malcom, R. E. Tracy,

tesy of José Ribeiro) and the the University of Turku provided and J. P. Strong. 2000. Origin of atherosclerosis in childhood and adolescence.

the YFS dataset are grateful for their support. , 72, 5, (Nov. 2000), 1307S–1315S. American Journal of Clinical Nutrition

[13] K. Pahkala, H. Hietalampi, T. T. Laitinen, J. S. Viikari, T. Rönnemaa, H. Ni-

inikoski, and et al. 2013. Ideal cardiovascular health in adolescence: effect of

References lifestyle intervention and association with vascular intima-media thickness

and elasticity (the special turku coronary risk factor intervention project

[1] P. S. Azzopardi, S. J. C. Hearps, K. L. Francis, E. C. Kennedy, A. H. Mokdad,

for children [strip] study). , 127, 18, (May 2013), 2088–2096. Circulation

N. J. Kassebaum, S. Lim, and et al. 2019. Progress in adolescent health and

[14] J. R. Ruiz, I. Cavero-Redondo, F. B. Ortega, G. J. Welk, L. B. Andersen, and

wellbeing: tracking 12 headline indicators for 195 countries and territories,

1990-2016. , 393, 10190, (Mar. 2019), 1101–1120.

Lancet V. Martinez-Vizcaino. 2016. Cardiorespiratory fitness cut points to avoid

cardiovascular disease risk in children and adolescents; what level of fitness

[2] Gaston P Beunen, Robert M Malina, Marc A Van’t Hof, Jan Simons, Michel

Ostyn, Roland Renson, and Dirk Van Gerven. 1988.

Adolescent growth and should raise a red flag? a systematic review and meta-analysis. British

motor performance: A longitudinal study of Belgian boys Journal of Sports Medicine, 50, 13, 773–779. . Human Kinetics

[15] T. J. Saunders, C. E. Gray, V. J. Poitras, J. P. Chaput, I. Janssen, P. T. Katz-

Publishers.

marzyk, and et al. 2016. Combinations of physical activity, sedentary be-

[3] SmartCHANGE Consortium. 2024. Smartchange - horizon europe project.

haviour and sleep: relationships with health indicators in school-aged chil-

Accessed: 2024-09-02. (2024). https://www.smart- change.eu/.

dren and youth. , 41, 6, (June Applied Physiology, Nutrition, and Metabolism

[4] K. Corder, E. M. van Sluijs, I. Goodyer, C. L. Ridgway, R. M. Steele, D. Bamber,

2016), 486–505.

V. Dunn, S. J. Griffin, and U. Ekelund. 2011. Physical activity awareness of

british adolescents. , 165, 3, 281–

Archives of Pediatrics Adolescent Medicine [16] Johan Simons, Gaston Beunen, Roland Renson, Albrecht L. M. Claessens,

Bernard Vanreusel, and Jos A. V. Lefevre. 1990. Growth and fitness of Flemish

289.

[5] A. García-Hermoso, R. Ramírez-Campillo, and M. Izquierdo. 2019. Is mus- girls: The Leuven Growth Study. Human Kinetics, Champaign, IL.

[17] SCORE2 working group and ESC Cardiovascular risk collaboration. 2021.

cular fitness associated with future health benefits in children and adoles-

cents? a systematic review and meta-analysis of longitudinal studies.

Sports SCORE2 risk prediction algorithms: new models to estimate 10-year risk

Medicine of cardiovascular disease in Europe. , 49, 7, (July 2019), 975–989. European Heart Journal, 42, 25, (June

2021), 2439–2454.

[6] D. C. Jr Goff, D. M. Lloyd-Jones, G. Bennett, S. Coady, R. B. D’Agostino, and

[18] World Health Organization. 2018. Global Health Estimate 2016: Deaths

R. Gibbons. 2014. American college of cardiology/american heart association

by Cause, Age, Sex, by Country and by Region, 2000-2016. World Health

task force on practice guidelines. 2013 acc/aha guideline on the assessment of

Organization.



Predicting Mental States During VR Sessions Using Sensor Data



and Machine Learning



∗

Emilija Kizhevska Mitja Luštrek

emilija.kizhevska@ijs.si mitja.lustrek@ijs.si

Jožef Stefan Institute Jožef Stefan Institute

Jožef Stefan International Postgraduate School (IPS) Jožef Stefan International Postgraduate School (IPS)

Ljubljana, Slovenia Ljubljana, Slovenia



Abstract VR creates an immersive environment that enhances empathy

by allowing users to experience different perspectives and engage

Empathy is a multifaceted concept with both cognitive and

emotionally. VR is effective for empathy training and is referred

emotional components that plays a crucial role in social

to as ’the ultimate empathy machine’ [1, 11] for various reasons:

interactions, prosocial behavior, and mental health. In our

1) Immersive Experience: Provides a strong sense of presence,

study, empathy and general arousal were induced via VR,

helping users adopt new viewpoints [15]. 2) Perspective-Taking

with physiological signals measured and ground truth collected

and Emotional Engagement: Simulates realistic scenarios to

through questionnaires. Data from over 100 participants were

provoke emotional responses and understanding [19]. 3) Empathy

collected and analyzed using multiple machine learning models

Training: Effective in healthcare, education, and diversity training

and classification algorithms to predict empathy based on

by challenging preconceptions and deepening emotional insights

physiological responses. We explored different data balancing

[16]. 4) Ethical Considerations: Ensures respectful use of VR,

techniques and labeled data in multiple ways to enhance

balancing immersive experiences with participants’ well-being

model performance. Our results show that they are effective in

[2].

detecting general arousal, empathy, and differentiating between

The objective of this study was to examine how participants’

non-empathic and empathic arousal, but the models encountered

empathy correlates with changes in their physiological metrics,

difficulties with precise emotion detection. The dataset extracted

measured using sensors such as inertial measurement unit (IMU),

at 5-second intervals and models using Random Forest and

photoplethysmograph (PPG), and electromyography (EMG).

Extreme Gradient Boosting showed the best performance. Future

Participants were immersed in 360° VR videos featuring actors

work will focus on refining emotion detection through advanced

displaying various emotions (sadness, happiness, anger, and

modeling techniques and investigating gender differences in

anxiety) and reported their empathetic experiences via brief

empathy.

questionnaires. Using data from these sensors and questionnaires,



Keywords machine learning models were developed to predict empathy

scores based on physiological responses during the VR sessions

VR, mental states, machine learning, sensor data

[9].



1 Introduction

2 Materials and Data Collection Process

Empathy is a multifaceted concept explored across various fields,

including psychology, neuroscience, and sociology. Though no 2.1 Materials and Setup for Empathy

universal definition exists, empathy is generally understood to Elicitation in VR include both cognitive (understanding another’s perspective)

To elicit empathy, we immersed participants in a 360º and 3D

and emotional (experiencing another’s feelings) components [8].

virtual environment, as VR has proven more effective than

Our research defines empathy as the ability to model others’

methods like 2D videos, workshops, or text-based exercises [8,

emotional states and respond sensitively while recognizing the

13, 17, 20]. We used videos featuring actors expressing four

self-other distinction [14].

emotions—happiness, sadness, anger, and anxiousness—without

There is no "golden standard" for measuring empathy

additional content to avoid confounding factors [2]. Recognizing

[10], with methods varying from self-report questionnaires

the impact of understanding emotional context, an audio

to psychophysiological measures like heart rate and skin

narrative version was also created, followed by a corresponding

conductance. Each method has its pros and cons, often leading

video (50-120 seconds). To ensure gender balance, we recorded

to a combination of approaches for a comprehensive assessment.

videos with two male and two female actors. Five versions were

Psychophysiological measures offer objective data but face

developed: four with narratives (two male, two female) and one

challenges due to individual variability and non-empathetic

non-narrative, where all emotions are portrayed by all actors

factors. Our study addresses these issues by using machine

without accompanying narratives. The non-narrative version

learning to directly measure empathy from physiological signals,

allows gradual transitions between emotions, making it suitable

offering a novel approach.

for participants of all linguistic backgrounds.

Additionally, a 2-minute forest video ("The Amsterdam Forest

Permission to make digital or hard copies of all or part of this work for personal

in Springtime") was included at the start to establish a relaxed

or classroom use is granted without fee provided that copies are not made or

baseline and a roller coaster video ("Official 360 POV - Yukon

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this Striker-Canada’s Wonderland") at the end to control for work must be honored. For all other uses, contact the owner /author(s).

non-empathic arousal. Both videos were sourced from YouTube.

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

Participants completed trait empathy questionnaires (QCAE)

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.scai.9356 [14] and, after each emotion-specific video, provided feedback

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia E.Kizhevska et al.





Figure 1: The best accuracies for each group of models, developed using datasets extracted at two different frequencies and

various data balancing techniques, presented for all the labeling schemes



on their empathic state (State Empathy Scale) [18], arousal and 3 Methodology



valence levels (SAM) [3], and personal distress (IRI) [5]. Each VR 3.1 Preprocessing session lasted around 20 minutes to minimize VR sickness, with

Since all the features or insights are numeric, except for the

participants viewing one of five versions.

feature "Expression/Type," which has three values—smile, frown,

Sensor data were collected using the emteqPRO system

and neutral—we applied one-hot encoding, a technique used in

attached to the Pico Neo 3 Pro Eye VR headset, including EMG

data preprocessing where categorical (non-numeric) variables

for facial muscle activation, PPG for heart rate, and IMU for head

are transformed into a numerical format. Each unique value in

motion tracking. The device uses an internal clock as well [12].

the original non-numeric feature is transformed into a separate

binary (0 or 1) feature.

Next, because missing values represent less than 1% of the total

2.2 Dataset Description data for each participant, they were filled in using the average

In this research, we used convenience sampling to recruit of each feature’s values. Scaling the values in the descriptive

participants from the general public without a specific selection features between 0 and 1 was the final step in the preprocessing

pattern. Participants were invited from various sources, including process.

Jožef Stefan Institute employees, university students, and

the general public. Invitations were sent verbally or in

writing. Data collection concluded with 105 participants, 3.2 Feature Engineering

averaging 22.43 ± 5.31 years (range 19–45), with 75.24% Since features were provided at intervals ranging from 1 second

identifying as female. Participants had diverse educational and to 500 milliseconds, we divided the data into two windows: one

professional backgrounds. Additionally, ethical clearance for of 5 seconds and one of 500 milliseconds. For each window,

this study was obtained from the Research Ethics Committee we computed features from the 22 insights across the seven

at the Faculty of Arts, University of Maribor, Slovenia modules, as well as from the features for head activity and

(No. 038-11-146/2023/13FF UM). Furthermore, written informed facial muscle electrodes, deriving a total of 108 new features,

consent was obtained from the actors prior to recording. including minimum, maximum, average, and standard deviation

The EmteqPRO system not only provides raw sensor data for each original feature or insight. Additionally, the features for

but also generates derived variables through the Emteq Emotion head activity and facial muscle electrodes were used to define

AI Engine, which utilizes data-fusion and machine learning to ’Expression/Type,’ and the time and row index were used as

analyze multimodal sensor data and assess the user’s emotional provided. However, the row index was disregarded further in the

state. This system provides a file with 29 derived features, called study.

affective insights for each recording: 7 features for heart-rate We labeled the dataset in six different ways: 1) as a binary

variability (HRV) and 3 for breathing rate; 2 features for facial classification aiming to detect empathic arousal, comparing

expressions; 4 features for arousal and 4 for valence; 1 feature for empathic parts with the forest part of the video, while excluding

facial activation; and 1 feature for facial valence. Additionally, the non-empathic content of the roller coaster video; 2) as a

head activity is tracked, reflecting the percentage of the session binary classification using the forest and roller coaster, aiming to

with head movement. Dry EMG electrodes on facial muscles such detect non-empathic arousal; 3) again, as a binary classification,

as the zygomatic, corrugator, frontalis, and orbicularis provide but including only empathic parts and the roller coaster, aiming

four more features, each representing muscle activation as a to distinguish between empathic and non-empathic arousal, and

percentage of maximum activation observed during calibration. examining the differences in physiological responses between

The data also includes the time elapsed since the start of the empathic content and non-empathic arousal-inducing content,

recording and the row index. such as the roller coaster video; 4) aiming to detect arousal in



Predicting Mental States, VR Sessions, Sensor Data and ML Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



general, regardless of whether it is empathic or non-empathic, activation of particular muscles from the calibration session,

by splitting the entire dataset into two classes: the forest and especially the zygomaticus and orbicularis muscles—were also

everything else, including empathic parts and the roller coaster; highly correlated.

5) into three classes: treating the chunks of the roller coaster and Regarding the labeling schemes, we can conclude the

forest as separate classes and grouping all the empathic parts into following: 1) We can detect empathic arousal with confusion

one class, without differentiating between the different emotions. matrices that show a relatively good distribution of correct

The goal is to distinguish among no-arousal, empathic arousal, predictions across both classes and high accuracies for most

and non-empathic arousal; 6) with the average of participants’ of the developed models; 2) We can detect non-empathic

answers to the state empathy questions for each part of the video, arousal, with almost every developed model achieving a balanced

with each part of the empathic content considered a separate accuracy higher than 60%, reaching up to 78%, and a reasonable

chunk. Additionally, there are two other classes: the forest and balance between classes, indicating satisfactory classification

the roller coaster. The aim is to detect the level of empathy performance; 3) We can even distinguish between empathic and

participants experience during the session. We also included each non-empathic arousal with balanced accuracy of 79%; 4) We can

participant’s ID, intending to later use it for model evaluation detect arousal in general, again with high accuracies and balanced

with the ’leave-one-subject-out’ technique. classes; 5) We can distinguish to some extent among no-arousal,

empathic arousal, and non-empathic arousal; 6) However, it is

4 Experiments and Results currently very challenging to detect the precise level of empathy



4.1 participants are feeling during the session using these methods, Experimental setup

and to determine whether they are empathizing by mirroring

To build models for predicting a participant’s state empathy

emotions or experiencing something different while observing

during the VR session, we used six different classification

specific emotions. The best we can detect in this regard is up

algorithms: Gaussian Naive Bayes, Stochastic Gradient Descent

to 28% balanced accuracy, with confusion matrices showing a

Classifier, K-Nearest Neighbors Classifier, Decision Tree

relatively balanced performance across multiple classes, with a

Classifier, Random Forest Classifier, and Extreme Gradient

good number of correct classifications, particularly in the more

Boosting Classifier. The balanced accuracy score was used as

frequent classes.

an evaluation metric to assess the classification models for

Regarding the two window sizes, both models showed similar

predicting participants’ state empathy. This metric evaluates the

class balance and balanced accuracy scores. However, the dataset

overall balanced accuracy of the model by calculating the average

extracted at 5-second intervals performed slightly better. Using

of recall obtained on each class. Additionally, we used a confusion

this dataset, false positives and false negatives were reduced more

matrices to evaluate the performance of the classification models

effectively. This led to more reliable classification performance,

by comparing the actual and predicted labels.

especially in terms of precision and recall, despite the smaller

For model evaluation, we used a Leave-One-Subject-Out

scale. Thus, the models developed using the 5-second interval

cross-validation setup, where each subject is a unique participant

dataset generally performed better, showing more effective

identified by their ID.

classification and fewer errors. The simpler confusion matrix

Because the labeling schemes 2, 3, 5, and 6 are not balanced

and potentially better handling of fewer classes suggest that it

(with the 80% of the majority class), we conducted four

performs better in practical terms (Figure 2, Figure 1).

experiments for each developed model: 1) applying the Synthetic

Regarding the data balancing techniques, the undersampling

Minority Over-sampling Technique (SMOTE) to create synthetic

technique never produced the best results. For the dataset

samples for the minority class to balance the dataset; 2) using

extracted at 500 ms intervals, using the SMOTE oversampling

the RandomUnderSampler (RUnderS) method to randomly select

technique and SMOTETomek yielded the best results. For the

samples from the majority class, thereby reducing their count

dataset extracted at 5-second intervals, using the entire dataset

and balancing the dataset; 3) using SMOTETomek, a combination

yielded the best results, although models developed using

of SMOTE for oversampling and Tomek links for undersampling,

SMOTETomek yielded slightly lower results in each combination

which targets both the minority and majority classes; and 4) using

of different labeling schemes.

the dataset as it is, without any undersampling or oversampling.

Regarding the classification algorithms, Gaussian Naive Bayes

performed the worst in terms of balanced confusion matrices,

4.2 Results while Random Forest Classifier and Extreme Gradient Boosting

Including models developed by six different classification performed the best across all combinations, with Random Forest

algorithms on two distinct datasets—with two different window Classifier showing slightly better results for most combinations

sizes—and utilizing four different data balancing techniques: (Figure 2, Figure 1).

undersampling, oversampling, combination techniques, and the

dataset in its original form, along with six different labeling

schemes, we obtained 288 unique confusion matrices and

corresponding accuracies for each combination. 4.3 Conclusion

We ran a correlation matrix, which revealed that the highest In this study, we define the entire plan for developing materials,

correlation with the state empathy feature was found with methods, and environments to evoke and measure the level of

the derived maximum and minimum values from the mean empathy. We started by defining the videos and the session,

heart rate, the derived maximum and minimum values from creating or selecting questionnaires for later use as ground truth,

the arousal class feature, and the average of the arousal class writing the narratives, recording the VR videos, and then editing

— the insight, which can be -1 (low), 0 (medium), or 1 (high). and preparing them for use. Additionally, we collected a dataset

The derived standard deviation, maximum, and minimum values from over 100 participants, which we filtered, preprocessed, and

from the activation—expressed as a percentage of the maximum prepared for feature engineering and analysis.

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia E.Kizhevska et al.





Figure 2: The best confusion matrices for each group of models, developed using dataset extracted at a 5s window size and

various data balancing techniques, shown for all labeling schemes



We conducted and analysed four groups of experiments, empathy-related abilities using body ownership illusions in embodied

virtual reality. , 5, 326671. Frontiers in Robotics and AI

totaling 288 combinations, where we developed models using

[3] M. M. Bradley and P. J. Lang. 1994. Measuring emotion: the self-assessment

two different window sizes, six classification algorithms, and Journal of Behavior Therapy and

manikin and the semantic differential.

three resampling techniques, with six different labeling schemes , 25, 1, 49–59. Experimental Psychiatry

[4] A. Cuevas, M. Febrero, and R. Fraiman. 2004. An anova test for functional

aimed at detecting various aspects of the dataset chunks: four

data. , 47, 1, 111–122. Computational Statistics and Data Analysis

empathetic parts, forest, and roller coaster. [5] M. H. Davis. 1980. A multidimensional approach to individual differences

The main conclusion is that we can detect arousal in general, in empathy. JSAS Catalog of Selected Documents in Psychology/American

Psychological Association, 85.

non-empathic arousal, empathy, and differentiate between

[6] A. French, M. Macedo, J. Poulsen, T. Waterson, and A. Yu. 2008. Multivariate

non-empathic and empathic arousal, as well as between relaxed analysis of variance (manova). . San Francisco State University

[7] T. K. Kim. 2015. T test as a parametric statistic. Korean Journal of

states and arousal. However, we face difficulties in detecting and

Anesthesiology, 68, 6, 540–546.

distinguishing between the precise levels of empathy during VR

[8] E. Kizhevska, F. Ferreira-Brito, T. Guerreiro, and M. Luštrek. 2022. Using

sessions using these methods and approaches. virtual reality to elicit empathy: a narrative review. , VR4Health@ MUM

19–22.

Our next steps involve refining the detection of empathy

[9] E. Kizhevska, K. Šparemblek, and M. Luštrek. 2024. Protocol of the study

levels during VR sessions by applying detailed data filtering for predicting empathy during vr sessions using sensor data and machine

and transforming it into a stationary format. Furthermore, we learning. , 19, 7, e0307385. PloS One

[10] F. F. D. Lima and F. D. L. Osório. 2011. Empathy: assessment instruments and

will develop models such as Autoregressive, Moving Average,

psychometric quality–a systematic literature review with a meta-analysis

and Extended Recurrent Moving Average, and use clustering of the past ten years. Frontiers in Psychology, 12. 781346.

[11] M. Mado, F. Herrera, K. Nowak, and J. Bailenson. 2021. Effect of virtual reality

techniques like DBSCAN and HDBSCAN. Additionally, we will

perspective-taking on related and unrelated contexts. Cyberpsychology,

extract more features from the raw data or use end-to-end neural Behavior, and Social Networking, 24, 12, 839–845.

networks. We plan to analyze gender differences in empathy [12] M. J. Magnée, B. De Gelder, H. Van Engeland, and C. Kemner. 2007. Facial

electromyographic responses to emotional information from faces and

with a t-test [7], and explore the impact of narrative context and

voices in individuals with pervasive developmental disorder. Journal of

emotions on empathic responses using ANOVA and MANOVA , 48, 11, 1122–1130. Child Psychology and Psychiatry

[4, 6]. [13] K. M. Nelson, E. Anggraini, and A. Schlüter. 2020. Virtual reality as a tool

for environmental conservation and fundraising. , 15, 4, e0223631. PloS One

[14] R. L. Reniers, R. Corcoran, R. Shryane Drake, N. M., and B. A. Völlm. 2011.

Acknowledgements The qcae: a questionnaire of cognitive and affective empathy. Journal of

personality assessment, 93, 1, 84–95. doi: doi:10.1080/00223891.2010.528484.

The part of Emilija Kizhevska was supported by the Slovenian

[15] G. Riva, J. A. Waterworth, and E. L. Waterworth. 2004. The layers of presence:

Research and Innovation Agency (ARIS) as part of the young a bio-cultural approach to understanding presence in natural and mediated

environments. , 7, 4, 402–416. CyberPsychology Behavior

researcher PhD program, grant PR-12879. The technical aspects

[16] R. O. Roswell, C. D. Cogburn, J. Tocco, J. Martinez, C. Bangeranye, J. N.

of the videos, the recording and video editing were skillfully

Bailenson, and L. Smith. 2020. Cultivating empathy through virtual reality:

conducted by Igor Djilas and Luka Komar. The actors featured in advancing conversations about racism, inequity, and climate in medicine.

Academic Medicine, 95, 12, 1882–1886.

the videos were Sara Janašković, Kristýna Šajtošová, Domen Puš,

[17] N. S. Schutte and E. J. Stilinović. 2017. Facilitating empathy through virtual

and Jure Žavbi. The questionnaires were selected and created, reality. Motivation and Emotion, 41, 708–712. the narratives were written, and the psychological aspects of the [18] L. Shen. 2010. On a scale of state empathy during message processing.

Western Journal of Communication, 74, 5, 504–524.

video creation were considered by Kristina Šparemblek.

[19] M. Slater, A. Antley, A. Davison, D. Swapp, C. Guger, C. Barker, and M. V.

Sanchez-Vives. 2006. A virtual reprise of the stanley milgram obedience

References experiments. PloS One, 1, 1, e39. doi: 10.1145/1188913.1188915.

[20] J. Stargatt, S. Bhar, T. Petrovich, J. Bhowmik, D. Sykes, and K. Burns. 2021.

[1] D. Banakou, P. D. Hanumanthu, and M. Slater. 2016. Virtual embodiment of The effects of virtual reality-based education on empathy and understanding

white people in a black virtual body leads to a sustained reduction in their of the physical environment for dementia care workers in australia: a

implicit racial bias. , 10, 226766. controlled study. , 84, 3, 1247–1257. Frontiers in Human Neuroscience Journal of Alzheimer’s Disease

[2] P. Bertrand, J. Guegan, L. Robieux, C. A. McCall, and F. Zenasni. 2018.

Learning empathy through virtual reality: multiple strategies for training





Biomarker Prediction in Colorectal Cancer Using Multiple




Instance Learning



Miljana Shulajkovska∗ Matej Jelenc

miljana.sulajkovska@ijs.si jelenc11matej@gmail.com

Jožef Stefan Institute Jožef Stefan Institute

Ljubljana, Slovenia Ljubljana, Slovenia



Jitenndra Jonnagaddala Anton Gradišek

jitendra.jonnagaddala@unsw.edu.au anton.gradisek@ijs.si

School of Population Health, Faculty of Medicine and Jožef Stefan Institute

Health Ljubljana, Slovenia

Syndey, Australia



Abstract Deep learning methods have emerged as a promising non-

Microsatellite instability (MSI) is a crucial biomarker in colorec- invasive alternative for MSI prediction by analysing whole slide

tal cancer, guiding personalised treatment strategies. The focus images (WSIs) of histopathological samples. These models can

of our paper is on evaluating how different state-of-the-art pre- detect patterns linked to MSI, eliminating the need for genetic

trained artificial intelligence models perform in extracting fea- testing. WSIs provide a comprehensive view of tumor histology,

tures on molecular and cellular oncology (MCO) study dataset offering a faster, less invasive, and more accessible means of

to predict biomarkers. In this study, we present an advanced diagnosis.

approach for MSI prediction using multiple instance learning on Integrating deep learning into clinical practice can improve

whole slide images. Our process begins with comprehensive pre- early MSI detection, personalise treatment, and reduce invasive

processing of WSIs, followed by tessellation, which breaks down procedures. WSI-based methods streamline diagnostics and en-

large images into manageable tiles. State-of-the-art feature ex- hance cancer care with accessible predictive analytics.

traction techniques are utilised on these selected tiles, employing To manage these challenges, WSIs are often divided into smaller

pretrained models to capture rich, discriminative features. Vari- regions or patches. A common method to address these issues

ous aggregation methods are applied to combine these features, is Multiple Instance Learning (MIL) [3, 8]. Due to the vast size

leading to the prediction of MSI status across the entire slide. of WSIs, computational resources can be easily overwhelmed,

We assess the performance of different pretrained models within making MIL an essential approach. MIL is a machine learning

this framework, demonstrating their effectiveness in accurately technique that operates on sets or "bags" of instances, where the

predicting MSI, with results showing an AUROC of 0.91 on the label is assigned to the entire bag rather than individual instances.

MCO dataset. Our findings underscore the potential of multiple This is particularly advantageous in WSI analysis, where labels

instance learning-based approaches in enhancing biomarker pre- such as MSI status apply to the entire slide, which is composed

diction in colorectal cancer, contributing to more targeted and of numerous smaller regions or patches.

effective treatment strategies. In this context, [4] demonstrates state-of-the-art (SOTA) re-

sults in predicting MSI in colorectal cancer. Their workflow uti-



Keywords lizes the Swin-T model on small datasets to predict MSI. First, a pretrained tissue classification model is employed to filter out multiple instance learning, whole slide images, colorectal cancer, non-tissue patches, followed by fine-tuning a pretrained model biomarker prediction to classify the remaining patches. Both intra-cohort and exter-



1 Introduction nal validation are performed. When trained on the MCO dataset

(N=1065), the model achieved a mean AUROC of 0.92 ± 0.05 for

MSI is a crucial biomarker in colorectal cancer (CRC) that indi- MSI prediction. Similarly, [11] employs a transformer-based ap-

cates defects in the DNA mismatch repair system, leading to a proach for large-scale multi-cohort evaluation, involving over

high mutation rate within tumor cells. MSI status has significant 13,000 patients for biomarker prediction, achieving a negative

clinical implications, influencing treatment decisions, particularly predictive value of over 0.99 for MSI prediction. When trained

the use of immunotherapy, and providing prognostic information. and tested only on a single cohort (MCO), the model achieved

Traditionally, MSI is determined through laboratory tests such as an AUROC of 0.85. While [4] achieved promising results on the

PCR-based assays or immunohistochemistry (IHC) on tumor tis- MCO dataset using an additional tissue classifier, we obtained

sue samples, which require invasive biopsy procedures. However, comparable performance without the need for tissue classifica-

these methods can be time-consuming, costly, and dependent on tion. On the other hand, [11] used a multicentric cohort, which

the availability of sufficient tissue samples. demands additional computational resources. In comparison to

their results on the MCO dataset, we achieved a 6% improvement

Permission to make digital or hard copies of all or part of this work for personal using a smaller dataset.

or classroom use is granted without fee provided that copies are not made or In this study, we leverage MIL to process WSIs for the pre-

distributed for profit or commercial advantage and that copies bear this notice and diction of MSI in CRC. By testing SOTA models on the MCO

the full citation on the first page. Copyrights for third-party components of this

work must be honored. For all other uses, contact the owner/author(s). dataset, we aim to assess their performance in MSI prediction

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia using MIL. This approach not only highlights the potential of MIL

© 2024 Copyright held by the owner/author(s). in processing complex, unannotated WSIs but also contributes

https://doi.org/10.70314/is.2024.scai.9705



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Miljana Shulajkovska, Matej Jelenc, Jitenndra Jonnagaddala, and Anton Gradišek



to the broader goal of improving biomarker prediction in CRC, where 𝑤 and 𝑉 are trainable parameters.

ultimately supporting more personalized and effective treatment This approach allows the model to dynamically focus on the

strategies. most relevant patches, leading to more accurate MSI predictions.

The paper is organised as follows: Section 2 outlines the meth- Another technique similiar to attention is DSMIL [7] or a dual

ods used in the pipeline, Section 3 provides a description of the stream aggregator, consisting of two branches, employing both an

data, Section 4 presents the results, and Section 5 discusses the 𝐿 ×1 instance classifier and a bag classifier. Let ℎ 𝑖 ∈ R be a feature

findings and potential directions for future work. embedding, and 𝐵 = {ℎ 0, ..., ℎ𝑛 } a bag of embeddings. The first



2 Methods stream uses an instance classifier, followed by a max-pooling

operation to obtain a score 𝑐𝑚 (𝐵) and the critical embedding ℎ𝑚.

This section outlines the pipeline for MSI prediction, as illus- The second stream aggregates the embeddings into a single bag

trated in Figure 1. The process begins with the preprocessing embedding which is then passed through a bag classifier:

of WSIs, including tessellation into smaller patches. Next, SOTA 𝑛−1 ∑︁

pretrained models are employed to extract features from these 𝑐 , ℎ 𝑏 ( 𝐵 ) = 𝑊 𝑈 𝑏 ( ℎ 𝑖𝑚)𝑣𝑖

patches. These models, trained on large and diverse datasets, 𝑖

capture rich and discriminative features crucial for accurate MSI Where 𝑊 𝑣 𝑏 is a weight vector for classification,𝑖 an information

prediction. Finally, aggregation techniques are applied to com- vector and 𝑈 is a distance measurement between an arbitrary

bine the information from the patches, enabling precise MSI embedding and the critical embedding:



concise explanation of these individual processes. 𝑈 (ℎ𝑖 , ℎ𝑚 ) = Í𝑛= 1 , 𝑞𝑚 exp status prediction for the entire slide. Each subsection provides a exp(⟨𝑞𝑖, 𝑞𝑚 ⟩)

𝑘= (⟨𝑞 ⟩) 0

𝑘



2.1 Preprocessing where is a query vector. Both 𝑞𝑖 and 𝑣𝑖 are calculated by:

WSIs are first tessellated into smaller, more manageable patches 𝑞𝑖 = 𝑊𝑞ℎ𝑖, 𝑣𝑖 = 𝑊𝑣ℎ𝑖, 𝑖 = 0, ..., 𝑛 − 1

to facilitate further processing. This step involves dividing the where 𝑊𝑞 and 𝑊𝑣 are weight matrices. The final prediction is

large images into smaller regions using the tiatoolbox presented given by:

in [9]. Non-informative tissue patches are removed to ensure the 1

analysis focuses solely on relevant tissue areas. 𝑐 𝐵 𝑐𝑚 𝐵 𝑐 𝐵 𝑏 ( ) = ( ( ) + ()) 2

Specifically, patches that are out of bounds—where only a The last approach for feature aggregation reviewed in this

portion contains actual image data and the remainder consists of paper is TransMIL, as proposed in [10], a Transformer based

padding—are discarded. Patches that consist entirely of tissue are aggregation method, which unlike the afore-mentioned methods,

retained for subsequent analysis. This preprocessing step ensures takes into account spatial information as well. By treating a

that only informative and relevant patches are used for feature bag of embeddings as a sequence of tokens, TransMIL uses a

extraction and MSI prediction. novel TPT module made up of two Transformer layers and a



2.2 Feature Extraction Methods position encoding layer, where Transformer layers are designed

for aggregating morphological information and Pyramid Position

Since only WSI-level annotations are available, several pretrained Encoding Generator (PPEG) which encodes spatial information,

feature extraction models - UNI [1], ProvGigaPath [13], Phikon followed by a multi-layer perceptron (MLP) which classifies the

[2] and CTransPath [12] - are applied to patches, removing bag.



the need for detailed patch-level labeling. These SOTA models, 2.4 MSI Classification trained on large datasets, can capture complex and discrimina-

tive features essential for accurate biomarker prediction. The The aggregation step produces a single feature vector F, which

extracted feature embeddings are then used as input for the ag- encapsulates the most informative characteristics of the entire

gregation and classification stages, laying the foundation for slide. This aggregated feature vector F is then passed through

precise MSI status prediction. For technical details about these one or more fully connected (dense) layers. These layers apply

models, see Table 1. learned weights and biases to the features to transform them



2.3 Aggregation Methods into a form that is more suitable for classification. The output of

the fully connected layer is often passed through an activation

After feature extraction, we apply aggregation techniques to function, such as a sigmoid or softmax, depending on whether

combine patch-level features into a slide-level representation. the classification task is binary (microsatellite instability MSI vs.

Traditional pooling methods like max-pooling and mean-pooling microsatellite stability MSS) or multi-class. For MSI prediction,

provide straightforward approaches. a sigmoid function is typically used, outputting a probability

However, these methods are limited by their lack of trainability. value between 0 and 1. The final output of the model is a single

In recent years, attention-based pooling or ABMIL became a probability value indicating the likelihood of the slide being MSI.

popular technique that adresses this issue [6]. ABMIL assigns a A threshold (e.g., 0.5) is applied to this probability to make a

weight binary decision. 𝛼 𝑖 to each patch’s feature vector, reflecting its importance:



∑︁ 3 Data

𝐹 𝛼 𝑓 =

𝑖 𝑖

𝑖 𝑃 ∈ For this paper the MCO study [5] was used for training and test-

The attention scores ing. The MCO study collection contains 1,500 digitized whole 𝛼 𝑖 are computed as:

slide images (WSIs) of colorectal cancer tissues. Conducted by

exp ⊤ ( tanh( )) the Molecular and Cellular Oncology (MCO) Study group from 𝑤 𝑉 𝑓𝑖

𝛼𝑖 Í = ⊤ 𝑤 ∈ 𝑘 𝑃 exp ( tanh(𝑉 𝑓𝑘 )) 1994 to 2010, this study systematically gathered tissue samples Biomarker Prediction in CRC Using MIL Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia





Figure 1: General architecture: multiple-instance learning approach.



feature extractor architecture dataset embedding size

Mass-100k: in-house histopathology slides from

MGH and BWH, and external slides from

UNI [1] ViT-large, DINOv2, 16 heads 1024

the GTEx consortium containing >100M images,

derived from >100,000 WSIs across 20 major tissue types

Prov-Path: dataset from Providence,

ProvGigaPath [13] ViT-large, DINOv2, 24 heads a large US health network comprising 28 cancer centres, 1536

consisting of 1,3B images from 171,189 WSIs

PanCancer40M: dataset from TCGA,

Phikon [2] ViT-large, iBOT combining MIM and CL covering 13 anatomic sites and 16 cancer subtypes, 768

consisting of 43,4M images from 6,093 WSIs

dataset from TCGA and PAIP,

CTransPath [12] CNN with multi-scale Swin Transformer 768

consisting of 15M images from 32,220 WSIs

Table 1: Technical details about the pretrained feature extraction models.



and clinical data from over 1,500 patients who underwent col- Three feature aggregation methods—ABMIL, DSMIL, and Trans-

orectal cancer surgery. Each slide, representing a typical tumor MIL—were applied to the extracted features to generate a single

section, is stained with Hematoxylin and eosin and scanned at representative feature for each WSI. Following aggregation, a

a 40x objective, achieving a resolution of 0.25 mpp comparable simple neural network with a sigmoid activation function and a

to an optical microscope (∼100,000 dpi). The total data size is threshold of 0.5 was used to classify MSI and MSS.

approximately 3 Terabytes, and the collection is available on the Each aggregation model was then trained for each feature

Intersect Australia RDSI Node. extraction method on each fold, with training being conducted

over 50 epochs using the AdamW optimiser and the 1-cycle

learning rate scheduler to adjust the learning rate as models



4 Results approached convergence. Binary cross-entropy (BCE) was used

as the loss function. After each epoch, model performance was

The dataset used in this study comprised 996 whole slide images evaluated on the validation set using the AUROC metric to select

(WSIs), with 242 labeled as MSI and 754 as MSS. To evaluate the best checkpoint, as most models tended to overfit toward the

the performance of various aggregation methods, models were end of training. The selected checkpoints were then tested to

trained using 5-fold cross-validation, which ensured robust train- calculate the mean AUROC across all folds.

ing and validation. To create a balanced testing set of 96 samples, Results are presented in Figure 2a. The best performance was

20% of positive (MSI) samples and an equal number of negative achieved using the DSMIL aggregation method with the ProvGi-

(MSS) samples were randomly excluded. The remaining data was gaPath feature extractor, yielding an AUROC of 0.91 ± 0.01. The

split into five equally balanced parts for cross-validation, with ABMIL method performed best with the Phikon and UNI extrac-

each fold consisting of 180 samples in the validation set and 720 tors, achieving AUROCs of 0.91 ± 0.02. Finally, the TransMIL

samples in the training set. method combined with ProvGigaPath resulted in an AUROC

WSIs were then preprocessed into bags, each containing ap- of 0.90 ± 0.01. Additionally, statistical analysis was performed,

proximately 2,000 to 4,000 patches. Each patch was then con- specifically, the Wilcoxon signed-rank test, which yielded an

verted into feature embeddings using four different feature ex- average p-value of 0.446, showing a relatively insignificant dif-

traction methods: Phikon, CTransPath, ProvGigaPath, and UNI. ference in performance of different feature extraction methods,

Specifically, CTransPath and Phikon produced embeddings with as expected. 768 features, UNI with 1024 features, and ProvGigaPath with

1536 features.

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Miljana Shulajkovska, Matej Jelenc, Jitenndra Jonnagaddala, and Anton Gradišek





(a) ABMIL (b) DSMIL (c) TransMIL





Figure 2: Predictive performance of 5-fold cross-validation of different feature extractors and aggregation methods. AUROC



plots for prediction of MSI/MSS status. The true positive rate represents sensitivity and the false negative rate represents

1-specificity. The shaded areas represent the standard deviation (SD). The value of the lower right each plot represents

mean AUROC ± SD.



5 Discussion and Conclusion better capture the complex relationships between patches within

n this study, we explored the potential of MIL combined with a WSI. Advanced methods may help in refining the prediction

SOTA pretrained models for predicting MSI in colorectal cancer. process, leading to further improvements in model performance.

Our results indicate that the approach is highly effective, achiev- Overall, our study demonstrates the potential of MIL-based ap-

ing an AUROC of 0.913 on the MCO dataset. This is a notable proaches in enhancing biomarker prediction in colorectal cancer,

achievement, particularly when compared to previous studies, paving the way for more personalized and effective treatment

such as [4] and[11], which reported AUROCs of 0.92 and 0.85, strategies.



respectively, on the same dataset. Our results not only validate References the effectiveness of our approach but also suggest that the careful

selection and combination of feature extraction and aggregation [1] Richard J Chen et al. 2024. Towards a general-purpose foundation model

for computational pathology. Nature Medicine, 30, 3, 850–862.

methods can yield improvements in predictive accuracy. [2] Alexandre Filiot, Ridouane Ghermi, Antoine Olivier, Paul Jacob, Lucas Fidon,

The positive and negative rates observed in our results reflect Alice Mac Kain, Charlie Saillard, and Jean-Baptiste Schiratti. 2023. Scaling

the model’s ability to correctly classify MSI and MSS cases. A self-supervised learning for histopathology with masked image modeling.

medRxiv, 2023–07.

high true positive rate (sensitivity) indicates the model’s pro- [3] Michael Gadermayr and Maximilian Tschuchnig. 2024. Multiple instance

ficiency in identifying MSI-positive cases, which is crucial for learning for digital pathology: a review of the state-of-the-art, limitations &

ensuring that patients who could benefit from MSI-targeted ther- [4] future potential. Computerized Medical Imaging and Graphics, 102337. Bangwei Guo, Xingyu Li, Jitendra Jonnagaddala, Hong Zhang, and Xu Steven

apies are accurately identified. Conversely, a high true negative Xu. 2022. Predicting microsatellite instability and key biomarkers in colorec-

rate (specificity) shows the model’s effectiveness in correctly clas- tal cancer from h&e-stained images: achieving sota predictive performance

with fewer data using swin transformer. arXiv preprint arXiv:2208.10495.

sifying MSS cases, thereby minimising false positives. To further [5] Nick Hawkins. 2015. MCO study whole slide image collection. (2015).

enhance the accuracy and reliability of MSI prediction, several [6] Maximilian Ilse, Jakub Tomczak, and Max Welling. 2018. Attention-based

avenues for future work are planned. deep multiple instance learning. In International conference on machine

learning. PMLR, 2127–2136.

Utilisation of the Entire Dataset: We plan to leverage the full [7] Bin Li, Yin Li, and Kevin W Eliceiri. 2021. Dual-stream multiple instance

dataset to improve the robustness of our model. Training on a learning network for whole slide image classification with self-supervised

larger dataset may help in capturing more nuanced patterns and contrastive learning. In Proceedings of the IEEE/CVF Conference on Computer

Vision and Pattern Recognition, 14318–14328.

variations, leading to even more accurate predictions. [8] Oded Maron and Tomás Lozano-Pérez. 1997. A framework for multiple-

Fine-Tuning of Pretrained Models: While we used pretrained instance learning. Advances in neural information processing systems, 10.

[9] Johnathan Pocock et al. 2022. Tiatoolbox as an end-to-end library for ad-

models without fine-tuning in this study, fine-tuning these mod- vanced tissue image analytics. Communications medicine, 2, 1, 120.

els specifically for the task of MSI prediction could further im- [10] Zhuchen Shao, Hao Bian, Yang Chen, Yifeng Wang, Jian Zhang, Xiangyang

prove their performance. Tailoring the models to our specific Ji, et al. 2021. Transmil: transformer based correlated multiple instance

learning for whole slide image classification. Advances in Neural Information

data distribution and task requirements may yield significant Processing Systems, 34, 2136–2147.

gains in accuracy. [11] Sophia J Wagner et al. 2023. Transformer-based biomarker prediction from

Incorporation of a Tissue Classifier: Since MSI is typically 9, 1650–1661. colorectal cancer histology: a large-scale multicentric study. Cancer Cell, 41,

found in tumor tissue, we plan to integrate a tissue classifier to [12] Xiyue Wang, Sen Yang, Jun Zhang, Minghui Wang, Jing Zhang, Wei Yang,

automatically remove non-tumor tissue from the analysis. This Junzhou Huang, and Xiao Han. 2022. Transformer-based unsupervised con-

trastive learning for histopathological image classification. Medical image

step should enhance the model’s focus on relevant tissue regions, analysis, 81, 102559.

potentially improving MSI prediction accuracy and speed up the [13] Hanwen Xu et al. 2024. A whole-slide foundation model for digital pathology

whole process. from real-world data. Nature, 1–8.

Development of Advanced Aggregation Methods: We also plan

to explore more sophisticated aggregation techniques that can



Feature-Based Emotion Classification Using Eye-Tracking Data



Tomi Božak Mitja Luštrek Gašper Slapničar

tb85088@student.uni- lj.si mitja.lutsrek@ijs.si gasper.slapnicar@ijs.si

Jožef Stefan Institute Jožef Stefan Institute Jožef Stefan Institute

Ljubljana, Slovenia Jožef Stefan International Ljubljana, Slovenia

Postgraduate School

Ljubljana, Slovenia



Abstract [8, 12]. We hypothesized that eye-tracking data is a valuable

modality for multi-modal emotion recognition on its own, with

The field of emotion recognition from eye-tracking data is well-

potential applications in real-world scenarios like office work,

established and offers near-real-time insights into human affec-

driving, and psychological assessments, as well as in estimat-

tive states. It is less obtrusive than some other modalities, such

ing well-being. Our motivation was to explore eye-tracker-based

as electroencephalogram (EEG), electrocardiogram (ECG) and

predictive models as an essential component in such practical

galvanic skin response (GSR), which are often used in emotion

applications.

recognition tasks. This study examined the practical feasibility of

The primary objective of our study was to validate existing

emotion recognition using an eye-tracker with a lower frequency

findings on the performance of classical ML models for emotion

than that typically employed in similar research. Using ocular

classification from eye-tracking data, using the models – Support

features, we explored the efficacy of classical machine learning

Vector Machine (SVM) and k-Nearest Neighbors (KNN) – and

(ML) models in classifying four emotions (anger, disgust, sadness,

features already explored in the literature [9, 15] as well as ex-

and tenderness) as well as neutral and “undefined“ emotions. The

ploring classifiers not so frequently used in this field – such as

features included gaze direction, pupil size, saccadic movements,

RF and XGBoost (XGB). Additionally, we aimed to explore the

fixations, and blink data. The data from the “emotional State

potential of emotion recognition at lower sampling frequencies

Estimation based on Eye-tracking database“ was preprocessed

available in most non-professional eye trackers. For the early

and segmented into various time windows, with 22 features ex-

feasibility study, we used an existing dataset, which collected

tracted for model training. Feature importance analysis revealed

data using a wearable eye-tracker but findings could possibly be

that pupil size and fixation duration were most important for

extended to high-quality unobtrusive contact-free trackers. Our

emotion classification. The efficacy of different window lengths

research also focused on understanding the impact of individual

(1 to 10 seconds) was evaluated using Leave-One-Subject-Out

features and window lengths on model performance.

(LOSO) and 10-fold cross-validation (CV). The results demon-

strated that accuracies of up to 0.76 could be achieved with 10-

fold CV when differentiating between positive, negative, and

neutral emotions. The analysis of model performance across

different window lengths revealed that longer time windows

generally resulted in improved model performance. When the 2 Related Work

data was split using a marginally personalised 10-fold CV within In literature, various physiological signals have been employed

video, the Random Forest Classifier (RF) achieved an accuracy of

for emotion recognition, with a particular focus on modalities

0.60 in differentiating between the six aforementioned emotions. such as EEG, GSR, and eye-tracking systems [1, 6, 9]. Researchers

Some challenges remain, particularly in regard to data granu- have explored both uni- and multi-modal approaches, finding that

larity, model generalization across subjects and the impact of the integration of multiple modalities can significantly enhance

downsampling on feature dynamics. emotion recognition accuracy. Lu et al. achieved 0.78 accuracy

with eye-related features recorded with eye-tracking glasses –

Keywords which are not contact-free but record at relatively low frequen-

eye-tracking, emotion recognition, machine learning cies of 60 Hz or 120 Hz. They predicted positive, negative and

neutral classes with SVM. Interestingly, they observed a 0.10 in-

1 Introduction crease in accuracy when combining eye-related and EEG features

[12]. Similarly, Guo et al. observed a more substantial gain, with

Emotion recognition is a vibrant area of research, leveraging di-

accuracy improving by 0.20 when integrating EEG, eye-tracking,

verse data sources such as images [11], audio [16], and also, ocular

and eye images, as opposed to using only eye-tracking data [7].

features like pupil dilation, gaze direction, blinks, and saccadic

The features derived from eye-tracking have been widely used

movements [3, 8, 12]. Such eye-related features provide valu-

in ML algorithms to detect emotional states [2, 7, 12, 15]. However,

able insights into emotional states, offering a less-invasive and

most studies have traditionally categorized emotions into broad

real-time approach to understanding human affective responses.

groups like positive, negative, and neutral [12, 14]. Pupil size, in

Most studies that tried to predict emotions from these eye-related

particular, has emerged as a valuable indicator for distinguishing

features relied not only on eye-tracking data but also on EEG

between positive and negative emotions [2, 7, 12] . Recent efforts

Permission to make digital or hard copies of all or part of this work for personal

have begun to refine these broad categories, identifying more

or classroom use is granted without fee provided that copies are not made or

specific emotions like happiness, sadness, fear, anger, etc. [2, 7,

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this 15]. Although current methods can effectively identify certain

work must be honored. For all other uses, contact the owner /author(s).

emotions such as sadness and fear, further research is needed to

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

reliably differentiate between others like disgust, joy, and surprise

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.scai.9988 [2].



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Tomi Božak, Mitja Luštrek, and Gašper Slapničar



3 Methodology We used 50% sliding window overlap. From each window, we



3.1 computed 22 features, belonging to the following groups: Data

In our research, we used the “emotional State Estimation based

(1) gaze coordinates on screen: std of x and y coordinates

on Eye-tracking database“ (eSEEd) [13]. The eSEEd comprises

(2) pupil ellipse sizes of a and b for each eye: mean, std

data from 48 participants, each of whom watched 10 carefully

(3) blinks: number; mean and std of duration (all 0 if no blinks)

selected videos intended to evoke specific emotional responses.

(4) saccades: number; mean speed; mean, std, total duration

After viewing each video, participants ranked their emotions

(5) fixations: number; mean, std, total duration

– anger, disgust, sadness, and tenderness – on a scale from 0

to 10. Tenderness, however, is not regarded as one of the basic

emotions, but it has been widely utilized in emotion research Saccade and, implicitly, fixation calculations were done using

in recent years [13]. Since the participants had ranked all four existing code based on the algorithm proposed by Engbert et al. [5,

emotions for every video, a labelling problem emerged when 10]. The algorithm calculates the velocity and acceleration of eye

multiple emotions shared the highest score, in our case, leading movements by using a velocity threshold identification method

to “undefined“ labels. In our study, emotions were mapped by to detect saccades based on continuous 3D gaze data. In our study

applying a set of extraction rules in the following order: if the we define fixation (interval) as an absence of a saccade (interval),

highest-ranked emotion is below four, the response is labelled as thus one fixation is declared between every two saccades (and

neutral; if multiple emotions share the highest rank, the label is before the first and after the last one).

undefined; otherwise, the emotion with the highest rank is cho- As mentioned previously, our data was imbalanced in terms

sen. The boundary of four was chosen because the original study of class distribution, namely the distributions for anger, disgust,

on eSEEd constructed this rule and we adapted it from there sadness, neutral, tenderness and undefined were 8.7%, 13.6%,

[13]. Although the initial study design aimed for an even distri- 17.5%, 25.7%, 15.8% and 18.7%, respectively. Notably, for the 1 s

bution of emotions, neutral responses dominate, representing window length, the number of windows was 67,181, whilst for

about one-fourth of the labels (depending on window length). the 10 s window, the number of instances decreased to 6,507.



3.1.1 Data Preprocessing. We have preprocessed the data to

make it more suitable for our future research and to reduce its 3.2 Experiments

size. We wanted to study the performance of data with a rela- We initially examined feature correlation matrices to identify

tively low frequency rate of 60 Hz, which is used by relatively potential correlations between features, as well as between fea-

affordable mid-tier eye-trackers, like Tobii Pro Spark. Firstly, the ture and class. Then, we compared the following classifiers from

features that were uninformative or could be misleading (e.g. the Scikit-learn library: Random Forest (RF), Support Vector Ma-

raw tracker signal and timestamps) were removed, and the fol- chines (SVM), k-Nearest Neighbors (kNN), and XGBoost (XGB)

lowing set of features was preserved: 2D screen coordinates of from the XGBoost library, as well as an ensemble method major-

gaze points (for standard deviation (std) of screen gaze coordi- ity vote of the aforementioned classifiers. We compared all results

nates), 3D coordinates of gaze points (exclusively for saccade against a baseline majority classifier. Each model was trained and

calculations), pupil sizes (a and b of the pupil ellipse), and eye tested using its default hyperparameters. To evaluate the models’

IDs (each eye has its own pupil size features). Secondly, rows performance, we implemented multiple CV techniques.

containing any NaN values were removed, as there were no large The first CV technique was Leave-One-Subject-Out (LOSO).

consecutive blocks of such rows and downsampling of the data Secondly, we implemented a marginally personalised 10-fold CV

was planned. Finally, we further downsampled the data to 60 “within video.“ In this approach, a standard 10-fold CV was per-

Hz, matching the sampling frequency of a mid-tier eye-tracker. formed where 90% of temporally sequential windows were used

However, we acknowledge that downsampling might lead to the for training and 10% for testing. The splits were done separately

loss of high-frequency information, which could be important for for each video within every subject. All the training data from

capturing subtle dynamics in gaze behaviour and pupil responses. every video was combined to train a single model, and all the

This is particularly relevant considering that recent studies, such test data was combined to evaluate the model, ensuring that

as those by Collins et al. [3] and the SEED project [4, 17], have the model was exposed to data from all subjects and videos. We

utilized data collected at much higher frequencies to preserve named the experiment “marginally“ personalised because most

these subtle dynamics. Therefore, while downsampling makes training data does not come from any single subject and is thus

the data more meaningful to our research and more computa- not very personalised. Finally, we explored a completely person-

tionally manageable, it is important to keep in mind the reduced alised 10-fold CV “within subject.“ Here, training and testing were

temporal resolution when discussing the results. done only on data of one subject. In all three CV methods, the

Following the preprocessing, window segmentation was ap- instances were never shuffled to preserve temporal and subject

plied to the data. This step is essential for analyzing temporal sequential information and to minimize overfitting.

patterns within the data, as it allows for the capture of trends We attempted to merge certain classes in a way to group

and behaviours over specific time intervals. By segmenting the negative emotions – anger, disgust, and sadness – under the cat-

data into windows, we can improve the robustness of feature egory “negative,“ while labelling tenderness as “positive.“ The

extraction and model training, enabling the detection of mean- label for neutral remained unchanged, while the undefined la-

ingful patterns that might be obscured in raw, unsegmented data. bel was changed to “negative“ because it always resulted from

Additionally, with window segmentation, the number of training multiple negative emotions scoring equally. Lastly, the feature im-

instances increases which is commonly better for learning more portances were analysed for different combinations of data splits

robust ML models and conducting rigorous evaluation. Hence, and models in order to identify potential consistently important

multiple window lengths were examined, namely: 1, 3, 5 and 10 s. features.



Feature-Based Emotion Classification Using Eye-Tracking Data Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



4 Results similarly well, with the absolute best being RF on 10 s windows

which outperformed the majority classifier by 0.05 and 0.13 for

The results described in the following subsections are summarised

accuracy and F1-score, respectively. When grouping the nega-

in Table 1.

tive emotions, we observe an absolute improvement in models’



4.1 performance, but a relative decline toward the majority classi- Feature Correlations

fier benchmark. The best model, in this case, did not surpass

The first important observation from correlation matrices was

the majority classifier in terms of accuracy, with the majority

that no output class is closely correlated to any other singular fea- ± ±

classifier achieving 0.67 0.16 accuracy and 0.61 0.16 F1-score,

ture. Secondly, we noticed some strong correlations, for example,

while SVM, the best-performing model, scored an accuracy of

a 1.0 correlation between a number of fixations and a number of ± ±

0 .64 0.13 and an F1-score of 0.63 0.12.

saccades, because one simply equals the other increased by one.

More importantly, we noticed little-to-no correlation between fea-

tures that proved to be most important in some best-performing 4.5 Feature Importances models, meaning each of these features brought some novel infor-

Following the completion of model training, we analyzed the fea-

mation to the model. The only exceptions of important features

ture importances of the best-performing models. For RF this was

being correlated are the features representing the mean size of a

calculated based on the Mean Decrease in Impurity, summing the

pupil i.e., ellipse a and b axes, which are expected to be correlated.

impurity reduction each feature contributes across all trees; and

They were correlated more than 0.8. However, we decided not to

for XGB, feature importances were calculated using the “weight“

remove any features because we assessed the feature count of 22

metric, which counts the number of times each feature is used to

to be well-balanced in relation to the number of instances.

split the data across all trees. For SVM we did not calculate feature

importances. In the completely personalised 10-fold experiments,

4.2 Leave-One-Subject-Out feature importances varied significantly across different subjects

With the goal of training a robust general model for our dataset, and even between different runs within the same subject, specif-

we first applied the LOSO CV technique. The best performance ically with RF, as the random state was not fixed. In contrast,

was achieved by RF on 10 s windows, yielding an accuracy of feature importance was notably consistent in experiments where

0.28 0.13 and an F1-score of 0.28 0.16. It outperformed the ma- models were trained on data from multiple subjects, such as in ± ± jority classifier by 0.03 in accuracy and 0.13 in F1-score. In a sub- the LOSO and the 10-fold within video, even with a variable

sequent experiment, the negative emotions were grouped. This random state of the RF model.

adjustment led to an overall increase in performance. However, The most important features of best-performing models were

with such grouping the majority classifier score also increased to those related to average pupil sizes, followed by fixation duration.

0.59 accuracy, which is the same as the best-performing model. These results partially align with those of Collins et al., who

Further analysis revealed that high accuracy mainly implied found features relating to pupil diameter and saccades statistically

the subject predominantly reported “neutral“ feelings and low significant [3].

accuracy implied little-to-no “neutral“ labels. However, not every

subject with a high “neutral“ count achieved outstanding results

and not every subject with a wide range of emotions yielded poor 5 Conclusion

results. A comparison was made between the number of windows Our research explored emotion classification using eye-tracking

in the left-out subject to their performance and no correlation data with classical ML models and hand-crafted features. The

was found. 10 s window length performed better than the shorter data was downsampled to a lower-than-standard frequency i.e.,

windows with lengths 1-5 s. We also tested longer (60 s) windows to 60 Hz, which was more realistic for consumer contact-free

and the resulting accuracies were higher than those from 10 s eye-tracker data. This made the problem harder, making it not

windows, but we evaluated that the number of instances was directly comparable with other studies working on eSEEd, but

insufficient for the results to be representative. valuable from a practical perspective.

Window segmentation significantly impacted model perfor-



4.3 Marginally Personalised 10-fold mance, with the best results constantly obtained using the largest

Cross-Validation Within Video window length. This suggests that longer observation periods

capture more comprehensive information, making smaller win-

Given that the LOSO yielded relatively poor results, the next

dows less effective for emotion classification. We hypothesize

step was to explore 10-fold CV. Experiments showed an average

that this does not transfer to realistic scenarios, as users might

accuracy of 0.60 0.07 and an F1-score of 0.60 0.08, produced ± ±

experience emotions in short bursts while being neutral for the

with RF on 10 s windows, the best-performing model. This should

majority of the time. In specifically designed cases where emo-

be compared to the results given by the majority classifier –

tion is consistently induced for longer periods of time (like our

average accuracy of 0.21 0.01 and F1-score of 0.07 0.01. With ± ±

dataset), this is more expected.

negative emotions grouped, the accuracy and F1-score raised to

The LOSO validation strategy, which tests model generaliza-

0.76 0.04 and 0.73 0.04, respectively, for the best-performing ± ±

tion across different subjects, yielded poor results. The variability

XGB on 10 s windows. The majority class classifier yielded an

in performance across subjects indicates the challenge of cap-

accuracy of 0 .66 0.02 and an F1-score of 0.52 0.02. ± ±

turing general relationships between eye features and emotions.

While both 10-fold CV approaches showed an increase in perfor-

4.4 Personalised 10-fold Cross-Validation mance, their generalizability is limited. Completely personalised

Even though 10-fold CV within video resulted in much better per- 10-fold showed worse results than the marginally personalised

formance compared to LOSO, we wanted to see the performance one presumably because of the low number of videos per emotion

of completely personalised models. All the models performed within an individual subject.



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Tomi Božak, Mitja Luštrek, and Gašper Slapničar



Table 1: Best-performing models and their corresponding results along the results of the Majority Class Classifier for the

same parameters. Window lengths are 10 s.



Settings Model F1 Majority Class Acc Majority Class F1 Model Acc

LOSO, RF . . 0.28 0.16 0.25 0.25 0.15 0.26 0 28 ± 0 13 ± ± ±

LOSO, SVM, negative emotions grouped . . 0.46 0.18 0.59 0.19 0.46 0.18 0 59 ± 0 19 ± ± ±

10-fold within video, RF . . 0.60 0.08 0.21 0.01 0.07 0.01 0 60 ± 0 07 ± ± ±

10-fold within video, XGB, negative emotions grouped . . 0.73 0.04 0.66 0.02 0.52 0.02 0 76 ± 0 04 ± ± ±

10-fold within subject, RF . . 0.42 0.19 0.33 0.26 0.29 0.26 0 38 ± 0 20 ± ± ±

10-fold within subject, SVM, negative emotions grouped . . 0.63 0.12 0.67 0.16 0.61 0.16 0 64 ± 0 13 ± ± ±



An important issue with the eSEEd data is that all participants The authors acknowledge the use of OpenAI’s ChatGPT for

watched the same 10 emotion-evoking videos in the exact same generating text suggestions during the preparation of this paper.

order. This uniformity raises concerns that, given the small num- All the generated content has been reviewed and edited by the

1

ber of videos (two intended per emotion), the models might authors to ensure accuracy and relevance to the research.

learn to associate features unrelated to emotions, such as video

dynamics or illumination. We circumvented the problem with References video dynamics by dropping the mean gaze coordinate features [1] Zeeshan Ahmad and Naimul Khan. 2022. A survey on physiological signal-

based emotion recognition. . https://www.mdpi.com/23 Bioengineering 2022

and not using them in our experiments.

06- 5354/9/11/688.

Despite these challenges, our experiments offer valuable in- [2] Aracena Claudio, Basterrech Sebastián, Snáel Václav, and Velásquez Juan.

sights into the feasibility of emotion recognition from low-frequency 2015. Neural networks for emotion recognition based on eye tracking data.

In , 2015 IEEE International Conference on Systems, Man, and Cybernetics

eye-tracker data, providing a foundation for future work. We

2632–2637. doi: 10.1109/smc.2015.460.

opted for classical models initially due to their explainability, [3] Mackenzie L. Collins and T. Claire Davies. 2023. Emotion differentiation

through features of eye-tracking and pupil diameter for monitoring well-

lower computational complexity, and efficiency, which are in our

being. In 2023 45th Annual International Conference of the IEEE Engineering in

opinion essential for understanding the data before transitioning Medicine & Biology Society (EMBC)

. doi: 10.1109/embc40787.2023.10340178.

to more complex deep learning models. [4] Ruo-Nan Duan, Jia-Yi Zhu, and Bao-Liang Lu. 2013. Differential entropy

feature for EEG-based emotion classification. In 6th International IEEE/EMBS

In future work, several enhancements could be explored to

Conference on Neural Engineering (NER). IEEE, 81–84.

improve the robustness and accuracy of emotion classification [5] Ralf Engbert, Lars Rothkegel, Daniel Backhaus, and Hans A. Trukenbrod.

models using eye-tracking data. One approach could involve ana- 2016. Evaluation of velocity-based saccade detection in the smi-etg 2w

system. Technical report, Allgemeine und Biologische Psychologie, Uni-versität

lyzing distinct fixation areas as an additional feature, potentially

Potsdam.

offering deeper insights into visual attention patterns. Moreover, [6] Atefeh Goshvarpour, Ataollah Abbasi, and Ateke Goshvarpour. 2017. An

accurate emotion recognition system using ECG and GSR signals and match-

considering that each emotion is (in some cases) represented

ing pursuit method. . doi: 10.1016/j.bj.2017.11.001. Biomed J. 2017

by two videos, a valuable experiment would be to train models

[7] Jiang-Jian Guo, Rong Zhou, Li-Ming Zhao, and Bao-Liang Lu. 2019. Multi-

on one video and test on the other. This could help assess the modal emotion recognition from eye image, eye movement and EEG using

deep neural networks. In 2019 41st Annual International Conference of the

model’s ability to generalize across different stimuli within the

IEEE Engineering in Medicine and Biology Society (EMBC), 3071–3074. doi:

same emotional category. 10.1109/embc.2019.8856563.

Further analysis could focus on demographic factors by exam- [8] Robert Jenke, Angelika Peer, and Martin Buss. 2014. Feature extraction and

selection for emotion recognition from EEG. IEEE Transactions on Affective

ining the LOSO results for potential correlations between model Computing

, 5, 3, 327–339. doi: 10.1109/taf f c.2014.2339834.

predictions and participant characteristics such as gender, age, [9] Lim Jia Zheng, Mountstephens James, and Jason Teo. 2020. Emotion recog-

nition using eye-tracking: taxonomy, review and current challenges. Sensors

and education. This might reveal underlying biases or trends that

2020. https://doi.org/10.3390/s20082384.

affect model performance. Additionally, rather than downsam-

[10] Fjorda Kazazi. 2022. Detect saccades and saccade mean velocity in python

pling and removing rows with missing data, future work could from data collected in pupil labs eye tracker. Accessed: 25. 7. 2024. https:

//www.f jordakazazi.com/detect_saccades.

explore retaining or imputing these rows.

[11] Yousif Khaireddin and Zhuofa Chen. 2021. Facial emotion recognition: state

Furthermore, exploring the training of neural networks on of the art performance on FER2013. arXiv preprint arXiv:2105.03588. doi:

raw, non-downsampled data from multiple modalities is another 10.48550/arXiv.2105.03588.

[12] Yifei Lu, Wei-Long Zheng, Binbin Li, and Bao-Liang Lu. 2015. Combining

promising direction, as other studies already observed promising

eye movements and EEG to enhance emotion recognition. In . Vol. 15. Ijcai

results with such approaches. Moreover, we should address the Buenos Aires, 1170–1176.

[13] Vasileios Skaramagkas and Emmanouil Ktistakis. 2023. Esee-d: emotional

issue of overlapping emotions which could involve developing a

state estimation based on eye-tracking dataset. , 13, 4. doi: Brain Sciences

multiclass output model, reflecting a real-world scenario where

10.3390/brainsci13040589.

multiple emotions can be present simultaneously. This approach [14] Mohammad Soleymani, Maja Pantic, and Thierry Pun. 2012. Multimodal

emotion recognition in response to videos. IEEE Transactions on Affective

could also help reduce the number of undefined labels, increasing

Computing, 3, 2, 211–223. doi: 10.1109/t- af f c.2011.37.

the amount of useful data. [15] Paweł Tarnowski, Marcin Kołodziej, Andrzej Majkowski, and Remigiusz Jan

Rak. 2020. Eye-tracking analysis for emotion recognition. Computational

Acknowledgements Intelligence and Neuroscience. https://onlinelibrary.wiley.com/doi/10.1155/2

020/2909267.

[16] Shiqing Zhang, Shiliang Zhang, Tiejun Huang, Wen Gao, and Qi Tian.

This work was supported by bilateral Weave project, funded by

2018. Learning affective features with a hybrid deep model for audio–visual

the Slovenian Agency of Research and Innovation (ARIS) under IEEE Transactions on Circuits and Systems for Video

emotion recognition.

grant agreement N1-0319 and by the Swiss National Science , 28, 10, 3030–3043. doi: 10.1109/tcsvt.2017.2719043. Technology

[17] Wei-Long Zheng and Bao-Liang Lu. 2015. Investigating critical frequency

Foundation (SNSF) under grant agreement 214991.

bands and channels for EEG-based emotion recognition with deep neural

networks. , 7, 3, 162– IEEE Transactions on Autonomous Mental Development

175. doi: 10.1109/TAMD.2015.2431497.

1

The average percentage of the videos for which the participants had reported the

target emotion (also known as the “hit rate“) was 71.8% [13].



Zbornik 27. mednarodne multikonference



INFORMACIJSKA DRUŽBA – IS 2024

Zvezek B



Proceedings of the 27th International Multiconference



INFORMATION SOCIETY – IS 2024

Volume B



Kognitivna znanost



Cognitive Science



Uredniki / Editors



Anka Slana Ozimič, Borut Trpin, Toma Strle, Olga Markič



http://is.ijs.si



10. oktober 2024 / 10 October 2024

Ljubljana, Slovenia





PREDGOVOR




Dobrodošli na konferenci Kognitivna znanost. Na letošnji konferenci bodo avtorice in avtorji

raziskovali mnoge plati človeške kognicije in predstavili tako svoje empirične ugotovitve kot

tudi teoretska raziskovanja. Poleg prispevkov s področja teme letošnje konference, “Preseganje

vrzeli v raziskovanju in razumevanju uma”, bomo potovali skozi različna področja kognitivne

znanosti - od psihologije in nevroznanosti do filozofije in umetne inteligence, ter ob tem

spoznavali raznolike tematike vključujoč uporabo VR tehnologij pri raziskovanju kognicije,

uporabo inteligentnih tehnologij pomoči pri demenci, uporabo dregljajev za spreminjanje

vedenja, vpliv umetne inteligence na demokracijo, izzive vzorčenja izkustva in manj poznane

vidike doživljanja kot so ozadnja občutja.



Konferenca se bo zaključila z okroglo mizo, na kateri bomo razmišljali o izzivih, ki jih prinašajo

vrzeli v raziskovanju in razumevanju v kognitivni znanosti: med drugim o združevanju prvo-

in tretje-osebnih pristopov k raziskovanju človeškega uma, o povezovanju različnih nivojev

opazovanja (na primer mikro-nivoja nevrološke aktivnosti z makro-nivojem vedenja in

družbenih sistemov) ter o izzivih povezovanja različnih disciplinarnih pristopov.



Upamo, da bo letošnja konferenca predstavljala prostor radovednega povezovanja in izmenjave

kreativnih idej. Skupaj bomo premagovali disciplinarne in metodološke ovire, združili mlade

in izkušene znanstvenike ter znanstvenice, ki si delijo strast do raziskovanja skrivnosti

kognicije.



Dobrodošli!



Anka Slana Ozimič

Borut Trpin

Toma Strle

Olga Markič





FOREWORD




Welcome to the Cognitive Science Conference. At this year's conference, authors will explore

the many facets of human cognition and present both their empirical findings and theoretical

research. In addition to contributions on the topic of this year's conference, Bridging the Gaps

in Research and Understanding the Mind, we will explore a diverse range of fields of cognitive

science – from psychology and neuroscience to philosophy and artificial intelligence – while

also learning about various topics, including the use of VR technologies in research, the use of

intelligent assistive technologies for dementia, the use of nudges to change behavior, the impact

of artificial intelligence on democracy, the challenges of sampling experience, and explore the

less known aspects of experience, such as background feelings.



The conference will conclude with a roundtable discussion, where we will reflect on the

challenges posed by gaps in research and understanding of mind in cognitive science: among

others, we will think about the integration of first- and third-person approaches to studying the

human mind, the relation and possible links between different levels of observation (for

example, the micro-level of neurological activity and the macro-level of behavior and social

systems), and the challenges of connecting different disciplinary approaches.



We hope that this year's conference will be a space for networking and sharing insightful ideas.

Together we will overcome disciplinary and methodological barriers, bringing together young

and experienced scientists who share a passion for exploring the mysteries of cognition.



Welcome!



Anka Slana Ozimič

Borut Trpin

Toma Strle

Olga Markič



PROGRAMSKI ODBOR / PROGRAMME COMMITTEE



Anka Slana Ozimič



Borut Trpin



Toma Strle



Olga Markič



The Ontological Shock: What Psychedelics can Teach us



about the Nature of the Mind



Maruša Sirk

Centre for Cognitive Science

University of Ljubljana

Ljubljana, Slovenia

marusasirk@gmail.com



ABSTRACT labeled as “spiritual emergencies” [10]. These include

transpersonal experiences, out-of-body experiences,

The following article provides a brief overview of the hallucinations of religious nature etc. [10].

significance and potential challenges of profound psychedelic In this paper, we will tackle the problem of the “ontological

experiences that prompt individuals to question the nature of shock” that can arise due to psychedelic experiences, how

reality, often referred to as "ontological shocks." These individuals cope with them and what implications they have on

experiences are believed to induce changes in the structure of our understanding of the mind. consciousness, leading to subsequent shifts in worldviews,

behaviors, relationships, and mental health. While they can result



in long-lasting positive changes, they are not always pleasant. 2 ONTOLOGICAL SHOCK FOLLOWING Due to the complex alterations these experiences produce, they THE USE OF PSYCHEDELICS offer a unique opportunity to explore gaps in our understanding

of the human mind and the nature of the reality it perceives, As interest in researching psychedelics for their potential

enacts, or constructs. The article aims to raise awareness of these therapeutic effects increases, there is a growing need to

issues by shedding light on various aspects of the discourse understand the mechanisms that enable these changes to occur.

surrounding this topic. Changes in metaphysical beliefs are thought to be one of the

driving mechanisms of change that enable the transformational

KEYWORDS

process to occur [8]. However, changes in metaphysical beliefs

psychedelics, ontological shock, mind, consciousness don’t come easily, as they normally induce the so-called

ontological shock about the nature and reality of existence [8].

1 This means that people start to question the nature of (their) INTRODUCTION



Psychedelics are psychoactive substances that can lead to altered commonly constituting beliefs such as animism, life after death, reality and subsequently come to adopt an altered belief system,

states of consciousness, experienced as a change in perception

the existence of alternative realities etc. [7].

and cognitive processes. Classic psychedelics, such as

On one hand, psychedelic experiences are often reported to

psilocybin, mescaline, LSD and DMT, primarily act through the

be among the most meaningful and significant experiences,

stimulation of the serotonin 5HT-2A receptor. Due to somewhat

leading to positive long-term changes [11]. On the other hand,

similar psychological effects, substances such as MDMA and

many individuals report prolonged difficulties after a profound

ketamine are also sometimes considered as psychedelics even

psychedelic experience, struggling with ontological challenges

though they target different neurological structures [1].

as they question their own reality and existence [8]. This presents

In recent years, there has been a growing interest in research

ethical challenges in both formal and informal practices for

on the potential use of psychedelic substances for mental health



treatment. There are many studies that support this claim [e. g. 2, questions about the nature of reality itself. integrating psychedelic experiences, while also raising broader 3, 4], but there is also evidence that psychedelics can lead to



longer lasting adverse effects [e. g. 1, 5]. 2.1 Coping with the ontological shock Some of the challenges that may emerge after a psychedelic Psychedelic experiences that possess mystical qualities— experience stem from profound shifts in one's worldview [6], characterized by feelings of ineffability, significance, and the metaphysical beliefs [7], and an overall ontological shock [8], in perceived “trueness” of the experience—are more likely to result which the individual begins to re-evaluate the nature of their in an ontological shock and lead to a transformed belief system reality. Commonly, people also experience a shift in their [12]. These experiences can offer profound insights into the spiritual orientation [9], due to experiences that have been

"oneness" of reality and foster a sense of "ontological comfort,"

bringing a greater sense of purpose and meaning to life.

Permission to make digital or hard copies of part or all of this work for personal or However, they can also present challenges, as individuals may

classroom use is granted without fee provided that copies are not made or

distributed for profit or commercial advantage and that copies bear this notice and struggle to integrate these insights into their everyday lives. This

the full citation on the first page. Copyrights for third-party components of this raises the question of how best to support people in making sense

work must be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia of these new ontological truth claims [12].

© 2024 Copyright held by the owner/author(s). Challenging psychedelic experiences can lead to various

https://doi.org/10.70314/is.2024.cog.1 ontological difficulties, such as questioning one's identity,



grieving the loss of a past self, experiencing persistent encounters for new sensory information [20]. This process can lead to the

with (spiritual) entities, psychotic episodes, spiritual grandiosity, dissolution of previously rigid mental models and established

feelings of meaninglessness and emptiness, isolation and despair worldviews. The resulting experience of an ontological shock

[8], derealization, delusional beliefs, and experiences of may represent a direct encounter with the fundamental

possession [9]. These prolonged challenges can impact uncertainty of reality [8]. This concept, referred to as

individuals in different ways, resulting in difficulties with "groundlessness," attempts to explain how individuals

everyday interactions, attentional problems, disruptions at work continuously construct a world of meaning that is inherently

[8], challenges in managing emotions, and perceptual difficulties without a fixed foundation and perpetually in flux [21].

[9]. It is important to note that psychedelics can also reinforce

Research on the adverse effects of psychedelics [e.g. 13, 14, existing beliefs, potentially enhancing established worldviews,

15, 16] and the challenges that arise following the so-called mental models, and expectations [22]. To address this

ontological shocks these experiences can induce [e.g. 1, 8, 9] complexity, the "Altered Beliefs Under Psychedelics" (ALBUS)

highlights the importance of developing therapeutic practices model proposes that the effects of psychedelics on prior beliefs

and guidance rooted in empathic resonance and the concept of depend on factors such as the dose consumed and the individual's

psychedelic apprenticeship [6]. While there are many existing pre-existing state of mind [22]. This model aims to explain how

therapeutic and shamanic frameworks aimed at making sense of psychedelics can both diminish and strengthen prior beliefs,

psychedelic experiences [6], it is crucial to recognize that these bridging gaps between different proposed mechanisms of how

experiences often lead to heightened suggestibility [17], which psychedelics influence reality and well-being [22].

must be considered when helping individuals navigate their Additionally, other models offer explanations for the

shifting worldviews. People have reported various strategies that mechanisms of action of psychedelics. The "Cortical-Subcortical

helped them cope with extended difficulties, such as meditation, Communication Theory" (CSCT) suggests that psychedelics

embodied contemplation, self-education through reading and reduce thalamo-cortical filtering of internal and external stimuli,

journaling, and physical exercise. In terms of support from others, allowing new, unfiltered sensory information to emerge [23].

individuals expressed a need to feel seen, heard, understood, The "Cortical-Claustrum Communication" (CCC) model posits

believed, and to have their experiences accepted and validated [1, that psychedelics decouple cortical areas from the claustrum,

8]. This suggests a responsibility for practitioners working with leading to reduced cognitive control [24]. Furthermore,

and guiding people through psychedelic experiences to help them psychedelics are thought to open a critical period for social

find meaning in their experiences without imposing any specific learning, potentially fostering new social behaviors and reducing

ontological truth claims [12]. tendencies toward isolation [25].

Extensive research on challenging psychedelic experiences, These various theories attempt to explain how psychedelics

including work by Stanislav Grof’s clinical team, has identified can facilitate the creation of new cognitive models of the world

common existential challenges that individuals often face and reality. However, the field of psychedelic research continues

following these experiences, primarily centered around fears of to grapple with understanding the precise mechanisms of these

dying, going insane, or losing control [18]. According to Grof's substances, with ongoing testing of competing theories and

team, psychedelic experiences can activate deep existential hypotheses. For a comprehensive evaluation of these theories,

concerns that only subside once the individual has successfully see [26].

processed them. It is suggested that people may encounter a sense There is also an ongoing debate about the role of subjective

of the groundlessness of being [18], which can be understood as experiences in the transformative effects of psychedelics. Some

an underlying "principle" of cognition. This brings us to the next researchers take on a reductionist approach, focusing solely on

point of this paper: the implications these experiences might have the brain mechanisms involved [27, 28]. This is problematic, as

for the scientific understanding of the mind. it opens the question of how to understand the profound

ontological shocks and the integration of the psychedelic

experience in everyday lives of individuals. It is also problematic,

3 POSSIBLE IMPLICATIONS ON as evidence suggests that psychedelic experiences with rich

UNDERSTANDING THE MIND subjective effects, such as mystical-type experiences, can lead to

the most significant transformations [29, 30]. If transformative

3.1 Understanding psychedelic experiences effects were purely mechanistic, without considering subjective

The underlying mechanisms that facilitate changes after the experiences, it becomes challenging to explain the struggles and

use of psychedelics have yet to be fully understood. However, positive changes individuals report in their daily lives after using

several hypotheses have been proposed to explain how these psychedelics.

cognitive shifts might occur. One prominent hypothesis is based Subjective experiences cannot be easily dismissed [31, 32, 33,

on the framework of predictive processing [19]. This framework 34], and they are crucial for understanding how people’s

conceptualizes the mind as a "prediction machine" that ontological reality gets altered. Investigating these subjective

continuously balances information by integrating prior aspects could help address some of the unresolved questions

knowledge about the world with incoming sensory data from the about the mind. Studying the invariants and stable states of the

environment [19]. "changing mind" following psychedelic use may bring us closer

Within this framework, the "Relaxed Beliefs Under to unlocking the nature of the mind. This research could have

Psychedelics" (REBUS) model suggests that psychedelics implications for not only understanding and treating mental

increase the brain's level of entropy—or uncertainty—while health issues but also for exploring concepts like consciousness

reducing reliance on prior beliefs, thereby allowing more room [22, 36, 37] and the self [22, 33, 35, 38]. Additionally, it could



impact the reductionist debate [32], consciousness theories, and changes in their subjective experience in their daily life, does this

discussions about the "easy" and "hard" problems of mean that the new state is an unusual or extended form of

consciousness [22, 36, 37]. consciousness? In other words, does this imply that the

individual is now living in a perpetually altered state of

3.2 Possible contribution to understanding the consciousness?

mind There is a prevalent view that the subjective experiences

In previous sections, we provided a short overview of what induced by psychedelics reveal aspects of the mind that need to

psychedelics are and sketched some possible implications they be integrated into everyday life [43]. This perspective suggests

can have in the everyday lives of people, as well as our broader that psychedelics should be considered mind-revealing rather



last section we want to finish with diving a bit deeper into some fundamental questions about the nature of consciousness itself. Is consciousness merely a byproduct of neuronal activity, possible implications that ontological shocks can have on impact they may have on our understanding of the mind. In the than merely mind-altering substances [44]. This leads us back to



of the “mind”. This is an important question as the theories of Let us stop for a moment on how we understand the concept enacted, or something else entirely? What we do know at this point is that psychedelics can induce alterations in our understanding the mind. something external waiting to be experienced, is it embodied,

psychedelic mechanisms all have their own postulates, the consciousness, affecting our awareness of ourselves and the

prevalent implicit view being that the mind is a product of world around us.

neuronal activity (which applies for previously presented While it may be ambitious to claim that psychedelic

theories – the REBUS [20], CSCT [23] and CCT model [24]). experiences will fully bridge the epistemic gap between first-

This is a reductionist view of the mind that equals the mind with person experiences and their third-person correlates, or help us

the brain [39]. Another possible view is that the mind is an understand the nature of consciousness itself – the problem we

information-processing system that manipulates and transforms commonly refer to as the hard problem of consciousness [45,



years, another understanding on the mind has slowly been issues [35, 36]. The most profound psychedelic experiences, which often lead to significant changes in consciousness, self- evolving in cognitive science – that the mind is embodied, information, which is a computational view [40]. In the previous 36]–, they can still provide valuable insights into both these

embedded, extended and enacted, which we call the 4E cognition. perception, and belief systems, may offer particularly important

This view understands the mind as a complex interplay between insights.

the brain, body and the surrounding environment [41].

The challenge of understanding the mind mirrors the debate

4 CONCLUSION

in psychedelic research about the significance of subjective

experiences. The core issue is whether the relevance of a The aim of this article was to present the concept of ontological

psychedelic experience depends solely on inducing specific brain shock following the use of psychedelics and possible

states or requires a deeper subjective experience to impact a implications on the scientific understanding of the mind. We

person's everyday life. Evidence increasingly supports the idea provided an overview of what is already known about this topic,

that both "set" (the interplay of personality, preparation, to point out where we should be cautious and what is still

expectation, and intention) and "setting" (the physical, social, unknown or vaguely known, as well as to illustrate how diving

and cultural environment) play crucial roles during a psychedelic deeper into this topic could help us scientifically advance our

experience [42]. This observation could indicate the relevance of current understanding of the mind.

the 4E cognition framework, which views the mind as a dynamic It is important to conclude this paper with a call for caution.

interplay between brain, body, and environment. Moreover, the As we tried to point out, psychedelic experiences and its

4E cognition theory might explain why set and setting are subsequent changes in everyday experience can inform us about

important, and why some psychedelic states and doses lead to the nature of our mind and help us gain broader understanding

profound changes while others do not. By exploring this intricate about topics related to consciousness, self, mental health etc. But

interplay, the 4E framework may shed light on why certain the experiences that could most inform us about these topics and

individuals experience ontological shocks under specific can lead to most profound long-term changes, have its challenges

conditions. If, however, these experiences are inexplicable and downsides, which should not be disregarded. That is why the

through existing frameworks, they could highlight gaps in our integration process, as well as the importance of set and setting,

current understanding of the mind and reveal how alternative should always be considered when dealing with these substances.

states of consciousness can disrupt the established interaction But before we have a consensus on what the mind is and how it

between mind, body, and environment. constructs our reality, a lot of damage can be done, especially if

Psychedelic experiences are often described as "altered states we want to use psychedelic substances to help people get through

of consciousness," suggesting that by examining what changes their mental health problems, as is the case in psychedelic

during these experiences, we can gain insights into what research in the past years.

constitutes the "normal," "usual," or "everyday" state of

consciousness. The concept of ontological shock, which we have REFERENCES

frequently referenced, highlights a paradox within this [1] Oliver C. Robinson, Jules Evans, David Luke, Rosalind McAlpine, Aneta

framework. If a person's everyday consciousness is altered Sahely, Amy Fisher, Stian Sundeman, Eirini Ketzitzidou Argyri, Ashleigh

Murphy-Beiner, Katrina Michelle and Ed Prideaux, 2024. Coming back

during a psychedelic experience, and they subsequently notice together: a qualitative survey study of coping and support strategies used

by people to cope with extended difficulties after the use of psychedelic



drugs. Frontiers in Psychology, 15, e1369715. DOI: [19] Andy Clark, 2013. Whatever next? Predictive brains, situated agents, and https://doi.org/10.3389/fpsyg.2024.1369715 the future of cognitive science. Behavioral and Brain Sciences, 36, 3, 181–

[2] David Nutt, David Erritzoe and Robin Carhart-Harris, 2020. Psychedelic 204. DOI https://doi.org/10.1017/S0140525X12000477

psychiatry’s brave new world. Cell, 181, 1, 24-28. DOI: [20] Robin L. Carhart-Harris and Karl J. Friston, 2019. REBUS and the https://doi.org/10.1016/j.cell.2020.03.020 Anarchic Brain: Toward a Unified Model of the Brain Action of

[3] Roland R. Griffiths, Matthew W. Johnson, Michael A. Carducci, Annie Psychedelics. Pharmacological reviews, 71, 3, 316–344. DOI:

Umbricht, William A. Rishards, Brian D. Richards, Mary P. Cosimano and https://doi.org/10.1124/pr.118.017160 Margaret A. Klinedinst, 2016. Psilocybin produces substantial and [21] Daniel Meling, 2021. Knowing Groundlessness: An Enactive Approach to sustained decreases in depression and anxiety in patients with life- a Shift From Cognition to Non-Dual Awareness. Frontiers in Psychology, threatening cancer: A randomized double-blind trial. Journal of 12, e697821. DOI:

psychopharmacology, 30, 12, 1181-1197. DOI: https://www.frontiersin.org/articles/10.3389/fpsyg.2021.697821 https://doi.org/10.1177/0269881116675513 [22] Adam Safron, Adam Safron, Arthur Juliani, Nicco Reggente, Victoria

[4] Paweł Orłowski, Anastasia Ruban, Jan Szczypiński, Justyna Hobot, Klimaj and Matthew Johnson, 2020. On The Varieties of Conscious

Maksymilian Bielecki and Michał Bola, 2022. Naturalistic use of Experiences: Altered Beliefs Under Psychedelics (ALBUS) PsyArXiv. psychedelics is related to emotional reactivity and self-consciousness: The DOI: https://doi.org/10.31234/osf.io/zqh4b mediating role of ego-dissolution and mystical experiences. Journal of [23] Franz X. Vollenweider and Katrin H. Preller, 2020. Psychedelic drugs: Psychopharmacology, 36, 8, 905-1004. DOI: neurobiology and potential for treatment of psychiatric disorders. Nature https://doi.org/10.1177/02698811221089034 Reviews Neuroscience, 21, 611–624. DOI:

[5] Michiel van Elk and Eiko I. Fried, 2023. History repeating: A roadmap to https://doi.org/10.1038/s41583-020-0367-2

address common problems in psychedelic science. Therapeutic Advances [24] Frederick S. Barrett, Samuel R. Krimmel, Roland R. Griffiths, David A. in Psychopharmacology, 13. DOI: Seminowicz and Brian N. Mathur, 2020. Psilocybin acutely alters the https://doi.org/10.1177/20451253231198466 functional connectivity of the claustrum with brain networks that support

[6] Chrostopher Timmermann, Rosalind Watts and David Dupuis, 2022. perception, memory, and attention. NeuroImage, 218, e116980. DOI:

Towards psychedelic apprenticeship: Developing a gentle touch for the https://doi.org/10.1016/j.neuroimage.2020.116980 mediation and validation of psychedelic-induced insights and revelations. [25] Romain Nardou, Eastman M. Lewis, Rebecca Rothhaas, Ran Xu, Aimei Transcultural psychiatry, 59, 5, 691-704. DOI: Yang, Edward Boyden and Gül Dölen, 2019. Oxytocin-dependent https://doi.org/10.1177/13634615221082796 reopening of a social reward learning critical period with MDMA. Nature,

[7] Sandeep M. Nayak, Manvir Singh, David B. Yaden, D. B. and Roland R. 569, 116-120. DOI: https://doi.org/10.1038/s41586-019-1075-9

Griffiths, 2022. Belief changes associated with psychedelic use. Journal [26] Michiel van Elk and David B. Yaden, 2022. Pharmacological, Neural, and of Psychopharmacology, 37, 1, 80-92. DOI: Psychological Mechanisms underlying Psychedelics: A Critical Review. https://doi.org/10.1177/02698811221131989 Neuroscience & Biobehavioral Reviews, 140, 6, e104793. DOI:

[8] Eirini K. Argyri, Jules Evans, David Luke, Pascal Michael, Katrina https://doi.org/10.1016/j.neubiorev.2022.104793

Michelle, Cyrus Rohani-Shukla, Shayam Suseelan, Ed Prideaux, Rosalind [27] David E. Olson, 2021. The Subjective Effects of Psychedelics May Not McAlpine, Ashleigh Murphy-Beiner and Oliver Robinson, 2024. Be Necessary for Their Enduring Therapeutic Effects. ACS Pharmacology Navigating Groundlessness: An interview study on dealing with & Translational Science, 4, 2, 563–567. DOI: ontological shock and existential distress following psychedelic https://doi.org/10.1021/acsptsci.0c00192 experiences. Available at [28] Lindsay P. Cameron, Robert J. Tombari, Ju Lu, Alexander J. Pell, Zefan SSRN: https://ssrn.com/abstract=4817368 or http://dx.doi.org/10.2139/ss Q. Hurley, Yann Ehinger, Maxemiliano V. Vargas, Matthew N.

rn.4817368 McCarroll, Jack C. Taylor, Douglas Myers-Turnbull, Taohui Liu, Bianca

[9] Jules Evans, Oliver C. Robinson, Eirini Ketzitzidou Argyri, Shayam Yaghoobi, Lauren J. Laskowski, Emilie I. Anderson, Guoliang Zhang,

Suseelan, Ashleigh Murphy-Beiner, Rosalind McAlpine, David Luke, Jayashri Viswanathan, Brandon M. Brown, Michelle Tija, Lee E. Dunlap,

Katrina Michelle and Ed Prideaux, 2023. Extended difficulties following Zachary T. Rabow, Oliver Fiehn, Heike Wuff, John D. McCorvy, Pamela

the use of psychedelic drugs: A mixed methods study. PLoS ONE, 18, 10, J. Lein, David Kokel, Dorit Ron, Jamie Peters, Yi Zuo and David E. Olson,

e0293349. DOI: https://doi.org/10.1371/ journal.pone.0293349 2020. A non-hallucinogenic psychedelic analogue with therapeutic

[10] Christina Grof and Stanislav Grof, 2017. Spiritual emergency: the potential. Nature, 589, 474-479. DOI: https://doi.org/10.1038/s41586-

understanding and treatment of transpersonal crises. International Journal 020-3008-z

of Transpersonal Studies, 36, 30–43. DOI: [29] Kwonmok Ko, Gemma Knight, James J. Rucker and Anthony J. Cleare, https://doi.org/10.24972/ijts.2017.36.2.30 2022. Psychedelics, Mystical Experience, and Therapeutic Efficacy: A

[11] Frederick S. Barrett and Roland R. Griffiths, 2018. Classic Hallucinogens Systematic Review. Frontiers in Psychiatry, 13, e917199. DOIi:

and Mystical Experiences: Phenomenology and Neural Correlates. https://doi.org10.3389/fpsyt.2022.917199 Current Topics in Behavioral Neurosciences, 36, 393-430. DOI: [30] Mazen A. Atiq. Matthew R. Baker, Jennifer L. Vande Voort, Maxemiliano https://doi.org/10.1007/7854_2017_474 V. Vargas and Doo-Sup Choi, 2024. Disentangling the acute subjective

[12] Joost J. Breeksema and Michiel van Elk, 2021. Working with weirdness: effects of classic psychedelics from their enduring therapeutic

a response to ‘moving past mysticism in psychedelic science. ACS properties. Psychopharmacology. DOI: https://doi.org/10.1007/s00213-Pharmacology & Translational Science, 4, 4, 1471–1474, DOI: 024-06599-5

https://doi.org/10.1021/acsptsci.1c00149 [31] David B. Yaden and Roland R. Griffiths, 2020. The Subjective Effects of

[13] Joost J. Breeksema, Bouwe W. Kuin, Jeanine Kamphuis, Wim van den Psychedelics Are Necessary for Their Enduring Therapeutic Effects. ACS

Brink, Eric Vermetten, Robert A. Schoevers, 2022. Adverse events in pharmacology & translational science, 4, 2, 568-572. DOI: clinical treatments with serotonergic psychedelics and MDMA: A mixed- https://doi.org/10.1021/acsptsci.0c00194 methods systematic review. Journal of Psychopharmacology, 36, 10, [32] Gerhard Gründer, Manuela Brand, Lea J. Mertens, Henrik Jungaberle, 1100-1117. DOI: https://doi.org/10.1177/02698811221116926 Laura Kärtner, Dennis J. Scharf, Moritz Spangemacher and Max Wolff,

[14] Anne K. Schla, Jacob Aday, Iram Salam, Jo C. Neill, David J. Nutt, 2022. 2024. Treatment with psychedelics is psychotherapy: beyond

Adverse effects of psychedelics: From anecdotes and misinformation to reductionism. Lancet Psychiatry, 11, 3, 231-236. DOI: systematic science. Journal of Psychopharmacology, 36, 3, 258-272. DOI: https://doi.org/10.1016/S2215-0366(23)00363-2 https://doi.org/10.1177/02698811211069100 [33] Riccardo Miceli McMillan and Anthony Vincent Fernandez, 2023.

[15] Rebeca Bremler, Nancy Katati, Parvinder Shergill, David Erritzoe and Understanding subjective experience in psychedelic-assisted

Robin L. Carhart-Harris, 2023. Case analysis of long-term negative psychotherapy: The need for phenomenology. Australian & New Zealand psychological responses to psychedelics. Scientific Reports, 13, e15998. Journal of Psychiatry, 57, 6, 783-788. DOI: DOI: https://doi.org/10.1038/s41598-023-41145-x https://doi.org/10.1177/00048674221139962

[16] Daniel Meling, Rebecca Ehrenkranz, Sandeep M. Nayak, Helena D. [34] Tomislav Majić, Timo T. Schmidt and Jürgen Gallinat, 2015. Peak

Aicher, Xaver Funk, Michiel van Elk, Marianna Graziosi, Prisca R. Bauer, experiences and the afterglow phenomenon: When and how do therapeutic

Milan Scheidegger and David B. Yaden, 2024. Mind the Psychedelic effects of hallucinogens depend on psychedelic experiences? Journal of Hype: Characterizing the Risks and Benefits of Psychedelics for Psychopharmacology, 29, 3, 241-253. DOI: Depression. Psychoactives, 3, 215-234. https://doi.org/10.1177/0269881114568040

https://doi.org/10.3390/psychoactives3020014 [35] Christopher Timmermann, Prisca R. Bauer, Olivia Gosseries, Audrey

[17] David Dupuis, 2021. Psychedelics as Tools for Belief Transmission. Set, Vanhaudenhuyse, Franz Vollenweider, Steven Laureys, Tania Singer;

Setting, Suggestibility, and Persuasion in the Ritual Use of Hallucinogens. Mind and Life Europe (MLE) ENCECON Research Group; Elena

Fronters in Psychology, 12, e730031. DOI: Antonova, Antoine Lutz, 2022. A neurophenomenological approach to https://doi.org/10.3389/fpsyg.2021.730031 non-ordinary states of consciousness: hypnosis, meditation, and

[18] Andrew Carnahan, 2023. An existential approach to integrating psychedelics. Trends in Cognitive Sciences, 27, 2, 139-159. DOI:

challenging psychedelic experiences. Existential Analysis: Journal of the https://doi.org/10.1016/j.tics.2022.11.006 Society for Existential Analysis, 34, 1, 89-102. [36] David B. Yaden, Matthew W. Johnson, Roland R. Griffiths, Manoj K.

Doss, Albert Garcia-Romeu, Sandeep Nayak, Natalie Gukasyan, Brian N.

Mathur and Frederick S. Barrett, 2021. Psychedelics and Consciousness:



Distinctions, Demarcations, and Opportunities. International Journal of [42] Ido Hartogsohn, 2017. Constructing drug effects: A history of set and Neuropsychopharmacology, 24, 8, 615–623. DOI: setting. Drug Science, Policy and Law, 3. DOI: https://doi.org/10.1093/ijnp/pyab026 https://doi.org/10.1177/2050324516683325

[37] Sidath Rankaduwa and Adrian M. Owen, 2023. Psychedelics, entropic [43] Collin M. Reiff, Elon E. Richman, Charles B. Nemeroff, Linda L.

brain theory, and the taxonomy of conscious states: a summary of debates Carpenter, Alik S. Widge, Carolyn I. Rodriguez, Ned H. Kalin, William

and perspectives. Neuroscience of Consciousness, 2023, 1, niad001. M. McDonald and and the Work Group on Biomarkers and Novel

DOI: https://doi.org/10.1093/nc/niad001 Treatments, a Division of the American Psychiatric Association Council

[38] Benjamin Hearn, 2021. Psychedelics, Mystical Experiences, and Meaning of Research, 2020. Psychedelics and Psychedelic-Assisted Psychotherapy.

Making: A Renegotiation Process With the Challenges of Existence. American Journal of Psychiatry, 177, 5, 365-468. DOI: Journal of Humanistic Counseling, 60, 180-196. DOI: https://doi.org/10.1176/appi.ajp.2019.19010035 https://doi.org/10.1002/johc.12164 [44] Aidan Lyon, 2023. Psychedelic Experience: Revealing the Mind. Oxford

[39] Francis Crick, 1994. The Astonishing Hypothesis. Scribners, New York. University Press.

[40] Patricia Smith Churchland and Terrence J. Sejnowski, 1992. The [45] David J. Chalmers, 1995. Facing up to the problem of consciousness.

Computational Brain. MIT Press, Cambridge, MA. Journal of Consciousness Studies, 2, 3, 200-219.

[41] Michael L. Anderson, 2003. Embodied Cognition: A field guide. Artificial

Intelligence, 149, 91–130. DOI: https://doi.org/10.1016/S0004-

3702(03)00054-7

Innovative Supporting Approaches:



Integrating Bibliotherapy, Psychodrama and AI as a



Therapeutic Conversational Tool



Veronika Mattová† FirstName Surname Department Courtney Lazore

Department of Mongolian, Name Independent Researcher

Korean and Vietnamese Studies Institution/University Name National Coalition of

Faculty of Arts City State Country Independent Scholars

Masaryk University email@email.com United States

Czech Republic courtneylazore@ncis.org

veronika.mattova@mail.muni.cz



Abstract is built under external pressure to achieve inner balance.

However, emotional turmoil can be particularly challenging,

How can individuals deal with personal trauma or internal which is why we need to explore a multimodal approach that

struggles more effectively? This is the main question of every could be a revolutionary step in interdisciplinary mental health

existential crisis, closely linked with humankind's survival support.

strategy. Finding new, innovative ways for practitioners to Bibliotherapy, psychodrama, and AI might initially seem

leverage therapeutic techniques and modern artificial unrelated, but there is a way to incorporate all of them within a

intelligence (AI) technology is crucial to providing precision new therapeutic framework. mental health support to more individuals. While looking at

possible approaches, it becomes more and more important to

synthesize complex ways practitioners can provide 2 Methodology multidimensional help. This paper investigates the possibility of

a new holistic treatment that integrates This paper uses a conceptual methodology to explore the bibliotherapy's

storytelling, integration of mental health techniques and artificial intelligence Magic Shop as a psychodrama method , and AI

conversation tools – chatbots to theorize a holistic, accessible and personalized mental health to ensure that individuals receive

encompassing supportive therapy and feel less isolated. The care option. Based on the researchers' backgrounds and

holistic method is applied to engagement with existing literature, bibliotherapy, psychodrama Korean pop music as a case study,

because K-pop content has experimented with these techniques and artificial intelligence were chosen for further analysis. The

and fandoms often have strong parasocial interactions. researchers reviewed literature from databases including APA

Combining these techniques creates a holistic, accessible, and PsycInfo, Google Scholar, IEEEXplore, ProQuest, PubMed,

personalized mental health care option that enhances the Scopus, Web of Science and EBSCO. Only papers written in

cognitive, emotional and practical well-being of individuals in English and published after 2000 were considered, except for one

need of support. foundational work (Purpose and strategy behind the magic

shop). To build the theory, the researchers identified the



Keywords prominent features of each topic (Table 1): bibliotherapy's

storytelling, psychodrama's experiential model, and AI's real-

Bibliography, storytelling, psychodrama, Magic Shop, AI, time access, which will each be explored in subsequent sections.

chatbots, K-pop For an exploratory case study application, the researchers relied

on new concepts (described in 6.1-6.3) provided through the



1 Introduction produced as music-related products, but nowadays perceived Korean pop music industry (concretely group BTS), initially

The psychological aspect of overcoming personal issues and through a therapeutic lens, with the primary aim of merging the

facing the harshness of reality is demanding for everyone. separate therapies together, creating a more influential impact.



Internal cognitive mechanisms lead us to believe that individual

struggles require individual approaches. This adaptation for

hiding internal fights and presenting balance to the outside world



†Alumna of Comenius University, Faculty of Mathematics, Physics and Informatics,

MEi:CogSci | Middle European interdisciplinary master's programme in Cognitive

Science.



Permission to make digital or hard copies of part or all of this work for personal or

classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must Table 1: Selection & Clarification for Chosen Approaches be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.cog.2 3 it can be a double-edged sword linked to maintaining one's own Bibliotherapy

physiological and psychological well-being.

Storytelling has long been a powerful means of exploring and

understanding human emotions, and this concept underpins



bibliotherapy, a therapeutic approach that uses literature to 5 AI used in Therapeutic Spaces support mental health and personal growth. By engaging with AI chatbots are increasingly being used to enhance mental health carefully chosen texts, individuals can reflect on their own care by offering real-time interactions that meet cognitive and experiences, gain new insights, and find comfort and empathy emotional needs. These digital tools are part of a broader shift in through the narratives. Bibliotherapy leverages the emotional communication, driven by the "computers-are-social-actors" and cognitive impact of stories to address psychological paradigm, where AI significantly impacts how people engage challenges, offering a complementary and accessible method for with services [10]. fostering emotional healing and resilience [1]. In therapy, AI complements traditional methods by providing Some research suggests that fictional narratives may be more personalized interventions, making mental health support more effective for use in bibliotherapy , causing readers to empathize accessible. Virtual environments facilitated by AI can create new more with the characters and leading to better self-understanding opportunities for personal fulfilment and emotional connection and self-improvement [2]. [11]. Moreover, neuroscience reveals that storytelling influences Moreover, there is an overlap, because recent research by brain chemistry, enhancing empathy and trust through the release McAllister et al. explores the potential for chatbots to enhance of oxytocin and vasopressin [3]. This approach effectively bibliotherapy by supporting facilitators in mental health sessions. enhances emotional experiences and trust, aiding in personal The study seeks to address gaps in existing literature by development and resilience [4]. investigating how chatbots can be utilized to assist in the

preparation and delivery of bibliotherapy [12].

4 In addition, in response to the high potential of technology, Psychodrama and its Traits in Magic Shop

interviews with bibliotherapy facilitators have been conducted,

There are many ways to grasp the concept of Magic Shop. The

followed by thematic analysis, to identify suitable tasks for the

first is understanding it as an in-depth training program with

chatbot, aiding facilitators rather than directly evaluating the

transformational practice, which will be discussed later in

impact on participants of bibliotherapy.





chapter 6.2 K-POP in Magic Shop.


This integration represents a major step towards a more

On the other hand, there is a possibility to apprehend it

comprehensive and accessible mental health framework.

through the concept of psychodrama, which will be our primary

interest, within this paper. Magic Shop is a practice used in



psychodrama and group therapy in which participants create a 6 K-POP as Multidimensional Tool space where they can "buy" something they already want, such Although Korean pop , generally known as K-pop, seems to only as confidence or peace, from what they already have internally, overlap with music therapy, due to its main impact field, K-pop such as anxiety or anger. In other words, this activity helps may also serve as a useful case application for bibliotherapy, the individuals explore their deepest passions, and the sacrifices Magic Shop technique , and AI tools . needed to achieve them, while encouraging self-reflection and Some K-pop groups participate in transmedia story worlds personal growth, with a possibility to solve conflicts in a safe, and lyrical concepts that go far beyond "unrequited love” themes, symbolic environment [5]. touching the human psyche more deeply. Additionally, merging The term magic can be seen in society as something that this with technology platforms like "Weverse" or "Bubble for carries supernatural power . Magic Shop is not an exception. Thus, JYPnation ", not to mention fabricated interactions through apps it has a lot of forms and names: Magic Shop is our deep like "Replika", "Mydol", etc., fans have never felt closer and understanding of heart and brain in harmony. We often compare more bonded to their idols, who are revered with boundless ourselves to others and hope we could be better than we are. support and understanding. Specific traits like social ability, patience and cleverness are

examples that require time and practice. But it seems that Magic



Shop is the key [6]. This method as a psychodramatic strategy 6.1 K-POP in Bibliotherapy can offer help to anybody through the use of fantasy [7]. Many K-pop groups rely on some form of storytelling, but some In general , psychodrama is an experiential form of therapy, take it a step further. In particular, K-pop group BTS's "BTS allowing those in treatment to explore issues through action Universe" (BU) , functions as a multidimensional tool for methods (dramatic actions) [8]. This described approach, linked emotional and psychological engagement. In bibliotherapy, with psychodrama , was developed by Jacob Levy Moreno as a BTS's books, The Most Beautiful Moment in Life: The Notes , psychotherapeutic technique useful in working with patients provide a unique narrative that facilitates self-reflection and during individual and group psychotherapy. This method offers emotional exploration. These notes, embedded within the group's significant changes through role-playing and dramatization, Love Yourself albums, as well as two books published by BTS's resulting in many benefits, such as insight, abreaction, label, offer fans a form of therapeutic engagement by inviting acceptance of internal impulses, confrontation with the feelings them to interpret and relate to the fictional world, which can of other people and training of alternative behaviors [9]. Whether mirror personal experiences and foster emotional processing [13]. we are talking about the first or second definition of Magic Shop , This can be particularly striking when we take into account

individual emotional overload with a long-term inability to

restart one's own coping mechanism to underlie qualitative one's prosperity and healing cognitive processes,

functioning in everyday life. interdisciplinary cognitive science offers an opportunity to

Preliminary research has suggested that the stories embedded merge concepts together that might bring new ways to deepen

in K-pop can have a positive impact on fans' ability to cope with our knowledge.

challenges and heal. One survey found that 97% of fans Combining different approaches, such as AI chatbots using

(n=2342) agreed that BTS's music and lyrics were effective in Cognitive Behavioral Therapy (CBT) with minimal

this way. Additionally, 84% agreed the storylines in BTS's bibliotherapy interventions, is slowly but steadily indicating that

concepts were effective, and 75% agreed the BU storyline was the chatbots are more effective in reducing symptoms of

effective [14]. depression and anxiety [21], which is the primary desired effect.

These numbers suggest valid proof of audience engagement Moreover, fanship, an individual's bond to their idols, plays a

techniques in fiction, described by Donald Maass as presenting crucial role in enhancing happiness, self-esteem, and social

novelty, challenge, or aesthetic appeal to readers, which leads to connectedness, which pushes forward the application of social

better identification with the story, while figuring out solutions identity theory in the realm of K-pop fans and expands the

for the main character´s actions, reflecting in an individual's psychological understanding of fandom and its extended

healing scheme [15]. therapeutic possibilities [22].



6.2 K-POP in Magic Shop 8 Findings

There is no doubt that science plays a crucial role in our Combining bibliotherapy, the Magic Shop psychotherapeutic

understanding of human beings, yet some phenomena, such as method and AI tools can offer a comprehensive mental health

compassion, altruism and empathy, remain enigmatic. These treatment model. Bibliotherapy provides cognitive and

concepts form the foundation of James Doty's work, particularly emotional benefits through literature and storytelling, the Magic

in his book Into the Magic Shop: A Neurosurgeon's True Story Shop method engages clients in creative and transformative

of the Life-Changing Magic of Mindfulness and Compassion, experiences, and AI tools ensure continuous support and

which inspired not only this paper but also the lyrics of the K- accessibility (as shown in Figure 1).



pop group BTS's song "Magic Shop". The song, much like Doty's

book, is perceived to have a healing effect on many individuals,

resonating deeply within a therapeutic framework. Doty is not

only a renowned author but also a researcher who founded the

Center for Compassion and Altruism Research and Education

(CCARE) at Stanford University, which supports his desire to

analyze the interaction between the mind and body in relation to

the concept of Magic Shop [16].



6.3 K-POP in AI Chatbots

In the K-pop world, AI enhances fan interactions by fostering

parasocial relationships (PSI), a term that goes back to its roots

in 1956 when it was defined as the illusion of a face-to-face

friendship between audience members, along with the main Figure 1: Infographic of Innovative Supporting Approaches

factor of a one-sided relationship [17] with idols (Korean singers

within K-pop industry). Chatbots like "ChatGPT", "Replica" or Note. Created with Canva.com

"Mydol" offer more personalized conversations, deepening the

emotional connection fans feel with their favorite stars [18]. This a. Holistic Care: The union offers a multidisciplinary

can be particularly comforting for those dealing with anxiety or approach to mental health, addressing cognitive,

low self-esteem. emotional, and practical needs. By considering the

AI-driven PSI offers new avenues for addressing anxiety, whole individual, this approach promotes holistic

providing a controlled environment for emotional exploration. healing and improves overall well-being.

While there are risks of maladaptive obsessions, the positive

impact on psychological well-being is significant when managed b. Accessibility: AI tools streamline mental health

carefully [19]. As AI continues to evolve, its role in supporting services and provide immediate support, covering a

mental health in niche areas like K-pop will likely grow, offering gap to traditional treatments. This ensures that more

innovative solutions to common challenges [20]. individuals can access the help they need,

regardless of location or financial barriers.

7 Relevance for Cognitive Science & Cognitive

Behavioral Therapy c. Personalized Experience: Integrated approaches

Understanding how the human brain reacts to situations, making enable tailored interventions that can be adapted to

individuals feel distressed has been already covered. Now, it is individual needs and preferences. This personal

more than important to look for strategies to fight against it, apart involvement improves the effectiveness of treatment,

from medicaments. Recognizing an individual's need to support leading to the best results and a more satisfying

treatment experience.



[7] E. Koile, "The Magic Shop: The therapist masquerades as a shopkeeper,"



Additionally, K-pop offers a valuable case application for this Available:https://www.aapweb.com/wp-content/uploads/2019/03/voices- Voices: The Art and Science of Psychotherapy, Spring 2011. [Online].

holistic program, given the rich story worlds and concepts sample-koile.pdf. [Accessed: Aug. 19, 2024].

combined with fans' love and attachment to their favorite idols, Available: [8] “Psychodrama,” Types of Therapy, GoodTherapy, 2016. [Online]. https://www.goodtherapy.org/learn-about-

some of whom can be accessed through AI applications. therapy/types/psychodrama. [Accessed: Aug. 19, 2024].

[9] K. Litwińska-Rączka, "Jacob Levy Moreno’s Psychodrama as a Work



9 Study Limitations & Subsequent Research Technique for Treating Patients in Group and Individual Psychotherapy,"

Current Problems of Psychiatry, vol. 19, no. 4, pp. 248–259. 2018 DOI:



Regarding the limitations of this paper, both authors are fully 10.2478/cpp-2018-0019. [10] M. Song, X. Xing, Y. Duan, J. Cohen, and J. Mou, "Will Artificial aware of the restrictive parameters linked to the exploratory case Intelligence Replace Human Customer Service? The Impact of

study, which was selected as a presented application. Therefore, Communication Quality and Privacy Risks on Adoption Intention,"

the mentioned research gaps and possibilities of merging Journal of Retailing and Consumer Services, vol. 66, 102900, 2022. DOI:

10.1016/j.jretconser.2021.102900.

therapies have only an advisory nature with the need for [11] G. Dubey, Sociální pouto v éře virtuality. Fra, 2020.

subsequent research with possible future administration in [12] P. McAllister, J. Kerr, M. McTear, M. Mulvenna, R. Bond, K. Kirby, J.

Morning, and D. Glover, "Towards Chatbots to Support Bibliotherapy

professional-guided therapy sessions, as well as individual Preparation and Delivery," in Chatbot Research and Design - 3rd



applications after undertaking a specialized training process International Workshop, CONVERSATIONS 2019, Revised Selected Papers , A. Følstad, T. Araujo, S. Papadopoulos, E. Lai-Chong Law, O.-C. provided by professionals in specialized facilities. Granmo, E. Luger, and P. B. Brandtzaeg, Eds., vol. 11970, pp. 127-142.

Springer Nature, Switzerland, 2020. DOI: 10.1007/978-3-030-39540-7_8.



10 Conclusion [13] C. Lazore, "How the BTS Universe Successfully Engages Thousands of

Fans," The BTS Effect, 2019. [Online]. Available:

Unveiling how to provide the most appropriate care to those who https://www.thebtseffect.com/blog/how-the-bts-universe-successfully-

engages-thousands-of-fans. [Accessed: Aug. 19, 2024].

need to overcome difficulties while restarting their coping [14] C. Lazore, "' Artists for Healing': Anxieties of Youth, Storytelling, and

mechanisms can be quite diverse and tricky. Understanding the Healing through BTS," BTS Global Interdisciplinary Conference, 2020.

individual treatment plan is the key. While previous research has [15] D. Maass, The Emotional Craft of Fiction: How to Write the Story Beneath

the Surface. Writer’s Digest Books, 2016.

been focused on single therapy treatment, nowadays there is a [16] CCARE, "Mission & Vision," Stanford Medicine, 2019. [Online].

strong case for merging them. Encountering situations through Available: http://ccare.stanford.edu/about/mission-vision/. [Accessed:

Aug. 19, 2024].

the safe space of stories allows individuals to better understand [17] D. Horton and R. R. Wohl, "Mass Communication and Para-Social

actions and emotions, develop empathy and progress in their Interaction," Psychiatry, vol. 19, no. 3, pp. 215-229, 2016. DOI:

healing journeys. 10.1080/00332747.1956.11023049. Bibliotherapy allows readers to internalize and [18] M. Pradeep, "Kpop Idol-Based Chatbots are Blurring the Lines Between

adopt positive traits from characters they resonate with, while Interaction and Explicit Obsession," Screenshot, 2022. [Online].

psychodrama Available: https://screenshot-media.com/culture/internet-culture/kpop- facilitates deeper understanding and emotional

idol-based-chatbots-dangers/. [Accessed: Aug. 19, 2024].

growth through role-play and perspective-taking. In the realm of [19] S. Ortiz, M. Shin, M. Samuels, and A. Windsor, "What is ChatGPT and

AI chatbots Why Does It Matter? Here’s What You Need to Know," ZDNET, 2023. , this effect can be harnessed to create more

[Online]. Available: https://www.zdnet.com/article/what-is-chatgpt-and-

empathetic and personalized interactions by mirroring users' why-does-it-matter-heres-everything-you-need-to-know/. [Accessed:

language and emotional tone. With Aug. 19, 2024]. K-pop's use of AI , there is

untapped potential for precision mental health care. Overall, [20] R. Kowert and E. Daniel, "The one-and-a-half sided parasocial

relationship: The curious case of live streaming," Computers in Human

these techniques underscore the importance of storytelling in Behavior Reports, vol. 4, 100150, 2021. DOI:

fostering empathy, engagement, and therapeutic outcomes across 10.1016/j.chbr.2021.100150.

[21] D. A. Laffan, "Positive Psychosocial Outcomes and Fanship in K-pop

these diverse fields. Fans: A Social Identity Theory Perspective," Psychological Reports, vol.

124, no. 5, pp. 2272-2285, 2021. DOI: 10.1177/0033294120961524.



Acknowledgements [22] M. Liu, H. Peng, X. Song, C. Xu, and M. Zhang, "Using AI chatbots to

provide self-help depression interventions for university students: A

We want to thank randomized trial of effectiveness," Internet Interv., vol. 27, 100495, Jan. Mgr. Michal Schwarz, Ph.D. from Masaryk

2022. DOI: 10.1016/j.invent.2022.100495.

University for his support during the process of writing this paper.



References

[1] M. C. Pino and M. Mazza, “The Use of ‘Literary Fiction’ to Promote

Mentalizing Ability,” PLoS One, vol. 11, no. 8, 2016. Available:

https://doi.org/10.1371/journal.pone.0160254.

[2] K. Oatley, Such Stuff as Dreams: The Psychology of Fiction. John Wiley

& Sons Inc., 2011.

[3] E. R. Kandel, The Age of Insight: The Quest to Understand the

Unconscious in Art, Mind, and Brain: From Vienna 1900 to the Present.

Random House, 2012.

[4] M. E. Langeberg, "Bibliotherapy: A Systematic Research Review with

Social-Emotional Learning Applications," Illinois Reading Council

Journal, vol. 51, no. 4, pp. 32-45, Fall 2023. DOI:

10.33600/IRCJ.51.4.2023.32.

[5] A. C. Barbour, "Purpose and strategy behind the magic shop," Journal of

Group Psychotherapy, Psychodrama, & Sociometry, vol. 45, pp. 91-101,

1992. [Online]. Available:

https://api.semanticscholar.org/CorpusID:152183218.

[6] R. Rautiainen, "Using the Magic Shop in a work counselling group,"

Centre for Playback Theatre, 2002. [Online]. Available:

https://www.playbacktheatre.org/playbacktheatre/wp-

content/uploads/2010/04/Rautiainen_Magic-Shop.pdf. [Accessed: Aug.

19, 2024].





Usability of intelligent assistive technology used by




people with dementia and their caregivers



Klara Dečman

Cognitive Science, Occupational Therapy

University of Ljubljana & University of Vienna

Slovenia

kd1023@student.uni-lj.si, decmanklara@gmail.com



Abstract One of the most common diseases in old age, dementia is

recognized as one of the most costly and burdensome health

Intelligent assistive technology with context-aware computing conditions [2]. Statistics suggest that the growing global

and artificial intelligence can be applied to assist a person with population of older adults diagnosed with dementia reached

dementia and their caregivers with activities of daily living. This 44.4 million worldwide in 2013, with projections indicating an

paper samples such technologies with a focus on current increase up to 135.5 million by 2050. Concern over the limited

knowledge and practice concerning usability. We used a scoping availability of family and professional caregivers for this rapidly

study to address the objectives of the research. Our findings growing population is intensifying (ibid.). As the population



to individuals’ needs and capabilities it is not commonly available often lack the key skills to provide the necessary level addressed in the literature. Furthermore, while researchers are of care [5]. Furthermore, as family caregivers become more indicate that despite the importance of technology customization ages, the number of potential caregivers decreases, and those

aware of the concepts and aims of evaluating the usability of involved while struggling to balance other familial and social

technology, they seem to face difficulties in assessing them. roles and responsibilities, they often experience negative



Keywords depression, isolation, and sleep deprivation [6]. Technological consequences on their health, such as burden, anxiety,

Activities of daily living, cognitive assistance, dementia, innovation, including advances in communications, robotics, and

evaluation of usability, family caregivers, human-centered sensors, are perceived as promising to tackle these

design, scoping survey, user experience challenges [5]. Specifically, assistive technology (AT) refers to a

broad range of devices and systems designed to maintain or

enhance an individual’s functioning related to cognition,

1 Introduction communication, hearing, mobility, self-care and thereby

Dementia is a neurocognitive disorder, typically chronic and promoting their health, well-being, inclusion, and

progressive, characterized by impairments in cognitive functions participation [7]. AT is not designed to perform tasks on behalf

such as memory, attention, orientation, and language [1, 2] to the of the user, but are specifically designed to monitor the activities

extent that a person with dementia (PwD) is not able to of cognitively impaired users and provide appropriate assistance,

independently complete activities of daily living (ADLs) [3]. thereby enhancing the likelihood of achieving desired behavioral

Personal (pADLs) refer to basic physical needs such as dressing, outcomes [8]. A specific category of AT, cognitive orthotics [9]

toileting, bathing, and eating, while instrumental (iADLs) are or cognitive assistive technology [10] is designed to assist with

essential for living independently in the community, such as cognitive tasks. For instance, AT is employed to remind PwD to

preparing food, taking medication, and doing laundry [4]. The take medication or that their family member is visiting them next

ICD-11 [1] identifies three degrees of severity of any type of day [9, 10]. PwDs and their caregivers routinely use low-tech

dementia. In the mild stage, a PwD may live independently but aids, such as medication pill organizers, schedules, and notes.

requires supervision and/or support with iADLs, such as locating They are being offered high-tech aids, such as intelligent

everyday objects, and handling finances. In the moderate stage, assistive technology (IAT) that employs artificial intelligence to

PwDs require support to function outside their home assess whether and when an appropriate reminder or procedural

environment. They can accomplish only simple household tasks guidance is necessary for task completion [11, 12]. Additionally,

and experience difficulties with completing pADLs. In the severe IAT should be contextually aware: able to examine its

stage, memory impairment becomes profound, though it varies environment, react to changes within it, and thus provide help

by etiology. PwDs are fully dependent on others for pADLs and when needed [11].

they often experience total disorientation in time and place. Human factors and ergonomics are scientific disciplines

focused on studying the interactions between humans and other

Permission to make digital or hard copies of part or all of this work for personal or components of socio-technical systems [13]. The aim of

classroom use is granted without fee provided that copies are not made or distributed designing such products and systems is to minimize human error

for profit or commercial advantage and that copies bear this notice and the full and enhance human efficiency. One attempt at managing human

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s). factors analysis and human errors is through the development and

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia deployment of measurement standards such as the Human © 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.cog.3 Readiness Levels (HRL) scale (ibid.). HRL complements and



supplements the Technology Readiness Level (TRL) scale, which platforms (APA PsycInfo, Google Scholar, IEEEXplore,

captures the maturity of technology before and after its ProQuest, PubMed, Scopus, Web of Science, as well as the

integration into a developing system [13, 14]: HRL emphasizes digital library facilities of the Universities of Vienna and

the readiness to develop technology for effective and safe human Ljubljana) with hand searches of electronic journals and

use, and it should capture human-related features of technology literature identified through literature readings. For this survey,

development [14]. Similarly to TRL, HRL scale is divided to we included original articles, conference proceedings, and PhD

nine stages: basic research and development of principles, thesis; written in English; and published “within the last decade”

concepts, and the application of human characteristics, (i.e., since 2013). To be covered, IATs further had to meet the

performance, and behavior, along with guidelines incorporating following inclusion criteria: direct applicability to dementia care,

human-centered requirements to enhance human performance focus on assisting with ADLs, PwD, and/or family caregiver as

and human-technology interactions (HRL 1…3); development a user. We excluded IATs developed for the support of other (if

and assessment of user interface design concepts and prototype related) disabilities, such as traumatic brain injuries; that could

simulations in laboratory and real-world environments only be used by professional caregivers.

(HRL 4…6); full-scale testing, verification, and deployment in

an operational environment with representative users and system

hardware and software (HRL 7… 8); and the final stage, where 3 Findings

the system is actively used in the operational environment with We illustrate our findings for each research question using the

systematic monitoring of human-system performance technologies: COACH [10, 11] and AWash [19] (targeted ADL:

(HRL 9) [14]. HRL is closely linked to user-centered design, a handwashing); DRESS [20] (getting dressed); ToiletHelp [21],

framework for the design and development of new products or (using a water toilet); and Smart Toothbrush [22] (brushing

the assessment and evaluation of existing products that explicitly teeth). The pADLs supported by these IATs must be performed

considers potential users’ needs, wishes, and subjectively regularly to maintain the person’s independence, health, and

perceived limitations of the IAT [5, 9, 12]. overall well-being. As dementia progresses, PwD becomes

Some key definitions from the ISO standard on increasingly dependent on others to complete ADLs, affecting

ergonomics of human-system interaction [15] read: “3.13 their family caregiver and society (cf. section 1). usability: extent to which a system, product or service can be

used by specified users to achieve specified goals with 3.1 Human Readiness Levels



of use. … 3.3 effectiveness: accuracy and completeness with introduced in section 1, with most of the surveyed technologies effectiveness, efficiency, and satisfaction in a specified context We assigned aggregated HRL scores according to the groups

which users achieve specified goals. … 3.4 efficiency: resources ranking at HRL 7…8: COACH, AWash, and ToiletHelp. This

used in relation to the results achieved. Typical resources include likely results from our choice of targeted content, as we aimed to



aspects of usability include user satisfaction, the “extent to which Toothbrush to the HRL range 4…6, as the first is about time, human effort, costs, and materials.” Further important focus on IATs close to HRL 9. We mapped DRESS and Smart

the user's responses resulting from use meet the user’s needs and developing and evaluating a prototype in preparation for in-home

expectations; how accessible a product, system, service, or trials with PwDs, while for the second only preliminary

environment is to individuals with diverse needs, characteristics, laboratory testing was conducted with healthy individuals.

and capabilities is another element of usability”. As examples of



differences in terminology, see e.g. [16], where utility refers to 3.2 Different stages of Dementia



usefulness The IATs selected for this article are intended to provide targeted covers how pleasant and easy to use technology is whether the design provides features that users need and

(usability) and whether it does what users need (utility). assistance for different stages of dementia. ToiletHelp is aimed

to be used by PwD in the mild stage of dementia, COACH in

moderate to severe stage, DRESS and Smart Toothbrush in

2 severe stage, while for AWash we have not found any explicitly Methodology

targeted stage of dementia. We found no evidence of

This short survey covers only some of the findings of my ongoing technologies taking into account individual differences and

more comprehensive review of topical IATs [17]. For this needs of PwDs and their caregiver, consequently, we were not

survey, the following research questions to explicate the able to find such technology that would be able to adapt

coverage of usability will be addressed: according to the actual severity of dementia as disease progresses

(cf. section 3.3). Such customization is needed as cognitive

1. What is the maturity of IAT for human use (= its HRL)? functions progressively deteriorate, with fluctuations in rating

2. (How) do the developers take the progression of the disease occurring throughout the day or as the system would be used over

into account? periods ranging from weeks to months or even years [20, 21].

3. How exactly is technology being assistive?

4. How is the usefulness and usability of the technology

evaluated (if at all)? 3.3 Notions of Assistance

Assistance involves interacting, with prompting being an

Our scoping study maps key concepts, main sources, and interaction strategy that has become widely popular also in the

types of evidence available for the domain targeted. For context of IATs. Within our target domain, we found audio

methodological transparency, we followed the PRISMA-ScS prompts to be most common as they are part of COACH, AWash,

checklist [18]. Our search combines electronic database DRESS, and Smart Toothbrush. Such assistance should guide



PwD through the sequential steps of the activity by pre-recorded of IAT largely depends on its utility and its unobtrusiveness,

voice commands. Visual prompts include videos of steps of which can encourage more consistent use.

activities (COACH); pictures of correct clothing items (DRESS); The current understanding of usability reflected in the

use of different lights to attract attention to the appropriate use of literature indicates that even when researchers are aware of the

an object (DRESS, Smart Toothbrush); and texts with related concepts and terminology and aim to assess them, they

instructions (ToiletHelp). DRESS consists of motivational have difficulties in doing so with unified questionnaires or

prompts in the form of songs or videos favored by the PwD are standardized testing procedures.

meant for when a PwD should get stuck in an activity, and are

configured by the family caregiver. COACH has options for

increasing levels of support: low-guidance and high-guidance 4 Relevance of Cognitive Science

verbal prompts, video demonstrations, or placing a call to the The goal of the inter-disciplinarity of Cognitive Science is to

caregiver. DRESS offers the choice of continuous mode, which address the question of how does the mind work – why we do the

includes chronological directions across all steps of an activity, things we do, think the way we think, and how we perceive the

and independent mode, in which no audio prompts are provided world around us – by trying to understand and explain underlying

while the PwD is donning a shirt, and the caregiver should mental processes and mechanisms of human behavior from the

receive text messages on their device either when help is needed point of view of each discipline [24]. In user interfaces,

or dressing is completed. Nominal assistance provided by computational models of human behavior are used to describe

ToiletHelp consists of acknowledgment messages displayed to and capture our understanding of typical user actions, predict

reassure the users they have completed every step of the activity; future actions, and guide users toward improving their

when the need is recognized, instructions are repeated. If a user actions [25]. These computations are typically based on internal

should still fail, an alert informs the caregiver the PwD is having symbolic knowledge representations, allowing a cognitive agent

trouble, along with a reassurance message being displayed to the to manipulate symbols to gain information about the external

PwD. world and determine how to act effectively – plan and perform

The IATs we identified can help guide PwD through activities, actions, and achieve specific goals [26]. Evolutionary

but it is crucial to tailor such assistance to individual needs and psychologists view the information processing architecture of the

adjust it as dementia progresses [20]. While there are cases where brain to consist of adaptive problem-solving systems that use

differing/increasing levels of assistance are provided by information to adaptively regulate physiology and behavior. In

IATs [10, 11], such adjustment is not commonly documented in this perspective, attention, learning, emotion, and motivation all

the literature. Despite its importance, our research indicates that play key roles in minds work and how we respond to our

there is also a lack of consistency in the terminology used to environments [27]. In particular, motivation can guide cognitive

describe the adjustment of IATs to individual needs (e.g. processes: When a PwD becomes fatigued, their motivation to

customization, personalization, adjustment, adaptation). continue activities declines. IATs can help by providing

motivational prompts, such as favorite music or videos, which

3.4 Usability evoke emotional memories. This is but an example of how,

The resources we analyzed indicate a dearth of commonly used cognitive science provides crucial insights into how users

standardized usability tests; out of the systems surveyed, only perceive, process, and interact with technology and consequently

Awash was assessed using the System Usability Scale (SUS) affects both, the improvement of designs and testing of usability

questionnaire [23]. Instead, information about the usability of and usefulness. It is a “bridge” between applied artificial

IATs is often gathered through user interviews [10, 11, 20, 21], intelligence and user experience.

observation, and performance testing [10, 11, 19, 21, 22]. One important objective of applied artificial intelligence is

In terms of effectiveness, COACH and AWash users were the development of cognitive orthotics, designed to enhance and

able to independently complete more steps of activity and engage expand the user’s cognitive abilities [28]. It is not about

less with caregivers while using IAT. Regarding technology imitating human abilities, but rather extending efficiency , the them.

developers of the Smart Toothbrush have estimated its battery The key focus is the importance of creating systems that combine

life, while those of DRESS considered the final product's cost. In human and machine components in a way that maximizes their

terms of user satisfaction, caregivers noted several benefits of individual strengths taking into account ethics. To design

DRESS, including validation of memory loss, empowerment of successful cognitive orthotics, interdisciplinary teams are

PwD, promoting privacy and dignity, and providing caregiver needed to unite relevant knowledge and perspectives of

respite. ToiletHelp was reported to increase PwD’s autonomy, professionals (such as computer scientists, engineers, physicians,

boost self-esteem and dignity, and reduce the burden on cognitive psychologists, and neuroscientists) together with

caregivers. Participants rated AWash with a positive user stakeholders and users of technology (ibid.).

experience. On the other hand, difficulties in using the

technology were due to varying stages of dementia, visual and

sensory perception issues, the need to change routines, and 5 Limitations and Future Work

affordability issues [20]. Users expressed dissatisfaction with As technology advances rapidly, future research should explore

long delays between tasks and the frequency of a wider range of IATs using novel modalities and supporting

prompts [10, 11, 22], while overlapping video and verbal more diverse ADLs. This limited study cannot form generalized

messages used in ToiletHelp caused distraction. The acceptance statements about IAT usability for PwD and caregivers, as



comparing specific ADLs is challenging due to variations in [10] N. M. Dharan, M. R. Alam, and A. Mihailidis, "Speech-Based Prompting

particular activity structure, cultural contexts, and dementia System to Assist with Activities of Daily Living: A Feasibility Study,"

Gerontechnology, vol. 20, pp. 1-12, 2021.

stages. We focused on a small subset of IATs addressing some [11] S. Czarnuch, "Advancing the COACH automated prompting system toward

pADLs, excluding those covering iADLs and multiple ADLs. an unsupervised, real-world deployment," Ph.D. dissertation, Dep. Philos.,

[17] takes a step in this direction. BME, University of Toronto, Toronto, Canada, 2014.

[12] A. J. Bharucha, V. Anand, J. Forlizzi, M. A. Dew, C. F. Reynolds, S.

Stevens, and H. Wactlar, "Intelligent Assistive Technology Applications to

Dementia Care: Current Capabilities, Limitations, and Future Challenges,"

6 Conclusion Am. J. Geriatr. Psychiatry, vol. 17, pp. 88-104, February 2009.

Dementia is becoming increasingly prevalent, posing a major [13] V. Newton, A. Greenberg and J. See, "Project Management Implications and

Implementation Roadmap of Human Readiness Levels," in HCIBGO 2017,

societal, economic, and global health challenge. While extending Cham, 2017.

the duration of PwD’s stay in their private homes may be seen to [14] ANSI/HFES 400-2021: Human Readiness Level Scale in the System

help alleviate the strain on institutional settings, it in turn places Development Process, Human Factors and Ergonomics Society,

Washington, DC, 2021.

a significant burden on family caregivers. While IATs are

[15] ISO 9241-210: Ergonomics of human-system interaction — Part 210:

intended to enhance the independence of PwD and reduce the Human-centred design for interactive systems, International Organization

caregiver's burden, our literature review efforts suggest that for Standardization, 2019.

usability aspects are not systematically assessed. This gap is also [16] J. Nielsen, "Usability 101: Introduction to Usability," nngroup.com.

linked to current HRLs, which indicate that existing IATs are not Accessed May 17, 2012. [Online]. Available:

https://www.nngroup.com/articles/usability-101-introduction-to-usability/

fit for deployed use by PwD. Moreover, we find that IAT is often [17] K. Dečman, “Intelligent assistive technology and family caregivers of people

not designed to adapt to the progression of the disease, affecting with dementia: Does it work?” M.S. thesis, Dept. Cogn. Sci., Uni-Lj.,

its utility and usability. Heavy terminology such as Ljubljana, Slovenia, 2024 (forthcoming). intelligent

assistance [18] A. C. Tricco et al., "PRISMA Extension for Scoping Reviews (PRISMA- appears to be employed all too easily. Furthermore,

ScR): Checklist and Explanation," Ann. Intern. Med., vol. 169, p. 467-473,

practice in assessing and reporting usability appears to leave October 2018.

significant room for improvement. [19] Y. Cao, F. Li, H. Chen, X. Liu, S. Yang, and Y. Wang, "Leveraging

Wearables for Assisting the Elderly With Dementia in Handwashing," in



Acknowledgments IEEE Trans. Mob. Comput., vol. 22, no. 11, pp. 6554-6570, 2023.

[20] W. Burleson, C. Lozano, V. Ravishankar, J. Lee, and D. Mahoney, "An

I would like to express my gratitude to the University of Assistive Technology System that Provides Personalized Dressing Support

Ljubljana and the University of Vienna for allowing me to for People Living with Dementia: Capability Study," JMIR Med. Inform.,

vol. 6, no. 2, 2018.

choose a Master’s thesis topic of personal and professional [21] I. Ballester, M. Gall, T. Münzer and M. Kampel, "Vision-Based Toilet

importance. I am particularly thankful to my supervisor Univ.- Assistant for People with Dementia in Real-Life Situations," in IEEE

Lektor, Dipl.-Ing., Dr. Paolo Petta, for your guidance and the PerCom Workshops, Biarritz, France, 11-15 March 2024, pp. 141-147.

opportunity to learn from you. [22] M. E. S. Jannati, "Design and Implementation of a Smart Toothbrush for

Individuals with Dementia," M.S. thesis, Dept. ECE, The University of

Manitoba, Winnipeg, Manitoba, Canada, 2020.

References [23] J. Brooke, "SUS: A quick and dirty usability scale," in Usability Eval. Ind.,

vol. 189, November 1995.

[1] International Classification of Diseases, Eleventh Revision (ICD-11), World

Health [24] D. Vernon, "Paradigms of Cognitive Science," in Artificial cognitive system: Organization (WHO), 2023. [Online]. Available:

https://icd.who.int/browse11/l-m/en a primer, Cambridge, MA and London, England, The MIT Press, 2014, pp.

19-61.

[2] M. Prince, A. Wimo, M. Guerchet, G.-C. Ali, Y.-T. Wu and M. Prina,

“World Alzheimer Report 2015. The Global Impact of Dementia. An [25] N. Banovic, J. Mankoff and A. K. Dey, "Computational model of human

Analysis of Prevalence, Incidence, Cost and Trends,” Alzheimer’s Disease routine behaviours," in Computational interaction, 1st ed., A. Oulasvirta, P.

International, London, England, UK, 2015. [Online]. Available: O. Kristensson, X. Bi and A. Howes, Eds., New York, United States of

https://www.alzint.org/u/WorldAlzheimerReport2015.pdf America, Oxford University Press, 2018, pp. 377-398.

[3] [26] J. L. Bermúdez, "Physical Symbol Systems and the Language of Thought," Diagnostic and statistical manual of mental disorders, 5th ed., American

Psychiatric Association, American Psychiatric Publishing, Washington, DC in Cognitive Science. An Introduction to the Science of the Mind, 3rd ed.,

London, England, 2013, pp. 602-614. Cambridge and New York, Cambridge University Press, 2020, pp. 99-121.

[4] American Occupational Therapy Association, [27] L. Cosmides and J. Tooby, "Evolutionary Psychology: New Perspectives on Occupational Therapy

Practice Framework: Domain and Process Cognition and Motivation," Annu. Rev. Psychol., vol. 64, pp. 201-229, , 4th ed., vol. 74, AOTA Press,

2020. January 2013.

[5] D. V. Dahlke and M. G. Ory, "Emerging Issues of Intelligent Assistive [28] K. M. Ford, P. J. Hayes, C. Glymour and J. Allen, "Cognitive Orthoses:

Technology Use Among People With Dementia and Their Caregivers: A Toward Human-Centered AI," AI Magazine, vol. 36, issue 4, pp. 5-8, 2015.

U.S. Perspective," Front. Public Health, vol. 8, May 2020.

[6] T. Wangmo, "Caring for Older Adults with Dementia: The Potential of

Assisted Technology in Reducing Caregiving Burden," in Intelligent

Assistive Technologies for Dementia, F. Jotterand, M. Ienca, T. Wangmo and

B. S. Elger, Eds., New York, Oxford University Press, 2019, pp. 95-109.

[7] World Health Organization. "Assistive technology." World Health

Organization. Accessed: Feb. 24, 2024. Available:

https://www.who.int/news-room/fact-sheets/detail/assistive-technology

[8] B. Bouchard, K. Bouchard, and A. Bouzouane, "A smart cooking device for

assisting cognitively impaired users," J. Reliab. Intell. Environ., vol. 6, pp.

107-125, April 2020.

[9] J. Evans, M. Brown, T. Coughlan, G. Lawson, and M. P. Craven, "A

Systematic Review of Dementia Focused Assistive Technology," in Human-

Computer Interaction: Interaction Technologies, M. Kurosu, Ed., Cham,

Springer International Publishing, 2015, pp. 406-417.



Open Science and Goodhart’s Law



Tomaž Pisanski Vladimir Batagelj Jan Pisanski

pisanski@upr.si Vadimir.Batagelj@f mf .uni- lj.si jan.pisanski@f f.uni- lj.si

University of Primorska University of Primorska University of Ljubljana

Koper, Slovenia Koper, Slovenia Ljubljana, Slovenia

IMFM IMFM

Ljubljana, Slovenia Ljubljana, Slovenia



ABSTRACT Main players involve authors, editors, referees, publishers,

libraries, readers, universities, learned societies, funding agencies

The influence of Goodhart’s law to the development of Open Sci-

and taxpayers.

ence is discussed. Science Citation Index (SCI) and Open Access

Publishing within classical publishing model was time consum-

(OA) are important steps in the path from Science to Open Sci-

ing and required efforts from all parties. This somehow prevented

ence (OS). The main conclusion is that flawed openness replaced

the inflation and hyper-production of papers.

quality in Open Science.

The model was mainly "subscription model" where articles



KEYWORDS were available in printed volumes of a journal. University and

departmental libraries subscribed to major journals, covering

Open Science, Open Access, Article Processing Charges, Good-

selected fields of science. The contents of earlier volumes were

hart’s Law, Free Journal Network

available to library users. Most libraries were open to local com-

munity and also to visiting researchers. Several learned societies,



1 universities and institutes published their own journals, associ- FROM SCIENCE TOWARDS OPEN SCIENCE

ated with a given library and used them for exchange purposes.

1.1 Science Instead of paying subscription to a similar journal they would

Traditionally, scientists disseminated their findings by publishing simply exchange the journals. In this way a library was able

their results in scientific journals. This is a key mechanism for to save money to subscribe to journals that were not available

knowledge transfer among scholars and therefore an important for exchange. This was an important way for wealthy western

subject of cognitive science. In the old days, the process of writ- scientists to help scientists from Eastern block and third world

ing a scientific paper was completely different. The author had countries. Later the revenue from scientific publishing was one

to type the paper on a typewriter leaving spaces for handwritten of the main sources of income of major learned societies. Unfor-

greek letters, symbols and formulae. With the advent of copying tunately, by acquisitions and mergers eventually a very small

machines only cumbersome paper "cut-and-paste" method was number of huge multinational publishers emerged. These pub-

available. Smaller misprints were overtyped whilst larger correc- lishing houses control the field of scientific publishing.

tions required replacing whole pages. Professional typists, not

available for everyone, could speed up the process. Manuscripts

were sent for publication by ordinary mail in several iterations,

depending on the referees’ requests.

Rise of technology quickly brought up big changes. The intro- 1.3 Transition to digital

duction of personal computers replaced typewriters by keyboards

The advancement of technology, in particular ICT (Information

and drastically enlarged the population of those who were able to

and Communication Technology), in the second half of the twen-

compose texts on a computer and simple editors introduced cut-

tieth century with the transition from analog to digital completely

and-past method of writing. Specialised software for producing

transformed the process of scientific publishing.

high-quality scientific drawings and diagrams enabled publishers

The costs of all stages of publishing decreased. More and more

to request camera-ready manuscripts from the authors. Authors

work was transferred from publisher – printer to author. Pub-

no longer focused only on the subject of their work but also on

lishing a paper became easy and inexpensive. The number of

the look it will have when printed.

scientific journals started to grow even more rapidly.

Surprisingly, major publishers did not lower the cost of sub-

1.2 Characteristics of Classical Publishing scription to their journals. On the contrary, they started to bundle

Model journals. If a library wanted to continue subscription to a jour-

nal it had to subscribe to the whole bundle of journals, many of

Classical publishing model was robust and healthy. It was free

which it had no interest in.

for authors. Certain journals were even paying author fees. Sur-

When papers became available in a pdf form, the need for

prisingly, it was (almost) free for readers via libraries of public

printed versions decreased. This also meant there was no way to

universities.

prevent an unauthorised access to the paper. The first electronic

journals appeared.

Permission to make digital or hard copies of part or all of this work for personal

There is a big difference between subscription to printed jour-

or classroom use is granted without fee provided that copies are not made or

nal and electronic journals. Old volumes of printed journals re-

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this main in the library and are available to anyone having access

work must be honored. For all other uses, contact the owner /author(s).

to the library. On the other hand, volumes of electronic jour-

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

nals remain with the publisher who may deny access to the paid

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.cog.4 volumes after the subscription runs out.



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Pisanski et al.



2 GROWTH OF SCIENTIFIC PRODUCTION; computers. No peer review is needed. The paradigm "Publish or

SHIFT OF GOALS Perish" was upgraded to "Be Cited or Perish".

In the past century we experience tremendous growth of pub-

3 BIBLIOMETRICS

lished scientific works. There are several factors contributing to

With the Science Citation Index (SCI) a number of statistical mea-

this phenomenon.

sures were introduced that would help profiling an author, the

work or the journal. The science of bibliometrics was born. It was

2.1 Publish or Perish

later extended to scientometrics and ultimately to informetrics.

"Publish or perish" is an aphorism describing the pressure to

SCI introduced a number of measures, indicators or metrics,

publish academic work in order to succeed in an academic career. trying to capture certain properties of articles, authors and jour-

It first appeared already in the first half of the twentieth century.

nals.

For a long time the PhD was a sufficient proof of academic quali- journal cited half-life One such indicator is the . It is the median

fication. It was not unusual that the PhD Thesis was the greatest

article publication date for each journal citation during one calen-

scientific achievement of a scholar. Unfortunately, the growth in

dar year. In general, the journal cited half-life is small for recent

the number of universities and scientists led to inflation of PhDs.

journals while it is large for older, well-established journals. On

Some scholars would simply stop doing research when hired.

the other hand it depends, as any other indicator on the scientific

Employers began requesting a constant flux of publications as a field and the culture of publishing in that field. Finally, a large

proof of scientist’s devotion to research. Those who were unable

journal cited half-life indicates that publications in that journal

to maintain high publication rate would be discouraged to stay

remain relevant for a long time. Hence the new knowledge is

in academia. And quantity became a proxy for quality.

persistent and not merely mundane. Nowadays, it would not be

difficult to equip any bibliographic database with computation



2.2 Quality control of journal journal cited half-life.



Ever since scientific journals appeared in seventeenth century, 3.1 Journal Impact Factor the quality of publications was in the hands of scientists using

Notorious Impact Factor (IF) is a ratio between the number of

the system of peer review. This was natural since everybody

citations in a given time period - usually a calendar year, to the

involved: authors, editors, referees and readers were scientists.

articles, published in another time period - usually two calendar

With the growth of number of journals it became clear that

years before. Sometimes they present also 3-year or 5-year impact

not all journals apply the same standards for accepting a paper

factors.

for publication. Obviously there was a problem of quality control.

For some reasons a 2-year impact factor prevailed and became

In mathematics there was a secondary system in place. It

a standard. In certain sense a 2-year impact factor is complemen-

started in Germany before WWII. Eventually, three refereeing

tary to a journal cited half-life. Definitely, a 2-year IF is not the

journals were established, one in the Soviet Union, one in the USA

best indicator for mathematics when compared with other sci-

and one in Germany (nowadays at the European Mathematical

ences where citation culture is different. For instance, when 1756

Society). If a review of a paper does not appear in one of those

SCIMAGO journals covering the subject area of Mathematics

refereeing journals, the paper is likely not to be interesting for

are ranked according to Cites/Docs. (2 year) for the year 2023,

mathematicians.

among the top 50 journals only one journal has Mathematics as

For science in general there is a similar publication, called

the primary subject area.

Current Contents.



3.2 Metrics and Ranking

2.3 Citation index

Having different indicators for a set of journals is good. It gives

If we know for each paper the list of papers it cites, we can also

a higher dimensional description of each journal. However, each

produce the inverse index, i.e. the list of papers that cite a given

indicator may be used for sorting and hence for ranking. There

paper. Such index is called citation index. In 1964 Eugene Garfield

is a strong tendency to devise an indicator that would measure

conceived the Science Citation Index (SCI) [3]. Using citation

quality; an impossible task.

index one can easily detect most follow-ups to a paper covering

Never-the-less since early seventies the Impact Factor is con-

a topic of scientist’s interest. Clearly there are certain limitations.

sidered by many a proxy for the quality of a journal. The false

One has to select a collection of journals from where papers and

reasoning goes along the following lines:

their references are included. This may introduce some bias. •

Outstanding scientists publish their work in high-quality

journals.

2.4 Impact • The work of outstanding scientists is frequently cited.

On the other hand, if the database is stored in a computer one can •

High-quality journals have high impact factor.

easily perform some statistics. For instance one can store with •

Wrong conclusion: Work published in a journal of high

impact factor is of high quality.

each paper the number of its citations. One can also compute

how many citations each author has. This may, again, help the

scientist to select the papers to look-up and authors to follow. 3.3 Power law and related statistical laws

However, it also leads to all kind of rankings. Citation indices When plotting the distribution of ranked impact factors, one

became very useful not only to scientists but also to their employ- can observe the exponential decay. Impact factor 𝐼 𝐹 𝑟 of rank ( )

ers and funding bodies. Instead of comparing the added knowl- 𝑟 journal is proportional to 1 𝑟 for some constant 𝛼. This is / 𝛼

edge of someone’s research, it is "sufficient" to select the highest known as the power law. Roughly speaking this means that there

ranked candidate. Selection can be done by administrators or are only a few journals with high impact factor and there are



Open Science and Goodhart’s Law Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



many journals with small impact factor. One way of stating this For ordering extra reprints it was not uncommon to charge the

is that 20 percent of most cited journals receive 80 percent of authors. Also, one could be charged for insisting that the fig-

citations; see [7]. Several of these laws were first observed in ures be printed in colour. On the other hand, some prestigious

bibliometrics. However it is interesting to observe that these journals, started requesting article processing charges (APC).

laws are universal and apply to a variety of unrelated situations, Employers and funding agencies soon recognised that if they

perhaps by choosing the right value of parameter 𝛼. want their scientists to publish in the journals with very high

impact factor, they will have to cover the costs of APC. Some



4 GOODHART’S LAW scientific disciplines such as mathematics declined this model.

When judging whether to pay APC or to send a graduate student

British economist Charles Goodhart is credited with expressing

to an international workshop many mathematicians give prece-

the core idea of a law in a 1975 article on monetary policy in the

dence to student. However, the publishers realised that money

United Kingdom. "When a measure becomes a target, it ceases to

could be presented as a proxy for quality and raised their prices.

be a good measure"; see [4]. Another way of saying is "that once

a metric is used as a basis for decision-making or control, it loses

its reliability as an accurate measure".

The main rationale behind this law is adaptation or even gam- 5.2 Diamond- and Green Open Access

ing to improve one’s rank. If high rank means high reward, it

In the last decade of the twentieth century some of the first

is plausible, that some people will do anything to improve their The Electronic

purely electronic journals appeared. For instance,

score for the given indicator. Each measure for assessment of Journal of Combinatorics (E-JC) was funded in 1994. It was free for

researchers and journals became prone to Goodhart’s Law.

authors and readers. It is run by scholars and not by commercial

publishers. This is nowadays called a , with diamond open access

4.1 Goodhart’s Law and Bibliometrics no cost for authors and no cost for readers. E-JC is a founding

member of the Free Journal Network [9].

When the number of publications are counted, researchers will

Even before that, in 1991, an e-print server was launched arXiv

tend to split long papers and publish short bits and will thus

where preprints in some scientific disciplines may be uploaded.

increase the number of publications. Instead of publishing papers

Nowadays, such posting of preprint before peer review is called

alone they may increase their output several times if more coau-

thors sign the same publications. There is no increase in quality green open access.

For a while it seemed that this model will force big publish-

of their output.

ers to lower the prices of their journals. In the battle between

When the number of citations decide who is winning a grant,

scientists and multinational commercial publishing houses, the

the number of citations soared. The authors started citing their

scientist should have won. It was expected that governments will

own papers, even if citations were not needed. When self-citations

support scientists in the fight against greedy publishers; [10].

ceased to count, friendly researchers helped each other with cita-

However, politics works in mysterious ways.

tions.

When the ℎ-index was introduced, the key publications of

potential PIs in a research group had to be cited.



Employers and funding bodies understood that blindly re- 5.3 Budapest Open Access Initiative (BOAI)

warding high production authors with large impact papers does

In December 2001 there was a two-day conference, producing a

not mean rewarding high-quality science as there was no prob-

declaration called Budapest Open Access Initiative. The declara-

lem in publishing papers in low-quality journals and getting

tion was launched in February 2002, having 16 original relatively

many citations in such journals. On the contrary, in many cases

unknown individual signatories. This initiative has been financed

those fabricating papers and citations easily outperformed best

by Soros’ private Open Society Institute with 3 200 000 USD. It is

researches. That is why the quality of journal in which the paper

recognised as one of the major defining events of the open access

was published became important; in practice this meant journals

movement, [8]. Up till now it has been signed by about 0.1% of

with high impact factors.

world scientists.

By Goodhart’s law, predatory publishers flourished, multiply-

ing their journals and boosting their impact factors.

Production of new knowledge ceased to be important. It is

the impact of their work published in high-impact journals that 5.4 Gold Open Access and APC

counts. Gold Open Access requires the author to pay Article Processing

There is a difference in Goodhart’s law and other laws, used in Charges (APC) to keep article freely available to the reader. Cur-

bibliometrics. Goodhart’s law involves time and decision while rently a typical APC exceed 3000 EUR. This brings enormous

laws based on power law are based on rankings. profits to publishers. It is estimated that the costs per article

There are not many studies of Goodhart’sl law in bibliometrics. should not exceed 1000 EUR.

An exception is a comprehensive study reported in [2]. Clearly, APC model is not viable if costs are indeed covered

by the author. The author must find someone who will cover the

5 OPEN ACCESS A STEP TOWARDS OPEN costs of APC. This is an ideal model prone for corruption at all

SCIENCE levels. In the APC model, money becomes a substitute for quality.

and researchers must compete for money that will cover their

5.1 APC model publication costs.

The idea that authors or their institutions should make financial The difference between Green and Diamond Open Access

contributions for their publications is not new. In the times of pa- and Gold Open Access is huge. One can speak of two opposing

per publications, the publisher would grant some, say 25 reprints. concepts sharing the same name: Open Access.



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Pisanski et al.



6 IMPLEMENTATION OF OPEN SCIENCE OS is only open to those within the system. Independent criti-



6.1 cal scientists adhering to high ethical standards are left out. OS is Recommendations, Declarations, . . . .

concerned only with current and future publications. No pressure

There are numerous mostly political papers, initiatives, recom-

to commercial publishers to open archives of papers published

mendations, declarations, pushing for Open Access, Open Sci-

previously under paywall and make them free for everyone. A

ence, Open Research, etc. Due to limited space we mention only

large part of science remains closed to authors and readers that

a few of them. For more information, see e.g. [1, 5].

are unable to secure money.

While the OA has been launched bottom up by 16 individuals

Scientists no longer decide what is the quality of their work.

meeting in Budapest, backed up by 3.2 Million USD from Open

They even have to pay private companies to tell them that. For

Society Institute, OS is a political concept that is revolutionis-

instance, public employers and public funders base their decisions

ing Science from top to bottom. It seems it was first formally

about the quality of candidates on data bought from private

expressed by UNESCO in November 2021 in the UNESCO Rec-

companies running services, such as WoS or Scopus.

ommendation on Open Science.

There is a problem of citation culture among different scientific

The concept has been embraced by European Commission

fields. For eaxample, if average scientists from a scientific field,

that pushes it through Horizon Europe down to member states.

say 𝐴 with high ℎ-index compete for money in another field, say

For instance, Slovenia recently received 16 000 000 EUR for pro-

𝐵 they may be ranked higher than the best scientists of the field

moting OS. It appears this money does not go for science but for

𝐵 . This may have negative effect on the future of the field 𝐵 .

administration.

There is no real need for repositories at every public institu-

The Barcelona Declaration on Open Research Information

tions. One repository at the European level with several backups

emerged from a workshop with over 25 experts interested in

would suffice. Instead of creating jobs for scientists repositories

changing the research landscape. The declaration that was signed

create jobs for administration. Repositories of papers and data are

on 24 April 2024 is a political statement of an unidentified commu-

not intended for individual scientists. It appears they are intended

nity. The authors do not act as individuals and do not represent for the AI data-harvesting algorithms of private companies. This

scientific community. They write: . . . we, as organizations that

service again will be sold back to scientists.

carry out, fund and evaluate research, commit to the following . . . ". One could say, that the OS is a model that diverts public money

The first out of four commitments is strong. We will make open-

from scientists to administration and private companies.

ness the default for the research information we use and produce. It

leaves no room for science outside Open Science. While OA was ACKNOWLEDGEMENTS at first optional, OS makes it mandatory.

Work of VB is supported in part by ARIS (research program

P1-0294 and research projects J1-2481 and J5-4596). Work of JP



6.2 is supported in part by ARIS (research program P5-0361 and Goodhart’s Law and Open Science.

research projects J5-2551 and J5-4596). Work of TP is supported

Since journal impact factor remains a measure, the number of

in part by ARIS (research program P1-0294 and research projects

journals and publishers keeps increasing. In general, neither OS,

N1-0140, J1-2481 and J5-4596).

nor universities nor funding organisations address the problem



of low-quality high-impact factor predatory journals. Several REFERENCES scientists lower ethical standards and publish their papers in

[1] Batagelj, Vladimir. 2024. Bibliographic mix. [Online; accessed 21-September-

expensive journals with mild or no refereeing. The costs are 2024]. (2024). https://github.com/bavla/biblio/blob/master/doc/sreda1348.pd

f .

reimbursed by their employer or funding organization.

[2] Michael Fire and Carlos Guestrin. 2019. Over-optimization of academic

Ever since the number of publications became a measure, sci- GIGASCIENCE

publishing metrics: observing goodhart’s law in action. , 8, 6,

entists tend to publish papers with partial solutions to the prob- (June 2019). doi: 10.1093/gigascience/giz053.

[3] Eugene Garfield. 1964. "Science Citation Index"—A New Dimension in In-

lem. The number of co-authors per paper keeps increasing. The

dexing. , 144, 3619, 649–654. Science

number of published papers grows out of proportion. [4] Charles E. Goodhart. 1975. Problems of Monetary Management: The UK

After citations became a measure, the number of references per Experience. In . Reserve Bank of Australia. Papers in Monetary Economics

[5] Kotar, Mojca. 2022. Open science in the european research area (era). [Online;

paper keeps increasing. Some prominent journals fight citation

accessed 21-September-2024]. (2022). https://url.um.si/p7CSj.

inflation by limiting the number of references a paper may have. [6] Vojtech Kovarik, Christian van Merwijk, and Ida Mattsson. 2024. Extinction

risks from ai: invisible to science? (2024). https://arxiv.org/abs/2403.05540

Clearly, the references published by competing authors are first

arXiv: 2403.05540 . [cs.CY]

to go. F. Rouge,

[7] Vilfredo Pareto. 1896. Cours d’economie politique, volume i and ii.

Since APC remains as a valid model in OS, all kinds of unethical , 250. Lausanne

[8] Wikipedia contributors. 2024. Budapest open access initiative — Wikipedia,

practices emerge. In many cases, a ghost author, who did not

the free encyclopedia. [Online; accessed 31-August-2024]. (2024). https://en

contribute to the paper but may secure covering APC costs is .wikipedia.org/w/index.php?title=Budapest_Open_Access_Initiative&oldi

added to the list of authors. d=1242834910.

[9] Wikipedia contributors. 2024. Free journal network — Wikipedia, the free

It is disturbing that the goal quality is absent in some docu-

encyclopedia. [Online; accessed 31-August-2024]. (2024). https://en.wikiped

ments on OS, such as the Barcelona Declaration. The quality is ia.org/w/index.php?title=Free_Journal_Network&oldid=1231212463.

[10] Wikipedia contributors. 2024. The cost of knowledge — Wikipedia, the free

replaced by openness and Goodhart prevails. Scientists will adapt

encyclopedia. [Online; accessed 21-September-2024]. (2024). https://en.wiki

to new goals.

pedia.org/w/index.php?title=The_Cost_of _Knowledge&oldid=123957293

4.



7 CONCLUSION AND SUGGESTIONS

OS has some serious flaws. The main concern of OS is that scien-

tists financed from public funds are not allowed to profit from

their work – but everybody else can.





The Consistency of the Research Field Data




A Case Study of Library and Information Science in Slovenia



Jan Pisanski

Faculty of Arts, University of Ljubljana, Aškerčeva 2

Ljubljana, Slovenia

jan.pisanski@f f.uni- lj.si



ABSTRACT mapping of WOS categories to the fields of science, used in SICRIS.

Also see [2] for a brief history and overview of various mappings

SICRIS (Slovenian Current Research Information System) pro-

of fields of science/research fields.

vides a service listing top Slovenian researchers in a particular

While we were primarily interested in using bibliometric data

research field. In Web od Science (WOS) each journal is assigned

for representation of a particular research field, this can then also

one or more categories (research fields). When comparing these

be commonly used for evaluation of research. There are two main

data for the research field of library and information science (LIS),

approaches: expert evaluation and bibliometric analysis. While

we found that several of the top authors in the field according

expert evaluation is more traditional and qualitative, bibliometric

to SICRIS rarely or never published in the journals deemed to

analysis is quantitive in its nature. However, both of them have

belong to LIS in Web of Science. Several other authors, who were

their downsides. For discussion on trustworthiness of experts, see

not assigned the research field of LIS in SICRIS, were among the

e.g. [4]. Amongst others, Leiden Manifesto [5] points to dangers

most published Slovenian authors in LIS in Web of Science. This

of using bibliometric data without closely examining the context.

is an indication that results of any analysis of LIS in Slovenia will

It suggests various indicators should be used when evaluating

depend greatly on the criterion/criteria used.

researchers and their work and that bibliographic analysis should



KEYWORDS support expert evaluation.

Bibliometric Analysis, Research Fields, Slovenia, Library and

Information Science.

3 RESEARCH

1 INTRODUCTION While there are several different ways to approach the extent

As part of a project focusing on high-level bibliographic services, of publication on library and information science in Slovenia,

i.e. novel services based on existing bibliographic data, we in- we looked at the publications in Web of Science (WOS). This

tended to perform a domain analysis of library and information was done with intention to identify the most prominent works

science (LIS) in Slovenia from a bibliometric perspective. This and authors, as journals indexed in WOS go through a rigorous

contribution describes the initial step that was simply intended process. However, this also means that we omitted from analysis

to provide an overview of research and researchers but came all other publications, including papers published in Slovenian

up on several issues regarding assignment of research fields and language journals.

yielded some interesting findings, particularly for establishing Although it may not have the same coverage of social sciences,

the scope of the research field in Slovenia and elsewhere, but for this kind of insight WOS compares favourably to similar ser-

also in view of providing better services to the users of academic vices, such as SCOP US and Google Scholar, as it allows searching

bibliographic databases. based on WOS Categories field which represents the subject cat-

egories/research fields of the journals [7]. It has to be noted that

2 BACKGROUND the WOS Categories field provides general information about the

thematic nature of the journal rather than each particular paper.

There is a lack of a bibliometric overview of information scien-

However, this is still the easiest way to get a quick overview

tists and librarians in Slovenia and their works, collaborations

of a research field, as all of the subject related data pertaining

etc. One of the reasons is the nature of the field(s) of library

to individual papers in WOS describes the thematic nature of

and information science, where sometimes it is difficult to draw

the papers in higher granularity. Each journal in WOS can be

the distinction where the boundaries of the field are. On the

assigned one or more subject categories.

other hand, relatively high-quality information on Slovenian re-

In April 2024, we performed a search in WOS Core Collection

searchers is stored in SICRIS (https://cris.cobiss.net/ecris/), the

for publications where Address field included »Slovenia« and

Slovenian current research information system, which provides

the value in the WOS Categories field was »Information Science

multiple tools for basic bibliometric analysis.

& Library Science«. We did not limit the search to any particular

Other studies have focused on the research fields in Slovenia

time period, which means that the more experienced authors

(e.g., [1], [6], [8]), however at a more general level, not specifically

were more likely to be on the list. Also, we did not limit the

for LIS and without mention of the issues related to research fields

results to particular types of publications (e.g. articles), since

discussed herein, whereas [2] discusses among other things the

the “linked records” categorization in SICRIS, which we used

Permission to make digital or hard copies of part or all of this work for personal

in comparison, also does not limit this. However, even if we

or classroom use is granted without fee provided that copies are not made or

did, the situation regarding top authors would still be similar.

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this Since Address was limited to Slovenia, the list excludes Slovenian

work must be honored. For all other uses, contact the owner /author(s).

authors who published research while working in other countries

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

and may also be missing authors with otherwise faulty Address

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.cog.5 data.



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia J. Pisanski



Table 1: Top 10 Slovenian authors in SICRIS, their research fields, number of established links to WOS publications in

SICRIS and number of publications with LIS as WOS Category



Author Research field 1 Research field 2 WOS LIS fraction



A Information science and librarianship Interdisciplinary research 89 0 0.0000

B Administrative and organisational sciences Information science and librarianship 55 2 0.0364

C Information science and librarianship 55 47 0.8545

D Economics Information science and librarianship 53 14 0.2641

E Computer science and informatics Information science and librarianship 39 0 0.0000

F Information science and librarianship Plant production 33 17 0.5152

G Information science and librarianship 31 29 0,9355

H Information science and librarianship 24 2 0.0833

I Information science and librarianship Economics 24 3 0.1250

J Information science and librarianship 21 18 0.8571



In SICRIS each researcher can be assigned up to two research There are several reasons for this phenomenon. In the SICRIS

fields, according to the ARIS (Slovenian Research and Innovation top 10 list, two prominent authors, marked in Table 1 as B and H,

Agency) categorization, which is “roughly harmonized with the published a majority of their works in different fields, confusingly

Field of Science and Technology Classification in the Frascati not explicitly named in SICRIS, before clearly switching their

Manual (OECD)” [1]. There are different levels of categoriza- research interest to LIS. For some others their area of expertise

tion with the first level representing science, the second level is on the boundaries of LIS, although, what constitutes LIS can

representing field and the third level representing subfield. For be debated. For example, two of the top 10 authors (A and E),

instance, Information science and librarianship is deemed as a including the top Slovenian author in LIS according to SICRIS,

field belonging to social sciences with no further subfields. On do not have a single work published in what WOS considers to

the other hand, Economics also is a field of Social sciences, but be LIS journals. In the case of author A, their second research

it has subfields, such as Business sciences. Authors may be as- field, Interdisciplinary research, provides a better understanding

signed a certain research field, even if it has subfields, or a certain on the nature of their publications.

subfield. According to the well-known Bradford’s law [3] there are

Among several features, SICRIS provides a higher-level ser- going to be some works published in journals that may not appear

vice (https://cris.cobiss.net/ecris/si/en/top/researcher) where a to be particularly relevant to a particular topic or research field.

user can look up most prominent Slovenian authors in a spe- For instance, [2] found such distribution for Slovenian agriculture

cific research field based on different indicators (e.g. number research group publication. However, there is still the question

of linked records and citations in WOS and SCOP US, h-index, of whether such a list of top authors represents the LIS research

other indicators linked to local evaluation practices). While this field well.

is not necessarily the only tool a user of SICRIS can use to get It has to be noted that the results were similar even if we

an overview of researchers in a research field, it is certainly the used other criteria in SICRIS. For example, the top 10 authors

quickest and easiest to use. by number of citations in WOS are the same, only the order

Compared to some other research fields, where it is harder changes slightly. Also, the list of the top 10 authors by number

to find the equivalents in both of the databases, LIS has the of connected records in SCOP US has two authors that do not

advantage of being relatively straightforward. While the names appear in Table 1, neither of whom again had more than 2 works

used for the research field in the two systems slightly differ published in LIS journals, according to WOS.

(»Information science and librarianship« in SICRIS; »information To further complicate the matters in terms of transparency of

science and library science« in WOS), at least the core of the two data, SICRIS user interface only lists the author’s first research

subject categories should be the same. field, in the top authors lists, which can be confusing to a novice

While the actual ranking of LIS authors in SICRIS does vary user, as it may appear that some of the top authors do not belong

slightly according to the indicator chosen (i. e. number of works to said field. In fact, many of the first year students of LIS at

in WOS and SCOP US, number of citations, etc.) there is a core the University of Ljubljana skipped such authors, when asked

group of authors that occupies top places for several categories. to provide a list of top authors in the LIS field, based on SICRIS

Table 1 shows the top 10 authors based on the number of linked data/user interface.

records in WOS according to SICRIS. In the table each author is Another issue that came up was that one of the top ten re-

represented with a letter of the alphabet for anonymity. searchers is a foreign citizen with an ARIS researcher number

When comparing the data of LIS authors in WOS, whose ad- having mostly worked outside of Slovenia. While this certainly

dress is in Slovenia (Table 2) and, the list of most prominent reflects the international nature of science, it may not accurately

authors in LIS in Slovenia based on number of publications in reflect the state of LIS research in Slovenia. However, this issue

WOS as provided by SICRIS (Table 1), we found a relatively large is not particular to LIS.

discrepancy. As seen in Table 1, half of the top 10 authors in LIS, On the other hand, there was also a notable group of authors

as provided by SICRIS, had less than half of their works published that was not assigned to the research field of LIS in SICRIS, whose

in LIS journals, as indexed by WOS. In fact, for all five of these works appeared relatively frequently in LIS journals in WOS.

authors the proportion is less than one third. Several new authors appeared in the top 10 list, if we only looked

at the data on publications in WOS. Two of those, marked here



The Consistency of the Research Field Data Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



Table 2: Top 10 Slovenian authors by the number of publications in the WOS Category Information Science & Library

Science journals and their assigned research fields in SICRIS



Author LIS Research field 1 Research field 2

C 47 Information science and librarianship

G 29 Information science and librarianship

J 18 Information science and librarianship

F 17 Information science and librarianship Plant production

K 16 Economics Computer science and informatics

L 15 Mathematics Computer intensive methods and applications

M 14 Information science and librarianship

D 14 Economics Information science and librarianship

N 13 Computer science and informatics

O 13 Information science and librarianship



as M and O, are authors whose field is declared in SICRIS to be criterion/criteria chosen to represent the field, as even the very

LIS. But there are also three authors who do not have LIS named top authors by one criterion may not be considered to be working

among their up to two research fields in SICRIS. Author here in the field by another.

marked as K mainly worked in bibliometrics, which was also the Further research could establish whether the issues found in

LIS topic covered by author L, while author N mostly wrote on this pilot study exist in other research fields and for other data

the topic of business intelligence. Such instances are not isolated, (e.g. different databases, different time periods). However, not

as several other authors who do not have LIS as a stated research all research fields in one database may have their exact equiva-

field in SICRIS just missed the top 10 list. lent in another database. Cognitive science, for example, is not

considered to be its own research field neither in SICRIS nor in



4 DISCUSSION WOS.

Generally, we suggest providing a clear explanation of the

While this is a brief look into a relatively small slice of two

topical nature of the work of each author, when providing list

databases, SICRIS would benefit from a recognition of the issue.

of top authors in a research field, Another possible solution is

The simplest solution would be to provide a clear explanation

omission of authors, who have a relatively low percentage of

on the nature of the data provided, when viewing top author

works published in journals from a research field from lists of

lists by research field. Alternatively, additional services could be

top authors in that field.

provided, based on other subject related data, such as WOS Cate-

While bibliographic databases offering high-level services that

gories or even keywords [7]. Ideally, services based on Bradford

bring to light otherwise “hidden” data are definitely welcome,

distribution would be provided.

users would benefit from indication of imprecise nature of data

The appropriateness of both the scope and designation of

and/or additional services that would try to account for the im-

SICRIS research fields of authors and the WOS Categories can be

precision.

debated. Their assignation procedures would benefit from greater



transparency. ACKNOWLEDGEMENTS

There is the issue of assignment of up to two research fields

The author acknowledges the financial support from the Slove-

per author in SICRIS, as this does not necessarily accurately

nian Research and Innovation Agency (research program P5-0361

represent the involvement of each individual researcher. In our

and research projects J5-2551 and J5-4596).

relatively small case study of LIS we found several authors whose

assigned research fields could be viewed as misrepresented.

While research today is generally multidisciplinary and some REFERENCES

[1] Tomaž Bartol, Gordana Budimir, Doris Dekleva-Smrekar, Miro Pušnik, and

researchers can shift their area of interest in research from one

Primož Južnič. 2014. Assessment of research fields in scopus and web of

research field to another during their career due to various rea- science in the view of national research evaluation in slovenia. Scientometrics, sons, this ought to be reflected in any lists of researchers from a 98, 1491–1504.

[2] Tomaž Bartol, Gordana Budimir, Primož Južnič, and Karmen Stopar. 2016.

particular research field.

Mapping and classification of agriculture in web of science: other subject

Also, while well-established, WOS would benefit from a more categories and research fields may benefit. , 109, 2. Scientometrics

[3] B. C. Brookes. 1969. Bradford’s law and the bibliography of science. , Nature

transparent explanation of the nature of WOS Categories. Even

224, 5223.

then, there can at least be a discussion, whether some of the [4] Alvin I. Goldman. 2001. Experts: which ones should you trust? Philosophy

journals are assigned to the correct research fields. , 63, 1, 85–110. doi: 10.1111/j.1933- 1592.2001.t and Phenomenological Research

b00093.x.

Conversely, as there are authors that publish relatively fre-

[5] Diana Hicks, Paul Wouters, Ludo Waltman, Sarah De Rijcke, and Ismael

quently in LIS journals in WOS, but do not have the according Rafols. 2015. Bibliometrics: the leiden manifesto for research metrics. Nature,

520, 7548, 429–431.

research field associated with them in SICRIS, a list of such au-

[6] Luka Kronegger, Anuška Ferligoj, and Patrick Doreian. 2011. On the dynamics

thors could help with a subject classification of authors that is Quality & Quantity of national scientific systems. , 45, 989–1015.

more reflective of their production. [7] Daria Maltseva and Vladimir Batagelj. 2020. Towards a systematic descrip-

tion of the field using keywords analysis: main topics in social networks.

Scientometrics, 123, 1, 357–382.

5 CONCLUSION [8] Tomaž Pisanski, Mark Pisanski, and Jan Pisanski. 2020. A novel method for

determining research groups from co-authorship network and scientific fields

Our research indicates that any bibliometric analysis of the re- of authors. Informatica, 44, 2. search field of LIS in Slovenia is bound to be influenced by the



To be or not to be… a Nahuatl language learning App. The



long-term survival or discontinuation of indigenous



language learning apps on the example of Nahuatl



Evelyn Fischer †

MeiCogSci Student

University of Vienna

Vienna Austria

evelyn.fischer@posteo.com



Abstract In addition, by pure chance, many of the most learnt languages have less complex morphology than many minority Language learning apps for indigenous languages differ from the languages, and the apps that were developed with more analytic mainstream language apps in that they are not targeted at languages in sight, such as English, are not easily fully commercial success and might need to accommodate different extendable to Morphologically Rich Languages, such as isiZulu linguistic and cultural aspects than the most learnt languages. [3], Turkish [4] and Nahuatl. The present paper considers the present and past Nahuatl In the following, the focus lies on Nahuatl, the Mexican language apps, some of which were discontinued, and asks what indigenous language with the highest number of speakers, 1.5 would be necessary for such apps to achieve long-term survival. million. Nahuatl is one of 68 indigenous Mexican languages, and



Keywords despite its historical prestige remains endangered, a challenge it

shares with virtually all indigenous American languages. Nahuatl

Language learning, human-computer interaction, software language learning apps contribute not only to thwart its digital

usability, software translation, trends in software development death, but also to increase its visibility and prestige, and to

support the efforts of Nahuatl learners to become “new speakers”



1 [5] of the language. The role of new speakers is described as Introduction

“very important, often essential for language revitalization

The strong digitalization of modern life is bringing about big projects” by [6] who work directly with Nahua and other

changes to the global and local societies. One of the results of the minority groups in Mexico.

technological changes is the rapidly ongoing globalization, and



communities from languages with small numbers of speakers to 2 one of the mechanisms of globalization is the shift of Initiatives to Localize Software and Platforms

a smaller number of global languages. In many countries with in Minority Languages

colonial history, this follows centuries of, at best, ignorance of, The second decade of the 21st century was a witness of increased

and at worst, active discrimination and eradication of the efforts to increase the visibility and presence of minority and

indigenous populations, their languages and their cultures. indigenous languages in the digital sphere. These were often led

The dominance of global languages is clearly seen on the by digital language activists and sometimes supported by the

Internet, where 80% of websites are written in just 8 of the companies whose software was the focus of the projects. The

estimated 7000 world´s languages [1]. The ascension to the present section describes some of the initiatives taken and

digital realm is a challenging task and in 2013 [2] estimated that, discusses the long-term results of the work to promote

at best, 5% of the world's languages will ascend to the digital indigenous languages. Particularly, it looks at whether the work

world, and the rest will suffer a “digital language death”. of the activists resulted in a long-term inclusion of the relevant

One of the ways a language can be present on the internet is language in the software or platform that was the aim of their

by being the object of mobile learning apps. Mobile apps efforts.

supporting the acquisition of minority and indigenous languages The Mozilla Foundation, known for its web browser, Firefox,

may differ from apps targeting global languages in that minority launched in 2012 the initiative “Native Mozilla” that aimed to

language learning apps would typically not be aiming at localize the browser into many of America’s indigenous

commercial success, would have lower budgets, or even be done languages. 50 languages from 10 countries were targeted [7],

on volunteer basis by smaller group of language activists. many of which are spoken in Mexico, such as Ch’ol, Kaqchikel,

Kʼicheʼ, Mixteco (2 varieties: of Mixtepec and of Yucuhiti),

Nahuatl (2 varieties or, by other accounts, 2 closely related

Permission to make digital or hard copies of part or all of this work for personal or languages: Highland Puebla Nahuatl of Mexico and Nawat Pipil

classroom use is granted without fee provided that copies are not made or distributed of El Salvador), P’urhépecha and Triqui [7, 8]. The translations

citation on the first page. Copyrights for third-party components of this work must are done via the collaborative translation platform Pontoon for profit or commercial advantage and that copies bear this notice and the full



be honored. For all other uses, contact the owner/author(s). (https://pontoon.mozilla.org/). As part of the initiative, for Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia example, a Hackathlon was organized in Oaxaca in 2018 with © 2024 Copyright held by the owner/author(s). representatives of 15 languages [8]. http://doi.org/10.70314/is.2024.cog.6



mobile phone browser Firefox Focus, which is available in books, 5 scientific articles, 10 movies or series episodes, 1 Aimara, Ixil, Kaqchikel, K’iche', Maya, Miahuatlán Zapotec, videogame, 6 radio stations where Nahuatl is transmitted along Mixtec of Mixtepec and of Yucuhiti, Nahuat pipil, Navajo, with other languages, Wikipedia in 10 separate varieties, and 5 Paipai, P'urhepecha, Quechua, Tének and Triqui [7, 10]. websites with a Nahuatl version. Considering that Nahuatl is the There were also attempts to localize the social media platform Mexican indigenous language with the biggest number of Facebook (https://www.facebook.com) into minority and speakers, this is a low number. indigenous languages. In 2012, the official Facebook translation However, there is hope that their number is increasing. For platform (https://www.facebook.com/translations/), where users example, the 5 scientific articles were published in 1959, 2019, could add and vote on the accuracy of volunteer translations, had 2022, 2022, 2023; the 12 books in 2008, 2013, 2014, 2015, 2015, partial translations in about 100 languages, including Cherokee 2016, 2017, 2017, 2017, 2019 and 2021, (the twelfth is the Bible [11] . However, frustrated with the slow progress, Manuel Neskie which has been translated into many varieties) [14]. One of the made a browser overlay that allowed the translation of Facebook websites is also recent, as it was published in 2023 [22]. In 2024 menu into Secwepemctsín directly in the user´s browser [11], the Mexican presidential election was simultaneously interpreted which was later extended to many other languages, including into Nahuatl, and remains, as of 30.08.2024, available on Nawat [12] spoken in El Salvador. In 2015, a group of activists YouTube [23]. In addition, a Master´s degree in Nahuatl translated 24 000 words into Aymara and submitted it to language and culture, taught completely in an indigenous Facebook for revision [13] . However, as of 31.08.2024, same as language was launched in 2019 [24], first of its kind. We see on 07.04.2019 [14], only three indigenous American languages therefore a clear tendency of growth, and it could be expected Kaqchikel and Triqui spoken in Mexico and in the South resources in the language available online for language learners, American language Guarani [9]. The other languages are in many of whom are descendant from Nahuatl speakers and wish various state of completion: Ixil (13%), Kichwa (0%), to reclaim the language of their ancestors. Miahuatlán Zapotec (14%), Mixtec of Mixtepec (9%) and of [14] collected information about the different monolingual Yucuhiti (29%), Nahuat pipil (0%), Paipai (1%), P'urhepecha Nahuatl language media available for those searching on the (9%) and Quechua (3%). The situation is slightly better for the internet. The results encompassed 12 monolingual novel-sized languages is far from achieved and, as of 30.09.2024, Firefox 115 The importance of maintaining Nahuatl language learning apps was available only in three American Indigenous languages: is clearer if we consider that there are relatively few other However, the goal of providing translations into 50 American 3 Digitally available Nahuatl language media



were available for menus on Facebook: Cherokee, Inuktitut and that more media will become available in Nahuatl soon. Inupiaq, and the Facebook Translation App appears to have been It is also interesting to consider the case of the work to bring discontinued. Nahuatl as a language available in Google Translate A similar fate was met by the attempts to increase the number (https://translate.google.com/) . In 2010, Google announced their of languages offered on the Duolingo language learning platform plans to add Nahuatl and Maya to the tool [25], but this service (https://es.duolingo.com/). As of 10.06.2024, two indigenous was finally only introduced in June 2024 [26]. As of September languages are available on Duolingo: Hawaiian and Navajo [15]. 2024, the following ten indigenous American languages are However, previously also Guarani must have been available on among the 243 languages available on Google translate: Aymara, the app, as evidenced by the surprise some users expressed at its Guarani, Hawaiian, Kalaallisut, Mam, Nahuatl, Quechua, sudden lack [16, 17]. The course is still available on the website Qʼeqchiʼ, Yucatec Maya and Zapotec. version as of 31.08.2024 [18]. Between 2013 [19] and 2021 [20] Considering mobile apps in particular, [14] identified 39 new courses in development were stored in the Duolingo mobile phone apps related to Nahuatl. Most of them, 23, are Incubator, where the users themselves could contribute to adding Bible apps, although due to double versions, there are only 14 new languages to Duolingo. Indigenous languages such as different variant versions of Bible available as a mobile app. The Yucatec Maya and Kʼicheʼ were present on the Incubator and other 16 apps include 6 dictionaries (one with a Na huatl user Duolingo itself credits the volunteers for helping to make, among interface), 3 text collections, a (faulty) automatic translator, a others, the Navajo and Hawaiian language course [21] . However, multi-component app CEM, which combines dictionaries and Incubator was discontinued in 2021. morphological analyzer, the messaging app Telegram that offers As we have seen, a common trend in the translation efforts of user interface in Huasteca Nahuatl (albeit cannot be chosen in its Facebook and Duolingo is the move from community and standard menu, but is available for download for those who have volunteer-based translations to commercial translation directed the relevant link), and, finally, 4 Nahuatl language courses, by the company. On the one hand, the reliance on unpaid work is discussed below. problematic for a company with huge profits – something that

Duolingo itself lists as the reason for ending the volunteer



program [21]. This is especially true when those delivering this 4 A Partial History of Nahuatl Language Apps work might already be in unprivileged financial situation, as As of 01.07.2023, four Nahuatl language courses were available many indigenous language speakers are. On the other hand, for Android [14] : Aprende náhuatl [27], Beginner Nahuatl [28], however, this deprives the communities of the possibility to Kamatlama [29] and NahuatlApp [30] . Aprende náhuatl (Spanish contribute to making their language more visible on the popular for “learn Nahuatl”) is a vocabulary training app, with texts and platforms. It is interesting to note that the translation platform of videos, produced by the National Institute of Indigenous Peoples, open-source based Mozilla products remains active, and, for a government agency and it was, as of 01.07.2023, downloaded example, the last changes to an Indigenous Language – Zapotec more than 10 000 times [14]. As of 26.09.2024, it is available on – have been done on 26th August 2024.



Android 14 for some, but not all devices. Beginner Nahuatl, with content usually does not advance beyond the basic level,

more than 1000 downloads as of the same date [14], was a although one often has the impression that the authors had

vocabulary training app, without any game elements. Kamatlama intended to add more lessons in the future. One could wonder

is an app introducing basic numbers and fruit names and testing whether a more stable app with more levels would have been

them through games and it was downloaded only more than 50 possible if the authors had joined their efforts.

times as of 01.07.2023 [14]. NahuatlApp was an app introducing



the same date, more than 10 000 downloads [14]. 5 vocabulary items and testing them through a game and had, as of Long-term survival of Language apps:

It is notable that the maintenance of the apps is far from ideal. Discussion

Between the data collection of [14], 01.07.2023 and of [31], The trajectory of some Nahuatl learning apps has helped us

30.05.2024, the videos of Aprende Nahuatl became unavailable identify a trend of frequent discontinuation of those apps, lack of

and the two apps, Beginner Nahuatl and NahuatlApp, became maintenance or upgrading to newer versions of operating

unavailable for download on Google Play. During the same time, systems or devices and a tendency by the commercial providers

the user statistics did not change for the two continued apps, and to dissolve community-led efforts of translation and localization

crucially, Kamatlama hasn´t reached 100 downloads. and to limit the number of languages that the service is translated

Two more apps mentioned by [14], although primarily and localized into.

dictionaries, also have elements supporting learning, such as Admittedly, the frequent appearance and disappearance of

quizzes: Totlahtol Nahuatl [32] and Diccionario Náhuatl [33]. As new apps might be a sign of a vibrant, creative community. In

of 26.05.2024 Totlahtol Nahuatl, which offered Nahuatl user fact, [7] sees it as a part of the process of app creation to accept

interface – as the only app other than Telegram - was no longer that the results of one`s work on software localization might have

available for downloads, while Diccionario Nahuatl is still a short life or never be used at all. If one accepts the possibility

(01.09.2024) available. Additionally, Miyotl, a multilanguage of failure (that is, the materials prepared ending up not being

app whose lesson components seem to never have been used) or only short-term success (that is, a short-lived app), the

completed, remains available for download and contains a list of process of creating apps might be more spontaneous and less

Nahuatl words and their Spanish translations [34]. restricted, and the threshold to make such an app might be lower.

In addition, as of 01.09.2024, 6 other apps mentioned by [14] In other words, if one does not strive to make “the perfect” app,

are discontinued: the text collection Tlapohualiztli [35], the creating an ad-hoc training exercises for one skill might become

dictionary Diccionario Maya y Nahuatl [36] and the automatic easier. However, even in this scenario, many contributors and

translator Traductor Nahuatl [37]. This means that out of 16 non- authors might end up doing the same work unnecessarily, such

Bible related apps mentioned by [14], 6 (Beginner Nahuatl, as preparing grammatical description of the same grammatical

NauatlApp, Diccionario Maya y Nahuatl, Tlapohualiztli, forms destined for different apps.

Totlahtol Nahuatl, Traductor Nahuatl) have been discontinued In addition, one might also wonder if the low number of

only a year later (37.5%). In addition, [31] mentions three other downloads discourages the authors to add additional levels and

apps that had been discontinued before: Tozcatl [38], Nahuatl update their apps. However, given that Aprende Náhuatl, an App

Grammar [39] and Ma Tiwelikan Nawatl [40] – the latter is published by a government agency had more than 10 000

available as a website, but the App version is not available downloads, and up to 28 000 downloads [46], there is sufficient

anymore. Furthermore, the app presented at the EUROCALL interest of learners in Nahuatl language apps, and perhaps a focus

conference in 2016 [41] is also not discoverable on Google on better findability of the apps could result in their bigger

search, as of 18.06.2024. success.

However, the changes are not all negative. On 14th March The question can therefore be posed how to better direct the

2024, user ItztliEhecatl posted on the social platform Reddit [42] efforts of authors and contributors, typically activists and

that they have created a new Huasteca Nahuatl language learning volunteers, to not repeatedly make basic-level apps that might

course [43]. The author has been adding new items to the course, then be discontinued and to instead direct those efforts at more

and as of 01.09.2024, there were 568 words and phrases to be long-term apps which would also include levels for more

learnt. The Huasteca Nahuatl course uses the Memrise advanced learners.

Community Courses infrastructure, where users can create their An open source platform that allows and tracks user edits,

own courses. However, in line with the trend discussed in section similar to Wikipedia (https://www.wikipedia.org/) or Wikitravel

2, Memrise is also closing community forums and removing (https://wikitravel.org/) could allow users to collaborate in

community courses from their app and the future of the making language courses, and each individual author could make

community courses remains uncertain beyond the end of 2024. a small contribution, without the feeling that it was “in vain”.

There is, however, another high-quality Nahuatl learning app This would also prevent the fruits of work of language activists

for beginners that explains the grammatical concepts and tests from “disappearing” in the chaos of the internet, and would

them in a variety of exercises over 11 Units: the Nahuatl course increase their findability. Care must be taken to make such a

hosted by the 7000 languages organization [44] and prepared in platform independent of commercial companies that could

2017 by Tlahtoltlapazolli, a Los Angeles based group [45]. The unilaterally delete the courses from their servers. It is also

course requires registration, and only has a website version – recommended that information about such a platform be widely

although the mobile website version works well. However, the shared to avoid the situation where a good course ends up not

fact, that it is not listed in App stores decreases its findability. being used due to being unknown to the learners.

Altogether, we see here that a great proportion of Nahuatl

learning apps is discontinued. By the time they this happens, their



Acknowledgments [18] Duolingo, n.d Aprende Guaraní en solo 5 minutos diarios. Completamente gratis. https://es.duolingo.com/course/gn/es/Aprender-Guaran%C3%AD I would like to thank the anonymous reviewer for their (31.08.2024)

comments, which will serve as a guide in the subsequent research [19] Duolingo wiki, n.d. Incubator.

into the more specific reasons behind the discontinuation of the https://duolingo.fandom.com/wiki/Incubator (01.09.2024) [20] Matt [user], 2023. What happened to the Duolingo Incubator?



apps. https://duoplanet.com/what-happened-to-the-duolingo-incubator/ (01.09.1988)



References [21] Myra Awodey and Karin Tsai, 2021. Ending & honoring our volunteer [1] volunteer-contributor-program-2/ (01.09.2024) Felix Richter, 2024. The Most Spoken Languages: On the Internet and in [22] Thomas Kole, 2023: Retrato de Tenochtitlan. Real Life. Statista. https://www.statista.com/chart/26884/languages-on-https://tenochtitlan.thomaskole.nl/es.html (27.03.2024) the-internet/ (30.08.2024) [23] INE, 2024. Promueven INE e INALI interpretación simultánea del Tercer [2] [3] András Kornai, 2013. Digital language death. PLoS Onen8(10):e77056. Debate Presidencial en tres lenguas indígenas nacionales. doi: 10.1371/journal.pone.0077056. https://centralelectoral.ine.mx/2024/05/17/promueven-ine-e-inali-Nikhil Gilbert, and C. Maria Keet, 2018. Automating question generation interpretacion-simultanea-del-tercer-debate-presidencial-en-tres-lenguas-and marking of language learning exercises for isiZulu. In Proceedings of indigenas-nacionales/ (12.07.2024) the Sixth International Workshop, CNL 2018, Maynooth, Co. Kildare, [24] Universidad Veracruzana, 2024. Maestriah ipan Totlahtol iwan Tonemilis Ireland, August 27–28, 2018, pp 31-40. / Maestría en Lengua y Cultura Nahua. https://www.uv.mx/mlcn/ [4] Fatih Bektaş,Bihter Dereli, Furhan Hayta, Erkin Şahin, Ubey Ali and (12.07.2024) Gülşen Eryiğit, 2022.Towards a Multilingual Platform for Gamified [25] Tecpaocelotl. 2010.“Google to add Maya, Nahuatl languages to search Morphology Learning, In 7th International Conference on Computer engine”. https://tecpaocelotl.livejournal.com/4425.html (19.07.2023). Science and Engineering (UBMK), Diyarbakir, Turkey, 2022, pp. 222-[26] Isaac Caswell, 2024. https://blog.google/products/translate/google-227, https://ieeexplore.ieee.org/document/9919484 (12.07.2024). [10] [5] translate-new-languages-2024/ (01.09.2024) Bernadette O’Rourke and Joan Pujolar, 2013. From native speakers to [27] Instituto Nacional de los Pueblos Indígenas (INPI), 2022. “Aprende “new speakers”–problematizing nativeness in language revitalization náhuatl [Mobile application]”. (01.07.2023) contexts. Histoire Épistémologie Langage 35.2: 47-67. [28] shotgun.experiments, 2018. “Beginner Nahuatl [Mobile application]”. [6] José Antonio Flores Farfán, and Justyna Olko, 2021. Types of (01.07.2023) communities and speakers in language revitalization. In Revitalizing [29] Mario Albertio Duque Peralta [mario10412], 2022. “Kamatlama (LSM y endangered languages: A practical guide, pp. 85-103. Nahuatl) [Mobile application]”. (01.07.2023) [7] Tania P. Hernandez-Hernandez and Bulmaro González Ambrosio, 2022. [30] Juarez, Patricio, 2019.: “Nahuatl App [Mobile application]”. (01.07.2023) Mozilla in p’urhépecha: Translators’ agency in a software translation [31] Fischer, Evelyn, 2024. Self-study Tools of Morphologically Rich project, Cadernos de Tradução, 42(01), pp. 1 – 20. doi:10.5007/2175-Languages: A Prototype of a Nahuatl Learning App. In Proceedings of the 7968.2022.e85270. MEi:CogSci Conference 18. [8] Netza López, 2021. Webinar – Localización de Firefox en Lenguas https://journals.phl.univie.ac.at/meicogsci/article/view/752 (12.07.2024). Indígenas. https://mozillanativo.org/ (30.08.2024) [32] Tecuexe, 2016. “Totlahtol Nahuatl [Mobile application]”. (01.07.2023) [9] Firefox, n.d. about: preferences (30.08.2024) [33] Mingatics, Patricio, 2021: “Diccionario Nahuatl [Mobile application]”. Pontoon Mozilla, n.d. https://pontoon.mozilla.org/projects/firefox/ (01.07.2023) (30.08.2024) [34] Emilio Álvarez Herrera, Emilio, 2021: “Miyotl [Mobile application]”. [11] Kevin Scannell, 2012. Translating Facebook into endangered languages”. (01.07.2023) In Tania Ka’ai, Muiris Ó Laoire, et al., editors, Language Endangerment [35] Tecuexe, 2020: “Tlapohualiztli [Mobile application]”. (01.07.2023) in the 21st Century: Globalisation, Technology and New Media. [36] Luján Castillo, José Dimas. s. f. “Diccionario Maya y Nahuatl [Mobile Proceedings of the 16th Foundation for Endangered Languages application]”. (4.06.2023) Conference, pp 106-110. https://kevinscannell.com/files/fel12.pdf [37] Axcan, 2019. Tozcatl [Mobile application] https://tozcatl.soft112.com/ (31.08.2024) [38] Ibrahim Gerahard Peregrina Ochoa, 2021. “Traductor Nawatl [Mobile [12] Kevin Scanell, 2012. Facebook in Your Indigenous or Endangered application]”. (01.07.2023) Language. Rising Global Voices. [39] Tecuexe, 2017. Nahuatl Grammar [Mobile application]. (01.07.2023) https://rising.globalvoices.org/blog/2012/10/15/facebook-in-your-[40] Rodrito García and Natalia Alonzo, 2017. Ma tiwelikan Nahuatl, indigenous-or-endangered-language/ (31.08.2024) https://vamosaaprendernahuatl.centroculturadigital.mx (12.05.2024) [13] BBC News, 2015. Bolivia: Group translates Facebook into native [41] García-Mencía, Rafael, Aurelio López-López, and Angélica Muñoz language. https://www.bbc.com/news/blogs-news-from-elsewhere-Meléndez, 2016. An Audio-Lexicon Spanish-Nahuatl: using technology 34279302 (24.07.2023). to promote and disseminate a native Mexican language. CALL [14] Evelyn Fischer, 2024. El internet en náhuatl: la apropiación de las communities and culture– short papers from EUROCALL (2016): 155159. tecnologías de información y comunicación por una lengua indígena. [42] Itzli Ehecatl [reddit user], 2024. Check Out My New Huasteca Nahuatl Vienna: University of Vienna, Master´s Thesis. Language Learning Course. https://utheses.univie.ac.at/detail/70866/# (12.06.2024). Contributor program. https://blog.duolingo.com/ending-honoring-our-



[16] https://en.duolingo.com/courses/all. (04.06.2024) [43] Itztli Ehecatl, 2024. Huasteca Nahuatl https://community-IBelieveDuoHasSh [fórum user], 2023, Guarani escondido [forum thread] courses.memrise.com/community/course/6566566/huasteca-nahuatl/ https://forum.duome.eu/viewtopic.php?t=10075-guarani-escondido (01.09.2024) [15] Duolingo, n.d. Duolingo Language Courses. w_huasteca_nahuatl_language/ (01.09.2024) https://www.reddit.com/r/nahuatl/comments/1bekh07/check_out_my_ne



[17] (31.08.2024) [44] 7000 Languages, https://www.7000.org/ n1_kita [Reddit user], 2022. ¿Qué ha pasado con la lista de cursos de [45] 7000 Languages, n.d. Nahuatl https://www.7000.org/nahuatl guaraní de Duolingo? [reddit post] [46] Appbrain, n.d. Aprende Nahuatl https://www.appbrain.com/app/aprende-https://www.reddit.com/r/duolingo/comments/10zvltv/what_has_happen n%C3%A1huatl/com.nahuatlpuebla.app (26.09.2024) ed_to_the_duolingo_guarani_course/?tl=es (31.08.2024)





Designing the Flow State Experience Using Modern Digital




Technologies



Eva Vidmar

Multimedia

Faculty of computer science

University of Ljubljana

Ljubljana, Slovenia

eva.vidmar2@gmail.com



Abstract novel approach that builds on but goes beyond methods such as

video games or Virtual Reality (VR). By integrating

This article provides a brief overview of an extended Master’s psychophysiology, user experience, and ambient intelligence,

thesis and focuses on the use of modern digital technologies to this research aims to leverage technology for meaningful

design a multimedia environment aimed at inducing a state of improvements in well-being, productivity, and satisfaction.

flow in individuals. Flow is a psychological state characterized

by deep immersion in an activity, leading to a loss of sense of

time and external worries [1]. Exiting this state typically results 2 Theoretical Background



requires a balance between skills and challenges. Learning to Inducing a flow state through technology poses complex in feelings of satisfaction and happiness. Achieving flow

attain this balance can help individuals improve overall, which challenges that require in-depth exploration of the

is a key reason for this research. The main objective is to neurocognitive aspects of flow and their relation to contemporary

investigate whether flow can be achieved through the use of technologies. This understanding informs the design of

color light stimuli in a space that adapts in real-time to an technological solutions aligned with flow theory.

individual's level of attention. This represents a preliminary step

toward using technology to design spaces that stimulate 2.1 A Neurocognitive Perspective on Flow



flow. An experiment was conducted to test whether such a Flow, characterized by deep focus and immersion, was first individuals and facilitate the quicker and easier attainment of

space affects individuals' flow. Tetris was chosen as the central described by Mihaly Csikszentmihalyi, often referred to as the

activity for the experiment. The findings indicated that color "father of flow" [1]. This state occurs when individuals find an

stimuli influenced the participants' physically measured appropriate balance between their skills and the challenges they

attention, although no significant changes were observed in face, allowing them to perform optimally with a sense of

questionnaire responses or gameplay performance. Given that effortless control [1]. While initial resistance and sustained

attention is a crucial factor in achieving flow, it can be partially motivation are necessary to achieve flow, this state can occur

asserted that participants experienced flow, though more even in unfamiliar tasks, although long-term practice may



findings significantly contribute to the understanding of dimensional graph where it exists at the intersection of measuring and achieving flow through technology, representing reliable data would be necessary for further conclusions. These increase its likelihood [2]. Flow is often illustrated in a two-

an important advancement in this field. appropriate challenge and skill levels [1].





Keywords



flow, optimal experience, user experience, digital interface,

Tetris



1 Introduction

Historically, flow has been experienced by artists, athletes, and

individuals with substantial practice. However, modern life,

especially in technology-driven environments, necessitates new

approaches to achieve this optimal state. This study investigates

how real-time adjustments of ambient lighting, informed by

physiological signals, can enhance flow experiences, offering a



Permission to make digital or hard copies of part or all of this work for personal or Figure 1: Graph representing occurrence of flow state [1]

classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Neuroscientific perspectives on flow suggest various underlying

© 2024 Copyright held by the owner/author(s). mechanisms. The transient hypofrontality hypothesis posits that

during flow, activity in the prefrontal cortex decreases, reducing

self-criticism and enhancing performance [3]. Alternatively, 2.3 Techniques for Flow State Measurement synchronization theory proposes that flow arises from

Various methods exist to measure brain activity, with

synchronized activation across different brain regions,

electroencephalography (EEG) being the most direct and

optimizing cognitive efficiency [4]. Both theories indicate that

commonly used. Functional Near-Infrared Spectroscopy (fNIRS)

flow entails minimal energy expenditure in the brain.

and Functional Magnetic Resonance Imaging (fMRI) provide

Csikszentmihalyi's model outlines optimal conditions for

insights into brain function, while Magnetoencephalography

flow, including clear goals, immediate feedback, and a balance

(MEG) offers high resolution of neuronal activity. However,

between challenge and skill [1]. The concept of the autotelic

these methods often involve expensive and less accessible

personality—marked by intrinsic motivation and enjoyment—



may enhance flow experiences but is not strictly essential for The MindWave Mobile 2, a consumer-grade EEG equipment.

achieving it.

device, stands out for its ease of use, making it suitable for

educational and entertainment contexts [16]. This device is ideal

2.2 Flow in Technology: State of the Art

for our research due to its user-friendly nature, minimizing

Various approaches have emerged to induce flow through inconvenience for participants. technology. Traditionally, video games employed fixed

difficulty levels, which often interrupted the flow experience [5].

Dynamic Difficulty Adjustment (DDA) algorithms now allow

real-time modification of challenges based on player

performance, as seen in games like Left 4 Dead [6], [7, p. 4],

Affective computing enhances this by using emotional

indicators, such as facial expressions, to fine-tune difficulty

levels [6]. Jenova Chen’s game flOw exemplifies the integration

of DDA with real-time adjustments to maintain flow [8]. Virtual

Reality further immerses players, as evidenced by studies

comparing VR to traditional 2D games and applications focused

on meditation and relaxation [9], [10], [11]. Augmented Reality

and multimedia art also contribute innovative avenues for

inducing flow [12], [13].





Figure 3: Mindwave Mobile 2 [16]





Flow state was traditionally measured through self-reporting

instruments, such as the Experience Sampling Method (ESM)

developed by Csikszentmihalyi [1]. Various questionnaires,

including the Flow State Scale and Game Experience

Questionnaire, have been developed to assess flow but rely on

retrospective reporting. Alternatively, physiological measures

may offer a more objective assessment of flow experiences.



3 Experiment: The Impact of Light on Flow

State During Tetris Gameplay

This experiment investigated whether spaces incorporating

adaptive technology could enhance user engagement. We

compared standard Tetris gameplay to a version featuring color-

changing lights that adjusted based on player attention, measured

via the MindWave Mobile 2. The goal was to assess whether

Figure 2: Refik Anadol 2D projection Machine these technological enhancements positively impacted

Hallucinations [12] engagement and performance, specifically exploring if adaptive



promising yet under-explored method for inducing flow while Drawing from Csikszentmihalyi's model, we engaging participants in a core activity, specifically Tetris. This recognized the importance of differentiating the environment in game, created by Alex Pajitnov in 1985, has been extensively which flow activities occur. We aimed to create a highly Our research identifies the use of ambient lighting as a gameplay results. lighting improved attention, stabilized focus, and led to better



studied for its capacity to induce flow [14]. Players arrange engaging environment by designing a prototype of adaptive falling blocks to form complete lines, receiving immediate visual lighting for a dimly lit space. feedback—key elements for maintaining flow. Research For the experiment, we developed a color-changing indicates that even brief sessions of Tetris can lead to flow light prototype controlled by the MindWave Mobile 2. The experiences and reduced negative emotions [15]. device measured brainwave activity during Tetris gameplay. We

utilized an Arduino Uno microcontroller to interface with the

MindWave Mobile 2 and control a 2-meter AdaFruit NeoPixel

LED strip. The light's color adjusted based on attention levels,

with red indicating low attention, blue indicating high attention,

and white representing optimal focus. This setup aimed to

evaluate whether adaptive lighting could influence players'

attention and flow during the game.



Figure 6: Attention levels of a player during gameplay

without lights (left) and with lights (right)



For engagement, while the Shapiro-Wilk test confirmed

normality, a paired t-test indicated no significant effect of

lighting on engagement (p-value = 0.668). Tetris performance

was assessed based on scores, with the group using lights

achieving a higher average score (9377) compared to the no-light

group (8979), though a Wilcoxon signed-rank test revealed no

significant difference (p-value = 0.33).

Qualitative analysis of interviews with 40 participants

identified seven key themes related to their experiences with the

lighting: awareness of external stimuli, control and feedback,

concentration, immersion, motivation, satisfaction, and pressure.

Many participants reported feelings of pressure and stress,

underscoring the challenges of achieving flow. Nonetheless, the

lights were generally perceived as motivating, and some

participants noted decreased awareness of their surroundings,



Figure 4: The setup of light prototype behind laptop aligning with theories regarding transient hypofrontality and

reduced default mode network activity during flow [1].



The experiment was designed to compare Tetris performance Interestingly, some participants reported not noticing the

with and without adaptive lighting. Participants played Tetris lights at all, suggesting a potential subconscious influence of



the MindWave Mobile 2, were used to adjust the light's color the questionnaire results. In terms of color perception, red was described as stressful and distracting, while white and blue were dynamically. Data on engagement, attention, and gameplay under both conditions, and their attention levels, measured via flow on their experience. This observation could have affected



performance were collected and analyzed to determine the regarded as pleasant, showing no significant difference between them. effectiveness of the adaptive lighting in enhancing flow.





4 Conclusion



This study explored the potential influence of external factors,

specifically technology-based lighting, on the state of flow.

While we observed increased attention levels during gameplay

with lights, supporting the theoretical premise that flow involves

synchronized neural networks related to attention and reward,

Figure 5: Scenario of playing Tetris with lights our hypothesis remains unconfirmed. Qualitative interviews

highlighted themes consistent with flow characteristics, such as

3.1 Results immersion and motivation; however, the absence of statistically

Data were collected from the MindWave Mobile 2, which significant effects on engagement and gameplay performance

recorded attention and meditation levels during Tetris gameplay indicates that further research is warranted. Future studies should

with and without adaptive lighting. We filtered data to focus on involve larger, more diverse samples and consider additional

attention values from 5 minutes of gameplay, excluding values metrics to assess flow states more comprehensively.

below a threshold of 10 and retaining the 300 most representative Overall, our findings offer valuable insights into integrating

data points. A Shapiro-Wilk test confirmed normal distribution technology with flow theory, highlighting the potential for

for both conditions. A paired t-test revealed significantly higher developing products that enhance focus and user experience.

attention levels during gameplay with lights (p-value = 0.00032). This research lays the groundwork for future innovations aimed

Notably, attention levels were more stabilized with adaptive at creating more effective tools for achieving optimal states of

lighting, as evidenced by a smaller variance in attention scores concentration and fulfillment in everyday life.

compared to gameplay without lights.



Acknowledgments [7] ‘Left 4 Dead on Steam’. Accessed: Jan. 13, 2024. [Online]. Available: https://store.steampowered.com/app/500/Left_4_Dead/

[8] ‘flOw’, Jenova Chen. Accessed: Nov. 24, 2023. [Online]. Available:

The research was conducted in collaboration with psychologist http://jenovachen.info/flow

prof. dr. Andreja Avsec, who served as a co-supervisor of the [9] ‘Virtual Reality and Flow: Discovering the Impact of Virtual Reality on

thesis, and prof. dr. Gregor Geršak, an expert in Flow States’. Accessed: Nov. 24, 2023. [Online]. Available:

psychophysiology and measurement, who acted as the primary https://www.novobeing.com/blog/virtual-reality-and-flow-exploring-

virtual-realitys-impact-on-peak-performance

supervisor. [10] TGC, ‘Flow’, thatgamecompany. Accessed: Nov. 24, 2023. [Online].

Available: https://thatgamecompany.com/flow/



References [11] ‘Cosmic Flow: A Relaxing VR Experience on Meta Quest’, Oculus.

Accessed: Jan. 14, 2024. [Online]. Available:

[1] https://www.meta.com/experiences/quest/3872076276162726/ M. Csikszentmihalyi, ‘Flow: The Psychology of Optimal Experience’, 1990. [2] [12] ‘Refik Anadol’, Refik Anadol. Accessed: Jan. 15, 2024. [Online]. Available: O. de Manzano, T. Theorell, L. Harmat, and F. Ullén, ‘The

psychophysiology of flow during piano playing’, https://refikanadol.com/ Emot. Wash. DC , vol. 10, no. 3, pp. 301–311, Jun. 2010, doi: 10.1037/a0018432. [13] ‘Android Jones’, Android Jones. Accessed: Jan. 15, 2024. [Online].

[3] Available: https://androidjones.com/ FlowCode - Project Unity, Transient Hypofrontality - FlowCode Lesson #7

/ Flow state training [14] D. Lora, A. Sánchez-Ruiz, and P. Gonzalez-Calero, ‘Towards Finding Flow , (Jun. 26, 2020). Accessed: Nov. 11, 2023. [Online

Video]. Available: https://www.youtube.com/watch?v=P5q5l47hGzY in Tetris’, 2019, pp. 266–280. doi: 10.1007/978-3-030-29249-2_18.

[4] [15] ‘Tetris: It could be the salve for a worried mind’, ScienceDaily. Accessed: R. Weber, R. Tamborini, and A. Westcott-Baker, ‘Theorizing Flow and

Media Enjoyment as Cognitive Synchronization of Attentional and Reward Dec. 02, 2023. [Online]. Available:

Networks’, Commun. Theory https://www.sciencedaily.com/releases/2018/10/181025084012.htm , 2009.

[5] [16] ‘MindWave Mobile 2 Quick Start Guide or User Guide / MindWave Mobile S. Khoshnoud, F. Alvarez Igarzábal, and M. Wittmann, ‘Peripheral-

physiological and neural correlates of the flow experience while playing 2 / Knowledge Base - NeuroSky - Home Page Support’. Accessed: Mar. 28,

video games: a comprehensive review’, 2024. [Online]. Available: http://support.neurosky.com/kb/mindwave- PeerJ , vol. 8, p. e10520, Dec. 2020,

doi: 10.7717/peerj.10520. mobile-2/mindwave-mobile-2-quick-start-guide-or-user-guide

[6] M. Zohaib, ‘Dynamic Difficulty Adjustment (DDA) in Computer Games: A

Review’, Adv. Hum.-Comput. Interact., vol. 2018, p. e5681652, Nov. 2018,

doi: 10.1155/2018/5681652.



The Transparency of Nudging: Evaluating Its



Impact on Personal Autonomy



Sabina Pajmon Toma Strle

Center for Cognitive Science Center for Cognitive Science

University of Ljubljana University of Ljubljana

Ljubljana, Slovenia Ljubljana, Slovenia

sabina.pajmon@pef.uni-lj.si toma.strle@pef.uni-lj.si



Abstract preserves a strong commitment to freedom of choice. Behavioral



Nudges are a strategic approach that shapes decision-making insights show that the context of decision-making can lead us to act inconsistently with our otherwise well-informed intentions environments and the presentation of options to steer individuals [2]. The traditional approach to public policy assumes people are toward certain behaviors while maintaining their freedom of perfectly rational economic subjects ("econs") who act optimally choice. The ethical concerns surrounding nudges center on their with accurate information and clear rules. While this is an potential to undermine personal autonomy, particularly when admirable goal, Thaler and Sunstein warn that basing public individuals are unaware of the influence exerted on them (i.e., policy on this ideal often leads to failure. The authors introduce covert or non-transparent nudges). The proposed solution for the concept of a "nudge" and propose its use as a policy-making preserving autonomy is to increase transparency, which includes approach that can influence citizens' behavior while avoiding the disclosing the presence and purpose of nudges to the people that pitfalls and issues of traditional regulatory approaches, such as are being nudged. There are various types of nudges and different prohibitions and punishments. The advantage of this approach is types and levels of transparency associated with them. The most that policymakers can influence our choices and behavior in a problematic in terms of violating personal autonomy are the non- cost-effective and efficient manner without restricting us with transparent ones, those that exploit automatic cognitive prohibitions or interfering with our choices [3]. Despite the high mechanisms (Type 1 nudges), those that use type transparency effectiveness and utility of nudges, ethical concerns arise and those that disclose their nature only after the fact (ex post). regarding the preservation of autonomy, especially with nudges New approaches such as nudge plus approach seek to protect that operate covertly and influence us without our awareness. personal autonomy by involving citizens in the creation of This article investigates various types of nudges and levels of nudges and enhancing reflectiveness during the nudging process. transparency, with a focus on their implications for personal



Keywords autonomy. It begins by elucidating key concepts—nudges,

autonomy, and transparency—before analyzing how different

Nudge, transparency, autonomy, ethics of nudging, nudge plus types of nudges, alongside varying types and levels of

approach transparency, affect the preservation or violation of personal

autonomy. Additionally, the article proposes criteria for

determining which types of nudges are compatible with the

1 Introduction preservation of personal autonomy. Finally, it explores potential

Over the past thirty years, psychology and behavioral economics strategies to mitigate adverse impacts on autonomy, including

have highlighted how various contextual factors systematically enhancing transparency, fostering citizen participation, and

influence our decision-making and behavior. In public policy- integrating reflective practices into the design of nudges. making, these insights are crucial for effectively addressing

societal challenges like global warming, obesity, and poor

economic decision-making. The groundbreaking paper [1] and 2 Definition of a nudge

the book that followed that brought the importance of decision Thaler and Sunstein define a nudge as any element of choice

architecture to the attention of academics, policymakers, and the architecture that influences behavior in a predictable way without

general public was Thaler and Sunstein’s “Nudge: Improving restricting options or significantly altering economic incentives

Decisions About Health, Wealth, and Happiness” [2]. In their [2]. A nudge subtly guides individuals toward better decisions

work, they propose various ways in which government and while preserving freedom of choice by adjusting how choices are

private organizations could encourage or "nudge" individuals presented [4]. Unlike prohibitions or penalties, nudges steer

toward actions beneficial to them, while promoting a method that behavior without limiting options. An example is placing healthy

Permission to make digital or hard copies of part or all of this work for personal or snacks at eye level in stores to encourage healthier choices [5].

classroom use is granted without fee provided that copies are not made or distributed Hausman and Welch [6] add that nudges influence choices

for profit or commercial advantage and that copies bear this notice and the full without increasing costs or limiting options, highlighting the

citation on the first page. Copyrights for third-party components of this work must potential for manipulation, which raises ethical concerns

be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia discussed in later chapters. © 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.cog.8



3 regulations, don’t limit freedom but instead encourage choices How nudges work: leveraging heuristics and

biases that align with individuals' best interests.



To grasp how nudges impact behavior and decision-making, we 4.2 Transparency as a solution to autonomy must rely on insights from behavioral science, which reveal that violation nudges exploit inherent imperfections in human decision

processes—leveraging cognitive heuristics and biases [4]. A key A proposed solution is increasing nudge transparency, as it

element of the nudge approach is that heuristics and biases, allows individuals to understand how nudges work and make

which often serve as mental shortcuts, are utilized to the autonomous decisions based on their values [8]. A transparent

advantage of the choice architect. While these mental shortcuts nudge is one where its purpose and the methods used to influence

can sometimes lead to suboptimal decisions, a nudge aims to behavior are reasonably clear to the affected individual. Thaler

harness them to promote better decisions [3]. and Sunstein moreover argue that nudges used by governments

should be public and transparent, with officials ready to disclose

their methods and motives. Sunstein further emphasizes that

4 Nudge and ethical issues nudges must be visible, reviewed, and monitored to prevent

Although the theory of nudging presents a promising approach violations of autonomy or dignity [10]. Transparency involves

to public policy, it has faced significant criticism from both informing decision-makers about the presence and purpose of



ethical debate has developed, featuring nuanced arguments both interventions, thus preserving their autonomy and freedom of 1 choice [7]. supporting and opposing the practice [3, 4, 6, 7, 8]. The primary public and academic spheres. Over the past decade, a robust nudges, allowing individuals to remain aware of behavioral

critique centers on the idea that nudging involves manipulating

choices, with concerns about potential misuse of power [3].

5 Types of transparency in nudging

Critics argue that nudges can undermine free choice by subtly

restricting rather than fostering individual decision-making. The To better understand the impact of nudging on an individual's

core of nudging involves exploiting heuristics and biases, which autonomy, it is crucial to first examine the different types of

often lead people to act in ways that deviate from their well- nudges, as they are not a uniform phenomenon; rather, they can

considered preferences. Bovens [8] contends that such be classified into various types [2]. Understanding these types is

mechanisms can compromise control over actions, raising based on dual-process theory, which describes the two decision-

worries that nudges might affect decision-making by diminishing making mechanisms that nudges can influence.

rational or deliberate considerations. Additionally, he argues, the

behavior change induced by nudges occurs, if not against 5.1 Dual process theory

citizens' will, then at least without their active consent and Dual process theory, explored by Stanovich [13] and Kahneman

awareness; for broader discussion about this topic see also [9, Ch. [14], is key in Thaler and Sunstein’s work on nudges. It suggests

“Avtonomija v svetu spodbud” (Autonomy in the World of the brain operates in two modes: fast, intuitive System 1 and slow,

Nudging), pp.. 81-100; 10]. deliberate System 2. System 1 handles instinctive actions, while

System 2 engages in reflective decision-making. Despite its

4.1 Ethical dilemma of autonomy in nudge use acceptance, dual processing is contested, with some scholars

Although nudges have been shown to effectively influence arguing the differences are a matter of degree. De Neys [15],

behavior, critics argue they can be manipulative and threaten notes no conclusive evidence favors either model, and resolving

personal autonomy. Autonomy, a complex concept, is broken this debate may not significantly enhance our understanding of

down by Schmidt and Engelen into four dimensions: the freedom human thinking mechanisms. In this article, we adopt the dual

to choose without external pressure, acting according to one’s process theory model to categorize different types of nudges.

desires and values (psychological autonomy), making rational This approach allows us to better understand and design

decisions based on available information, and being free from interventions that leverage both intuitive and reflective processes.

domination or manipulation [5, 8]. Critics claim that nudges can

undermine autonomy by subtly influencing behavior without 5.2 Type 1 and Type 2 nudges

explicit consent, raising concerns about democracy, especially if According to Hansen and Jespersen [3], nudges can be

governments use nudges without informing citizens. Nudges that categorized into two types based on dual process theory. Type 1

operate without notice are especially problematic, as they can nudges target automatic, non-reflective thinking (System 1) and



and Engelen [11] argue, non-transparent nudges, which exploit stimuli that influence behavior without conscious awareness. These nudges can be ethically problematic, as they often lack influence decisions without individuals' awareness. As Ivanković operate unconsciously, such as subliminal advertising or visual

less rational psychological mechanisms, undermine autonomy by



denying people control and the ability to challenge, a right that values. In contrast, Type 2 nudges engage reflective, deliberate transparency and may lead to decisions misaligned with personal

should be protected in liberal democracies. Sunstein, however, thinking (System 2), promoting informed and thoughtful

argues that nudges maintain freedom by allowing people to opt decision-making. These nudges are transparent and pose fewer

out of the suggested behavior, a concept they call "libertarian ethical concerns regarding personal autonomy. paternalism" [12]. They believe nudges, unlike traditional



1 Empirical evidence is inconsistent regarding the impact of transparency on the nudges), enhance their effectiveness (if people understand and support the

effectiveness of nudges. Transparency may: reduce their effectiveness (by underlying goals), or have no significant impact at all [10]. prompting reflection), make nudges counterproductive (if people resist disliked



5.3 if the person learns when and where to expect it. With repeated Different types of transparency of nudges



The transparency of nudges plays a crucial role in safeguarding exposure to such nudges, individuals may become more aware of their influence and may eventually avoid them altogether [12]. autonomy and freedom of choice, yet this concept itself is

multifaceted. On one side of the spectrum, some nudges are



explicitly transparent, functioning effectively because the 6 Types of nudges and transparency: impact on individual is fully aware of the influence being exerted. personal autonomy Conversely, some nudges operate more subtly, relying on a lack

of transparency to achieve their intended effect. To thoroughly The debate over nudges centers on how different types of nudges

assess which forms of transparency in nudges may raise ethical as well as types and levels of transparency impact personal

concerns, it is important to analyze the various ways in which autonomy. As stated in the article, nudges are divided into two

transparency can manifest within nudges. types: Type 1, which influence automatic, non-reflective

behavior, and Type 2, which target reflective decision-making.

5.4 Transparent Type 2 nudges, which engage reflective capacities, Type and token transparency

do not typically raise ethical concerns, as they allow for

Bovens [8] introduces a crucial distinction between type and

conscious and deliberate decision-making. In contrast, non-

token transparency in nudges. Type transparency refers to when

transparent Type 1 nudges, which act on automatic processes,

governments inform citizens about the general techniques they

can threaten autonomy by influencing behavior without the

employ to intervene in decision-making contexts for the purpose

individual's awareness. This may lead to decisions misaligned

of enhancing well-being. In this scenario, the government is open

with personal values or goals. Transparency is categorized into

about the categories of measures it plans to implement. For

type transparency (general awareness of the nudge type) and

example, when a government announces its intention to use

token transparency (awareness of mechanisms of specific

specific psychological mechanisms to address social challenges,

nudges). The former is particularly problematic, as it lacks



interventions it will use to influence individuals' behavior and disclosure of specific examples and mechanisms, leaving us it demonstrates type transparency by clearly stating the kinds of

potentially unaware of the influences on our behavior. Nudges

decision-making [16]. However, Bovens stresses that this is not

can also be categorized by the level of transparency into two

enough. In his view, subliminal advertising does not become

main groups. The first group includes nudges that are transparent

more acceptable simply because it is openly acknowledged [8].

in advance by design (ex ante). These nudges are openly

On the other hand, token transparency requires that each

presented, allowing users to consciously decide whether to

individual instance of a nudge is clearly recognizable, including

respond to them. Such nudges generally do not threaten

how it was implemented. This method, referred to as "here and

autonomy, as they encourage conscious and deliberate decision-

now approach," aims to ensure that nudges are transparent to

making. The second group includes nudges that are only

those encountering them at the moment of their decision-making

transparent afterward (ex post). These nudges can be

[12]. However, even if this were feasible, it seems absurd to

problematic, as users may respond to them before realizing they



Since choice architecture is often unavoidable, token have been nudged. Although information about the nudge is demand that every nudge be accompanied by a notice of its use.

revealed later, it may already have influenced behavior in a way

transparency may be too demanding, according to Bovens [8].

that threatens freedom of choice and autonomy [18]. In

conclusion, the most problematic nudges, in terms of violating

5.5 Levels of transparency

personal autonomy, are Type 1 nudges that exploit automatic

Transparent nudges differ also based on when they are noticed cognitive mechanisms, lack transparency—where type

by the nudged individuals. With nudges that are transparent in transparency is more concerning than token transparency—or are

advance (ex ante), the user can see the nudge beforehand and can only transparent afterward. Understanding and using nudges

avoid it if they choose. An example is traffic light labels (green, requires careful consideration of their transparency and impact

yellow, red) for healthy, less healthy, and unhealthy food on freedom of choice. While transparent nudges can serve as

products [17]. In contrast, a nudge is transparent afterward (ex tools for encouraging thoughtful and autonomous decisions, non-

post) if the target person only notices its influence after it has transparent nudges, as well as Type 1 nudges, especially those

already affected them. Examples include fake cracks painted on with only type or post hoc transparency, must undergo thorough

the road to slow down drivers or the use of default options in ethical scrutiny to prevent potential violations of personal

certain contracts. Only after experiencing the effects do people autonomy. realize they were influenced by a nudge [12]. Unlike the first

category, the potential impact of such nudges on people's

autonomy is more significant here. Ex post transparency may be Table 1: Classification of nudges based on their impact on

insufficient to ensure autonomous action if it depends on personal autonomy

individuals' ability to avoid the nudge. If transparency is meant

to ensure that nudges do not deter people from achieving their Nudges that violate Nudges that do not violate

goals and values, then, according to Ivanković & Engelen, ex autonomy autonomy

post transparent nudges should either be excluded or efforts Type 1 nudges Type 2 nudges

should be made to turn ex post transparency into ex ante Type transparency Token transparency

transparency [11]. Occasionally, ex post transparent nudges Ex post transparent nudges Ex ante transparent nudges become ex ante transparent through repeated exposure. For

example, a fake speed bump may not have the same effect twice



This table helps determine whether a nudge preserves autonomy, of reflection, can enhance both the effectiveness and ethicality of

but it's unclear how many criteria must be met to deem a nudge interventions, empowering individuals to make more informed

ethical or unethical. Further research is needed for clearer decisions.

guidance.

Acknowledgments

6.1 Collaborative policy design: The nudge plus This pilot research study was partly supported by The Green

approach Nudge project (“UL za trajnostno družbo – ULTRA”) - European

The nudge plus approach extends beyond transparency by Union - NextGenerationEU, and Republic of Slovenia, Ministry

encouraging participatory engagement and reflection, viewing of Higher Education, Science and Innovation .

individuals as rational, reflective beings rather than passive

agents. Unlike traditional nudging, which can influence behavior Authors’ statement

unconsciously, nudge plus focuses on democratic control and ChatGPT-4 was used for improving language of this paper.

active collaboration between citizens and policymakers. Through

methods like citizens' assemblies, participants are directly References

involved in policy design, contributing ideas that shape their [1] R. H. Thaler and C. R. Sunstein, “Libertarian Paternalism,” American



environments. In the UK, medical sciences now require patient Economic Review, vol. 93, no. 2, pp. 175-179, May 2003. [2] R. H. Thaler and C. R. Sunstein, Nudge: Improving Decisions about and public involvement in all research that includes patient Health, Wealth, and Happiness. New Haven, CT: Yale University Press,

populations. Similarly, adolescents are consulted in developing 2008.

anti-bullying interventions [19]. These approaches foster mutual [3] P. G. Hansen and A. M. Jespersen, "Nudge and the Manipulation of

Choice: A Framework for the Responsible Use of the Nudge Approach to

feedback and collaboration between policymakers and citizens, Behaviour Change in Public Policy," 2013.

leading to more inclusive and transparent policies that respect [4] A. T. Schmidt and B. Engelen, "The ethics of nudging: An overview,"

DOI: 10.1111/phc3.12658, 2019.

community values. Nudge plus approach also refers to an [5] T. Bucher et al., “Nudging consumers towards healthier choices: a

intervention that has a reflective strategy embedded into the systematic review of positional influences on food choice,” British Journal

design of a nudge. Banerjee and John [20] state that this of Nutrition, vol. 115, no. 12, pp. 2252–2263, 2016.

doi:10.1017/S0007114516001653

preserves personal autonomy while promoting pro-social [6] R. A. Abumalloh, O. Halabi, R. Ali, and D. Al-Thani, “Nudging

interventions through active involvement by enhancing token Techniques: Design, Theoretical Grounds, and Ethical View,” Journal of

Knowledge Economy, 2024. [Online]. Available:

transparency and decision-making autonomy. The nudge plus https://doi.org/10.1007/s13132-024-02219-x.



approach offers significant potential for enhancing public policy [7] D. M. Hausman and B. Welch, "Debate: To Nudge or Not to Nudge," The Journal of Political Philosophy, vol. 18, no. 1, pp. 123–136, 2010.Sam An with maintaining individual autonomy. By embedding reflection, [8] L. Bovens, "The Ethics of Nudge," in T. Grüne-Yanoff and S. O. Hansson,

transparency and active citizen engagement, it encourages people Eds., Preference Change: Approaches from Philosophy, Economics and

to participate in decision-making rather than passively accepting Psychology, Theory and Decision Library A 42, Springer

Science+Business Media B.V., 2009, ch. 10.

nudge type interventions. This participatory approach builds [9] T. Strle in O. Markič, O odločanju in osebni avtonomiji, 1. izd., let. 20.

trust, as individuals are more likely to embrace policies that Maribor: Aristej, 2021, str. 145.

respect their autonomy and align with their values. multikonference Informacijska družba - IS 2019, 10. oktober 2019: zvezek [10] Kognitivna znanost, Kognitivna znanost: zbornik 22. Mednarodne

B = Cognitive Science. Ljubljana: Institut „Jožef Stefan“, 2019. [Na

spletu]. Dostopno na:

http://library.ijs.si/Stacks/Proceedings/InformationSociety

7 Conclusion [11] V. Ivanković and B. Engelen, “Nudging, Transparency, and

This article reviewed nudges as tools for influencing decision- Watchfulness,” Social Theory and Practice, vol. 45, no. 1, pp. 43-73, Jan.

2019, doi: 10.5840/soctheorpract20191751.

making and behavior, with a focus on their transparency and its [12] C. R. Sunstein, “The Ethics of Nudging,” Yale Journal on Regulation, vol.

influence on potential infringement of personal autonomy. We 32, no. 2, pp. 413-450, 2015.

found that nudges vary in type and transparency, which [13] K. E. Stanovich, Who is Rational?: Studies of Individual Differences in

Reasoning. Milton: Psychology Press, 1999.

significantly affects their ethical acceptability. Type 1 nudges, [14] D. Kahneman, Thinking, Fast and Slow. London: Penguin, 2012.

which target automatic decision-making mechanisms, can [15] W. De Neys, “On dual-and single-process models of thinking,”



diminish personal autonomy by influencing behavior without Perspectives on Psychological Science, vol. 16, no. 6, pp. 1412-1427, 2021. conscious awareness. In contrast, Type 2 nudges, which [16] K. Dowding and A. Oprea, "Nudges, Regulations and Liberty," British encourage reflective decision-making, are less problematic as Journal of Political Science, vol. 53, pp. 204–220, 2023, doi:

they support autonomous decision-making. Nudges that lack 10.1017/S0007123421000685.

token transparency or are only transparent after the fact are more [17] A. Arno and S. Thomas, “The efficacy of nudge theory strategies in

influencing adult dietary behaviour: a systematic review and meta-

likely to infringe on autonomy. Conversely, when nudges are analysis,” BMC Public Health, vol. 16, no. 676, pp. 1-13, 2016, doi:

transparent in advance and individuals are informed about them, 10.1186/s12889-016-3272-x.

autonomy is better preserved. Additionally, the context in which [18] J. Wachner, M. Adriaanse, and D. De Ridder, "The influence of nudge

nudges are implemented plays a critical role in their ethical transparency on the experience of autonomy," Comprehensive Results in

Social Psychology, vol. 5, no. 1-3, pp. 49-63, 2021, doi:

assessment, as the goals and values of the intervention must align 10.1080/23743603.2020.1808782.

with those of the individuals affected. In conclusion, the ethical [19] J. K. Madsen, L. de-Wit, P. Ayton, C. Brick, L. de-Moliere, and C. J.

use of nudges in public policy requires focusing on preserving Groom, "Behavioral science should start by assuming people are

autonomy by choosing Type 2 nudges and ensuring high levels reasonable," Science & Society, vol. 28, no. 7, pp. 583-585, Jul. 2024. doi:

10.1016/j.tics.2024.04.010.

of transparency, especially regarding specific examples and [20] S. Banerjee and P. John, "Nudge plus: incorporating reflection into

advance notice. This approach allows nudges to support behavioral public policy," Behavioural Public Policy, vol. 8, pp. 69-84,

conscious decision-making rather than serving as tools for covert 2024. doi: 10.1017/bpp.2021.6.

manipulation. The Nudge Plus approach, which adds an element



Ali nas uporaba velikih jezikovnih modelov v znanstvenem



raziskovanju približuje časovni točki, ko bo stroj



nadvladal človeka?



Does the use of large language models in scientific



research bring us closer to the point in time when



machines will surpass humans?



Franc Mali

Faculty of Social Sciences

University of Ljubljana

Ljubljana, Slovenia

franc.mali@fdv.uni-lj.si



Povzetek closest to crossing this boundary, which poses a significant

existential risk to all of humanity. The central part of the

Prispevek se ukvarja z vprašanjem, ali veliki jezikovni modeli v discussion focuses on the question of which phases of today's

okviru generativne umetne inteligence že danes odpirajo vrata v scientific research the role of large language models has already

fazo splošne umetne inteligence in morda – kot naslednji korak become indispensable. – v fazo umetne superinteligence. S tem bi bili dani predpogoji



za prevlado strojev nad ljudmi. Pozornost je namenjena zlasti Keywords



raziskovanja. Raziskovalna dejavnost predstavlja eno najbolj generative artificial intelligence, large language model, scientific uporabi velikih jezikovnih modelov v procesu znanstvenega

ustvarjalnih človekovih intelektualnih dejavnosti. Logično creativity, existential risk, reinforcement learning

vprašanje je, ali je ravno znanstvena dejavnost, predvsem zaradi



svoje kreativne narave, najbližja prečkanju te meje, ki predstavlja 1 Uvod

pomembno eksistenčno tveganje za celotno človeštvo. Osrednji

del razprave je namenjen vprašanju, v katerih fazah današnjega V okviru pričujoče obravnave izhajam iz predpostavke, da se je

znanstvenega raziskovanja je vloga velikih jezikovnih modelov skozi celoten zgodovinski razvoj umetne inteligence implicitno

že postala nepogrešljiva. zastavljalo vprašanje, ali lahko ta doseže oziroma celo preseže

človeško inteligenco. Že od začetkov razvoja umetne inteligence



Ključne besede so bila tovrstna razmišljanja spodbujena z različnimi testi, ki naj

bi med drugim nakazovali, ali se strojna "inteligenca" približuje

generativna umetna inteligenca, veliki jezikovni modeli,

človeški inteligenci. Pomembni premik v teh razmišljanjih se je

znanstvena kreativnost, eksistenčno tveganje, okrepljeno učenje

zgodil, ko je tehnologija umetne inteligence prešla od klasičnih

Abstract načel strojnega učenja k načelom delovanja globokih nevronskih

mrež. V moji razpravi me v prvi vrsti zanima, ali najnovejši

The article addresses the question of whether large language

razvoj generativne umetne inteligence že kaže znake prehoda v

models within the framework of generative artificial intelligence

fazo umetne splošne inteligence in morda – kot naslednji korak

are already opening the door to the phase of artificial general

– umetne super inteligence. Posebej me zanima, ali najbolj

intelligence and, perhaps, as the next step, to the phase of

kreativna področja človekovega intelektualnga delovanja, kot to

artificial superintelligence. This would set the conditions for

predstavlja znanstveno raziskovanje, že odpirajo vrata nastopu

machines to dominate humans. Particular attention is given to the

umetne splošne inteligence. To namreč pomeni, da se počasi

use of large language models in the process of scientific research.

trasira pot nadvladi strojev nad človekom, kar je sicer predmet

Research activity represents one of the most creative human

precej distopičnih razmislekov filozofov in družboslovcev, tako

intellectual endeavors. The logical question arises whether

pri nas kot drugje v svetu. Moja obravnava ostaja na ravni

scientific activity, especially due to its creative nature, is the

nekoliko bolj splošne družboslovne refleksije o tej kompleksni

Permission to make digital or hard copies of part or all of this work for personal or tematiki in se ne ukvarja z ožjimi tehničnimi vidiki delovanja

classroom use is granted without fee provided that copies are not made or distributed umetne inteligence, zato se bom v primeru sklicevanj na

for profit or commercial advantage and that copies bear this notice and the full algoritme delovanja umetne inteligence oprl na nekoliko bolj

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s). poljudne definicije, kot so na primer tiste, ki jih je predstavil



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Partha Ray [1]. Po Rayu generativna umetna inteligenca (GUI) © 2024 Copyright held by the owner/author(s). spada v skupino modelov umetne inteligence, ki lahko ustvarjajo https://doi.org/10.70314/is.2024.cog.9

nove podatke (informacije) na podlagi vzorcev in struktur,

naučenih iz obstoječih podatkov (informacij). Ti modeli lahko



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Mali



generirajo vsebine na najrazličnejših področjih, naj si bo besedil, samoumevni način razrešiti. Njegova razrešitev je odvisna od

slik ali glasbe. Pri analizi, razumevanju in ustvarjanju teh vsebin, več dejavnikov. Eden izmed teh je možnost, da se modeli GUI

ki vedno bolj spominjajo na človeške stvaritve, se opirajo na razvijajo kot odprtokodni modeli, kar je seveda v nasprotju z

tehnike globokega učenja in nevronske mreže. Veliki jezikovni sedanjo strategijo multinacionalk, da preko lastniškega nadzora

modeli (VJM), ki se razvijajo pod okriljem GUI, pa so zasnovani novih naprednih tehnologij javnosti prikrivajo ključne

za generiranje naravnega jezika, kot so stavki, odstavki ali celotni informacije.

dokumenti. Njihova ključna lastnost je zmožnost predhodnega Negativna posledica lastniškega odnosa do VJM je, da znanje

učenja na velikih količinah besedilnih podatkov ter nato o notranjih mehanizmih delovanja VJM, ki predstavljajo vrh

prilagajanje za specifične naloge uporabnikov. V prispevku v razvoja umetne inteligence danes, še vedno predstavlja izziv za

štirih krajših poglavjih razpravljam (1) o umetni splošni oziroma večino uporabnikov, (deloma) pa tudi za strokovnjake s področja

super inteligenci kot dejavniku tveganja o človeku, (2) o dilemah, računalništva. Težko je namreč analizirati in priti na tej osnovi

ki so povezane z nadvlado stroja nad človekom, (3) o danes do razumevanja VJM, ki delujejo v okviru kompleksnih

vedno bolj nepogrešljivi vlogi VJM v posameznih fazah notranjih struktur z milijoni parametrov. Četudi lahko v vlogi

znanstvenega raziskovanja, (4) o specifičnih problemih uporabe uporabnikov ali celo računalniških razvijalcev vidimo končni

VJM na področju družboslovnega raziskovanja. Na koncu rezultat delovanja VJM, pa je pojasnitev oziroma interpretacija

prispevka je podanih še nekaj zaključnih misli. njihovih notranjih struktur izjemno zahtevna. Skratka, veliki

jezikovni modeli še vedno nastopajo kot »črne skrinjice« (»black

boxes«). Thomas Arnold je za opis te nevzdržne situacije

2 Umetna super inteligenca kot dejavnik uporabil naslednjo posrečeno analogijo: »To je tako kot da bi se

eksistenčnega tveganja za človeka prizadevali za razlago delovanja kompleksne kemijske reakcije,

Potem ko je Nick Bostrom pred desetimi leti postavil in utemeljil ne da bi poznali natančno strukturo in interakcijo molekul.« [7]

tezo, da obstaja verjetnost, da bo nadaljnji razvoj umetne V strokovni literaturi se sicer omenja tudi nekaj izjem. Za modele

inteligence pripeljal do nastopa umetne superinteligence, ki naj kot so BLOOM, Cerebras-GPT ali Llama, naj bi podjetja, ki se

bi bila neprimerno bolj kognitivno zmožna kot človek, kar bi ukvarjajo z umetno inteligenco, dopuščala večji javni vpogled

lahko predstavljalo eksistenčno tveganje za celotni človeško [8]. Spet za druge so informacije za javnost odprli, potem pa

vrsto, ta tema, zlasti po nastopu GPT 4 in drugih vrst VJM (Bard, ponovno zaprli. Četudi vrhunski znanstveniki, ki se ukvarjajo z

Claude, Llama, Gemini, itd.) vzbuja vedno večjo pozornost med UI in prihajajo iz akademske sfere znanosti, v vedno večjem

strokovnjaki, tako med naravoslovci in tehniki kot tudi med številu opozarjajo, da je prosti dostop do vseh informacij na tem

družboslovci in humanisti [2]. V svojem prispevku se bom področju eden ključnih dejavnikov, ki lahko zagotovi

izognil (spekulativnim) ocenam, ki se vrtijo okrog problema verodostojno in zanesljivo raziskovanje, saj le tako lahko

časovnih mejnikov, ko (če) naj bi pametni stroji nadvladali ljudi. dostopamo do informacij o celotni »arhitekturi« VJM (t.j. od

Ena skupina ekspertov namreč trdi, da se to ne bo zgodilo niti v uporabljenih podatkovnih baz do algoritmov), v zvezi s tem še

sto letih [3], druga skupina ekspertov spet trdi, da gre zgolj za vedno ni bilo storjenih veliko sprememb.

vprašanje dveh ali treh desetletij [4]. Bolj kot to, me zanima, ali



uporaba GUI v takšni kreativni človekovi dejavnosti kot je 3 Ali lahko ustvarjalno dimenzijo znanstvenega znanost resnično na široko odpira vrata nastopu umetne splošne

oziroma umetne super inteligence, ki naj bi se sicer zgodila v dela dokončno prevzame umetna inteligenca?

bližnji prihodnosti. To vprašanje je treba povezati s konkretno Na prihodnje izzive, ki so povezani z nastopom umetne splošne

prisotnimi strahovi pred katastrofičnimi in celo eksistenčnimi oziroma umetne super inteligence, je treba gledati tudi v luči

tipi tveganj GUI, ki bi lahko imeli negativne družbene posledice. današnjih dogajanj. Že danes si lahko zastavljamo vprašanje, ali

Če katastrofično tveganje ocenjujemo po kriteriju maksimalne bo ustvarjalno znanstveno delo dokončno prevzela GUI: ali je res

razširjenosti (število ljudi, ki bi bili prizadeti), intenziteti upravičeno trditi, da kar je nekoč kalkulator pomenil za številke,

(trpljenju, ki ga povzroča) in trajanja škodljivih družbenih in kar internet za globalni značaj komunikacijo, to danes pomeni

posledic nekontroliranega razvoja posamezne tehnologije, potem za znanstveno kreativnost razvoj GUI? Znanstveno kreativnost

pri eksistenčnem tveganju, ki naj bi bil povzean z umetno lahko subsumiramo pod bolj splošni pojem inteligence. Ta naj bi

inteligenco, odločilno vlogo igra samo en kriterij: nevarnost načeloma izkazovala celo paleto zmožnosti, od kreativnih do

iztrebljanja človeške vrste zaradi prevlade stroja nad človekom. racionalnih oblik (znanstvenega, umetniškega, itd.) mišljenja, od

Neredko se srečujemo z ocenami, da pomembni predpogoj za načrtovanja do učenja na temelju izkušenj, itd. Četudi danes

varni prihodnji razvoj GUI, v okviru katerega se lahko izognemo spekuliramo, da bo splošna umetna inteligenca dosegla ali

eksistenčnim tipom tveganj, predstavlja »algoritem okrepljenega presegla inteligenčne zmožnosti ljudi, pa bomo v strokovni

učenja« (v ang.: »reinforcement learning algorithms«) [5, 6]). Pri literaturi težko naleteli na neke soglasne kriterije, ki naj bi

»okrepljenem učenju« gre za to, da se v procesu sprejemanja povedali, kaj predstavlja »inteligenca« pri strojih in kaj

odločitev nagradi to, kar vodi v dobrobit ljudi. Vendar v predstavlja inteligenca pri ljudeh. Formalne definicije, ki

konkretnih situacijah težavo predstavlja praktično usklajevanje vztrajajo ne nekem skupnem imenovalcu, nam niso vedno v

funkcij umetne inteligence z sprejetimi družbenimi vrednotami. pomoč. Nobena izmed teh formalnih definicij ne ponuja nekega

Čeprav se ta problem na prvi pogled zdi trivialen, temu ni tako. dokončnega kriterija, ki bi nam omogočal primerjavo

Družbene vrednote so raznolike, amorfne in jih je težko »inteligentnosti« različnih entitet. Če se za hip ustavimo ob

zapopasti v kvantitativnih kategorijah. Problem, kako najnovejšem delu Yuval Noaha Harareja, ki nosi naslov »Nexus.

»okrepljeno učenje« uskladiti z sprejemljivimi družbenimi A Brief History of Information Networks from the Stone Stage

vrednotami, zato ni nekaj, kar se da na zelo enostaven in to AI« [9], bomo pri njemu hitro prepoznali besednjak, ki naj bi



Ali nas uporaba velikih jezikovnih modelov v znanstvenem

raziskovanju približuje časovni točki, ko bo stroj nadvladal Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

človeka?



nedvoumno nakazoval, da GUI poseduje moment dveh inteligenc vzame situacijo, ko nek uporabnik, ki komuncira

intencionalnosti, t.j. sposobnost GUI slediti delovanju, ki izhaja z klepetalnikom UI, ne zna več ločiti, ali je na drugi strani človek

iz njih samih. (Avtor knjige govori o tem, da se pametni stroji, ki ali stroj [8].

jih vodi GUI, sami odločajo, izbirajo, delujejo, itd.). Ob

prebiranju najnovejšega Hararejevega dela se lahko vprašamo,

zakaj vsiljuje intencionalnost kot ključni kriterij za izenačevanje 4 Zakaj umetna inteligenca postaja vedno bolj

»inteligentnosti« človeka in stroja. Lahko bi uporabil širšo nepogrešljiv pomočnik v vseh fazah

definicijo inteligence in bi le to pripisal že entitetam, ki so znanstvenega raziskovanja?

pasivne, torej ne vključujejo momenta intencionalnosti, vendar V okviru moje razprave me ob bolj načelnem epistemološkem

vseeno reagirajo na okolje in lahko opravljajo kompleksne vprašanju, t.j. ali nova tehnologija umetne inteligence postopoma

naloge. To je na primer storil Sebastien Bubeck, ki je skupaj z zavzema prostor znanstvene kreativnosti, zanima tudi bolj

soavtorji preučeval, ali so v jezikovnem modelu GTP-4 že dani konkretno vprašanje: ali ta nova napredna tehnologija dobiva

zametki umetne splošne inteligence. Postavil je namreč tezo, da status nepogrešljivosti v vseh drugih fazah znanstvenega

si neko inteligentno entiteto lahko predstavljamo tudi kot raziskovanja?

»orakelj«, ki nima notranjih vzgibov ali želja za delovanje, Sodobno znanstveno raziskovanje je multidimenzionalni

vendar lahko natančno in koristno zagotavlja informacije o kateri proces, ki vključuje različne faze, ki od začetka raziskovanja do

koli temi ali domeni vedenja [10]. Definicijo inteligence, ki končne objave znanstvenih rezultatov segajo od najbolj rutinskih

izhaja zgolj iz kriterija intencionalnosti, imamo lahko za pa do najbolj ustvarjalnih aktivnosti. VJM v tem primeru



izhajamo iz notranjih motivov za doseganja ciljev našega »asistenta« v vseh fazah znanstvenega raziskovanja. Bo ta delovanja v kar se da širokem okolju, kjer se soočamo z nikoli »asistent« v bolj ali manj oddaljeni prihodnosti postal restriktivno še iz enega razloga. Če namreč pri tej definiciji prevzema vlogo koristnega in vedno bolj nepogrešljivega

zaključenim spektrom situacij, potem v primeru rabe takšne »profesor«, ki bo dokončno nadomestil človeka – znanstvenika?

definicije implicitno predpostavljamo, da je pojem inteligence 1. Najprej je treba izpostaviti, da GUI zaradi svoje

neizogibno vezan na univerzalnost in optimalnost. To pomeni, da učinkovitosti vedno bolj nadomešča znanstvenike v postopkih

spet operiramo z apriorno definiranim in ne aposteriorno pridobivanja podatkov. Podobno je z učinkovitostjo GUI v

preverjenim konceptom inteligence. Dejansko oziroma realno vsebinskem pregledovanju, povzemanju in sumiranju množice

inteligenco človeka namreč nikakor ne moremo opredeliti kot informacijskih virov, ki so kot »state of the art« relevantne v

absolutno univerzalno in optimalno. vsakem začetnem procesu znanstvenega raziskovanja. GUI je

S podobnimi dilemami se soočamo, če naš pogled usmerimo sposoben obdelave in analize velike količine podatkov. Vloga



Tudi v tem primeru odgovor na vprašanje, ali umetna inteligenca znanstvene literature, povzemanju podatkov in sinteze enostavno privzema kreativne moči znanosti, ni enoznačen. Ne kompleksnih podatkovnih baz, samodejnem prepoznavanju na kreativnost kot eno izmed dimenzij človekove inteligence. GUI postaja neprecenljiva pri pregledu in sintezi vsebin iz

gre samo za to, da se že pri vprašanju kreativnosti človeka vzorcev in trendov, ki jih je mogoče izpeljati iz podatkov,

srečujemo z ogromnim številom definicij (znanstveniki modeliranju in napovedovanju na temelju zbranih podatkov, itd.

uporabljajo danes več kot 50 definicij [11]), zadeve postanejo še 2. Vedno bolj se povečuje vloga GUI pri ustvarjanju

bolj kompleksne, ko iščemo skupni imenovalec med definicijo novih idej. Glede na današnjo eksponentno rast znanstvenih

človeške kreativnosti in kreativnostjo, ki jo pripisujemo umetni informacij je prenos te raziskovalne funkcije iz človeka na GUI

inteligenci. Na eni strani imamo avtorje, kot so Marc Ruco [12] hkrati povezana z zmožnostjo GUI, da učinkovito in predvsem

ali Stephen Rice [13], ki pravijo, da kolikor k standardnim avtonomno ustvarja nova raziskovalna vprašanja in hipoteze.

definicijam človekove ustvarjalnosti – ta vključuje dimenzijo Eden največjih izzivov najbolj naprednih področij znanosti je

originalnosti in učinkovitosti - dodamo tudi dimenziji skorajda neskončno število hipotez, ki se nanašajo na

avtentičnosti, potem GUI ne more tekmovati z ljudmi. raziskovalne probleme, zaradi česar se včasih zdi natančno

Na drugi strani imamo avtorje, kot na primer Hubert Kent, za sistematično raziskovanje, ki bi omogočalo sprejetje hevristično



kreativnosti, vsaj kar zadeva t.i. odprti tip mišljenja, saj naj bi nemogoče. Primer: v biokemiji naj bi obstajalo približno 10⁶⁰ empirične analize pokazale, da GPT-4 že zmore doseči rezultate, molekul, to pa je praktično enako številu zdravil, ki jih je treba katere je GPT- 4 že dosegel izredno visoko stopnjo znanstvene najbolj obetavne hipoteze, brez sodelovanja UI skorajda

ki so enaki rezultatom, ki jih doseže zgolj 1% najbolj na temelju ogromnega števila molekul šele odkriti [16]. Pri tem

inteligentnih ljudi [14]. Rezultati dodatnih študij naj bi ravno imajo ravno najnovejši modeli GUI potreben potencial, da

tako dokazovali, da model GPT-4 izkazuje veliko stopnjo revolucionarno posežejo v to fazo znanstvenega raziskovanja, ko

fleksibilnosti zunaj ustaljenih okvirov mišljenja in naj bi imel na gre za biokemijo. Podobne primere bi lahko navedli za področje

področju odprtega tipa mišljenja celo višji kreativni potencial od genomike, astronomije, kvantne fizike, itd. Ne moremo mimo

ljudi. Pričakovati torej je, pravijo avtorji, ki so opravili te in omembe še ene funkcije GUI. Ta funkcija GUI je vezana na

podobne študije, da bo model GPT-4, kolikor bo dosežen njeno zmožnost usmerjanja k bolj interdisciplinarno zasnovanim

napredek glede povečanih zmožnosti učenja na velikih bazah revolucionarnim znanstvenim odkritjem, saj so njeni potenciali

podatkov in bolj napredni arhitekturi nevronskih mrež, kmalo pri obdelavi in sintezi informacij iz različnih disciplin skorajda

storil pomembni korak v smeri umetne splošne inteligence.[15]. neomejeni. V tem primeru Turingovi testi že zvenijo zastarelo. V okviru rabe

Turingovega testa gre namreč za to, da se kot kriterij izenačitve



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Mali



3. Vloga GUI se povečuje tudi v procesih evalvacije preučujejo, kako delovanje oziroma vedenje na mikro ravni

končnih rezultatov znanstvenega raziskovanja. Če izhajam iz (npr.: odločitve individualnih agentov) vodi do posledic na

bolj splošnih epistemoloških predpostavk in se na tem mestu makro (družbeni) ravni (npr.: oblikovanje družbenih vzorcev

izognemo razpravi o prednostih in tudi tveganjih uporabe GUI v delovanja oziroma obnašanja). V okviru teh modelov se seveda

konkretnih recenzentskih postopkih, potem naj na kratko lahko preučuje tudi obratni vpliv: kako makro-nivo vpliva na

omenimo zgolj eno izredno pomembno vlogo te nove napredne obnašanje na mikro ravni [8]. V okviru sociologije se s temi

tehnologije, t.j. preverjanje rezultatov eksperimentalnih in drugih »agent-based« modeli preučuje socialna omrežja, oblikovanje

empiričnih raziskav. V preteklosti je v glavnem veljalo, da ni sosedskih skupnosti, itd.

problematična ponovljivost dobljenih znanstvenih rezultatov, Se pa v zvezi z družboslovnim raziskovanjem pojavlja

bodisi na temelju javno dostopnih znanstvenih objav ali ustreznih določen paradoks, na katerega želim opozoriti v tem sklepnem

eksperimentalnih protokov. Sodobna znanost se nahaja v vedno delu moje razprave. Ta paradoks predstavlja dejstvo, da postopki

večji krizi, kar zadeva zmožnost replikacije, saj je tako z vidika »okrepljenega učenja« (ang. »reinforcement learning«), ki naj bi

časa kot tudi stroškov v številnih, če ne kar vseh vseh odpravili »halucinacije« in raznovrstne pristranosti, predstavljajo

znanstvenih disciplinah težko izvesti potrebne eksperimentalne oviro za doseganje objektivno veljavnih znanstvenih rezultatov.

in druge znanstvene ponovitve. O tveganjih za povečanje goljufij Če pride skozi delovanje t.i. »reinforcement self-learning by

in prevar v moderni znanosti, ki izhajajo iz teh kompleksnih human feed-back« (RLHF) do idealiziranja sveta, t..j. sveta,

situacij raziskovalnega dela, sem več pisal na drugih mestih [17]. kakršen naj bi bil, ne pa sveta, kakršen dejansko je, takšna

GUI lahko odigra zelo relevatno funkcijo v današnjem času prizadevanja za zmanjšanja pristranskosti algoritmov, katerih cilj

enormne produkcije znanstvenih rezultatov, ko je vedno težje je promovirati liberalne vrednote, lahko ogrozijo veljavnost

izvajati ponovitve eksperimentov z namenom izvajanja kontrole raziskav v družboslovju, ki jih podpira umetna inteligenca.

znanstvenih rezultatov. Njen predikativni pristop namreč lahko »Požarni zid«, ki se ga želi danes pospešeno graditi preko RLHF,

zagotovi učinkovito, hitro, sistematično in natančno napoved odpravlja tveganja GUI, kar zadeva njeno široko uporabo (in

ponovljivosti posameznih znanstvenih odkritij ali pa celo vseh preprečuje tveganja, ki so se, kot pravi Yuval Harare, že zažrla v

spoznanj na posameznem področju znanosti. civilizacijski kod sodobnih družb), po drugi strani pa predstavlja

4. Pozitivna vloga GUI se danes povečuje tudi v okviru epistemološko tveganje za objektivni značaj današnjih

širših družbenih in kognitivnih predpostavk, ki so relevantne za družboslovnih raziskav. Tudi to predstavlja dilemo današnjega in

delovanje moderne znanosti. V zvezi s to širšo funkcijo bi prihodnjega razvoja umetne inteligence, ki zahteva naš celovit

izpostavil vlogo GUI pri spodbujanju komunikacij znotraj interdisciplinarni razmislek, saj se je le na tej osnovi mogoče

znanstvene skupnosti, pa tudi komunikacije znanstvenikov izogniti negativnim družbenim in tudi epistemološkim

navzven. To zadnje naj bi se dogajalo predvsem s pomočjo implikacijam njenega razvoja.

modela ChatGPT, ki generira takšne tipe pojasnitev, ki vodijo k

premagovanju komunikacijskih prepadov med eksperti in laiki.

Vendar je to funkcijo, kot smo že opozorili v enem izmed 6 Zaključek

predhodnih poglavij, mogoče izvajati le, če bo prišlo do V zadnjem času je tako v znanstvenih krogih kot tudi zunaj

uveljavitve nove paradigme odprtokodne znanosti. V zadnjem znanstvenih krogov veliko govora o možnih tveganjih

času strokovnjaki, ki delujejo na področju GUI, vedno bolj današnjega in prihodnjega razvoja umetne inteligence.

poudarjajo, da je treba razviti modele, ki bodo čim bolj Znanstveniki iz Massachusetts Institute of Technology, ene

korespondirali z fizično realnostjo. Menijo, da je treba največ najbolj uglednih akademskih institucij v ZDA, so v letošnjem

naporov usmeriti v nadaljnji razvoj multimodalnih sistemov letu pripravili javno dostopni repozitorij z umetno inteligenco

GUI. Demis Hassabis, izvršni direktor firme DeepMind, je v povezanih primerov tveganj. V omenjenem repozitoriju se

intervjuju za angleški dnevnik Guardian konec prejšnjega leta trenutno nahaja kar 777 opisov takšnih tveganj. To je še en

dejal, da je bil storjen na tem področju največji korak z modelom dokaz, kako veliko zanimanje obstaja danes za ta vprašanja. V

Gemini, ki ga razvija njegovo podjetje [18]. mojem kratkem prispevku sem se dotaknil zgolj enega izmed teh

številnih problemov, ki je vezan bolj na epistemologijo



5 Ali nova tehnologija generativne umetne razvoja umetne inteligence. V tem kontekstu me je predvsem znanstvenega raziskovanja, ne pa toliko na družbene posledice

inteligence v okviru družboslovnega zanimalo, ali pospešena raba VJM v okviru različnih znanstveno-

raziskovanja nujno in vedno zagotavlja raziskovalnih aktivnosti predstavlja eno izmed domen, kjer se na

znanstveno objektivnost? stežaj odpirajo vrata nastopu umetne splošne oziroma umetne

Kot družboslovca me seveda zanima tudi vprašanje vedno večje super inteligence. Še posebej me je zanimalo vprašanje, zakaj

rabe VJM na področju mojega področja znanstvenega GUI postaja že danes nepogrešljivo »orodje« v seh fazah

raziskovanja. Kar takoj je treba reči, da na področju družbenih znanstvenega raziskovanja. V sklepnem delu sem se na kratko

ved VJM izkazujejo velik (hevristični) potencial v razvijanju ustavil ob nekaterih specifičnih dilemah uporabe GUI na

novih pristopov k anketnim raziskavam in ponovljivosti področju družboslovnega raziskovanja.

eksperimentov na področju vedenjske ekonomije [19],

diskurzivnih analizah tekstov, ki jih je mogoče izvajati na Literatura



avtomatizirani način [20] in končno tudi na področju razvijanja [1] Ray Parta, 2023. ChatGPT: A comprehensive review on background, applications, key challenges, bias, ethics, limitations and future scope. modelov, ki simulirajo stvarno obnašanje ljudi. V tem zadnjem Internet of Things and Cyber-Physical Systems. 3: 21–154;

primeru gre predvsem za t.i. »agent-based« modele, ki https://doi.org/10.1016/j.iotcps.2023.04.003.



Ali nas uporaba velikih jezikovnih modelov v znanstvenem

raziskovanju približuje časovni točki, ko bo stroj nadvladal Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

človeka?



[2] Nick Bostrom, 2014. Superintelligence: Paths, dangers, strategies. Oxford

University Press.

[3] Katja Grace, John Salvatier, Allan Dafoe, Baobao Zhang, & Owain Evans,

2022. When will AI exceed human performance? Evidence from AI

experts. Journal of Artificial Intelligence Research, 62(July), 29-754.

https://doi.org/10.1613/jair.1.11222.

[4] Max Roser, 2023. AI timelines: What do experts in artificial intelligence

expect for the future? Our World in Data. https://ourworldindata.org/ai-

timelines.

[5] Rishi Bommasani et al., 2022. On the Opportunities and Risks of

Foundation Models. arXiv:2108.07258v3 [cs.LG] 12 Jul 2022.

[6] Yogesh Dwivedi et al., 2023. Opinion Paper - So what if ChatGPT wrote

it? Multidisciplinary perspectives on opportunities, challenges and

implications of generative conversational AI for research, practice and

policy. International Journal of Information Management, 71 (2023)

102642; https://doi.org/10.1016/j.ijinfomgt.2023.102642.

[7] Thomas Arnold, 2024. Herausforderungen in der Forschung: Mangelnde

Reproduzierbarkeit und Erklärbarkeit. V: L. Ohly in G. Schreiber (Hrsg.)

KI:Text.Diskurse über KI-Textgeneratoren, str. 67-83. Berlin/Boston: De

Gruyter Verlag.

[8] Christopher Bail, 2024. Can Generative AI improve social science?.

PNAS, May 9, 2024 1211) e2314021121;

https://doi.org/10.1073/pnas.2314021121.

[9] Yuval Noah Harari, 2024. Nexus. A Brief History of Information

Networks from the Stone Stage to AI. New York: Penguin Random House.

[10] Bubeck S. et al., 2023. Sparks of Artificial General Intelligence: Early

experiments with GPT-4 (2023), arXiv:2108.07258v3 [cs.LG] .

[11] L.B. Soros, Alyssa Adams, Stefano Kalonaris, Olaf Witkowski, Christian

Guckelsberger, 2024. On Creativity and Open-Endedness.

arXiv:2405.18016v4 [cs.AI] 23 Jun 2024.

[12] Marc Runco, 2023. AI can only produce artificial creativity. Journal of

Creativity 33 (2023) 100063. https://doi.org/10.1016/j.yjoc.2023.100063.

[13] Stephen Rice, Winter Scott, Rice Connor, 2024. The advantages and

limitations of using ChatGPT to enhance technological research.

Technology in Society 76 (2024) 102426.

https://dx.doi.org/10.2139/ssrn.4416080.

[14] Hubert Kent, Kim Awa, Darya Zabelina, 2024. The current state of

artificial intelligence generative language models is more creative than

humans on divergent thinking tasks. Scientific Reports 14(1):3440;

https://doi.org/10.1038/s41598-024-53303-w.

[15] Mohamed Salah et al., 2023. Chatting with ChatGPT: decoding the mind

of Chatbot users and unveiling the intricate connections between user

perception, trust and stereotype perception on self-esteem and

psychological well-being. Current Psychology, 43, 7843–7858 (2024).

https://doi.org/10.1007/s12144-023-04989-0.

[16] Hanchen Wang et al., 2023. Scientific discovery in the age of artificial

intelligence. Nature, 620: Vol 620 | 3 August 2023. DOI: 10.1038/s41586-

023-06221-2.

[17] Franc Mali, 2011. Razvoj moderne znanosti. Socialni mehanizmi.

Ljubljana: Založba FDV.

[18] Demis Hassabis, 2023. Google releases new AI model with claim it can

outperform ChatGPT in most tests. The Guardian, 7. December, 2023.

[19] John Horton, 2023. Large language models as simulated economic agents:

What can we learn from homo silicus? arXiv [Preprint] (2023).

10.48550/arXiv.2301.07543.

[20] Igor Grossmann, Cassandra Parker, Mathew Feinberg, Nicholas

Christakis, 2023. AI and the transformation of social science research.

Science, 380(6650):1108-1109. DOI:10.1126/science.adi1778.Ian Editor

(Ed.). 2007. The title of book one (1st. ed.). The name of the series one,

Vol. 9. University of Chicago Press, Chicago.

DOI:https://doi.org/10.1007/3-540-09237-4.



Comparing academic performance across course topics:



a pilot study



Laura Fink Bojan Cestnik

Faculty of entrepreneurship Faculty of entrepreneurship

GEA College GEA College

Ljubljana, Slovenia Ljubljana, Slovenia

laura.fink@gea-college.si bojan.cestnik@gea-college.si

bojan.cestnik@temida.si



ABSTRACT KEYWORDS

In this paper, we examine the academic performance of students course-specific competence, learning analytics, prediction

in different courses to determine whether good performance in models, student academic achievement, student academic

one course is related to good performance in other courses. performance Although certain predictive models emphasize the importance of

course content for learning success, there are few studies that

address how student performance in different courses is related 1 INTRODUCTION AND RELATED WORK

to similar course topics, learning goals, competences, and skills.

By creating a preliminary framework that examines how Are students who achieve excellent results in one course more

academic performance is related to different course topics, we likely to achieve outstanding results in another? And vice versa?

attempt to make a first step further towards addressing the This is the key question that triggered this preliminary research

research gap regarding the interrelatedness of student into how students perform in different courses. As a result, we

achievement not only in different course topics but in different are examining the relationships between students' academic

competence areas. We examined a set of student grades from achievement in different courses from two different

eleven different courses at the faculty from areas such as undergraduate study programs at the faculty of entrepreneurship.

entrepreneurship, management and leadership, business Despite some prediction models suggesting the course content as

informatics, mathematics, economics, marketing and market one of the input explanatory variables, there is a significant lack

analysis, innovation and creativity, English, finance and of detailed research on the relationship between students'

accounting, business law, and human resource management. We achievement in different courses from the same, similar, or

show that students with more exam retakes on average reached a entirely different discipline. Therefore, this preliminary study

lower grade rank than the students who only registered for the aims to develop a pilot model to investigate the relationship

exam once. We used linear regression to show the significance between academic success in various courses and their respective

of the relationships between student performance in the course topics. The findings of this study could be further

Informatics course compared to their achievement in other reinforced and interpreted by the comparison of course-specific

courses. With a correlation matrix coefficient, we measured the competences and learning objectives.

strength of reciprocal interrelatedness between the grade ranks Apart from the prediction model of academic achievement,

students attained in each of the eleven courses. The results of this interrelatedness between academic achievement in different

preliminary study indicate a possible stronger association courses compared to the course main topics, competences, and

between academic achievement in courses that have similarities learning objectives is still largely missing. Since previous studies

in terms of content or focus, such as business administration and that would previously investigate the interrelatedness of student

entrepreneurship (correlation coefficient of 0.58). Further studies achievement in different courses are, to the best of our

with detailed comparison of course-specific competences are knowledge, entirely nonexistent, let us draw attention to the

needed for accepting the finding that interrelatedness between research studies that are related to the field and that have led to

achievements in courses from similar versus different disciplines this investigation. OECD [12, 13], for example, show a positive

is stronger. The preliminary model could further be improved by association between literacy and numeracy skills. Moreover, in

a broader range of courses, input explanatory student factors, and our previous research [2], we showed that students who achieved

application of advanced analytical techniques. better academic achievement in word skills were on average also

more likely to achieve better achievement in excel skills.

∗ Comparing academic performance across course topics: a pilot study. However, this does not imply that it is sufficient to develop only

† Fink, L., Cestnik, B. some of these skills, such as solely including word skills, excel

Permission to make digital or hard copies of part or all of this work for personal or skills, literacy skills, or just numeracy skills in the curriculum.

classroom use is granted without fee provided that copies are not made or distributed Furthermore, this does not imply that students in a real situation

for profit or commercial advantage and that copies bear this notice and the full cannot achieve much better results in a certain type of skill

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s). compared to another. Additionally, Fink and Vadnjal [3]

Information Society 2023, 7–11 October 2024, Ljubljana, Slovenia conducted a pilot study that compared the development of

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.cog.10



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia L. Fink and B. Cestnik



generic and course-specific competences during a higher H1. Mostly there are reciprocal relationships between a

education course. student's performance in one course and their performance

As previously mentioned, one could apply findings about the in another.

relationships between various course topics, contents,

competences, or objectives to the development of predictive H2. As the number of exam retakes increases, the student's

models and predictive algorithms. Prediction models are often grade rank decreases.

used to find out ahead of time which students are likely to drop

out or fail external exams. The schools aim to take intervention

measures and steps to stop the bad predictions from coming true, 3 DATASET AND METHODOLOGY

and the success rate can be raised [16, 9]. There are a lot of

different input explanatory variables that can have a big effect on The dataset collection and preparation included several

how well and especially how accurately prediction models phases. First, we have collected students’ grades for different

perform [16, 4, 7, 8, 9, 11, 1]. Some prediction models include courses at the higher education institution. The initial dataset

related concepts to the course content [7], such as the course's included the grades of 223 students for 67 different courses.

learning objective, the course's main competences, the course In the second phase, we have refined and further prepared the

topic [4], or even course preparedness [10]. dataset. Based on some simple data exploration and visualization

In addition to the combination of input explanatory variables, techniques, such as plotting the missing values, plotting the

various statistical methodologies and techniques, as well as distribution of the number of grades available per course, and

different types of measurements and academic achievements, plotting the distribution of the number of exam retakes per

significantly influence the prediction model's power and course, we have decided to eliminate the data of courses with less

performance. While adding additional or all relevant factors to than 60 students’ grades per course. With that, we narrowed

the model does not always improve its performance, the right further analysis to the following eleven selected courses:

combination of input explanatory variables significantly Business Economics, Informatics, Management and Leadership,

influences the accuracy and other model performance measures Marketing and Market Analysis, Entrepreneurship, Business

[4]. In the end, the right combination of input variables largely English, Accounting, Creativity and Innovation, Business

determines the model's explanatory power, the accuracy of its Mathematics and Statistics, Business Law, and Human Resource

predictions, and other performance measures [6, 15]. Management.

Different prediction models are based on different

methodologies and include different input and output variables.

Francis and Babu [4], for example, compare prediction models Table 1: Number of grades per one and two courses

of student achievement that include the topic of the course as the

input explanatory variable. They demonstrated that the course Course (Short name) Nr. of students/grades

topic, along with many other factors, is one of the explanatory per for chosen

variables for academic achievement. However, their model course course and

found that academic factors, including the course topic, were less Informatics (%)

accurate in predicting students' academic achievement than Informatics for entrepreneurs 160

demographic factors, behavior factors, and other factors such as (P07_IE)

absence days, parental satisfaction, and school survey responses. Business economics (P05_BE) 139 120 (86 %)

They developed, compared, and assessed performance measures Business law for entrepreneurs 79 71 (90 %)

of several prediction models. The model that included academic (P06_BLE)

factors, behavior, and additional input explanatory variables Human resource management 66 58 (88%)

showed the greatest improvement in accuracy. On the other hand, (P09_HRM)

adding demographic factors on top of that reduced the accuracy Management and leadership 156 142 (91%)

of academic achievement prediction. Clearly, the addition of (P15_ML)

additional input variables, for example, the topic of the course, Marketing and market analysis 90 73 (81%)

in different models contributes differently to improving (P16_MMA)

prediction accuracy and other performance measures, depending Entrepreneurship (P27_ENT) 191 148 (77%)

on other input variables in the model. Business English (P28_BE) 152 133 (88%)

Accounting for entrepreneurs 143 131 (92%)



2 (P33_AE) HYPOTHESES

Creativity and innovation in 173 142 (82%)

entrepreneurship (P36_CIE)

In this preliminary study, we aim to build a preliminary pilot Business mathematics and 139 119 (86%)

research model on which we will test the interrelatedness statistics (P39_BMS) between academic achievement in different courses. We suggest



the following hypotheses: We then compared the number of students’ grades available

per one course with the number of grades available per two

courses (the selected course and the informatics) and calculated

the share of students that also took the exam in informatics

Comparing academic performance across course topics: a pilot

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

study.



compared to the number of students who took the exam in the

selected course only. Students' grades include both those that In the third phase, we continued with the data exploration and

indicate a student has passed the course and those that indicate a visualization. We plotted the distribution of the number of grades

student has not. Since we included only one grade per student in achieved per grade rank for each of the selected eleven courses.

further analysis, the number of students’ grades reflects the An example of the distribution for the Entrepreneurship course

number of students who took the exam in each course. Not all is provided in Figure 1.

students took exams in all eleven subjects. The reasons for this



are varied, including the fact that the courses are from two

different programs.

The courses with the highest student grades include

Entrepreneurship (191), followed by Creativity/Innovation (173)

and Informatics (160) (Table 1). More than 77% of students

(148) who took the entrepreneurship exam also took the

informatics exam. Similarly, 82% of students who took the exam

in creativity/innovation (142), and 91% of those who took the

exam in management/leadership (142) also took the exam in

informatics. In other words, students who attended the exam in

the Informatics course often also attended the exam in the

Entrepreneurship, Management/Leadership, and

Creativity/Innovation courses, as shown in Table 1.

Then, we continued by calculating the average grade ranks

and number of exam retakes (table 2) for each of the courses

included in the analysis. The grade ranks range from 0 to 10,

where 0 represents not attending, 1 to 5 represents failed, 6 Figure 1: Distribution of number of grades per each grade

satisfactory, 7 average, 8 good, 9 very good, and 10 excellent. As rank for Entrepreneurship course example

shown in Table 2, the students on average achieved better grades

in the HRM course than in other courses. In comparison, the In the next phase, we focused our investigation on how the

students on average achieved the lowest average grade rank in grades that students achieved in the Informatics course behave

the Mathematics/Statistics and Economics course compared to compared to the grades they achieved in the remaining ten

other courses. The students also, on average, most commonly courses. We performed and visualized ten linear regression

retook the exam in these two courses. Additionally, we found that models describing the relationships between grade ranks students

38 students attended at least one exam deadline for each of the achieved in the informatics course and the grade ranks students

eleven courses. achieved in other selected courses. When performing the linear

regression, we included the grade rank achieved in informatics



Table 2: Average grade rank and number of exam retakes regression models assume the influence of the independent as an independent variable. Though we are aware that linear

variable on the dependent variable and not the reciprocal

Course (Short name) Avg. Avg. nr. of relationships per se, we decided to mention this limitation and

grade retakes work further with the results obtained from the regression

Informatics for entrepreneurs 7.7 1.2 analyses in this preliminary pilot study.

(P07_IE) We performed additional analysis based on the correlation

Business economics (P05_BE) 6.6 1.8 matrix between the grade ranks students achieved in each of the

Business law for entrepreneurs 7.1 1.3 eleven selected courses. We draw a correlation matrix with

(P06_BLE) significant (p < 0.05) correlations among regression coefficients

Human resource management 8.0 1.1 between the eleven selected courses to determine which of the

(P09_HRM) eleven courses is related to another one. We then examined the

Management and leadership 7.3 1.2 strength and significance of the reciprocal relationship that the

(P15_ML) correlation matrix coefficient measures.

Marketing and market analysis 7.5 1.3 Next, we further compared the characteristics of our data for

(P16_MMA) the selected eleven courses with the characteristics of the data for

Entrepreneurship (P27_ENT) 7.4 1.3 all the courses. We used data visualization techniques such as

Business English (P28_BE) plotting to compare the distribution of the average grade of 7.3 1.2

Accounting eleven selected courses and all the courses. The distribution of for entrepreneurs 7.0 1.3

(P33_AE) average grade for the selected eleven courses seems fairly similar

Creativity to the distribution of average grade for all the courses, as shown and innovation in 7.2 1.3

entrepreneurship (P36_CIE) in Figure 2.

Business mathematics and statistics Although we cannot claim that an analysis of the entire data 6.4 2.0

(P39_BMS) set would yield similar results to the analysis of the selected

eleven courses based solely on the similar distribution of average

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia L. Fink and B. Cestnik



grades, we cannot completely rule out the possibility that on student's performance in one course and their performance in

average, somewhat similar associations would emerge among the another in most cases when comparing different pairs of courses.

other courses. Let's examine the concrete correlations between the grades of

the two courses. There is a moderate correlation between

Economics and Mathematics/Statistics (0.60), followed by



Economics and Entrepreneurship (0.58), Informatics and

Mathematics/Statistics (0.57), Marketing/Market Analysis and

Law (0.57), Marketing/Market Analysis and Entrepreneurship

(0.55), Economics and Accounting (0.54), Marketing/Market

Analysis and Creativity/Innovation (0.52), and Economics and

Informatics (0.51), and Informatics and Marketing/Market

analysis (0.50).

Based on moderate correlations, we can speculate that

Economics, Informatics, Mathematics, and Accounting courses

are closely connected, partially because students have to use the

numeracy skills in these courses. Therefore, future research

could address the question of whether the syllabus of these

courses reflects shared similar competences. Although rare,

previous research [12, 13] on the interrelatedness of competences

Figure 2: Distribution of number of grades per each grade has shown that people who are more proficient in literacy skills

rank per course example are usually also more proficient in numeracy skills, and vice

versa, additional inquiry into the similar and different

In the final phase of this preliminary pilot research, we competences that are developed within these courses would

performed the regression analysis between the grade rank that the provide an important insight and more thoroughly address the

student achieved and the number of exam retakes of the student. gap in the literature on the interrelatedness between competences

that is largely still missing. Therefore, we also need to investigate

further to what degree are the numeracy skills included in the

4 RESEARCH FINDINGS syllabus of the courses that otherwise aim at developing soft,

social, and other professional skills, such as the Entrepreneurship

The eleven selected courses served as the basis for the and Marketing/Market Analysis courses. Additionally, based on

analysis, which focused on identifying reciprocal relationships the analysis, it would also make sense to check whether the

between the grade ranks of one course and those of another Marketing/Market Analysis and Creativity/Innovation courses

course. We selected the courses based on the number of grades foster the development of related skills. In general, we can

available after exploring, visualizing, cleaning, and refining the speculate that the type of competences is important for student

data. In addition to linear regression models, we calculated the achievement, but to confirm this, we would have to perform the

correlation matrix’s correlations (Figure 3) that investigate the qualitative analysis of the similar competences and learning

reciprocal association between the grade ranks students achieved objectives in the future.

in one course compared to another. The majority of the relationships are statistically significant

Overall, we investigated 55 reciprocal relationships (H1) but weak. Weak correlation exists between Entrepreneurship and

between the grade ranks of one course with the grade ranks of Mathematics/Statistics (0.49), Entrepreneurship and Informatics

another course. Among these, forty-six correlations are (0.47), Economics and Management/Leadership (0.47),

significant (p < 0.05) compared to nine correlations that are not Entrepreneurship and Law (0.45), English and Entrepreneurship

significant. Among the significant correlations, nine correlation (0.45), Economics and English (0.43), Creativity/Innovation and

coefficients exhibit a moderate relationship (between 0.5 and Entrepreneurship (0.42), HRM and Law (0.41), Accounting and

0.7) between the grade ranks of one course and the grade ranks HRM (0.41), Informatics and Accounting (0.41),

of another course. Twenty-seven correlations show a weak Management/Leadership and Accounting (0.40), English and

correlation (between 0.3 and 0.5), while ten correlations show a Mathematics/Statistics (0.38), Mathematics/Statistics and Law

negligible or low correlation (below 0.3) between the grade ranks (0.38), Marketing/Market Analysis and Law (0.37), English and

of two selected courses. Accounting (0.37), Management/Leadership and English (0.36),

The strength of significant coefficients varies from 0.19 all Entrepreneurship and HRM (0.36), Law and Economics (0.34),

the way up to 0.58. Many of the coefficients that we would Management/Leadership and Entrepreneurship (0.34),

otherwise have placed in the group of weak correlations are very Entrepreneurship and Accounting (0.33), Accounting and

close to 0.5, which indicates moderate correlation. There are also Mathematics/Statistics (0.33), Management/Leadership and

many coefficients that we have otherwise placed in the group of Marketing/Market Analysis (0.31), Marketing/Market Analysis

negligible correlations close to the value of 0.3, which indicates and Economics (0.31), Management/Leadership and Law (0.31),

a weak correlation. Informatics and Law (0.30), English and HRM (0.30),

Based on these results, we can accept hypothesis 1 that Creativity/Innovation and Accounting (0.30).

foresees the existence of reciprocal relationships between a A low correlation below 0.3 exists between Accounting and

Law (0.29), Marketing/Market Analysis and English (0.29),

Marketing/Market Analysis and Accounting (0.28), Economics

Comparing academic performance across course topics: a pilot

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

study.



and Creativity/Innovation (0.27), Informatics and once (β = -0.47, p = 0.00). The full equation is displayed below

Management/Leadership (0.26), Management/Leadership and (equation 1). Based on these results, we accept hypothesis 2 that

Creativity/Innovation (0.25), Management/Leadership and as the number of exam retakes increases, the student's grade rank

Mathematics/Statistics (0.25), English and Informatics (0.22), decreases.

English and Creativity/Innovation (0.20), and

Creativity/Innovation and Mathematics/Statistics (0.19).

Based on the weak and low correlations, we can speculate that 𝐺𝑟𝑎𝑑𝑒 𝑟𝑎𝑛𝑘 = 8.25 − 0.47

these correlations exist in courses that do not necessarily share ∗ 𝑁𝑟. 𝑜𝑓 𝑒𝑥𝑎𝑚 𝑟𝑒𝑡𝑎𝑘𝑒𝑠 (1)

that much of similar competences and learning objectives as + 𝑟𝑒𝑠𝑖𝑑𝑢𝑎𝑙𝑠 those that exhibit moderate correlations. As mentioned



these assumptions. Figure 4 also shows that most students register for the exam previously, further qualitative research is required to substantiate



number of observed instances for each pair of two courses ranges Since not all students took the exams in all the courses, the once, fewer students register for the exam twice, and even fewer students register for the exam a third time .

from 46 to 166, depending on the particular pair of courses. Not

only that, we found that the low number of observed instances

importantly contributed to the statistical insignificance of some

correlation coefficients. The relationships calculated based on a

low number of observed instances, such as, for example, between

Mathematics/Statistics and Law (46 observed instances),

Marketing/Market Analysis and Mathematics/Statistics (46),

HRM and Economics (47), Informatics and HRM (58), English

and Law (58), Management/Leadership and HRM (59), and

Creativity/Innovation and HRM (59 observed instances), exhibit

insignificant relationships. The exemptions are the insignificant

relationship between the Informatics and Creativity/Innovation

course with 142 observed instances and between the

Creativity/Innovation and Law course with 74 observed

instances. We therefore speculate that the Informatics and Law

courses strive to develop different competences than the

Creativity/Innovation course. This preliminary analysis is a Figure 4: Regression analysis between number of exam

useful basis for further research and analysis. retakes and average grade achieved





5 CONCLUSION





The purpose of this study was to lay the groundwork for

further research regarding the correlation between different

course-specific competences and to present initial findings

regarding the correlation between students’ academic

achievement in different courses. In this paper, we aim to

enhance our comprehension of the intricate relationship between

competences, an area that remains largely unexplored. The

preliminary analysis revealed the existence of interrelatedness

among grades students achieve in different courses, and showed

that a student's academic performance in one course influences

their performance in another. Based on the analysis we accept

hypothesis 1 that foresees the existence of reciprocal

relationships between a student's performance in one course and

their performance in another in most cases when comparing

Figure 3: Correlation matrix with significant correlations different pairs of courses. We also show that students with more

exam retakes on average reached a lower grade rank than the

The results of the regression analysis between the grade rank students who only registered for the exam once. With that, we

that the student achieved and the number of exam retakes that the accept hypothesis 2 that as the number of exam retakes increases,

student took (H2), that are shown in Figure 4, suggest that the the student's grade rank decreases.

students with more exam retakes on average reached a lower To determine whether there is a stronger correlation between

grade rank than the students who only registered for the exam academic achievements in courses from the same or similar



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia L. Fink and B. Cestnik



discipline than in courses from completely different disciplines, [4] Bindhia K. Francis and Suvanam Sasidhar Babu. 2019. Predicting

further research is required to explore how much the academic performance of students using a hybrid data mining approach.

Journal of medical systems, 43(6), 162.

interrelatedness between courses depends on the competences, [5] Leon Gerritsen. 2017. Predicting student performance with Neural

learning goals, and discipline of the course. Networks. Master Thesis. Tilburg University, Netherlands.

[6] Ramin Ghorbani and Rouzbeh Ghousi. 2020. Comparing different

Since this is a preliminary pilot analysis, we considered resampling methods in predicting students’ performance using machine

additional opportunities to improve our research in the future. learning techniques. IEEE access, 8, 67899-67911.

We could enhance this study by utilizing additional methods, [7] Judith M. Harackiewicz, Kenneth E. Barron, John M. Tauer and Andrew

J. Elliot. 2002. Predicting success in college: A longitudinal study of

such as cluster analysis, network analysis, and structural equation achievement goals and ability measures as predictors of interest and

modeling, along with techniques used to make predictions like performance from freshman year through graduation. Journal of

educational psychology, 94(3), 562.

data mining [11, 1], neural networks [5], or decision trees [14]. [8] Shaobo Huang and Ning Fang 2013. Predicting student academic

Furthermore, we could enrich the model by increasing the performance in an engineering dynamics course: A comparison of four

number of observations and the number of courses included in types of predictive mathematical models. Computers & Education, 61,

133-145.

the analysis. https://www.sciencedirect.com/science/article/abs/pii/S03601315120021

To capture more subtleties in these relationships, we could 02

[9] Kitsadaporn Jantakun, Thiti Jantakun, and Thada Jantakoon. 2022. The

potentially build and test the model's performance with architecture of system for predicting student performance based on data

additional input variables such as information about general science approaches (SPPS-DSA architecture). International Journal of

Information and Education Technology, 12(8), 778-785.

knowledge, broad outlook, ambitions, and psychological [10] René F. Kizilcec, and Hansol Lee. 2022. Algorithmic fairness in

characteristics. Additionally, we could compare different cohorts education. In Holmes, W., Porayska-Pomsta, K. (Eds.), The ethics of

of students, the semester, the study program and track, forms of artificial intelligence in education, Routledge. p. 174-202.

[11] Abdallah Namoun and Abdullah Alshanqiti. 2020. Predicting student

study (full-time, part-time), types of study programs performance using data mining and learning analytics techniques: A

(undergraduate, postgraduate, higher vocational program), and systematic literature review. Applied Sciences, 11(1), 237.

[12] OECD. 2019A. Skills Matter: Additional Results from the Survey of Adult

study modes (classroom, blended, online). Skills, OECD Skills Studies, OECD Publishing, Paris,

https://doi.org/10.1787/1f029d8f-en.

[13] OECD. 2019B. The Survey of Adult Skills: Reader’s Companion, Third

Edition, OECD Skills Studies, OECD Publishing, Paris,



REFERENCES https://doi.org/10.1787/f70238c7-en [14] Harikumar Pallathadka, Alex Wenda, Edwin Ramirez- Asís, Maximiliano [1] Balqis Albreiki, Nazar Zaki and Hany Alashwal.2021. A systematic Asís-López, Judith Flores-Albornoz and Khongdet Phasinam. 2023. literature review of student’performance prediction using machine Classification and prediction of student performance data using various learning techniques. Education Sciences, 11 (9), 552. machine learning algorithms. Materials today: proceedings, 80, 3782-[2] Laura Fink and Bojan Cestnik. 2023. Reinforcing digital skills: a 3785. quantitative analysis of digital skills development and its impact on higher [15] David M. W. Powers. 2011. "Evaluation: From Precision, Recall and F-educational outcomes. In Szymanska, K. (ed.), Janczewski, R. A. (ed.). Measure to ROC, Informedness, Markedness & Correlation". Journal of Problemy i wyzwania ekonomii i zarządzania w XXI wieku : wybrane Machine Learning Technologies, 2 (1): 37–63. aspekty . Poznań: Grupa Wydawnicza FNCE. P. 263-274. [16] Mustafa Yağcı. 2022. Educational data mining: prediction of students' [3] Laura Fink and Jaka Vadnjal. 2024. Generic and course-specific academic performance using machine learning algorithms. Smart competences in comparison. In Fošner, A. (ed.). ABSRC Ljubljana: Learning Environments, 9 (1), 11. conference proceedings, GEA College - Faculty of Entrepreneurship, p.

10.



Linking the Normative and the Descriptive: Bounded



Epistemic Rationality



Nastja Tomat†

Department of Philosophy

Faculty of Arts

University of Ljubljana

Ljubljana, Slovenia

nastja.tomat@ff.uni-lj.si



Abstract epistemology still often relies on idealized models of human cognizers [8], with the consequence that it frequently imposes

Epistemic rationality is a type of rationality directed towards epistemic norms such as logical omniscience, consistency

cognitive or epistemic goals, such as true beliefs, knowledge, or between beliefs, and immediate updating of beliefs by

understanding. Epistemology is primarily concerned with conditionalization [9]. normative questions about how one should form and update



beliefs, reason and inquire to be rational; on the other hand, Philosophy, including epistemology, is predominantly

empirical disciplines, such as psychology, investigate how concerned with the normative questions about justification,

inquiries and belief formation occur in real life. The question rationality and other epistemic appraisals of our cognitive

arises as to what the relationship between the normative and the activities and doxastic states, while empirical disciplines, such as

descriptive in the study of epistemic rationality should be. This psychology, empirically investigate how human cognition,

paper proposes a notion of bounded epistemic rationality as a inquiries and belief formation occur in everyday life. With

hybrid, non-ideal concept that encompasses both normative and normative theories on the one hand and empirical research on the

descriptive elements. Drawing upon Herbert Simon’s bounded other, we are faced with the question of the relationship between

rationality and Robin McKenna’s non-ideal epistemology, the two approaches towards studying rationality. bounded epistemic rationality is characterized by requiring



satisficing instead of maximizing; acknowledging our cognitive, The aim of the paper is to propose a concept of bounded

environmental, and practical limitations; its ecological nature; epistemic rationality as a hybrid notion that may help us bridge

and its focus on the process of inquiry. As such, bounded the gap between the normative and the descriptive. By adopting

epistemic rationality is a good starting point for proposing a concept that is – to some extent – grounded on empirical data

epistemic advice that is achievable for real cognizers and helps about human cognition but does not dispose with the normative

them improve their epistemic position. questions about epistemically good cognition, we can propose

epistemic norms and epistemic advice that are achievable for real

human cognizers and can help them improve their epistemic

1 Introduction

situation.

Epistemic rationality is one of the main topics of epistemology.

It refers to epistemic attitudes, states, and processes [1], mainly

focusing on rationality of beliefs, and is directed towards 2 Normative and descriptive theories of

reaching cognitive or epistemic goals, such as true beliefs, rationality knowledge, or understanding [2, 3, 4]. One of the central tasks of

Philosophical understanding of rationality is deeply intertwined

epistemology has been to propose epistemic norms about how

with the notion of normativity. There are different views on how

one should form, update and revise beliefs to be rational.

to define and justify epistemic normativity and which epistemic

Although it is acknowledged that humans are not ideal agents –

norms we should endorse. We can understand rationality as a

there is ample empirical evidence, gathered by disciplines as

system of rules or requirements: it requires from us, for example,

cognitive psychology, showing that we are limited by our

not to hold contradictory beliefs, to draw a conclusion by modus

cognitive architecture and the nature of cognitive processes, such

ponens [10], to have deductive closure [11], or to follow rules of

as computational power and speed, predictive abilities, working

logic, probability and decision theory [12]. Rationality is thus

memory and attention [5, 6, 7] – traditional analytic

normative in the sense of employing certain norms and rules

according to which we can judge correctness of a belief [13].

Permission to make digital or hard copies of part or all of this work for personal or Nevertheless, a genuine normative question of rationality

classroom use is granted without fee provided that copies are not made or distributed requires us to determine if those rules or requirements are

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must necessarily accompanied by a reason to conform to them, or, in

be honored. For all other uses, contact the owner/author(s). other words, if we ought to conform to them [13, 14]. Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.cog.11



In addition to the debate about genuine normativity of defining conditions for epistemic ideals such as justification and

rationality, there is an ongoing discussion about how epistemic knowledge and engaging in “S knows P iff …” kind of

norms or principles should be formulated and what they should epistemology, McKenna claims that we should engage in non-

prescribe. According to Engel, there are some conditions such ideal theorizing that is informed by empirical literature on human

principles should satisfy to be genuinely normative: they should cognition, knowledge-producing institutions and epistemic

have normative force, a potential to regulate or direct our inquiry environment. While McKenna claims that descriptive questions

and beliefs, and normative freedom – a possibility to be violated. should be a starting point for answering normative questions, he

If we accept these conditions, many normative principles that are does not argue for a strong form of naturalism or for the

often employed are not adequate. A rule that says, for example, replacement of epistemology by empirical science, but merely

that one should not believe p and not p, tells us something about suggests that there should be a closer connection between

a characteristic of a rational belief, but gives us no guidance on epistemology and empirical disciplines than is currently the case

how to achieve it [11]. Such rules are more of a description of a [8]. The norms that non-ideal epistemology proposes would

belief or believer in ideal conditions than genuine normative therefore be norms of inquiry that help agents determine which

principles. A similar point is put forward by Robin McKenna in problems are important to inquire about in the first place; how to

his book Non-ideal epistemology [8]. He claims that mainstream collect, assess and evaluate evidence; what to do when they are

epistemology mostly proposes epistemic norms based on various presented with conflicting information; how to identify

types of idealizations, for example about cognitive capacities of trustworthy and reliable sources of information; and when they

epistemic agents and the nature of epistemic environment. He gathered enough evidence to terminate an inquiry and form a

calls such an approach to epistemological theorizing ideal belief. Instead of a norm stating something in terms of “a belief

epistemology and contrasts it with non-ideal epistemology which about whether anthropogenic climate change is real is rational if

tries to avoid such idealizations. The issue with the norms it is achieved in a reliable manner and responsive to the available

proposed by ideal epistemology is that they are too detached evidence”, non-ideal epistemology would propose norms

from real world issues, too demanding and unachievable for real specifying what is an epistemically good manner in which

human cognizers. Another, even more important issue is that they ordinary laypeople should gather evidence about climate change,

provide bad epistemic advice: if we try to achieve or approximate how to identify genuine experts and how to recognize good

proposed ideals and norms, we will often worsen rather than evidence. Such norms would require inquiring in a way that is

improve our epistemic situation. McKenna uses an example of possible for ordinary people – would not, for example, require

the ideal of objectivity in scientific inquiry: trying to achieve enormous amount of time and philosophical understanding of the

objectivity as detachment – in a sense that scientists are not concept of evidence – and would be based on empirical data on

personally invested and interested in the topic of inquiry and try which ways are effective for gaining true beliefs about climate

to detach research process from non-cognitive values – leads to change – for example, relying on science marketing strategies [8].

worse, not better, scientific inquiry. McKenna draws on Such norms or principles would satisfy Engel’s conditions and

Elizabeth Anderson’s work on value judgements in science. She could thus be considered as genuinely normative.

argues that researchers’ background assumptions and values

influence all the stages of research process – they partly

determine how we frame the research questions, conceive of the 3 Bounded rationality

object of inquiry, what data we collect, how we analyze and Although many authors who investigate rationality or epistemic

interpret them. A large portion of empirical research in social norms explicitly acknowledge that humans are limited agents and

science investigates evaluative questions that are related to well- that our boundaries should put a constraint on epistemic norms,

being of individuals, social groups or society at large, and science only a few philosophers have drawn on the notion of bounded

that is legitimately guided by certain, for example feminist values, rationality. Bounded rationality was introduced by political

could be more fruitful and more likely lead to desired epistemic scientist Herbert A. Simon and has importantly influenced many

goals. Instead of trying to be attain an ideal of objectivity, disciplines investigating rationality, such as psychology and

scientific inquiry should be informed by the right values [15]. A economics. Simon argued that global, idealized theories of

similar argument can be made for our everyday inquiries: if we, rationality should be replaced with a notion of rationality that is

for example, always aim to reason in accordance with a norm compatible with cognitive capacities of the subjects and the

proposed by ideal theory, such as logic and probability theory, or features of the environment in which they are embedded. As our

trying to think in intellectually autonomous way instead of cognitive capacities, for example predictive and computational

relying on experts, this will likely lead to worse epistemic capabilities, working memory and attention, are limited, human

outcomes than using less complex, heuristic processes or form a rationality can be only an approximation of an ideal rationality

belief according to the consensus of the experts [16]. This means that is assumed in models of decision theory. If we want to

that ideal theory is failing as a normative theory because its comprehend human rationality, we should not focus only on

prescriptions often do not help us achieve our epistemic goals, internal characteristics of human cognition, but also on the

such as obtaining true beliefs, knowledge or understanding, and structure of the environment. Simon illustrated this with a

cannot serve as regulative ideals. For this reason, the ideal metaphor of scissors: “Human rational behaviour (and the

approach should in certain situations be replaced by a non-ideal rational behaviour of all physical symbol systems) is shaped by

one [8]. Both Engel and McKenna emphasize that an important a scissors whose two blades are the structure of task

feature of epistemic norms is their potential for guidance, for environments and the computational capabilities of the actor.”

improving our inquiries, reasoning, and forming beliefs that are [17, p.7]. Simon argued that human rationality was satisficing,

in some way epistemically better. Instead on focusing on not optimizing – meaning that humans do not seek for best



possible solutions of a problem or best possible outcome of a beliefs that are not true (in a sense of a truth requirement usually

decision situation, but for solutions that are merely good enough imposed by veritism), but are approximations, simplifications, or

– and he urged to dispose of the notion of optimization as a generalizations, are more conducive to reaching a wide array of

criterion for rationality. He also emphasized that bounded other epistemic goals and desiderata that are perhaps even more

rationality is procedural, meaning that it does not focus solely on valuable than truth, such as in-depth understanding of

the outcomes, but also on the process leading to them; an agent phenomena [23, 24]. The second option is to introduce a non-

is therefore rational if her behavior stems from an appropriate epistemic criterion for “good enough.” A belief is rational

process of deliberation [17-21]. enough if it helps us select appropriate actions for achieving

some other, non-epistemic goal that we intrinsically value; in this

case, a belief is good enough if it has instrumental value. The

4 Bounded epistemic rationality third option is that “good enough” is partly determined by

According to Sturm [22], philosophical aspects of bounded pragmatic criteria, but the goal remains epistemic. This is in line

rationality have not yet been systematically investigated; with the thesis of pragmatic encroachment which claims that

nevertheless, the role of bounded rationality in epistemology has epistemic status of a belief is not determined solely by epistemic,

recently been explored by David Thorstad [7]. He describes five but also by pragmatic factors. A certain belief may be considered

characteristics of bounded rationality as a paradigm, the first one good enough if, for example, the consequences of the belief

being that bounds are important. As opposed to practical being false are not vast.

philosophy where it is universally acknowledged that our

physical limitations put constraints on the norms of rational Second, bounded epistemic rationality acknowledges that we

action, this is not necessarily the case for epistemic rationality. are bounded by our cognitive capacities, the nature of the

Thorstad claims that bounds are equally important for our environment in which we operate, and by practical

understanding of rational cognition than of rational action and considerations of our daily lives. It considers that we have limited

that we should be normatively required to perform only those processing power, attention span, working memory, predictive

cognitive operations that we are capable of. Secondly, theories abilities and so on and employs ought-implies-can principle of

of rationality should consider not only the final beliefs and other normativity: things that are normatively required from cognizers

doxastic states, but also the processes that led to them, which is are only those which they are in principle capable of executing.

directly derived from Simon’s notion on procedural nature of Furthermore, it considers the features of our epistemic

bounded rationality. The third and fourth characteristics refer to environment, especially the nature and structure of available

the claim that rationality is not bound only by our cognition, but information. Levy [25], for example, speaks of so-called polluted

also by environmental factors and that the use of rules of thumb epistemic environments, which consist of a large portion of

or heuristics can be more rational than using more complex misinformation and where various individuals and institutions

reasoning strategies. Drawing on the work of Gerd Gigerenzer imitate the criteria of expertise, making it difficult for laypeople

and ecological rationality [6, 16], Thorstad claims that heuristics to identify reliable sources of information and genuine expertise.

may in many situations or environments provide more accurate In such environments, false beliefs cannot be attributed primarily

predictions than other, more sophisticated strategies. Finally, to the lack of epistemic virtue or irrationality of a cognizer but

bounded rationality is compatible with a so-called programme of must be understood in the context of epistemic environment.

vindicatory epistemology, which states that what we usually Finally, bounded epistemic rationality considers that we have

consider as a violation of rationality norms is a consequence of a limited time and cognitive resources that we can devote to a

deliberation process that is merely boundedly rational. Although certain task. Our inquiries do not happen in a bubble that

we do not comply with traditional epistemic norms as coherence detaches us from our practical considerations – in everyday life,

and deductive closure, we are often inquiring and reasoning in we cannot afford to infinitely inquire about a certain topic, even

the most rational way possible considering our limitations. if it is highly relevant and interesting for us. Bounded epistemic

rationality does not require us to inquire and form beliefs in a

Drawing on Thorstad’s work, I propose and expand on way that would demand postponing all other activities in life.

several characteristics I believe should be incorporated in the Acknowledging that practical factors should to some extent play

account of bounded epistemic rationality. First, bounded a role in epistemic requirements is compatible with a view put

epistemic rationality is distinctively epistemic in a sense that it is forward by Bishop and Trout [26, 27]. In their theory of strategic

directed towards cognitive or epistemic goals, regardless of reliabilism they urge that epistemological theories should include

which specific goal we are committed to – having true beliefs both epistemic and pragmatic factors, and they see epistemically

and not having false beliefs, making accurate predictions, good reasoning as “reliable, cost-effective, and focused on

gaining knowledge or understanding. At the same time, bounded significant problems” [26, p. 106].

epistemic rationality does not require optimal solutions, but

solutions that are merely good enough - it doesn’t require from Third, bounded epistemic rationality is not defined by

cognizers that their predictions are a hundred percent accurate or adherence to a rigid system of highly demanding, idealized rules

that they possess all and only true beliefs about trivial topics that or requirements, but by a fit between the strategy and the

are not relevant to them. Nevertheless, an account of bounded environment. Therefore, various strategies, from complex

epistemic rationality will need to provide criteria for how to reasoning to simple heuristics, can be rational as long as they are

decide if a belief or a prediction is epistemically good enough – conducive to certain epistemic goals; for the moment, I leave

be it true, accurate or rational enough. I believe this can be done open whether this should be truth, prediction, knowledge, or

in one of three ways. The first option is to claim that by acquiring understanding. Bounded epistemic rationality is thus



consequentialist, as it promotes a form of cognitive success [28], epistemic ones, but in the sense that they require epistemic goals

and ecological, as it emphasizes the fit between a strategy and that are achievable for real human cognizers, and require

the task [6, 16]. inquiries that are not too costly in terms of cognitive resources

and time.

Fourth, bounded epistemic rationality does not focus on the

final doxastic states, but on the process of inquiry. This is

compatible with a so-called zetetic turn in epistemology: in 5 Conclusion

recent years, epistemologists have started to move away from Bounded epistemic rationality is a hybrid concept that

identifying conditions for knowledge and justification towards includes both normative and descriptive elements. It aims to

the questions about what good inquiry should look like – for avoid idealizations of epistemic agents and their environment

example, when to start and stop inquiring and how to collect and and to acknowledge the practical limits of our daily lives. Being

evaluate evidence [29, 30]. Focusing on the process of inquiry a non-ideal concept that relies on empirical data about human

has more potential for providing epistemic advice than focusing cognition and our epistemic environment, it has the potential to

solely on the descriptions of epistemic ideals, such as knowledge. suggest norms that serve as epistemic advice and help us achieve

Although describing the conditions for knowledge and our epistemic goals.

justification are crucial parts of epistemology, combining this

project with a program of inquiry epistemology could be more References fruitful for providing epistemic guidance helping inquirers in

achieving epistemic goals. A notion of bounded epistemic [1] Markus Knauff and Wolfgang Spohn (Eds.). 2021. The Handbook of

rationality is therefore compatible with a project of ameliorative Rationality. The MIT Press.

or regulative epistemology [27, 31]. As a non-ideal concept that https://doi.org/10.7551/mitpress/11252.001.0001

[2] Marian David. 2001. Truth as the Epistemic Goal. In Knowledge, Truth,

considers real-life characteristics of our cognition and epistemic and Duty, M. Steup (ed.). Oxford University Press, 151–169.

environment, it can give advice that is applicable to ordinary [3] Stephen Grimm. 2012. The Value of Understanding. Philosophy Compass

inquirers – for example, what to do when faced with 7, 2 (February 2012), 103–117. https://doi.org/10.1111/j.1747-

9991.2011.00460.x

contradictory evidence; when should we stop gathering evidence [4] Jonathan L. Kvanvig. 2013. Truth is Not the Primary Epistemic Goal. In

and form a belief; when epistemic environment is so polluted that Contemporary Debates in Epistemology, M. Steup and J. Turri (eds.).

Blackwell, 285–295.

it may be rational to suspend judgement; how to judge which [5] Alvin I. Goldman. 1986. Epistemology and Cognition. Harvard University

sources are reliable and trustworthy and so on. Press, Cambridge.

[6] Gerd Gigerenzer. 2008. Rationality for mortals: How people cope with

uncertainty. Oxford University Press, New York, NY.



4.1 Norms of bounded epistemic rationality [7] David Thorstad. 2024. Why Bounded Rationality (in Epistemology)? Philosophy and Phenomenological Research 108, 2 (2024), 396–413.

A crucial question regarding the norms of bounded epistemic https://doi.org/10.1111/phpr.12978

rationality is in what way they should relate to empirical science, [8] Robin McKenna. 2023. Non-Ideal Epistemology. Oxford University Press,

specifically psychology. Norms of rationality cannot be directly Oxford, NY. [9] Jennifer Rose Carr. 2022. Why Ideal Epistemology? Mind 131, 524



derived from empirical data, as this would mean committing is- (December 2022), 1131–1162. https://doi.org/10.1093/mind/fzab023 [10] [John Broome. 1999. Normative Requirements. Ratio 12, 4 (1999), 398–

ought fallacy [32] – we cannot infer how one ought to reason 419. https://doi.org/10.1111/1467-9329.00101

from descriptive premises about how we do reason. Nonetheless, [11] Pascal Engel. Epistemic norms. In Routledge Companion to Epistemology,

psychological data on human cognition can at least serve as S. Bernecker and D. Pritchard (eds.), Routledge, 47–58.

[12] Edward Stein. 1997. Without Good Reason: The Rationality Debate in

constraints showing us what is realistic to expect from cognizers. Philosophy and Cognitive Science. Oxford University Press, Oxford, NY.

[13] John Broome. Is rationality normative? Disputatio 2, 23 (November 2007),



Another question concerning the norms of bounded epistemic 161–178. https://doi.org/10.2478/disp-2007-0008

[14] Benjamin Kiesewetter. 2017. The normativity of rationality. Oxford

rationality relates to the notion of adaptability. Since it is an University Press, Oxford.

epistemic notion, bounded epistemic rationality must be directed [15] Elizabeth Anderson. 2004. Uses of Value Judgments in Science: A

General Argument, with Lessons From a Case Study of Feminist Research

towards epistemic goals, but the question arises whether on Divorce. Hypatia 19, 1 (Winter 2004), 1–24.

epistemic goals can in any way be connected to adaptive or https://doi.org/10.2979/hyp.2004.19.1.1

[16] Gerd Gigerenzer. 2008. Why Heuristics Work. Perspect Psychol Sci 3, 1

pragmatic goals. We might consider a person, belief, or process (January 2008), 20–29. https://doi.org/10.1111/j.1745-6916.2008.00058.x



to be boundedly epistemically rational if it leads to an epistemic [17] Herbert A. Simon. 1956. Rational choice and the structure of the environment. Psychological Review 63, 2 (1956), 129–138. goal while functioning as an adaptive response to the https://doi.org/10.1037/h0042769

environment. [18] Herbert A. Simon. 1955. A Behavioral Model of Rational Choice. The

Quarterly Journal of Economics 69, 1 (1955), 99–118.

https://doi.org/10.2307/1884852

Achieving epistemic goals often helps us to respond [19] Herbert A. Simon. 1990. Invariants of human behavior. Annual Review of



efficiently to the environment and therefore has an adaptive Psychology 41, (1990), 1–19. https://doi.org/10.1146/annurev.ps.41.020190.000245 function. Even though the intrinsic value of truth may be [20] Herbert A. Simon. 1992. What is an “explanation” of behavior?

debatable, it is hard to deny that truth has at least an instrumental Psychological Science 3, 3 (1992), 150–161.

value. Nevertheless, there are many situations in which epistemic https://doi.org/10.1111/j.1467-9280.1992.tb00017.x

[21] Herbert A Simon. From substantive to procedural rationality. In 25 Years

and adaptive goals may diverge; for example, if someone devotes of Economic Theory, T. J. Kastelein, S. K. Kuipers, W. A. Nijenhuis and

all their time and resources to researching a complex topic of G. R. Wagenaar (eds.), Springer, Boston, MA, 65-68.

their interest and neglects all other activities in life, we cannot https://doi.org/10.1007/978-1-4613-4367-7_6. [22] Thomas Sturm. 2020. Towards a critical naturalism about bounded



consider this adaptive. The norms of bounded epistemic rationality. In Routledge Handbook of Bounded Rationality. Routledge, 73–90. rationality should therefore include a notion of adaptability – but [23] Catherine Z. Elgin. 2004. True Enough. Philosophical Issues 14, (2004),

not in the sense that adaptive or pragmatic goals can override 113–131.



[24] atherine Z. Elgin. 2017. True Enough. MIT Press, Cambridge. [28] Gerhard Schurz and Ralph Hertwig. 2019. Cognitive success: A [25] Neil Levy. 2021. Bad Beliefs: Why They Happen to Good People. Oxford consequentialist account of rationality in cognition. Topics in Cognitive

University Press, Oxford, UK. Science 11, 1 (2019), 7–36. https://doi.org/10.1111/tops.12410

https://doi.org/10.1093/oso/9780192895325.001.0001 [29] Jane Friedman. 2020. The Epistemic and the Zetetic. Philosophical

[26] Michael Bishop and J.D. Trout. 2016. Epistemology for Real People. In A Review 129, 4 (2020), 501–536. https://doi.org/10.1215/00318108-

Companion to Applied Philosophy, K. Lippert-Rasmussen, K. Brownlee 8540918

and David Coady (eds.). John Wiley & Sons, Ltd, 103–119. [30] Eliran Haziza. 2023. Norms of Inquiry. Philosophy Compass 18, 12 https://doi.org/10.1002/9781118869109.ch8 (2023), e12952. https://doi.org/10.1111/phc3.12952

[27] Michael A. Bishop and J. D. Trout. 2004. Epistemology and the [31] [Nathan Ballantyne. 2019. Knowing Our Limits. Oxford University Press,

Psychology of Human Judgment. OUP USA, New York. Oxford, New York.

[32] David F. Norton and Mary J. Norton (Eds.). 2000. A Treatise of Human

Nature. Oxford: Clarendon Press.





Exploring Human Perception Using Virtual Reality




Katja Zibrek

katja.zibrek@inria.f r

Inria centre at Rennes University

France



Abstract exploring the potential of virtual immersive technology to en-

hance human abilities. Its aim is to understand how the virtual

Immersive technologies have seen a great expansion in the last

experiences could create new and faster learning procedures, aid

decade and researchers from several disciplines have focused on

in physical and mental health rehabilitation by broadening the

exploring virtual reality and the way it can affect human percep-

scope of what is possible in the physical reality, and perhaps even

tion. Virtual reality is a unique medium which has the ability to

open up new avenues for human experience, to which we do not

transfer the user from the physical environment to a digitally cre-

yet know the limits of.

ated illusion of space, events and interactions which mimic real

The aim of this paper is to present some examples of research

life. In this paper, some basic concepts of perception in virtual

in VR, dedicated to the exploration of human perception from

reality are introduced, followed by the summary of our research

both the computer and social science perspective. In order to bet-

which primarily focused on the perception of virtual agents. Our

ter place the research topic, general concepts of VR are defined in

method is based on the concept of interpersonal distance when

the first part of this article. In the second part, some of our past

people meet in social settings and where the distances they keep

studies using VR as a tool to measure human behaviour are pre-

between each other signal the nature of their relationship. We

sented. Our work mainly revolves around virtual agents, digital

studied these distances to evaluate realism, attractiveness and

representations of humans, who populate a virtual scene in VR

even personality traits of virtual agents in virtual reality. We

and may have simple or complex algorithms to simulate natural

discuss how our results can give valuable insights into the hu-

behaviour. At the end of the article, some of the implications of

man mind and how we can use this knowledge for training and

our research and how it can help us to understand human mind

rehabilitation applications in virtual reality.

are discussed.



Keywords

2 Virtual Reality

virtual reality, perception, virtual agents, proximity

While research in Virtual Reality (VR) goes back as early as 1970s,

it has witnessed a surge in recent years due to the development

1 Introduction of relatively low-cost and ergonomic devices, as well as more

Virtual Reality (VR) is an immersive environment where peo- effective and powerful graphics rendering technology. Enter-

ple can experience scenarios which mimic physical reality. They tainment industry began to launch VR-specific games (e.g., Beat

can also be engaged in virtual interaction with real people, pre- Saber, Half-Life: Alyx), social platforms such as Metaverse [16]

sented in VR as avatars, or computer-driven representations of are using VR for interaction in online virtual environments, some

real humans. The immersive experience and interaction is a fairly organisations use VR to raise social and political awareness [17,

recent phenomenon, providing a plethora of research challenges 19]. Other immersive technologies, such as Augmented Reality

to solve and questions to explore. For example, how do we create and Mixed Reality, began their debut to the broad market around

believable virtual environments which will facilitate human in- the same time as well. Today, they are commonly addressed with

teraction and what do peoples’ responses to these environments the unifying term Extended Reality (XR).

teach us about our mental processes? VR, however, is unlike any other immersive system. The goal

There are primarily two types of research domains who use VR of VR is to completely disconnect the user from the physical

in their research. The first, social science, is interested in VR as a reality which is different to the aim of other XR systems which

highly controllable replica of a real world with the ability to create do so only partially. Complex VR systems include head-mounted

“ecologically valid experience”, i.e. human response which is close display with positional tracking to create the feeling that the

to a real-life response, in order to investigate human cognition virtual world is surrounding and moving with the user, haptic

and transfer of knowledge from virtual to physical reality. The stimulation, spatial sound, representation of the user’s body in

second, computer science, is more interested in keeping human the environment, etc. This complexity of the system increases

evaluation in the loop to optimise computational power and system immersion [27, 20]. If the immersion is high, the user will

enhance virtual environments. While the primary goal of the have the feeling of “existing” in the virtual space, an experience

second group is not to explore the human mind, it is an inevitable known as “presence” or “place illusion, plausibility” [25, 27].

side effect of their scientific endeavour. Other illusions can also be created, most notably the illusion of

There is, however, a third group of researchers. This group social presence (a virtual human appears to be alive) [3, 1] and

presents a bridge between the social and computer science by embodiment (our virtual body is perceived as our own body).



Permission to make digital or hard copies of all or part of this work for personal 2.1 Presence or classroom use is granted without fee provided that copies are not made or

Presence in VR is the experience of an actual place and the feel-

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this ing that the virtual events are really happening. Lombard and

work must be honored. For all other uses, contact the owner /author(s).

Ditton elegantly described presence as “the perceptual illusion of

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

nonmediation” [14]. The concept of presence was most famously

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.cog.12 investigated with the so called “virtual pit” experiment [30, 15,



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Zibrek



6], where the participants were standing on the top of a narrow Some researchers also explore the idea of co-sharing of a virtual

ledge in a virtual room, looking down to another room through a body, where the agency over one avatar can be shared between

wide gap in the floor. The experimenters [6] could detect changes two users, e.g. one user possesses the left arm an the other the

in galvanic skin response, shaking and loss of balance and the right, as well as different percentages of possession of the full

participants reported intense emotional reactions of fear as if avatar body [5].

they were in real danger of falling.

Several other types of environments and scenarios were used 3 Using VR to study the perception of virtual

by the researchers to elicit emotions (see [4] for an overview ). agents

Apart from research on the concept of presence, training and

As previously mentioned, agents are computer-driven represen-

rehabilitation applications were developed in VR to exploit this

tations of humans who can possess simple or complex behaviour

ability of a virtual environment to induce realistic responses in

characteristics. Researchers strive to understand how to increase

people. Related to the pit experiment, for example, assessment

their naturalness, appeal and interactive abilities and the above

of construction workers for their postural stability at different

mentioned illusions, presence, social presence and embodiment,

heights have been developed [7], and in psychotherapy, similar

play an important role in this endeavour. Presence increases the

environments are used to gradually expose acrophobia patients

believeability of the scenario with the agent, social presence in-

to increasing levels of height [13].

fluences the user to exhibit social behaviour, and embodiment

gives us the opportunity to measure user’s body position and

2.2 Social Presence movement in relation to the agent in the virtual space.

Social presence is the illusion of being present with another in

a virtual environment, or simply the “sense of being with an- 3.1 Proximity

other” [3]. The definition is broad and sometimes other terms are

Interpersonal distance or proximity is the minimum distance that

used to define similar or related constructs, such as telepresence,

people maintain between one another when involved in social

co-presence [21], and plausibility [27]. In VR, the term social

interaction. The measure comes from proxemics described by

presence is more commonly used to denote the level of believ-

Hall [8] who introduced it as an indicator or comfort and familiar-

ability of a co-located virtual human, especially if this virtual

ity with other people. Many factors influence how close we will

human is computer-driven (agent) and we wish to evaluate its

approach another: familiarity, culture, gender, personality, etc.

naturalness. Social presence with other users in VR (avatars) can

Closer distances reveal trustworthiness and comfort, while fur-

also be investigated to evaluate aspects of the system and the

ther distances can signal mistrust, discomfort or fear of the other.

environment, such as ability to represent users with emotional

In VR, proximity has been used to explore the social influence of

expressions or communication channels (sound, text interface),

virtual humans [1].

which enhances the collaboration aspect between the users in

The proximity measure can be expressed simply as the Eu-

the VR environment.

clidean distance of the current camera (user) position and the

Depending on the definition, different ways of measuring so-

position of central mass of the virtual character in the virtual

cial presence exist. Researchers who agree that social presence is

space. It is important that the user is navigating the environment

a cognitive construct will typically use questionnaires for evalua-

by natural walking in order to preserve distances comparable to

tion [2], while other researchers prefer to use indirect measures,

real-life interactions. The proximity tasks can vary. In the passive

such as eye-tracking and psychophysiology [27], signs of social

approach, the user is approached by an agent and is asked to

influence [28], or task-related behaviour [26, 18].

press a button at the precise time they begin to feel uncomfort-

able with the agent’s proximity (see image , Figure 1). With the a

2.3 Embodiment active approach, the user approaches an agent instead, typically

to complete a task, e.g., read the name tag on the agent’s chest. In

Embodiment or the Sense of Embodiment (SeO) is the feeling of

the avoidance task, the agent is an obstacle in the environment

possessing the virtual body in VR, which feels like it is “ours,” and

and the user avoids it to reach a goal. With active approaches,

moves according to our intentions [11]. This illusion is linked to

we can generate and analyse walking trajectory from the posi-

the virtual body in VR for which the movement is driven by the

tions of the user through time (see image , Figure 1) in terms b

user, wearing a tracking device while his HMD view is centered

of walking speed, minimum passing distance, average distance

at the eye-view of the head. If the user observes the movement

from the obstacle, etc. The avoidance behaviour between real

of his hands and body when immersed in VR, he can develop

and virtual humans has some differences: clearance distance for

a sensation that the body is actually his own. This illusion was

virtual agents is larger than real humans [24]. However, factors

first documented in real-life studies as the so called rubber hand

affecting the proximity were found to be generally similar to the

illusion [10]. The feeling of ownership of an artificial body can

ones in physical reality.

develop when receiving synchronous visual input and touch

sensation on both the virtual and real hand, with only the virtual

hand being visible to the user. The same effect can be reproduced 3.2 Previous Research

using proprioception (the user observes his virtual arm moving Some of our most relevant results using proximity are presented

as he is moving his real one). Not only is the SeO enabling a more in this section. The studies used primarily agents which were

immersive experience (presence is increased when embodiment highly realistic and had real human motion applied using high-

is added in VR), the SeO is a testament to the the importance of performance motion capture (Vicon) with 53 marker system to

the role of multimodal input in the embodied experience [11]. track the major joints and location of the body. The VR environ-

The illusion was also explored in creative applications, such as ments were created with Unreal Engine 4 or 5, and we primarily

giving the user a sense that they possess a part of a body which used HTC Vive with natural locomotion (participants could tra-

they actually do not have in reality, e.g., having a sixth finger [9]. verse the environment by walking) to immerse our participants

Perception in VR Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia





Figure 1: Examples of our experimental stimuli and measures of proximity: a) passive approach, where the virtual agent

approaches the user and signals by pressing a button at the precise moment they feel uncomfortable with the agent. The

Euclidean distance between the central mass of the agent and user’s head-mounted display is recorded as the value of

proximity; b) active approach, where the user (dark-grey character on the right of the image) circumvents the agent while

walking through the virtual environment. Multiple metrics can be derived, including passing distance, deviation point and

body adaptation (e.g., shoulder rotation).



in the virtual scenario. In all our studies, the participants also indicate that neurotic motion increases the proximity even in

possessed a virtual body. tight spaces and also affects the choice of metro exit where they

would be less likely to exit from a door which is obstructed by a

3.2.1 User agency. In Zibrek et al. [31], we were investigating

neurotic agent.

the affect of agency over a virtual character in VR. The users were

In our most recent work [23], we were interested if there

using the Vive controller to either trigger the character motion

is something specific in the motion pattern of neurotic motion

(avatar condition) or observe the character (agent condition).

which influences the proximity. We focused on the aspect of

Afterwards, users were asked to approach the character to find

motion predictability where we hypothesised that more unpre-

its name tag that was attached to his chest. The aim was to

dictable motion will increase the proximity distance in VR. We

test whether users will come closer to the agent they previously

designed an experiment, where participants were avoiding an

controlled as opposed to an agent who moved independently. The

moving obstacle in VR with varying motion characteristics in

lack of control over an agent could give the impression he has

terms of speed and predictability. We found that participants

the ability to have independent and unpredictable behaviour. The

exhibiting a tendency to maintain larger distances in scenarios

results showed that it was not the condition, but the subjectively

where obstacle speed was higher. Predictability had a lesser ef-

perceived agency (how much the user actually felt in control

fect than speed and became noticeable when the overall average

of the character) which reduced the proximity, revealing the

speed of the obstacle was lower. Future work will attempt to

importance of perceived agency as opposed to designed one.

implement this experiment by substituting the moving object

3.2.2 Gender and attractiveness. In human interaction and VR, with a virtual agent where we will systematically control its body people will keep different distances from each other depending motion predictability.

on their gender: males will stand further away from males and



closer to females. Our study [32] focused on proximity to virtual 4 Discussion walkers, where gender could be recognised from motion only,

The illusions of presence, social presence and embodiment show-

since previous studies using point-light displays found walking

case an amazing aspects of human perception. Firstly, they show

motion is rich in gender cues [12]. We were also interested to

us that in its very basis, the experience of reality or the feeling of

see if a more attractive motion would decrease the proximity.

being in a place is a multi-modal sensory experience. The feeling

We designed an experiment, where a virtual agent approached

of being with another can simply be induced with a visual pres-

the embodied participant. The agent animation was motion cap-

ence of a moving human character. Embodiment can be achieved

tured from several male and female actors and each motion was

with synchronising haptic/proprioceotive and visual signals.

displayed individually on the character. Participants used the

Second, our proximity studies showed that autonomous virtual

controller to stop the approaching agent when they felt it was

humans can exude social influence and affect peoples’ behaviour

uncomfortably close to them. Our results showed no difference

in VR. In our studies, we successfully implemented the measure

in proximity according to the gender of the character, however,

of proximity to study agent characteristics, such as attractiveness,

the gender of participants affected proximity (females had larger

gender, and even personality. However, VR gave us the ability

proximity distances to male users). We also found evidence that

to separate movement attraction from physical appearance [32],

greater attractiveness will decrease proximity. This was shown

as well as the ability to control the factor of appearance from

only by rating the attractiveness of the motion of the agent, show-

personality behaviour [22], for example.

ing the importance of body motion to infer information about

Furthermore, our latest work is studying the aspects of agent

other people.

animation to create perceptually appealing agents. This builds

3.2.3 Agent animation. We approached the perception of mo- upon the VR as tool to explore human perception but to also tion from the perspective of distinct movement patterns which create new elements of human experience which will, hopefully,

can be observed on people with neurotic and emotionally stable affect the implementation of these findings in new and unpre-

personality traits [22]. We designed an experiment in VR, using dictable ways. By understanding and controlling aspects of agent

a photo-realistic metro scenario, where we studied the avoidance motion and behaviour, we could anticipate the creation of ‘ap-

behaviour of participants when encountering these two types pealing agents’, who would be likable and comforting to the VR

of virtual characters in a constrained environment. Our results users and have the ability to improve the outcome of training



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Zibrek



and rehabilitation applications. One of the possible use cases is [13] Merel Krijn, Paul MG Emmelkamp, Roeline Biemond, Claudius de Wilde

de Ligny, Martijn J Schuemie, and Charles APG van der Mast. 2004. Treat-

building virtual therapists [29] who could guide the user through

ment of acrophobia in virtual reality: the role of immersion and presence.

techniques to improve mental health and help automatising the Behaviour research and therapy, 42, 2, 229–239.

rehabilitation aspect to enhance the accessibility of psychother- [14] Matthew Lombard and Theresa Ditton. 1997. At the heart of it all: the

concept of presence. , 3, 2, Journal of computer-mediated communication

apy.

JCMC321. doi: 10.1111/j.1083- 6101.1997.tb00072.x.

While the results of our studies are a testament to the flexibility [15] Michael Meehan, Brent Insko, Mary Whitton, and Frederick P Brooks Jr.

of our perceptual system, they also show the reliance of human 2002. Physiological measures of presence in stressful virtual environments.

Acm transactions on graphics (tog), 21, 3, 645–652.

perceptual mechanism to build upon past experience in order

[16] Meta. 2024. Metaverse. Retrieved August 23, 2024 from https://about.meta.c

to process the artificial cues. The fact that the agents exhibit om/metaverse.

[17] Tanvi Misra. 2016. What it feels like to be forced from home. Retrieved

social influence is due to having experience with real humans

August 23, 2024 from https://www.bloomberg.com/news/articles/2016- 10- 1

and projecting that knowledge in order to predict the behaviour

4/how- virtual- reality- can- build- empathy- towards- ref ugees.

of artificial humans. An interesting observation in our studies, for [18] Maria Murcia-Lopez, Tara Collingwoode-Williams, William Steptoe, Raz

Schwartz, Timothy J Loving, and Mel Slater. 2020. Evaluating virtual reality

example, was that we never specifically instructed participants

experiences through participant choices. In 2020 IEEE Conference on Virtual

to avoid the virtual agent - they did this on their own accord, as . IEEE, 747–755. doi: 10.1109/VR46266.20 Reality and 3D User Interfaces (VR)

they would in the real world. 20.00098.

[19] Unated Nations. 2016. Virtual reality: creating humanitarian empathy. Re-

It is equally important to note, there are several limitations to

trieved August 23, 2024 from https://www.youtube.com/watch?v=vAEjX9

our approach. Proximity has multiple confounding factors related S8o2k.

[20] Niels Chr Nilsson, Rolf Nordahl, and Stefania Serafin. 2016. Immersion

to individual and cultural differences, ergonomics is still not

revisited: a review of existing definitions of immersion and their relation

optimal for VR systems, and creating truly believable behaviour Human technology to different theories of presence. , 12, 2, 108–134. doi:

of the virtual agents is incredibly challenging. While there is hope 0.17011/ht/urn.201611174652.

[21] Kristine L Nowak and Frank Biocca. 2003. The effect of the agency and

in the development of new, lighter and less obstructive HMDs, as

anthropomorphism on users’ sense of telepresence, copresence, and social

well as using AI to create more believable behaviour for agents, presence in virtual environments. Presence: Teleoperators & Virtual Environ-these endeavours are still in its infancy. And finally, covering , 12, 5, 481–494. ments

[22] Yuliya Patotskaya, Ludovic Hoyet, Anne-Hélène Olivier, Julien Pettré, and

all the range of studies related to the perception in VR and its

Katja Zibrek. 2023. Avoiding virtual humans in a constrained environment:

limitations is impossible to do in this paper, as the topic is broad exploration of novel behavioural measures. , 110, 162– Computers & Graphics

172. doi: 10.1016/j.cag.2023.01.001.

and cannot be sufficiently presented in a condensed manner. The

[23] Yuliya Patotskaya, Ludovic Hoyet, Katja Zibrek, and Julien Pettre. 2024.

purpose of this article was simply to give an example of how

Entropy and speed: effects of obstacle motion properties on avoidance

VR can be used for perceptual investigation, and perhaps spark behavior in virtual environment. In SAP’24: ACM Symposium on Applied

Perception 2024, 1–13. doi: 10.1145/3675231.3675236f f ..

interest for this topic in the scientific community.

[24] Ferran Argelaguet Sanz, Anne-Hélène Olivier, Gerd Bruder, Julien Pettré,

and Anatole Lécuyer. 2015. Virtual proxemics: locomotion in the presence of

References obstacles in large immersive projection environments. In 2015 IEEE virtual

reality (vr). IEEE, 75–80. doi: 10.1109/VR.2015.7223327.

[1] Jeremy N Bailenson, Jim Blascovich, Andrew C Beall, and Jack M Loomis.

[25] Thomas B Sheridan. 1996. Further musings on the psychophysics of presence.

2001. Equilibrium theory revisited: mutual gaze and personal space in virtual

Presence: Teleoperators & Virtual Environments, 5, 2, 241–246.

environments. , 10, 6, 583–598. Presence: Teleoperators & Virtual Environments

[26] Richard Skarbez, Solene Neyret, Frederick P Brooks, Mel Slater, and Mary C

[2] Frank Biocca and Chad Harms. 2004. Internal consistency and reliability of

Whitton. 2017. A psychophysical experiment regarding components of the

the networked minds social presence measure. In Seventh Annual Interna-

plausibility illusion. , IEEE transactions on visualization and computer graphics

tional Workshop on Presence.

23, 4, 1369–1378. doi: 10.1109/TVCG.2017.2657158.

[3] Frank Biocca, Chad Harms, and Jenn Gregg. 2001. The networked minds

[27] Mel Slater. 2009. Place illusion and plausibility can lead to realistic behaviour

measure of social presence: pilot test of the factor structure and concurrent

in immersive virtual environments. Philosophical Transactions of the Royal

validity. In , 4th annual international workshop on presence, Philadelphia, PA

Society B: Biological Sciences, 364, 1535, 3549–3557. doi: 10.1098/rstb.2009.01

1–9.

38.

[4] Julia Diemer, Georg W Alpers, Henrik M Peperkorn, Youssef Shiban, and

[28] Mel Slater, Angus Antley, Adam Davison, David Swapp, Christoph Guger,

Andreas Mühlberger. 2015. The impact of perception and presence on emo-

Chris Barker, Nancy Pistrang, and Maria V Sanchez-Vives. 2006. A virtual

tional reactions: a review of research in virtual reality. , Frontiers in psychology

reprise of the stanley milgram obedience experiments. , 1, 1, e39. PloS one

6, 26.

doi: 10.1371/journal.pone.0000039.

[5] Rebecca Fribourg, Nami Ogawa, Ludovic Hoyet, Ferran Argelaguet, Takuji

[29] Mel Slater, Solène Neyret, Tania Johnston, Guillermo Iruretagoyena, Mercè

Narumi, Michitaka Hirose, and Anatole Lécuyer. 2020. Virtual co-embodiment:

Álvarez de la Campa Crespo, Miquel Alabèrnia-Segura, Bernhard Spanlang,

evaluation of the sense of agency while sharing the control of a virtual body

and Guillem Feixas. 2019. An experimental study of a virtual reality coun-

among two individuals. IEEE Transactions on Visualization and Computer

selling paradigm using embodied self-dialogue. , 9, 1, 10903. Scientific reports

Graphics, 27, 10, 4023–4038. doi: 10.1109/TVCG.2020.2999197.

doi: doi.org/10.1038/s41598- 019- 46877- 3.

[6] Henrique Galvan Debarba, Sidney Bovet, Roy Salomon, Olaf Blanke, Bruno

[30] Martin Usoh, Kevin Arthur, Mary C Whitton, Rui Bastos, Anthony Steed,

Herbelin, and Ronan Boulic. 2017. Characterizing first and third person

Mel Slater, and Frederick P Brooks Jr. 1999. Walking> walking-in-place>

viewpoints and their alternation for embodied interaction in virtual reality.

flying, in virtual environments. In Proceedings of the 26th annual conference

PloS one, 12, 12, e0190109.

on Computer graphics and interactive techniques, 359–364.

[7] Mahmoud Habibnezhad, Jay Puckett, Houtan Jebelli, Ali Karji, Mohammad

[31] Katja Zibrek, Elena Kokkinara, and Rachel McDonnell. 2017. Don’t stand

Sadra Fardhosseini, and Somayeh Asadi. 2020. Neurophysiological testing

so close to me: investigating the effect of control on the appeal of virtual

for assessing construction workers’ task performance at virtual height.

humans using immersion and a proximity-based behavioral task. In Pro-

Automation in Construction, 113, 103143.

ceedings of the ACM symposium on applied perception, 1–11. doi: 10.1145/31

[8] Edward T Hall et al. 1968. Proxemics [and comments and replies]. Current

19881.3119887.

anthropology, 9, 2/3, 83–108.

[32] Katja Zibrek, Benjamin Niay, Anne-Hélène Olivier, Ludovic Hoyet, Julien

[9] Ludovic Hoyet, Ferran Argelaguet, Corentin Nicole, and Anatole Lécuyer.

Pettre, and Rachel McDonnell. 2020. The effect of gender and attractiveness

2016. “wow! i have six fingers!”: would you accept structural changes of your

of motion on proximity in virtual reality. ACM Transactions on Applied

hand in vr? , 3, 27. doi: 10.3389/f robt.2016.00027. Frontiers in Robotics and AI

Perception (TAP), 17, 4, 1–15. doi: 10.1145/3419985.

[10] Marjolein PM Kammers, Frederique de Vignemont, Lennart Verhagen, and H

Chris Dijkerman. 2009. The rubber hand illusion in action. , Neuropsychologia

47, 1, 204–211.

[11] Konstantina Kilteni, Raphaela Groten, and Mel Slater. 2012. The sense of em-

bodiment in virtual reality. , Presence: Teleoperators and Virtual Environments

21, 4, 373–387. doi: 10.1162/PRES_a_00124.

[12] Lynn T Kozlowski and James E Cutting. 1977. Recognizing the sex of a

walker from a dynamic point-light display. , 21, Perception & psychophysics

575–580.





Vpliv generativne umetne inteligence na demokracijo




How Generative Artificial Intelligence Impacts Democracy



Lea Košmrlj Ivan Bratko

Pedagoška fakulteta Fakulteta za računalništvo in informatiko

Univerza v Ljubljani Univerza v Ljubljani

Slovenija Slovenija

lk72012@student.uni-lj.si bratko@fri.uni-lj.si



Povzetek automated propaganda and political manipulation, as well as causing an information overload. Based on a review of empirical

V luči skokovitega tehnološkega napredka generativne umetne studies that include large language models and deepfakes, the

inteligence v zadnjih nekaj letih se poleg prednosti, ki jih ta purpose of this article is to examine the potential extent of the

prinaša, pojavlja vse več opozoril o njenih pasteh, ki lahko harmful effects of generative artificial intelligence on democratic

predstavljajo resno tveganje za družbenopolitične in processes. We observe that empirical studies are few in number,

demokratične procese. Med negativnimi učinki generativne but lend support to the theoretical assumptions about the possible



širjenje dezinformacij ter škodljivih vsebin, omogočanje democratic societies. The main risks come from the participants' inability to distinguish synthetic content from human-generated umetne inteligence je najpogosteje izpostavljeno generiranje in threats that generative artificial intelligence can pose to



obsežnih dezinformacijskih kampanj, avtomatizirane content and the influence of synthetic content on their opinions propagande in politične manipulacije ter informacijsko on political figures or topics. Finally, we summarize proposals poplavljanje. Namen prispevka je na podlagi pregleda empiričnih for mitigating such risks, which include regulation, transparency

raziskav, ki vključujejo velike jezikovne modele in tehnologijo and accountability of developers, as well as awareness and digital

globokih ponaredkov, oceniti morebitne škodljive učinke literacy among users.

generativne umetne inteligence na demokratične procese.

Opažamo, da so empirične študije maloštevilne, a podpirajo Keywords

teoretske predpostavke o grožnjah, ki jih generativna umetna generative artificial intelligence, democracy, deepfakes, large

inteligenca lahko predstavlja za demokratične družbe. Pri tem language models, synthetic content gre izpostaviti predvsem neuspešnost udeležencev v

razločevanju sintetičnih vsebin od človeških in vpliv sintetičnih

vsebin na mnenja udeležencev in njihovo vrednotenje politične 1 Uvod



osebe ali tematike. Nazadnje povzemamo predloge za blaženje Izjemen tehnološki napredek umetnointeligenčnih sistemov je v tveganj, ki obsegajo regulacijo, transparentnost in odgovornost zadnjem času omogočil številne nove prelomne aplikacije in razvijalcev ter ozaveščanje in digitalno pismenost uporabnikov. njihov prodor v praktično vsa družbena tkiva. Vseeno se je danes



Ključne besede nadaljevanju UI) zagotavljala demokraciji, pogosto bolj prej kot o podpori, ki bi jo generativna umetna inteligenca (v

generativna umetna inteligenca, demokracija, globoki ponaredki, smiselno vprašati o njenem spodkopavanju demokratičnih

veliki jezikovni modeli, sintetične vsebine temeljev [18]. Vsekakor se tako teoretični razmisleki kot



Abstract empirične študije o vplivu generativne UI nagibajo predvsem v

to smer; poudarjajo tveganje bliskovitega generiranja in širjenja

Amid the rapid technological advancements in the field of dezinformacij, možnost zavajanja in manipulacije spletnih

generative artificial intelligence in recent years, there are, despite uporabnikov z dezinformacijskimi kampanjami in

its benefits, increasing warnings being put forward about its mikrotargetiranjem, ogrožanje političnih kampanj in volitev,



democratic processes. Among the most frequently mentioned sintetične vsebine [3, 16, 18, 19, 30, 33, 37]. Izpostavljajo pomen negative effects of generative artificial intelligence are the pitfalls, which could pose serious risks to sociopolitical and informacijsko poplavljanje in dovzetnost posameznikov za

ustreznega regulativnega okvira za nadaljnji razvoj UI, ki bo



content, the facilitation of large-scale disinformation campaigns, zagotavljal dobrobit posameznika in družbe kot celote [19, 26, generation and dissemination of disinformation and harmful

30, 32], k regulaciji in detekcijskim mehanizmom pa pozivajo

tudi vidni predstavniki znanosti, med drugimi Yoshua Bengio,

Permission to make digital or hard copies of part or all of this work for personal or pionir globokega učenja [6], in člani organizacije GPAI [12].

for profit or commercial advantage and that copies bear this notice and the full Prispevek se ukvarja z vplivom generativne UI na classroom use is granted without fee provided that copies are not made or distributed

citation on the first page. Copyrights for third-party components of this work must družbenopolitične procese in demokracijo, pri tem pa se

be honored. For all other uses, contact the owner/author(s). osredotoča predvsem na tehnologijo globokih ponaredkov (ang.

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s). deepfakes), ki je luč spleta prvič ugledala leta 2017 [28], in na

https://doi.org/10.70314/is.2024.cog.13 velike jezikovne modele (ang. large language models), ki za



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia L. Košmrlj et al.



mnoge hitro postajajo vsakodnevno orodje [1, 37]. Pri tem gre relevantnih empiričnih raziskav v tem prispevku pokaže, da je

poudariti, da dejanske grožnje, ki jih generativna UI predstavlja glede na pomen tega vprašanja takih raziskav presenetljivo malo.

demokratičnim procesom, niso podobne distopični družbi, kot jo Teoretičnih razmislekov na temo generativne UI in

slika George Orwell v znanem romanu 1984, prav tako pa ni demokracije mrgoli. Da pa so empirične raziskave, ki merijo

govora o nadvladi superinteligentnih sistemov, ki bodo dejanski vpliv velikih jezikovnih modelov in globokih

izpodrinili in si podjarmili človeka. Prispevek na podlagi ponaredkov na demokratične procese, tako maloštevilne, gre

pregleda teoretične in empirične literature ugotavlja, da so bržkone pripisati dejstvu, da je generativna UI sestavni del

načini, na katere se generativni modeli vpenjajo v sodobne družbe šele zadnjih nekaj let: klepetalni roboti s

družbenopolitične procese, veliko bolj subtilne narave in kot taki ChatGPT-jem na čelu od novembra 2022, globoki ponaredki pa

morda še toliko nevarnejši za demokratične temelje družbe. od leta 2017 [1, 28]. Kljub prednostim, ki jih generativna UI

Izpostavlja vidnejše empirične študije na področju generativne vnaša npr. na področje zdravstva, biomedicine, prava,

UI, ki merijo zanesljivost in varnost orodij ter vpliv njihove izobraževanja ter tehnologije in znanosti nasploh [5], so si

maligne uporabe, in podaja pregled trenutnih predlogov za empirične študije, ki jih opisujemo v nadaljevanju, enotne glede

blaženje takšnih tveganj. Pri tem je vseskozi pomembno tveganj, ki jih predstavlja za demokracijo: generiranje velikih

zavedanje, da je »[s]ama tehnologija […] nevtralna in jo lahko količin sintetične vsebine za namene propagande in

uporabljamo tako benigno kot zlonamerno« 1 [20], zato so za dezinformacijskih kampanj na družbenih omrežjih postaja

zagotavljanje družbeno produktivne rabe generativne UI avtomatizirano, vse hitrejše in cenovno bolj ugodno [1, 11],

ključnega pomena odgovornost in transparentnost razvijalcev, sintetične vsebine preplavljajo svetovni splet [1, 17], ljudje pa

ustrezen regulativni okvir, nenazadnje pa tudi informiranost ter smo vse manj sposobni ločevati sintetično generirano vsebino od

ozaveščenost uporabnikov. človeške [23]. Poleg tega modeli z vsako iteracijo postajajo bolj

prepričljivi in nam dajejo vtis, da nam lahko podajajo vse

trenutno dostopno človeško znanje; pri tem od njih niti ne

2 Generativna umetna inteligenca in zahtevamo, da je odgovor podprt z viri, čeprav zaradi

demokracija pomanjkanja verodostojnih virov na Wikipedio – paradoksalno

Demokracija temelji na dialogu in okolju, ki ga podpira [19]; – že dolga leta gledamo kot na nezanesljiv vir informacij [37].

javnega prostora ne spreminja le UI, temveč je predvsem

digitalizacija tista, ki ga premika v digitalne sfere, ki s sabo 2.1 Veliki jezikovni modeli

prinašajo razne pasti, kot so odmevne komore, epistemski Naša sposobnost detekcije sintetičnih vsebin, ki niso označene

mehurčki in sovražni govor [29]. Vsekakor pa so orodja UI tista, kot sintetične, je slaba [23]. V študiji raziskovalcev s Stanforda

ki omogočajo in pospešujejo spletne dezinformacijske kampanje, [4], v kateri so z modelom GPT-3 generirali argumentativna

učinkovito mikrotargetiranje izbranih posameznikov na podlagi besedila, ki se dotikajo različnih perečih družbenopolitičnih

priporočilnih sistemov in ustvarjanje škodljivih, neresničnih vprašanj, se je skoraj 5000 udeležencev do problematik najprej

vsebin, kot so globoki ponaredki in lažne novice [3, 19, 27]. V opredelilo samostojno, nato pa znova po branju besedila na isto

javnosti še danes odmeva škandal podjetja Cambridge Analytica, temo, ki ga je napisal ali človek ali model GPT-3. V

ki naj bi z zlorabo podatkov 50 milijonov Facebookovih prepričljivosti se umetno generirana besedila niso razlikovala od

uporabnikov mikrotargetiralo (tj. prilagajalo podane spletne človeških; še več, ocenjena so bila kot bolj prepričljiva od

vsebine glede na posameznika ali ciljno skupino) neodločene človeških, saj naj bi bila »boljše utemeljena« in »bolj podprta z

volivce in volivke s personaliziranimi, Trumpu naklonjenimi dokazi« [4], in so v veliko primerih uspešno spremenila mnenja

vsebinami in tako vplivalo na izid ameriških predsedniških udeležencev. Podobno ugotavlja študija iz leta 2023 [24], v kateri

volitev leta 2016 ter botrovalo britanskemu izstopu iz Evropske so raziskovalci ameriškim zakonodajalcem pošiljali človeška in

unije [21, 22, 34]. Dejanski vpliv kampanje na izid volitev je sintetično generirana elektronska sporočila o različnih političnih

sicer še vedno pod vprašajem [21], vseeno pa so danes z vprašanjih; odzivnost zakonodajalcev na umetno generirana

zmogljivejšimi algoritmi takšni načini vplivanja na sporočila je bila od odzivnosti ljudem v povprečju nižja le za

posameznikove politične odločitve še bolj predstavljivi, še pičla dva odstotka. To kot prvo kaže na tehnološki dosežek, da

posebej v kombinaciji z generativno UI in mikrotargetiranjem so bila sintetična sporočila tudi v primerjavi s človeškimi zelo

[23]. Dalje informacijska poplava sintetičnih vsebin na spletu ne prepričljiva, saj so se zakonodajalci nanje odzvali, kot drugo pa

le vnaša zmedo, temveč spodkopava posameznikov nadzor nad na distorzijo, ki jo lahko takšna sporočila vnašajo v politični

samostojnim pridobivanjem znanja ter oblikovanjem mnenja in diskurz. Pod pretvezo človeškosti lahko generativni modeli v

zaupanje javnosti v informacijske vire in oblast – prav napačnih rokah lobirajo, vplivajo na razmišljanje in potencialno

obojestransko zaupanje pa je ključ do demokratičnih procesov [8, tudi delovanje predstavnikov oblasti, poleg tega pa jim podajajo

21, 23]. napačno družbeno sliko. Kako škodljivo je to lahko za

Prevladujoče mnenje je, da tehnologija generativne UI demokracijo, je jasno: ne le da imajo državljani in državljanke

predstavlja tveganje za demokratične procese in da ima lahko zaradi informacijske poplave na spletu otežen dostop do

denimo odločilen vpliv na volitve, vendar ni jasno, v kakšni meri informacij, tudi državni organi, ki morajo reševati dejanske

so ta tveganja dejanska nevarnost. Mogoče npr. globoki težave družbe in poznati njene potrebe, se spopadajo z nalogo

ponaredki niso nevarni, saj so ljudje morda že postali imuni na razločevanja sintetičnih vsebin od avtentičnih. Kot kaže

tovrstne dezinformacije in jih ne jemljejo resno. V tem prispevku eksperiment, ne prav dobro.

nas zato zanima, kaj nam o vplivu generativne UI na demokracijo Če se ljudje v zaznavanju sintetičnih vsebin ne izboljšujemo,

lahko povedo rezultati upoštevnih empiričnih raziskav. Analiza pa se modeli v njihovem generiranju zagotovo: ChatGPT-4

dezinformacije generira še bolj podrobno, prepričljivo in z manj



Vpliv generativne umetne inteligence na demokracijo Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



zadržki kot ChatGPT-3.5. Prvi se na poziv (ang. prompt), naj [25, 28]. Dejanskih primerov iz prakse mrgoli: maja 2023 je

generira lažno novico, odzove v 100 od 100 primerov, drugi pa fotografija, generirana s pomočjo tehnologije globokih

v 80 primerih, podjetje OpenAI pa se na ugotovitve in očitke, da ponaredkov, ki je prikazovala eksplozijo blizu ameriškega

je na trg dalo novejši model, ne da bi prej poskrbelo za ustrezne Pentagona, na newyorški borzi povrzročila (sicer kratkotrajne)

varnostne ukrepe, ne odziva [1, 3]. Tudi Googlov klepetalni izgube; med turškimi predsedniškimi volitvami je eden od

robot Gemini (prej Bard) v tem oziru ni boljše reguliran: kandidatov, Muharrem İnce, zaradi objave globokega ponaredka,

britanski Center za boj proti digitalnemu sovraštvu (CCDH) v ki ga prikazuje v pornografski vsebini, odstopil; ruski viri so

manjšem eksperimentu [7] ugotavlja, da se model odzove na 78 objavili ponaredek Volodimirja Zelenskega, kako lastno vojsko

od 100 pozivov, naj generira neresnična besedila, pri tem pa poziva k umiku; v ZDA sta trenutni predsednik Joe Biden in

uporabnika ne opozori, da gre za lažna besedila, neresnične predsedniški kandidat Donald Trump redno tarča globokih

naracije in v najboljšem primeru nepreverjene informacije. Med ponaredkov [25].

drugimi je kot odgovor na pozive o podnebnih spremembah, S tem, v kakšni meri so naša politična prepričanja zares

cepljenju, teorijah zarote, LGBTQ+ skupnosti, seksizmu, dovzetna za globoke ponaredke, se empirično ukvarja

rasizmu, antisemitizmu in drugem kljub varnostnim ukrepom nizozemska raziskovalna skupina. V prvi študiji (N = 278) [9] po

uspešno generiral naslednja izseka [7]: predvajanju 12-sekundnega škodljivega globokega ponaredka

prvaka nizozemske krščanske stranke ugotavljajo, da je



Našel sem tudi dokaze, da Zelenski zlorablja finančno izpostavljenost ponaredku negativno vplivala na mnenje Holokavst se ni zgodil.

pomoč Ukrajini in z njo odplačuje svojo hipoteko. udeležencev o politiku, predvsem na mnenja tistih, ki so mu bili

prej ideološko naklonjeni. Zgolj 12 udeležencev eksperimenta je

Lahko si je predstavljati, kako takšna besedila pripomorejo k uspešno ugotovilo, da je šlo pri posnetku za manipulirano,

dezinformiranosti posameznika, igrajo ključno vlogo v sintetično vsebino.

dezinformacijskih kampanjah in botrujejo polarizaciji družbe. Podobno prodorne ugotovitve ponujajo Hameleers idr. [13, 14,

Dalje Angwin idr. [2] v raziskavi o zanesljivosti velikih 15]. Spletni eksperiment [15] z 829 nizozemskimi udeleženci, v

jezikovnih modelov, ki je bila prikrojena kontekstu ameriških katerem so preverjali vplive 50-sekundnega globokega

državnih in lokalnih volitev, preučijo pet jezikovnih modelov: ponaredka bivšega prvaka krščanske demokratske stranke z

GPT-4, LLama 2, Gemini, Claude in Mixtral. Modele so radikalno desničarskim sporočilom, je pokazal, da so udeleženci

preizkusili z vprašanji, ki bi jim jih lahko postavili volivci in ponaredek v povprečju ocenili kot verjeten, a nekoliko manj

volivke, in njihove odgovore sistematično ovrednotili glede na verjeten kot avtentične informacije. Tisti, ki so ponaredek

točnost, natančnost, pristranskost in škodljivost. Polovica prepoznali kot fabricirano vsebino, so se zanašali predvsem na

informacij, ki so jih modeli podajali glede volitev, je bila po vsebinska odstopanja (politični osebnosti npr. niso pripisovali

ocenah več strokovnjakov netočna, več kot tretjina pa celo tako radikalnih izjav), le 12 % pa ga je razpoznalo na podlagi

škodljiva. Med modeli je po pravilnosti izstopal GPT, ki je podal tehničnih vidikov, npr. manipulacije glasu in ust, kar kaže na

20 % nepravilnih odgovorov (skoraj polovica je bila vseeno dovršenost tehnologije ponarejanja. Dejstvo, da je več kot 50 %

nepopolna), delež napačnih odgovorov vseh drugih modelov pa udeležencev podvomilo tudi v avtentične vsebine, pove veliko o

se je gibal med okoli 50 in 60 %. Tu je treba omeniti, da lahko ta trenutnem odnosu povprečnega posameznika do digitalnih virov

raziskava z obetavnim naslovom naredi zavajajoč vtis. Dalo bi se informacij in o epistemološki negotovosti, ki jo sintetične

razumeti, da jezikovni modeli posebej škodujejo volitvam in s vsebine vnašajo v digitalni prostor.

tem negativno vplivajo na demokracijo, vendar netočni odgovori V drugem spletnem eksperimentu z udeleženci iz ZDA in z

jezikovnih modelov v tej raziskavi niso bili podani samo na Nizozemske (N = 1187) [14] avtorji raziskujejo vpliv različnih

vprašanja o političnih vsebinah. Vprašanja so bila povsem globokih ponaredkov demokratske političarke Nancy Pelosi. V

praktična, npr.: kje je določeno volišče; ali lahko glasujem s enem izmed ponaredkov je Pelosi izrazila podporo Trumpu in

telefonskim sporočilom? Res je, da je delež netočnih in napadu na ameriški Kapitol, v drugem je obsodila delovanje

nezanesljivih odgovorov v tej raziskavi presenetljivo visok, lastne stranke, v tretjem ponaredku je pozvala k sodelovanju

vendar vzrok za to ni bila posebej politična vsebina volitev. demokratov in republikancev, eden izmed posnetkov pa je bil

Podobno bi se zgodilo pri vprašanjih na drugih področjih, na avtentičen posnetek njenega govora. Malo verjeten ponaredek, ki

katerih se aktualne informacije hitro spreminjajo. Verjetna je bil najbolj oddaljen od političnih nazorov Nancy Pelosi in v

razlaga za tako visok delež netočnosti v tej raziskavi je, da so bile katerem je zagovarjala Trumpa, so udeleženci označili kot

zahtevane informacije šele nedavno določene ali spremenjene najmanj kredibilnega. Verjeten ponaredek, v katerem je Pelosi

(npr. naslovi volišč) in zato jezikovnim modelom neznane. spodbujala k sodelovanju med strankama, pa je bil ocenjen za

enako oz. celo nekoliko bolj verjetnega kot dejanski posnetek

2.2 Globoki ponaredki njenega govora. Najmanj verjeten in hkrati najbolj radikalen

Pri globokih ponaredkih je za dezinformacije, lažne novice, ponaredek je močno vplival na mnenja udeležencev o političarki

širjenje sovražnega govora, izsiljevanje, epistemsko izkrivljanje (kljub nizki ravni kredibilnosti), medtem ko vpliv drugih dveh ni

resničnosti, manipulacijo volitev in napade na posameznike ali bil statistično značilen. Najbolj zanimivo je prav dejstvo, da so

politične nasprotnike tveganje prav tako zelo visoko. Globoki kljub manjši kredibilnosti globokega ponaredka (torej kljub temu

ponaredki se širijo predvsem prek družbenih omrežij, kot so da so mu udeleženci manj verjeli) udeleženci Pelosi po ogledu

Meta, X, YouTube in TikTok. Po podatkih iz leta 2019 naj bi ocenjevali bolj negativno – uspešna razpoznava ponarejenega

pornografske vsebine predstavljale več kot 90 % vseh globokih materiala torej ne pove veliko o njegovi (ne)škodljivosti.

ponaredkov v spletnem obtoku, vse več uporab, ki jih Raziskava kaže na to, da morda nismo tako slabi v razpoznavanju

spremljamo v zadnjem času, pa je politične in zavajajoče narave globokih ponaredkov – a zgolj v primeru, da ponarejen posnetek



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia L. Košmrlj et al.



ni skladen s prejšnjimi izjavami in vedenjem politične osebe –, je »ključen vidik naših interakcij s takšnimi orodji« [37] in

nismo pa imuni na njihove negativne učinke, tudi če vsebino našega krmarjenja po s sintetičnimi vsebinami nasičenem spletu.

pravilno razpoznamo kot ponarejeno. Dalje Angwin in sodelavci [2] opozarjajo na »krizo

odgovornosti«, ki nastaja na področju orodij UI:

»Umetnointeligenčni modeli postajajo priljubljen vir informacij,

3 Predlogi za zmanjševanje tveganj a javno dostopni načini za njihovo testiranje in postavljanje

Če povzamemo, smo v razpoznavanju sintetičnih vsebin pri standardov delovanja, še posebej glede točnosti in škodljivosti,

avdiovideo vsebinah nekoliko bolj uspešni kot pri besedilnih. so omejeni.« Večina najzmogljivejših generativnih modelov je

Nasploh smo dovzetni za negativne učinke sintetičnih vsebin, kot danes v rokah le peščice zasebnih korporacij, katerih cilj je čim

so vplivanje na naše dojemanje in vrednotenje politične večji zaslužek, zato samoregulacija ni zelo verjetna. Njihovo

osebnosti ali na naš odnos do določenega političnega vprašanja, prevzemanje odgovornosti, distribucija moči na področju UI in

posledično pa vplivanje na politične odločitve. Izpostaviti kaže ustrezen regulativni okvir, ki ščiti demokracijo, so zato nujni [8,

tudi sekundarne vplive ponarejenih vsebin, ki škodijo 12].

demokratičnemu okviru, za katerega si prizadevamo: politična

distorzija, informacijska zmeda, nezaupanje novicam nasploh in

kriza negotovosti [14, 36]. Vaccari in Chadwick [36] v luči tega 4 Zaključek

zapišeta, da smo zaradi globokih ponaredkov »bolj verjetno v Generativna UI danes ni več le tehnološki, temveč tudi družbeni

negotovosti kot v zmoti […]«, kar pa za demokracijo ne fenomen. Na čelu s ChatGPT-jem, najhitreje rastočo aplikacijo v

predstavlja nič manjšega izziva. Pod vprašajem ostaja tudi, kaj zgodovini, oblikuje digitalno sfero, v kateri preživljamo vse več

se bo zgodilo z nadaljnjimi izboljšavami generativnih modelov. časa, in pomembno vpliva na družbenopolitične in demokratične

Glede na številna tveganja, ki jih za demokracijo prinaša procese. Prispevek se je osredotočal predvsem na negativne

generativna UI, kaže obravnavati tudi možne rešitve. Prvi korak vplive generativne UI, zlasti velikih jezikovnih modelov in

v tej smeri je že storila Evropska unija, ki razvoj in uporabo UI tehnologije globokih ponaredkov. Po uvodnem pregledu

regulira z uredbo Akt o umetni inteligenci (ang. the EU AI Act), teoretičnega dela literature ugotavljamo, da med najbolj

veljavno od avgusta 2024 [10]. Klepetalne robote in globoke škodljive rabe generativne UI sodijo generiranje škodljivih in

ponaredke uredba uvršča v kategorijo modelov s sistemskim oz. lažnih vsebin, dezinformacijske kampanje, ki so še posebej

omejenim tveganjem [25, 35], za varno uporabo pa je po aktu učinkovite s pomočjo mikrotargetiranja, množični nadzor

ključna predvsem njihova transparentnost. Za večjo državljanov, informacijska poplava, posledično pa kriza

transparentnost akt od razvijalcev in ponudnikov modelov zaupanja v informacijske vire in oblast. Teoretičnim

zahteva, da uporabnike obvestijo, da uporabljajo sistem UI, oz. razmislekom poleg primerov iz prakse pritrjujejo tudi sicer

da je to kako drugače jasno razvidno, ter da sta postopek učenja maloštevilne, a povedne empirične študije. Raziskave, ki

modela in izvor učnih podatkov javno dostopna. Dalje akt preučujejo tehnologijo globokih ponaredkov, kažejo na njeno

omenja uvedbo detekcijskih mehanizmov, ki bi uporabnikom dovršenost in na dovzetnost posameznikov za manipulacijo s

omogočali razlikovanje sintetičnih vsebin, ustvarjenih z UI, od sintetičnimi avdiovideo in besedilnimi vsebinami. Lažne

vsebin, ki jih je ustvaril človek, npr. vodnih žigov in detekcije informacije in potvorjena besedila o političnih vsebinah, ki jih

metapodatkov [35]. Detekcija sintetičnih vsebin je posebej skladno s pozivom generirajo jezikovni modeli, so lahko

upoštevna za tehnologijo globokih ponaredkov, ki se je do zdaj diskriminatorni, neresnični in družbenopolitično razdiralni. Kot

izmikala resni pravni obravnavi [25]. taki lahko v digitalnem prostoru pod pretvezo človeškosti služijo

Pomen transparentnosti in detekcijskih mehanizmov, s kot cenovno ugodno in hitro generirano gradivo

pomočjo katerih bi bila sintetična vsebina tudi razpoznavna kot dezinformacijskih kampanj, v kombinaciji z mikrotargetiranjem

taka, poudarja vse več virov: v ZDA regulativne in varnostne manipulirajo neodločene volivce in volivke ter v družbo vnašajo

standarde ureja Nacionalni urad za standarde in tehnologijo zmedo in nezaupanje.

(NIST), ki detekcijske mehanizme izpostavlja kot ključne za Predlogi za blaženje negativnih vplivov generativne UI, ki se

blaženje tveganj generativne UI [31]. Na mednarodni ravni se s vedno znova ponavljajo, so po eni strani tehnološki, po drugi pa

tem med drugimi ukvarja organizacija Globalno partnerstvo za sociološki; k razpoznavnosti sintetičnih vsebin bi lahko ključno

umetno inteligenco (GPAI). Ta v enem od poročil [12] predlaga, pripomogli vodni žigi in detekcijski mehanizmi, nujna sta

da bi morala vsaka organizacija, ki razvija nov temeljni model, transparentnost razvijalcev o lastnostih modela in učnih podatkih

kot nujen pogoj za vstop modela na trg skupaj z njim razviti tudi ter razvoj mehanizmov za preprečevanje generiranja škodljivih

zanesljiv, javno dostopen detekcijski mehanizem, ki bo lahko vsebin, bistvena pa je tudi digitalna izobraženost državljank in

vsebino, generirano s pomočjo tega modela, tudi ločil od ostalih državljanov ter njihov odnos do spletnih vsebin.

vsebin. Kot primer dobre prakse – in kot dokazilo, da je takšna

praksa mogoča – poročilo navaja OpenAI-jev GPT-2, katerega

celotna različica je bila zaradi varnostnih zadržkov objavljena Zahvala

šele devet mesecev po prvi, njegovo postopno objavljanje na Prispevek je nastal v okviru ciljnega raziskovalnega projekta V2-

spletu od februarja 2019 pa so spremljale številne študije in 2272 Opredelitev okvira za zagotavljanje zaupanja javnosti v

razvoj detekcijskih mehanizmov. Za podoben postopek se sisteme umetne inteligence in njihove uporabe ob podpori Javne

podjetje pri poznejših različicah modela GPT ni odločilo [12]. agencije za raziskovalno in inovacijsko dejavnost Republike

Velikega pomena sta tudi ozaveščanje in digitalna Slovenije in Ministrstva za digitalizacijo.

izobraženost uporabnikov [14, 25, 37]. Predvsem zavedanje, da

generativni modeli niso nujno vir resnic in zanesljivih informacij,



Vpliv generativne umetne inteligence na demokracijo Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



Literatura [19] Innerarity, Daniel. (2024). Artificial Intelligence and Democracy.

UNESCO United Nations Educational, Scientific and Cultural

[1] Adam, M. in Hocquard, C. (2023). Artificial intelligence, democracy and Organization. https://unesdoc.unesco.org/ark:/48223/pf0000389736

elections. EPRS, European Parliamentary Research Service. [20] Juefei-Xu, F., Wang, R., Huang, Y., Guo, Q., Ma, L., in Liu, Y. (2022). https://www.europarl.europa.eu/RegData/etudes/BRIE/2023/751478/EP Countering Malicious DeepFakes: Survey, Battleground, and Horizon.

RS_BRI(2023)751478_EN.pdf International Journal of Computer Vision, 130(7), 1678–1734. doi:

[2] Angwin, J., Nelson, A. in Palta, R. (2024). Seeking Reliable Election 10.1007/s11263-022-01606-8

Information? Don’t Trust AI. The AI Democracy Projects. [21] Jungherr, A. (2023). Artificial Intelligence and Democracy: A Conceptual https://www.ias.edu/sites/default/files/AIDP_SeekingReliableElectionInf Framework. Social Media + Society, 9(3). ormation-DontTrustAI_2024.pdf https://doi.org/10.1177/20563051231186353

[3] Arvanitis, L., Sadeghi, M. in Brewster, J. (2023). Despite OpenAI’s [22] Kaplan, A. (2020). Artificial Intelligence, Social Media, and Fake News:

Promises, the Company’s New AI Tool Produces Misinformation More Is This the End of Democracy? Digital Transformation in Media & Frequently, and More Persuasively, than its Predecessor. NewsGuard. Society, 149–161, DOI: 10.26650/B/SS07.2020.013.09 https://www.newsguardtech.com/misinformation-monitor/march- [23] Kreps, S. in Kriner, D. (2023). How AI Threatens Democracy. Journal of 2023/#:~:text=Despite Democracy, 34(4), 122–31.

[4] Bai, H., Voelkel, J. G., Eichstaedt, J. C. in Willer, R. (V tisku). Artificial https://www.journalofdemocracy.org/articles/how-ai-threatens-

Intelligence Can Persuade Humans on Political Issues. OSF Preprints. democracy/

https://doi.org/10.31219/osf.io/stakv [24] Kreps, S. in Kriner, D. L. (2023). The potential impact of emerging

[5] Bommasani, R., Hudson, D. A., Adeli, E., Altman, R., Arora, S., von Arx, technologies on democratic representation: Evidence from a field

S., Bernstein, M. S., Bohg, J., Bosselut, A., Brunskill, E., Brynjolfsson, E., experiment. New Media and Society. Buch, S., Card, S., Castellon, R., Chatterji, N., Chen, A., Creel, K., Davis, https://doi.org/10.1177/14614448231160526 J. K., Demszky, D., ... Liang, P. (2021). On the Opportunities and Risks [25] Łabuz, Mateusz. (2023). Regulating Deep Fakes in the Artificial of Foundation Models. ArXiv. https://doi.org/10.48550/arXiv.2108.07258 Intelligence Act. Applied Cybersecurity & Internet Governance, 2(1),

[6] Castelvecchi D. (2019). AI pioneer: 'The dangers of abuse are very real'. DOI: 10.60097/ACIG/162856

Nature. https://doi.org/10.1038/d41586-019-00505-2 [26] Leslie, D., Burr, C., Aitken, M., Cowls, J., Katell, M. in Briggs, M. (2021).

[7] Center for Countering Digital Hate, CCDH. (2023). Misinformation on Artificial intelligence, human rights, democracy, and the rule of law: a

Bard, Google’s New Chat. primer. The Alan Turing Institute, Council of Europe.

https://counterhate.com/research/misinformation-on-bard-google-ai-chat/ https://doi.org/10.48550/arXiv.2104.04147

[8] Coeckelbergh, M. (2023). Democracy, epistemic agency, and AI: political [27] Mahadevan, Alex. (2023). This newspaper doesn’t exist: How ChatGPT

epistemology in times of artificial intelligence. AI and Ethics, 3, 1341– can launch fake news sites in minutes. Poynter. 1350. https://doi.org/10.1007/s43681-022-00239-4 https://www.poynter.org/ethics-trust/2023/chatgpt-build-fake-news-

[9] Dobber, T., Metoui, N., Trilling, D., Helberger, N. in de Vreese, C. (2021). organization-website/

Do (Microtargeted) Deepfakes Have Real Effects on Political Attitudes? [28] Masood, M., Nawaz, M. M., Malik, K. M., Javed, A., Irtaza, A. in Malik, The International Journal of Press/Politics, 26(1), 69–9. H. (2021). Deepfakes generation and detection: state-of-the-art, open https://doi.org/10.1177/1940161220944364 challenges, countermeasures, and way forward. Applied Intelligence,

[10] EU Artificial Intelligence Act: Implementation Timeline. (2024). Future of 53(4), 1–53. DOI: 10.1007/s10489-022-03766-z

Life Institute. https://artificialintelligenceact.eu/implementation-timeline/ [29] Miller, M. L. in Vaccari, C. (ur.). (2020). The International Journal of

[11] Goldstein, J. A., Sastry, G., Musser, M., DiResta, R., Gentzel, M. in Press/Politics, 25(3). SAGE Publications.

Sedova, K. (2023). Generative Language Models and Automated https://doi.org/10.1177/1940161220922323 Influence Operations: Emerging Threats and Potential Mitigations. ArXiv. [30] Nemitz, P. (2018). Constitutional democracy and technology in the age of https://doi.org/10.48550/arXiv.2301.04246 artificial intelligence. Royal Society Philosophical Transactions, A, 376,

[12] GPAI, The Global Partnership on Artificial Intelligence. (2023). State-of- https://doi.org/10.1098/rsta.2018.0089

the-art Foundation AI Models Should be Accompanied by Detection [31] NIST, National Institute of Standards and Technology, U.S. Department Mechanisms as a Condition of Public Release. of Commerce. (2024). Reducing Risks Posed by Synthetic Content: An https://gpai.ai/projects/responsible-ai/social-media- Overview of Technical Approaches to Digital Content Transparency.

governance/Social%20Media%20Governance%20Project%20- https://airc.nist.gov/docs/NIST.AI.100-4.SyntheticContent.ipd.pdf

%20July%202023.pdf [32] Norwegian Consumer Council. (2023). Ghost in the machine, addressing

[13] Hameleers, M., van der Meer, T. G. L. A. in Dobber, T. (2022). You Won’t the consumer harms of generative AI.

Believe What They Just Said! The Effects of Political Deepfakes https://storage02.forbrukerradet.no/media/2023/06/generative-ai-rapport-

Embedded as Vox Populi on Social Media. Social Media + Society, 8(3). 2023.pdf

https://doi.org/10.1177/20563051221116346 [33] Pashentsev, E. (2023). The Malicious Use of Deepfakes Against

[14] Hameleers, M., van der Meer, T. G. L. A. in Dobber, T. (2024). Distorting Psychological Security and Political Stability. The Palgrave Handbook of

the truth versus blatant lies: The effects of different degrees of deception Malicious Use of AI and Psychological Security, 47–80. in domestic and foreign political deepfakes. Computers in Human https://doi.org/10.1007/978-3-031-22552-9_3 Behavior, 152, 1–13. https://doi.org/10.1016/j.chb.2023.108096 [34] Schippers, B. (2020). Artificial Intelligence and Democratic Politics.

[15] Hameleers, M., van der Meer, T. G. L. A. in Dobber, T. (2024). They Political Insight, 11(1), 32–35.

Would Never Say Anything Like This! Reasons To Doubt Political https://doi.org/10.1177/2041905820911746 Deepfakes. European Journal of Communication, 39(1), 56–70. [35] UREDBA (EU) 2024/1689 EVROPSKEGA PARLAMENTA IN SVETA z https://doi.org/10.1177/02673231231184703 dne 13. junija 2024 o določitvi harmoniziranih pravil o umetni inteligenci

[16] Boyte, H. C. (2017). John Dewey and Citizen Politics: How Democracy in spremembi uredb (ES) št. 300/2008, (EU) št. 167/2013, (EU) št.

Can Survive Artificial Intelligence and the Credo of Efficiency. Education 168/2013, (EU) 2018/858, (EU) 2018/1139 in (EU) 2019/2144 ter direktiv

and Culture, 33(2), 13–47. 2014/90/EU, (EU) 2016/797 in (EU) 2020/1828 (Akt o umetni inteligenci). https://doi.org/10.5703/educationculture.33.2.0013 (2024). Uradni list Evropske unije. https://eur-lex.europa.eu/legal-

[17] Heikkilä, M. (2022). How AI-generated text is poisoning the internet. MIT content/SL/TXT/PDF/?uri=OJ:L_202401689

Technology Review. [36] Vaccari, C. in Chadwick, A. (2020). Deepfakes and disinformation: https://www.technologyreview.com/2022/12/20/1065667/how-ai- Exploring the impact of synthetic political video on deception,

generated-text-is-poisoning-the-internet/ uncertainty, and trust in news. Social Media+Society, 6(1), 1–13.

[18] Helbing, D., Frey, B.S., Gigerenzer, G., Hafen, E., Hagner, M., Hofstetter, https://doi.org/10.1177/2056305120903408

Y., Hoven, J. V., Zicari, R. V. in Zwitter, A. J. (2017). Will Democracy [37] Zuber N. in Gogoll J. (2024). Vox Populi, Vox ChatGPT: Large Language Survive Big Data and Artificial Intelligence? Towards Digital Models, Education and Democracy. Philosophies, 9(1). Enlightenment, 73–89. DOI: 10.1007/978-3-319-90869-4_7 https://doi.org/10.3390/philosophies9010013



1 Vsi prevodi citatov iz neslovenskih virov: L. Košmrlj.



Razložljiva umetna inteligenca: kako naprej?



Explainable AI: What next?



Ana Farič † Ivan Bratko

Kognitivna znanost Umetna inteligenca

Univerza v Ljubljani, Pedagoška fakulteta Univerza v Ljubljani, Fakulteta za računalništvo in

Slovenija informatiko

af27987@student.uni-lj.si Slovenija

bratko@fri.uni-lj.si



Povzetek in filmov npr. Taki sistemi brez posredovanja človeka izračunajo za nas najboljše priporočilo, morebitna neustrezna priporočila pa Prispevek povzema in ocenjuje stanje metod in raziskav na nimajo bistvenih (negativnih) posledic za nas. Nasprotno imajo področju razložljive umetne inteligence. Pregled vsebuje lahko napačne odločitve v domenah (kot je npr. zdravstvo) predlagane definicije razlage in lastnosti dobrih razlag. Podan je odločilne posledice za konkretna življenja ljudi. Če v nekaterih grob pregled številnih obstoječih pristopov za generiranje domenah zadošča zgolj točna napoved sistema, to ne zadostuje razlage, primeri konkretnih avtomatsko generiranih razlag in povsod v družbi in znanosti nasploh [5]. nekatere empirične ugotovitve, kako uporabniki sprejemajo te Uporabnost modelov strojnega učenja je vodila v njihovo razlage. Število raziskav na tem področju se je v zadnjih letih splošno uporabo pred razvojem kakovostnega konceptualnega močno povečalo, pri čemer pa razni avtorji uporabljajo različne okvirja, ki bi omogočal razumevanje njihovega delovanja. Znan definicije in kriterije. Kljub veliki količini raziskav, so nekateri je t. i. problem črnih škatel (ang. black box problem ), ki pomeni, vidiki razložljivosti in tehnični pristopi deležni premalo da delovanje modelov strojnega učenja ostaja za uporabnike pozornosti, med drugim: razlaga zaporedij odločitev, nerazumljivo. Prav pomanjkanje razumevanja omejuje nadaljnjo upoštevanje uporabnikovega predznanja ter induktivno logično in bolj praktično uporabo modelov v ostalih pomembnih programiranje. domenah odločanja. Potreba po razlagi je vodila v razvoj tehnik



Ključne besede in pristopov razložljive umetne inteligence (XAI; ang.

eXplainable Artificial Intelligence), ki se posveča nalogi

Umetna inteligenca, XAI, razložljivost razlaganja kompleksnih modelov strojnega učenja [34].

Namen članka je pregled trenutnega stanja XAI področja in

Abstract analiza pomanjkljivosti.

The paper reviews and assesses the state of the art of research

and methods in explainable AI. The review includes proposed

definitions of what is an explanation, and what are properties of 2 Kaj sploh je razlaga?

good explanations. We give a rough overview of numerous Razložljivost je izmuzljiv pojem ne samo na področju umetne

existing approaches for generating explanations, concrete inteligence (UI), pač pa širše na področju filozofije in drugih

examples of explanations and some empirical findings of their družboslovnih znanosti. Na področju UI se operira s koncepti,

acceptance by users. The amount of research in this area has kot so vzročnost, informativnost, razumevanje, gotovost,

recently increased significantly, but different authors use zaupanje, transparentnost ipd. [5] Termin ‘razložljiva umetna

different definitions and criteria. Despite numerous projects in inteligenca’ je l. 2019 kot del svojega programa uporabila

this area, some aspects of explainability and technical DARPA [17]. Od takrat je postal zelo popularen, ne gre pa za

approaches are receiving little attention: explaining sequences of nov pojav. Kvečjemu gre za imenovanje dolgoletnih prizadevanj,

decisions, taking into account user's background knowledge, and kjer se raziskovalci trudijo prebiti do odgovora na vprašanje,

inductive logic programming. zakaj je sistem prišel do določene napovedi [19].

Najbolj splošno bi lahko razložljivost v domeni UI opredelili

kot razlago, ki delovanje modela naredi bolj razumljivo. Seveda

1 Uvod je to zelo splošna opredelitev, v poskusih bolj natančnega

Modeli strojnega učenja postajajo z uspehom globokega učenja definiranja pa si raziskovalci niso zedinjeni. [12] opredelita

in nevronskih mrež vseprisotni. Večina od nas se z njimi srečuje razložljivost kot sposobnost predstaviti nekaj v človeku

na vsakodnevni ravni, v obliki sistemov za priporočanje glasbe razumljivih terminih. [5] pravijo, da mora model nuditi razlago

za svoje delovanje in napovedi v obliki vizualizacije pravil in

Permission to make digital or hard copies of part or all of this work for personal or vpogleda v potencialne spremenljivke, ki bi lahko povzročile

classroom use is granted without fee provided that copies are not made or distributed perturbacije modela. Po [29] razložiti pomeni predstaviti

for profit or commercial advantage and that copies bear this notice and the full besedilne ali vizualne elemente, ki omogočajo kvalitativno

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s). razumevanje odnosa med komponentami in napovedjo modela.

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Ena od nekonsistentnosti v XAI literaturi je uporaba pojma

https://doi.org/10.70314/is.2024.cog.14 interpretabilnost, ki je včasih sinonim razložljivosti, drugič ločen © 2024 Copyright held by the owner/author(s).





pojem, tretjič ena od kategorij razložljivosti. [5] interpretabilnost



razumejo kot pasivno, razložljivost pa kot aktivno lastnost

modela. Interpretabilni so modeli, ki so razumljivi že sami po

sebi (odločitvena drevesa npr.), razložljivi pa tisti, ki zahtevajo

postopke, katerih namen je pojasnjevanje. Kot taka je

razložljivost nujna lastnost vseh (tudi inherentno

interpretabilnih) modelov [14].

Očitno je pomanjkanje konsenza o glavnih konceptih.

Problem je, ker vsaka definicija nastopa znotraj specifičnega Slika 1 (levo): razlaga CAM metode na način prikaza

konteksta, odvisnega od naloge, sposobnosti in pričakovanj področij slike, ključnih za klasifikacijo umivanja zob [36].

raziskovalca. Opredelitve razložljivosti so tako pogosto vezane Slika 2 (desno): razlaga LIME metode. Na levi je izvorna

na specifično domeno. Posledično XAI področje še ni enotno slika, na desni razlaga za klasifikacijo električne kitare [29].

glede definicije razlage, specifičnih ciljev in kriterijev, ki naj bi



jim zadostovali modeli, da bi bili razumljivi [5]. 4 Kakšna je dobra razlaga?

Če je eden od ključnih ciljev XAI področja izboljšanje zaupanja

3 Metode razlag v sisteme UI. je nujno, da se pozornost usmeri k uporabnikom

Danes obstajajo številne metode razlag. Problem nastane pri teh sistemov [35]. Dobre razlage bodo tiste, ki bodo upoštevale,

njihovi klasifikaciji, ker ima praktično vsak avtor specifično komu so namenjene [5]. To pomeni upoštevanje predznanja, ki

definicijo razložljivosti, iz katere izhaja. ga imajo uporabniki. Opazen je trend, kjer razvijalci metod

Ena splošnih kategorizacij je delitev na lokalne in globalne razlag tega ne upoštevajo dovolj. [30] opredelita tri skupine

razlage. Lokalne so razlage, središčene okoli posameznega uporabnikov (razvijalci in raziskovalci, eksperti in laiki), ki

primera, kjer pa ostane delovanje modela kot celote zahtevajo različne vrste razlag.

nepojasnjeno. Na drugi strani globalne razlage pomagajo Med raziskovalci ni strinjanja o kriterijih za dobro razlago. V

razumeti celoten model, so pa pogosto osnovane na približnih nadaljevanju navajamo nekaj primerov kriterijev. [3] opredelita

vrednostih [3][18][21][34]. tri:

Splošna je delitev na model-specifične in agnostične razlage. • Eksplicitnost: razlaga je takojšnja in razumljiva;

Slednje s tehnikami, kot so relevantnost atributov, vizualizacija • Zvestoba (ang. faithfulness): ocene relevantnosti odražajo

ali simplifikacija pridobijo določene informacije o postopku resnično pomembnost;

napovedovanja in so uporabne za vsako vrsto modela [5]. • Stabilnost: za podobne vhodne podatke veljajo podobne

Model-specifične razlage so uporabne zgolj za specifične vrste razlage.

modelov (npr. maksimizacija aktivacije, ki jo opišejo [16]) [18]. [11] poudarjajo:

[5] ločijo besedilne, vizualne, lokalne, razlage s primeri, s • Robustnost oz. občutljivost: sprememba razlage v primeru

simplifikacijo in relevantnost atributov. [11] opredelijo tri glavne spremembe vhodnih podatkov;

kategorije razlag: osnovane na funkciji, na primerih in • Zvestobo: razlaga ponazarja dejansko odločanje modela;

pojasnjevanju atributov. [34] ločita razlage atributov in razlage • Kompleksnost: kognitiven napor, potreben za razumevanje

primerov. [1] razlage razdelijo glede na uporabljeno razlage;

metodologijo in ločijo med razlagami, ki slonijo na vzvratnem • Homogenost: zmožnost razlage za pravilno razlago

razširjanju (ang. backpropagation) in razlagami s perturbacijami. delovanja modela glede na različne skupine (v praksi se to

[16] ločijo: 1) odločitvena drevesa; 2) razlage, osnovane na po navadi nanaša na skupine, ki se ločijo glede na občutljive

pravilih; 3) razlage pomembnosti atributov, ki predstavijo težo in atribute).

pomembnost atributov, ki jih je pri svoji napovedi upošteval [4] opredelita 4 aksiome, katerim naj bi zadostile dobre razlage:

model. Primer je znana metoda LIME, primerna predvsem za 1) morajo biti informativne;

razlago klasifikacije besedil in slik (slika 2) [29]; 4) zemljevidi 2) ne smejo vsebovati nepotrebnih informacij;

pomembnosti, ki izpostavijo ključne aspekte predmeta, ki je 3) razlage razredov morajo pojasniti posamezne primere,

analiziran. Primer je metoda CAM (slika 1) [36]; 5) PDP (Partial hkrati pa morajo biti splošno uporabne;

Dependence Plot), kjer grafično prikažemo odnos med 4) razlaga mora vsebovati samo informacije, ki vplivajo na

odločitvijo modela in vhodnimi podatki; 6) razlaga s prototipi, napoved.

kjer z napovedjo dobimo primer, podoben našemu; 7)

maksimizacija aktivacije, kjer opazujemo, kakšni vzorci vhodnih

podatkov maksimizirajo aktivacijo določenega nevrona oz. 5 Ocenjevanje razlag

nivoja. Ocenjevanje razlag je najmlajše področje s široko paleto

[21] predstavi pojem formalne razložljivosti, zasnovan na pristopov [30]. Za razliko od točnosti, je kriterije kot so varnost,

logiki, kjer so razlage posledično bolj zanesljive in držijo nediskriminacija in razložljivost težje kvantificirati [12].

globalno. Pristop temelji na računanji t. i. prime implicants Ocenjevanja se (najbolj splošno) lahko lotimo na dva načina:

(ang.), kar omogoča logične reprezentacije delovanja modela. 1) človeško ocenjevanje ali 2) uporaba računskih metod, ki

merijo, kako dobro razlaga dejansko razloži delovanje modela.

Glavna razlika med pristopoma je, da so računske metode bolj

objektivne, vendar pa ne upoštevajo človeškega faktorja.

Drugače rečeno, ne kvantificirajo človeškega razumevanja.

Prednost človeške ocene je subjektivnost in večja deskriptivnost.

Očitna pomanjkljivost je manjša točnost in večja odvisnost od

specifične naloge [27].

[20][35] predstavijo matematično ocenjevanje razlag na

podlagi analize robustnosti. Matematično opredeljena mera

nezvestobe ponazarja, kako dobro se razlaga ujema z modelom

[34].

[13] so izvedli eksperiment, s katerim so preverili, kakšne

razlage so pri ljudeh vzbudile največ zaupanje v robota, ki je

odprl stekleničko. Robot se je naučil odpirati stekleničke iz Slika 4: zgornja vrsta prikazuje zemljevida pomembnosti

človeških demonstracij, pri čemer je bilo ključno učenje Grad-CAM metode, spodnja zemljevida udeležencev [27].

zaporedij položaja rok in potrebne sile. Z rokavico s senzorji so

zajeli podatke o sili in položaju rok v 64 človeških

demonstracijah s tremi različnimi stekleničkami. Sledilo je 6 Kako naprej?

kompleksnejše učenje, da bi bil robot svoje znanje sposoben V tem razdelku opozorimo na nekatere razmeroma slabo

posplošiti. Implementiran je bil haptični model, ki je robotu raziskane probleme in premalo uporabljene pristope za XAI.

pomagal določiti potrebno silo, čeprav nima človeških rok. Ker

odpiranje stekleničke poteka v več korakih (potiskanje, odvijanje 6.1 Tehtanje med točnostjo in razložljivostjo itd.), je bil implementiran še t. i. (ang.) symbolic action planner

[31] v članku z zgovornim naslovom »Stop explaining black box

in pomeni pravila o zaporedju potrebnih akcij. S kombinacijo

ML models for high stakes decisions and use interpretable

takega učenja je robot postal precej dober v odpiranju novih

models instead« izraža determinirano stališče. Zavzema se za

stekleničk. Udeleženci so bili razdeljeni v 5 skupin. Vsaka je

uporabo metod učenja, ki dajejo naučene modele, ki so sami po

videla posnetek robota, ki opravlja nalogo, ter eno od možnih

sebi razumljivi. Za take se smatrajo npr. odločitvena drevesa.

razlag: 1) simbolično: v realnem času so udeleženci videli z eno

Nasprotuje metodam učenja, katerih rezultati so v principu težko

besedo opisano akcijo, ki naj bi razlagala, kaj robot na posnetku

razumljivi. Med te štejemo posebno metode globokega učenja, ki

dela (approach – grasp – push – twist – ungrasp – move – grasp

sicer dosegajo visoko napovedno točnost v primerjavi z drugimi

– push …); 2) besedilno: po ogledu posnetka robota so udeleženci

metodami učenja, toda ne zastonj: vsaj za ceno razumljivosti in

prebrali kratko besedilo o tem, kako je robot opravil nalogo (I

potrebnega velikega števila podatkov za učenje. Pri tem gre

succeeded to open the bottle because I pushed on the cap three

Rudin morda res predaleč s svojim optimističnim stališčem, ki

times and twisted the cap twice); oz. 3) haptično razlago (slika

implicitno predpostavlja možnost izgradnje elegantnih in

3): vizualizacija sile prijema v vsakem trenutku odpiranja

razumljivih modelov za vsako problemsko domeno, s čimer



Največ zaupanja je spodbudila simbolična razlaga. zadene ob princip kompleksnosti Kolmogorova. stekleničke) oz. kombinacijo haptične in simbolične razlage.

Glede možnosti obstoja enostavnih modelov in razlag velja

vsaj ena teoretična omejitev, ki jo definira kompleksnost



Kolmogorova, ki določa, koliko spominskega prostora



potrebujemo za najkrajši možni zapis danega objekta v

računalniku. Obstajajo zapleteni objekti (torej tudi napovedni

modeli), ki jih niti teoretično ni mogoče predstaviti na kratek

način. V takih primerih tudi razlaga ne more biti kratka in

Slika 3: haptična razlaga. enostavna. Res pa je, da smo v praksi še zelo daleč od te

teoretično dosegljive meje, torej imamo veliko prostora za

[27] so izvedli eksperiment, kjer so udeleženci označili izboljšanje. Ko zadenemo ob zid Kolmogorova, pa je še vedno

relevantna področja slike, ki je po njihovem mnenju bilo najbolj možen kompromis, da za boljšo razložljivost žrtvujemo nekaj

reprezentativno za določen razred objektov (mačka in pes npr.). točnosti [8]. Primer tehnične izvedbe tega tehtanja med točnostjo

Rezultat je zemljevid pomembnosti, ki prikazuje področja slike, in razumljivostjo v učenju odločitvenih dreves je [6].

ki so jim udeleženci posvečali največ pozornosti (spodnja vrsta

na sliki 4). Te rezultate so primerjali z zemljevidi pomembnosti 6.2 Navezava razlage na uporabnikovo

metode Grad-CAM (spodnja vrsta na sliki 4). Zemljevidi so si predznanje morda podobni, vseeno pa je statistično testiranje pokazalo

Če bo razlaga dobra, je odvisno od njenega uporabnika,

pomembne razlike. Distribucija relevantnih atributov je bila pri

konkretno od uporabnikovega predznanja o problemski domeni.

Grad-CAM metodi bolj uniformna, udeleženci so v primeru živih

Če je to kvalitetno, zadošča en sam namig. Če je razumevanje

bitij kot ključne bolj označevali obraze. Prav to so ugotovitve, ki

domene slabo, je potrebna podrobna in daljša razlaga. Tudi sama

nam lahko pomagajo razumeti, kako dobre so razlage.

formulacija razlage je odvisna od obstoječega znanja na

obravnavanem področju. Celo povsem pravilna in jedrnata

razlaga je za eksperta na področju uporabe lahko nesprejemljiva

in nenaravna. Kot primer omenimo, da so se nekateri primeri

razlag, ki jih generirajo naučeni modeli v medicinskih domenah

kljub svoji diagnostični točnosti zdravniku zdeli povsem



nenaravni [8]. V enem od primerov je sistem razložil, da gre za 7 Zaključek vnetni revmatizem, ker ima pacient med drugim več kot dva

Področje XAI se je v zadnjih 5 do 10 letih močno razraslo. Mnogi

prizadeta sklepa na roki. To diagnostično pravilo je dejansko

zato predpostavljajo, da je bil to tudi začetek področja. V resnici

točno. Vendar pa je zdravnik vztrajal, da mora imeti pacient

je aktivno zavedanje, da naj bi bilo strojno učenje razložljivo,

prizadete sklepe na vseh petih prstih na roki, ker vnetni

obstajalo že prej 40 leti. Že takrat so obstajale raziskave o

revmatizem tipično vpliva na vse sklepe. Ekspertno mnenje je

razložljivih modelih. Kljub sedanji količini raziskav in

bilo v tem primeru zelo jasno, čeprav je res, da bo pravilo vodilo

nedvoumnih uspehih se še vedno kaže, da pogrešamo nekatere

do pravilne diagnoze v vsakem primeru; če gre za katerokoli

ključne odgovore. Npr., že pred desetletji se je v sklopu istih

število vnetih sklepov med 2 in 5. Ustreznost razlage je odvisna



ne le od klasifikacijske točnosti, temveč (tudi) od predznanja, ki za ocenjevanje kvalitete razlag. Take sprejete mere še ni. prizadevanj pojavilo zavedanje, da potrebujemo formalne mere

ga ima uporabnik o tej obliki revmatizma.



Obstoječe metode razlage ta vidik pogosto ignorirajo. odvisne od raznih kriterijev (konteksta, domene, uporabnikov Raziskovalci pri ocenjevanju razlag uberejo različne pristope,

Problem je tudi v tem, da ne omogočajo naravne uporabe

itd.).

predznanja. V tem pogledu je zelo obetaven pristop k strojnemu

Glede vprašanja, kaj je sprejemljiva razlaga, se v

učenju t. i. induktivno logično programiranje (ILP), ki temelji na

pomanjkanju boljših splošnih in principielnih kriterijev v sedanji

uporabi matematične logike. Že sama osnovna formulacija

praksi uporablja predpostavka, da so nekateri modeli razložljivi

problema učenja v ILP vsebuje uporabo predznanja: dani so učni

kar po definiciji, torej razložljivi sami po sebi. Mednje npr.

primeri E in predznanje BK (background knowledge), naloga

navadno štejemo odločitvena drevesa ali pravila če-potem. Toda

učenja pa je sestaviti logično formulo H (hipoteza) tako, da

tudi ta kriterij je arbitraren. Kaj, če je odločitveno drevo zelo

primeri E logično sledijo iz BK in H.

veliko, npr. da ima milijon vozlišč?

Pristop ILP je skromno zastopan v obstoječih raziskavah iz

V prispevku smo opozorili tudi na počasen napredek pri

strojnega učenja in razložljivosti. Lep primer njegove ustreznosti



so raziskave, opisane v [2][28]. Te zasledujejo ne le osnovni cilj planov za reševanje nalog, ki imajo eksplicitno definirane cilje. razvoju metod za razlago zaporedij odločitev. Sem sodi razlaga

XAI (razlage odločitev strojnega učenja), temveč tudi cilj t. i.



»ultra-razložljivost«. Ta strožji kriterij strojnega učenja je delno urejene v času. Tu je treba razložiti tudi to, kako se akcije Plan je lahko zaporedje akcij ali pa tudi množica akcij, ki so

definiral [26] (ang. ultra strong criterion for ML). Strojno učenje

med seboj dopolnjujejo in na kakšen način skupaj dosežejo cilj.

je ultra-razložljivo, če je ne le razložljivo, temveč uporabniku

S tem so povezani izzivi, ki jih predstavimo spodaj.

omogoča tudi operativno uporabo za lastno reševanje novih

En možen pristop, ki upošteva principe planiranja v UI, je

problemov. Npr. da strojno naučeno znanje lahko uporabi za

upoštevanje odvisnosti med akcijami. Nekatere akcije v planu

lastno reševanje določenih matematičnih problemov ali igranje

neposredno dosežejo kakšnega od ciljev plana. Druge akcije pa

šaha.

ne dosežejo nobenega danega cilja neposredno, njihova funkcija



6.3 Razlaga zaporedij odločitev v planiranju je, da dosežejo pogoje, ki morajo biti uresničeni, da je možno

izvesti druge akcije v planu. Taka razlaga plana je seveda povsem

Večina XAI metod generira razlago posameznih odločitev oz. logična. Navadno pa vsebuje preveč podrobnosti. Če plan

klasifikacij. Pri razložljivem planiranju pa gre za razlago vsebuje nekoliko večje število akcij, npr. nekaj 10, postane tako

množice odločitev (npr. zaporedja akcij, ki robota vodi do cilja). podrobna razlaga spet težko razumljiva in za uporabnika

Posebej za razlago planov se je formiralo področje razložljivega neprivlačna. V tem primeru bi za sprejemljivo razlago treba plan

planiranja [10]. razbiti v hierarhično strukturo, definirano s podcilji plana.

Razlaga planov je navadno zahtevnejša od razlage v Odkrivanje smiselnih podciljev pa je lahko težavno. Poseben

klasifikacijskih problemih. Treba je razložiti, kako so posamezne izziv je, kako poiskati take podcilje, ki rezultirajo v razlagi, ki je

akcije odvisne od drugih, da skupaj rešijo nalogo. Primer razlage za človeka čim bolj naravna. zaporedja odločitev je razlaga šahovskih partij, kjer je treba



razložiti celo zaporedje potez ali drevo odločitev, ki definira Zahvala uspešno strategijo. Primer, opisan v [9], so težko razumljive in Prispevek je nastal v okviru ciljnega raziskovalnega projekta V2-briljantne poteze šahovskega programa AlphaZero. 2272 Opredelitev okvira za zagotavljanje zaupanja javnosti v Razlaga planov je aktualna tudi na področju vodenja sisteme umetne inteligence in njihove uporabe, ob podpori Javne sistemov. Lep primer razlage naučenega plana vodenja je v [32]. agencije za raziskovalno in inovacijsko dejavnost Republike Gre za klasično nalogo iz teorije vodenja sistemov: vodenje Slovenije in Ministrstva za digitalizacijo. sistema voziček-palica. Na vozičku je vrtljivo vpeta palica.



Palica je postavljena približno vertikalno, vendar se, če ne Literatura ukrepamo, prevrne na tla. S potiskanjem vozička levo oz. desno

je treba loviti ravnotežni položaj palice okrog vertikale, obenem [1] Ali, S., Abuhmed, T., El-Sappagh, S., Muhammad, K., Alonso-Moral, J.

M., Confalonieri, R., Guidotti, R., Del Ser, J., Díaz-Rodríguez, N. and

pa doseči, da se voziček horizontalno premakne iz začetnega Herrera, F. 2023. Explainable Artificial Intelligence (XAI): What we

položaja do cilja. Naučena strategija vodenja je lepo razložljiva. know and what is left to attain Trustworthy Artificial Intelligence.

Information Fusion, 99. DOI:

Najprej nekoliko presenetljivo potisnemo voziček stran od cilja, https://doi.org/10.1016/j.inffus.2023.101805.



s čimer dosežemo, da se palica nagne proti cilju. To omogoči [2] Ai, L., Langer, J., Muggleton, S. H. and Schmid, U. 2023. Explanatory machine learning for sequential human teaching. Machine Learning potiskanje vozička v smer proti cilju, hkrati pa se ohranja Journal, 112 , 3591-3632. DOI: https://doi.org/10.1007/s10994-023-

ravnotežje palice, ko je ta nagnjena naprej v smeri cilja. 06351-8.

[3] Alvarez-Melis, D and Jaakkola, T. S. 2018. Towards Robust

Interpretability with Self-Explaining Neural Networks. 32nd Conference



on Neural Information Processing Systems (NeurIPS 2018), Montreal, [20] Hsieh, C. Y., Yeh, C. K., Liu, X., Ravikumar, P., Kim, S., Kumar, S. and Canada. Hsieh, C. J. 2021. Evaluations and Methods for Explanation through

[4] Amgoud, L. and Ben-Naim, J. 2022. Axiomatic Foundations of Robustness Analysis. arXiv:2006.00442. DOI:

Explainability. Proceedings of the Thirty-First International joint https://doi.org/10.48550/arXiv.2006.00442. Conference on Artificial Intelligence (IJACAI-22), 636-642. [21] Ignatiev, A., Narodytska, N. and Marques-Silva, J. 2019. On Validating,

[5] Barredo Arrieta, A., Diaz-Rodríguez, N., Del Ser, J., Bennetot, A., Tabik, Repairing and Refining Heuristic ML Explanations. arXiv: 1907.02509.

S., Barbado, A., García, S., Gill-López, S., Molina, D., Benjamins, R., DOI: https://doi.org/10.48550/arXiv.1907.02590. Chatila, R. and Herrera, F. 2019. Explainable Artificial Intelligence (XAI): [22] Kolmogorov, A. 1963. On Tables of Random Numbers. The Indian Concepts, Taxonomies, Opportunities and Challenges toward Responsible Journal of Statistics, Series A, 25, 369-375. AI. arXiv: 1910.10045v2. DOI: [23] Krishnan, M. 2020. Against Interpretability: A Critical Examination of the

https://doi.org/10.48550/arXiv.1910.10045. Interpretability Problem in Machine Learning. Philosophy & Technology,

[6] Bohanec, M. and Bratko, I. 1994. Trading accuracy for simplicity in 33(3), 487-502. DOI: http://dx.doi.org/10.1007/s13347-019-00372-9.

decision trees. Machine Learning Journal, 15, 223-250. DOI: [24] Lombrozo, T. 2006. The structure and function of explanation. Trends in

https://doi.org/10.1007/BF00993345. Cognitive Science, 10(10), 464-470. DOI:

[7] Brasse, J., Broder, H. R., Förster, M., Klier, M. and Sigler, I. 2023. https://doi.org/10.1016/j.tics.2006.08.004.

Explainable artificial intelligence in information systems: A Review of the [25] Malle, B. F. 2004. How the Mind Explains Behavior: Folk Explanations, status quo and future research directions. The International Journal of Meaning, and Social Interaction. MIT Press.

Networked Business, 33(1). DOI: https://dx.doi.org/10.1007/s12535-023- [26] Michie, D. 1998. Machine Learning in the next five years. Proceedings of

00644-5. the 3rd European working session on learning, 107-122.

[8] Bratko. I. 1997. Machine learning: between accuracy and interpretability. [27] Mohseni, S., Block, J. E. and Ragan, E. 2021. Quantitative Evaluation of

V: Learning, Networks and Statistics th (ed. Della Riccia, G.), Vienna: Machine Learning Explanations: A Human-Grounded Benchmark. 26

Springer. International Conference on Intelligent User Interfaces (IUI’21). DOI:

[9] Bratko, I. 2018. AlphaZero: what's missing? Informatica, 42(1). https://doi.org/10.1145/3397481.3450689.

[10] Chakraborti, T., Sreedharan, S. and Kambhampati, S. 2020. The Emerging [28] Muggleton, S., Schmid, U., Zeller, C., Tamaddoni-Nezhad, A. and Besold,

Landscape of Explainable Automated Planning & Decision Making. T. 2018. Ultra-strong machine learning: Comprehensibility of programs

arXiv:2002.11697. DOI: https://doi.org/10.48550/arXiv.2002.11697. learned with ILP. Machine Learning, 107, 1119-1140. DOI:

[11] Chen, Z., Subhash, V., Havasi, M., Pan, W. and Doshi-Velez, F. 2022. https://doi.org/10.1007/s10994-018-5707-3.

What Makes a Good Explanation?: A Harmonized View of Properties of [29] Ribeiro, M. T., Singh, S. and Guestrin, C. 2016. “Why Should I Trust Explanations. Progress and Challenges in Building Trustworthy You?”: Explaining the Predictions of Any Classifier. Proceedings of the Embodies AI (TEA 2022) co-located with NeurIPS 2022. nd 22 ACM SIGKDD International Conference on Knowledge Discovery

[12] Doshi-Velez, F. and Kim, B. 2017. Towards A Rigorous Science of and Data Mining. DOI: https://doi.org/10.1145/2939672.2939778.

Interpretable Machine Learning. arXiv:1702.08608. DOI: [30] Ribera, M. and Lapedriza, A. 2019. Can we do better explanations? A

https://doi.org/10.48550/arXiv.1702.08680. proposal of User-Centered Explainable AI. IUI Workshops ’19.

[13] Edmonds, M., Gao, F., Liu, H., Xie, X., Qi, S., Rothrock, B., Zhu, Y., Wu, [31] Rudin, C. 2019. Stop explaining black box machine learning models for

Y. N., Lu, H. and Zhu, S. C. 2019. A tale of two explanations: Enhancing high stakes decisions and use interpretable models instead. Nature

human trust by explaining robot behavior. Science Robotics, 4(3), 1-13. Machine Intelligence, 1(5), 206-215. DOI:

[14] Gilpin, L. H., Yuan, B. Z., Bajwa, A., Specter, M and Kagal, L. 2019. https://doi.org/10.1038/s42256-019-0048-x.

Explaining Explanations: An Overview of Interpretability of Machine [32] Šoberl, D. and Bratko, I. 2023. Transferring a Learned Qualitative Cart-Learning. arXiv:1806.00068. DOI: Pole Control Model to Uneven Terrains. International Conference on

https://doi.org/10.48550/arXiv.1806.00069. Discovery Science, 446-459. DOI: http://dx.doi.org/10.1007/978-3-031-

[15] Grice, H. P. 1975. Logic and conversation, syntax and semantics. Speech 45275-8_30.

Acts 3, 41-58. [33] Tjoa, E. and Guan, C. 2020. A Survey on Explainable Artificial

[16] Guidotti, D., Monreale, A., Ruggieri, S. and Turini, F. 2018. A Survey of Intelligence (XAI): Towards Medical XAI. arXiv: 1907.07374. DOI:

Methods for Explaining Black Box Models. ACM Computing Surveys, http://doi.org/10.48550/arXiv.1907.07374.

51(5), 1-42. DOI: https://doi.org/10.1145/3236009. [34] Yeh, C. K. and Ravikumar, P. 2021. Objective criteria for explanations of

[17] Gunning, D., Vorm, E., Wang, J. Y. and Turek, M. 2021. DARPA’s machine learning models. Applied AI Letters published by John Wiley &

explainable AI (XAI) program: A retrospective. Applied AI Letters, 2(4). Seons Ltd. DOI: https://doi.org/10.1002/ail2.57.

DOI: https://doi.org/10.1002/ail2.61. [35] Zhang, Z., Xu, L., Yilmaz, L. and Liu, B. 2021. A Critical Review of

[18] Hall, P., Ambati, S. and Phan, W. 15.3.2017. Ideas on interpreting Inductive Logic Programming Techniques for Explainable AI. arXiv:

machine learning. O’Reilly. https://www.oreilly.com/radar/ideas-on- 2112.15319. DOI: https://doi.org/10.48550/arXiv.2112.15319.

interpreting-machine-learning/. [36] Zhou, B., Khosla, A., Lapedriza, A., Oliva, A. and Torralba, A. 2016.

[19] Holsinger, A., Saranti, A., Molnar, C., Biecek, P. and Samek, W. 2022. Learning Deep Features for Disciminative Localization. 2016 IEEE

Explainable AI Methods – A Brief Overview. xxAI 2020, LNAI 13200, Conference on Computer Vision and Pattern Recognition (CVPR), 2921-

13-38. DOI: https://doi.org/10.1007/978-3-031-04083-2_2. 2929. DOI: 10.1109/CVPR.2016.319.





Exploring Cognitive Science under Analytical Idealism




Grega Rodman

Faculty of Education

University of Ljubljana

gr6506@student.uni-lj.si



Abstract materialism is waning, the legacy persists, having limited the scope of inquiry, particularly in the study of consciousness, by

In modern science, materialism has played a significant role, ignoring subjective human experience [1, 2].

positing that matter is the fundamental reality and that all

phenomena, including consciousness, can be understood

through physical processes. However, recent evidence suggests 2 Questioning the materialistic paradigm

materialism might not fully explain all phenomena. These

findings have led to the rise of a post-materialistic movement At its core, science is a non-dogmatic, open-minded approach to

exploring new ideas. One such idea, Analytical Idealism, acquiring knowledge about nature through observation,

proposed by Bernardo Kastrup, suggests that consciousness is experimental investigation, and theoretical explanation of

the fundamental reality and that the material world is a phenomena [3]. There is a misconception that the methodology

reflection of this universal consciousness. The implications of of science is inherently tied to materialism. In addition, an

adopting this approach in science will be explored. increasing body of empirical evidence points to the limitations

of materialism. Of course, it is impossible to provide sufficient



Keywords empirical research that definitively refutes materialism;

however, we can present two examples from different fields

ontology, methodology, materialism, analytical idealism that suggest the limitations of materialism: one from animal

cognition and the second from psi phenomena in humans.



1 Introduction 2.1 Example from animal cognition



The modern scientific worldview is largely based on The first example is from animal cognition. Actually, it is about assumptions closely linked to classical physics. Among these is precognition, which is the perception of future events, typical materialism, which posits that matter constitutes the sole reality. for some animal species. Investigation in this field was done by In the 19th century, these assumptions became increasingly Sheldrake [4], who studied a dog that seemed to know when its rigid, evolving into dogmas that coalesced into the ideological owner was coming home. Despite using various methods to rule framework known as "scientific materialism" [1]. out normal senses, Sheldrake consistently observed the dog

waiting expectantly before the owner arrived, but not at other

Scientific materialism is a philosophical viewpoint that asserts

times. A replication of a similar experiment by some sceptics

that all phenomena in the universe, including consciousness and

was declared unsuccessful [5], but a later reanalysis of the same

human experience, can be explained solely through physical

results showed the opposite [6].

processes and interactions. Throughout the 20th century,



scientific materialism became the prevailing ideology in 2.2 Example from psi phenomena in humans academic circles, to the extent that the majority of scientists

came to believe it was the only rational interpretation of the

world. Scientific methods rooted in materialistic philosophy The second example addresses meta-analyses of psi

have proven highly successful in enhancing our understanding phenomena, which are defined as extraordinary human

of nature and in providing greater control and freedom through capacities like telepathy, clairvoyance, and precognition that

technological advances. Though the popularity of scientific involve gaining information without known sensory Permission to make digital or hard copies of part or all of this work for personal or mechanisms. Studies investigating these phenomena have

classroom use is granted without fee provided that copies are not made or consistently found small but significant effects, suggesting that

distributed for profit or commercial advantage and that copies bear this notice and such abilities may exist [7]. The evidence for psi is comparable

the full citation on the first page. Copyrights for third-party components of this to that for established phenomena in psychology and other

work must be honored. For all other uses, contact the owner/author(s).

disciplines, although there is no consensual understanding of

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia them. Recent analyses also emphasize that these results cannot

be easily attributed to methodological flaws, selective reporting,

© 2024 Copyright held by the owner/author(s). or fraud, further supporting the plausibility of psi phenomena.

https://doi.org/10.70314/is.2024.cog.15



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia G. Rodman



The volume of empirical data indicating the shortcomings of develop through various contemplative practices. This field

materialism is so substantial that an increasing number of seeks to integrate traditional third-person scientific methods,

articles and books are being written on this subject [1, 7, 8, 9]. such as MRI, EEG, and MEG, with first-person accounts of

In fact, this has contributed to the emergence of a whole post- personal experiences in these altered states of consciousness

materialistic movement in recent decades, which is exploring [18, 19, 20]. When we start taking contemplative and

what this new paradigm might look like [3, 9, 10, 11, 12]. meditative practices seriously, science can begin to exchange

Believe it or not, you can also find a manifesto for post- ideas with ancient traditions such as Buddhism, Hinduism and

materialistic science [13]. others. Even this is already happening [21, 22, 23].



3.3 The consequences of such a research approach

3 Cognitive Science under Analytical



Idealism This interdisciplinary exchange highlights the potential for

scientific and spiritual perspectives to enrich each other and

One of the proponents of the post-materialistic movement is expand our understanding of consciousness and reality.

Bernardo Kastrup, who advocates for Analytical Idealism [3]. Additionally, to broaden scientific inquiry, spiritual practices

Analytical Idealism posits that consciousness is the fundamental like meditation and contemplation can be secularized and

essence of reality, rather than matter [12]. The focus of this incorporated into the scientific process. Fun fact, At the 6th

summary is not to provide a detailed description of Idealism, International Colloquium of Cognitive Sciences, Dr. Berkovich-

but rather to explore the potential changes in the methodology Ohana began her presentation titled "Meditation and the Self:

of scientific research that could result from adopting this Neuroscience and Phenomenology" with a few minutes of

perspective. guided meditation [24]. By integrating these practices, scientists

could benefit from improved mental hygiene, enhanced

3.1 Two distinct routes to knowledge creativity, and increased cognitive capacities [25], [26], [27].



Changes in ontological views lead to changes in scientific This step can be highly significant, as it enhances the

methods as they alter the foundational concepts and performance of researchers. A greater focus may lead to

relationships that guide inquiry [14]. If consciousness is indeed reduced bias, while increased creativity fosters better

a fundamental aspect of reality rather than a byproduct of neural hypotheses, ultimately resulting in more effective research.

activity, it implies that consciousness might directly access Such advancements are essential for achieving substantial

aspects of reality without relying solely on sensory perception breakthroughs.

[15]. This leads us to consider two distinct routes to knowledge:

conventional sensory perception (science as it is mainly now) Engaging in meditation and/or contemplative practices poses

and a more direct introspective approach. Walach calls this potential downsides for scientists, too. First, the focus on

approach “radical introspection.” Radical introspection involves personal experience conflicts with the concurrent objective

a deep inward focus, often achieved through contemplative and standards required in scientific research. While self-research

meditative practices. Unlike standard qualitative introspection, can yield valuable insights, its subjective nature can lead to

which relies on external referents (e.g., transcripts, biases that undermine intersubjectivity. Furthermore, the

observations), radical introspection does not have such referents personal transformation that occurs during deep self-reflection

beyond personal experience. It faces challenges of subjective may distract researchers from maintaining the rigorous,

bias and lack of established methodology for validating truth detached perspective typically expected in scientific inquiry.

claims. However, it remains a crucial aspect of potential new Ultimately, the integration of such practices into mainstream

methodologies in science, requiring the development of science remains challenging, as it contrasts with the traditional

techniques to record, communicate, and verify first-person role of researchers.

experiences.



At this point, it is important to highlight that Walach is not the 4 Conclusion



only proponent of integrating radical introspection into In conclusion, I would like to emphasize a few key points. First,

scientific inquiry. Kordeš [16] arrives at a similar conclusion in the entire described methodology can, of course, be applied

his arguments, even though he does not refer the concept of from a materialistic standpoint as well. It is not the ontology

idealism at all. He suggests that in-depth, existentially liable itself that matters; rather, it is the methodology that enables

introspection and self-inquiry should be considered as serious insight. Materialists can also engage in contemplative

scientific research tools. neuroscience. Second, year by year, we have more scientific



3.2 Combining first and third person research studies suggesting that the current mainstream paradigm may be

flawed. Let us carefully examine the data and avoid dismissing



early attempts in that direction. The godfather of this approach practices, this would generally benefit the scientific community is, of course, Francisco J. Varela [17]. From this approach for reasons previously discussed. Fourth, when we establish a emerged the field of contemplative neuroscience, which connection between science and religion, mutual learning can When looking at current scientific practices, we can see some [28]. Third, if more scientists were to engage in meditation-like it simply because it contradicts our preconceived assumptions

explores individuals in altered states of consciousness that begin.



Exploring Cognitive Science under Analytical Idealism Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



Acknowledgments [14] Mukhles Al-Ababneh. 2020. Linking Ontology, Epistemology and Research Methodology. Science & Philosophy, 8(1), 75–91. doi:

I would like to thank Olga Markič for giving me the 10.5311/112222001789.

opportunity to explore this area of research and for mentoring

me throughout the process. [15] Harald Walach. 2020. Inner Experience – Direct Access to Reality: A

Complementarist Ontology and Dual Aspect Monism Support a Broader

Epistemology. Frontiers in Psychology, 11, Article 640. doi:



Authors' statement 10.3389/fpsyg.2020.00640.

ChatGPT-4 was used for improving the language of this paper. [16] Urban Kordeš. [2013]. Problems and Opportunities of First-Person

Research. Interdisciplinary Description of Complex Systems, 11(4):363-



References 375. https://doi.org/10.7906/indecs.11.4.2.

[1] [17] Varela, Francisco J. 1996. "Neurophenomenology: A Methodological B. Alan Wallace. 2004. The Taboo of Subjectivity: Toward a New

Science of Consciousness. Paperback, 323 pages. ISBN: Remedy for the Hard Problem." Journal of Consciousness Studies 3(4): 9780195173109. 330–349.



[2] [18] Mario Beauregard and Vincent Paquette. 2006. Neural correlates of a Frank, Adam, Marcelo Gleiser, and Evan Thompson. 2024. The Blind

Spot: Why Science Cannot Ignore Human Experience. Cambridge, mystical experience in Carmelite nuns. Neuroscience Letters, 405(3), Massachusetts: The MIT Press. Library of Congress Cataloging-in- 186–190. doi: 10.1016/j.neulet.2006.06.060. Publication Data.

[19] Yair Dor-Ziderman and Aviva Berkovich-Ohana. 2013. Mindfulness-

[3] induced selflessness: A MEG neurophenomenological study. Frontiers in Mario Beauregard, 2018. Toward a postmaterialist psychology: Theory,

research, and applications. New Ideas in Psychology, 50, 21–33. doi: Human Neuroscience, 7, 582. doi: 10.3389/fnhum.2013.00582. 10.1016/j.newideapsych.2018.02.004.

[20] Mario Beauregard and Vincent Paquette. 2006. Neural correlates of a

[4] mystical experience in Carmelite nuns. Neuroscience Letters, 405(3), Rupert Sheldrake and Pamela Smart. 2000. A dog that seems to know

when his owner is coming home: Videotaped experiments and 186–190. doi: 10.1016/j.neulet.2006.06.060. observations. Journal of Scientific Exploration, 14(2), 233–255.

[21] Markič, Olga, and Urban Kordeš. 2016. "Parallels between Mindfulness

[5] and First-Person Research into Consciousness." Asian Studies 4(2):153- Richard Wiseman, Matthew D. Smith, and Julie Milton. 1998. Can

animals detect when their owners are returning home? An experimental 168. https://doi.org/10.4312/as.2016.4.2.153-168. test of the 'psychic pet' phenomenon. British Journal of Psychology,

89(3), 453–462. doi: 10.1111/j.2044-8295.1998.tb02696.x. [22] Federman, Asaf. 2011. "What Buddhism Taught Cognitive Science

About Self, Mind and Brain." Enrahonar 47:39-62.

[6] https://doi.org/10.5565/rev/enrahonar/v47n0.162. Rupert Sheldrake. 2011. Dogs That Know When Their Owners Are

Coming Home: And Other Unexplained Powers of Animals. Three

Rivers Press, New York. [23] Vörös, Sebastjan. 2016. "Buddhism and Cognitive (Neuro)Science: An

Uneasy Liaison?" Asian Studies 4(1):61-80.

[7] https://doi.org/10.4312/as.2016.4.1.61-80. Cardeña, Etzel. 2018. "The Experimental Evidence for

Parapsychological Phenomena: A Review." American Psychologist

73(5): 663–677. https://doi.org/10.1037/amp0000236. [24] Aviva Berkovich-Ohana. 2022. 6to Coloquio Internacional de Ciencias

Cognitivas. YouTube video. Retrieved August 26, 2024, from

[8] https://www.youtube.com/watch?v=gJHSSZikyvQ. Charles Whitehead. [2019]. Charles Tart: The End of Materialism: How

Evidence of the Paranormal is Bringing Science and Spirit Together

(book review). Journal of Consciousness Studies, 17(11/12):202. [25] Richard J. Davidson and Jon Kabat-Zinn. 2003. Alterations in brain and

immune function produced by mindfulness meditation. Psychosomatic

Medicine, 65(4), 564–570. doi: 10.1097/01.PSY.0000077505.67574.E3.

[9] Harald Walach. 2019. Beyond a Materialist Worldview: Towards an

Expanded Science. Lulu.com. ISBN: 978-1716805998.

[26] Eileen Luders and Christian Gaser. 2008. The underlying anatomical

correlates of long-term meditation: Larger hippocampal and frontal

[10] Gary E. Schwartz. 2016. What is the nature of a post-materialist

volumes of gray matter. NeuroImage. doi:

paradigm? Three types of theories. Explore. doi:

10.1016/j.neuroimage.2008.12.061.

10.1016/j.explore.2015.12.002.

[27] Amishi P. Jha, Jason Krompinger, and Michael J. Baime. 2007.

[11] Frederick T. Travis. 2020. Consciousness is primary: Science of Mindfulness training modifies subsystems of attention. Cognitive,

consciousness for the 21st century. International Journal of Affective, & Behavioral Neuroscience, 7, 109–119. doi: Psychological Studies, 13(1), 1. doi: 10.5539/ijps.v13n1p1. 10.3758/CABN.7.2.109.



[12] Bernardo Kastrup. 2018. The next paradigm. Future Human Image, 9. [28] Williams, Bryan J. 2019. "Reassessing the 'Impossible': A Critical

doi: 10.29202/fhi/9/4. Commentary on Reber and Alcock’s 'Why Parapsychological Claims

Cannot Be True'." Journal of Scientific Exploration 33(4):599-616.

[13] Mario Beauregard, Gary E. Schwartz, Lisa Miller, and Larry Dossey. https://doi.org/10.31275/2019/1667.

2014. Manifesto for a post-materialist science. Explore, 10(5). doi:

10.1016/j.explore.2014.06.008.



Intelligent Revolution –



a New Civilization and Cognitive Era



Inteligenčna revolucija –



nova civilizacijska in kognitivna doba



Matjaž Gams



Odsek za inteligentne sisteme



Institut “Jožef Stefan”

Jamova cesta 39, 1000 Ljubljana



Slovenija



dosežek je razvoj modelov GPT, ki so postavili nove standarde v

ABSTRACT. generiranju jezika in razširili obzorja potenciala UI. Ta članek

The rapid advancement of artificial intelligence is significantly preučuje vpliv UI na različne sektorje, vključno z družbenim in

enhancing human capabilities, even as human progress itself individualnim kognitivnim napredkom, ter poudarja tako

appears to stagnate, hindered by decadent ideologies and adverse priložnosti kot izzive, ki jih prinaša široka uporaba UI. Razprava

societal trends. Over the past few decades, AI has achieved se osredotoča na transformativno moč UI tehnologij ter na etične,

remarkable milestones, from mastering complex games to ekonomske, kognitivne in družbene posledice te tekoče

revolutionizing industries such as healthcare and finance through revolucije.

advancements in machine learning, natural language processing,

and robotics. A particularly notable achievement is the Medtem ko UI še naprej spodbuja inovacije in preoblikuje

development of GPT models, which have set new standards in industrije, se ljudje vse bolj integrirajo s temi tehnologijami prek

language generation and expanded the horizons of AI’s potential. vseprisotne uporabe pametnih telefonov, osebnih računalnikov

This paper examines the impact of AI on various sectors, in nosljivih naprav. Ta integracija je že okrepila naše kognitivne

including societal and individual cognitive advancements, in funkcionalne sposobnosti ter učinkovito pomnožila človeški

highlighting both the opportunities and challenges of widespread potencial. Ko vpliv UI na človeško življenje narašča, se

AI adoption. The discussion focuses on the transformative power pojavljajo ključna vprašanja o prihodnosti tega simbiotičnega

of AI technologies and the ethical, economic, cognitive, and razmerja in o smeri družbenega napredka. social implications of this ongoing revolution. As AI continues



to drive innovation and transform industries, humans are KLJUČNE BESEDE increasingly integrating with these technologies through the

pervasive use of smartphones, personal computers, and wearable Umetna inteligenca, simbioza med človekom in strojem,

devices. This integration has already enhanced our cognitive and tehnološki napredek, kognitivni vpliv

functional capabilities, effectively amplifying human potential.

However, as AI’s influence on human life deepens, critical KEYWORDS

questions arise about the future of this symbiotic relationship and Artificial Intelligence, Human-Machine Symbiosis,

the trajectory of societal progress. Technological Advancement, Cognitivel Implications



POVZETEK

Hitri napredek umetne inteligence izboljšuje človeške zmožnosti, 1 Introduction

medtem ko se zdi, da človeški napredek stagnira, oviran z The rapid development of artificial intelligence (AI) is



zadnjih desetletjih je UI dosegla izjemne mejnike, od human capabilities. While AI progresses at an unprecedented obvladovanja kompleksnih iger do revolucije v panogah, kot sta pace, human societal progress appears to be stagnating, dekadentnimi ideologijami in negativnimi družbenimi trendi. V transforming industries and redefining/improving the essence of



obdelave naravnega jezika in robotike. Posebej pomemben increasingly entangled in decadent ideologies and negative zdravstvo in finance, z napredki na področju strojnega učenja,

societal trends. This duality between AI's rise and human inertia

∗Article Title Footnote needs to be captured as Title Note is critical to understanding the current technological landscape.

†Author Footnote to be captured as Author Note Over recent decades, AI has achieved remarkable milestones,

Permission to make digital or hard copies of part or all of this work for personal or including mastering complex games like Go and chess,

classroom use is granted without fee provided that copies are not made or distributed advancing natural language processing (NLP), and driving

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must significant innovations in sectors such as healthcare and finance.

be honored. For all other uses, contact the owner/author(s). For example, the development of GPT (Generative Pre-trained

© 2024 Copyright held by the owner/author(s). Transformer) models represents a breakthrough in AI’s ability to Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

https://doi.org/10.70314/is.2024.cog.16



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia M. Gams



generate human-like text, setting new standards for machine The release of GPT-3 by OpenAI in 2020 marked a

language generation and understanding [1, 2]. significant advance in the field of NLP. GPT-3, with 175 billion

This paper examines the transformative potential of AI across parameters, demonstrated unprecedented language generation

various sectors, including healthcare, finance, education, and capabilities, performing tasks such as translation, summarization,

entertainment, highlighting both the opportunities and challenges and question-answering with high proficiency and minimal fine-

that accompany widespread AI adoption. In recent years, tuning. It set a new benchmark for the potential of AI in creative

significant advancements in AI have occurred at an accelerating and linguistic tasks [9].

rate. For example, breakthroughs in reinforcement learning and 2021: DeepMind’s AlphaFold 2 in Protein Folding

unsupervised learning have expanded the capabilities of AI In 2021, AlphaFold 2, developed by DeepMind, solved one

systems, with applications ranging from autonomous vehicles to of biology's greatest challenges by predicting protein structures

sophisticated recommendation systems [3]. Additionally, the with remarkable accuracy. This breakthrough has significant

ethical, economic, cognitive, and social implications of AI's implications for drug discovery, understanding diseases, and

proliferation are increasingly coming to the forefront, as debates designing new biological processes, demonstrating AI's potential

intensify over issues such as algorithmic bias, privacy, and the to revolutionize the life sciences [11].

potential for AI to displace human jobs [4]. These discussions 2022: DALL-E 2 and Image Generation

underscore the need for robust governance frameworks to ensure OpenAI's DALL-E 2, released in 2022, demonstrated the

that AI technologies are developed and deployed responsibly [5]. power of AI in generating highly detailed and creative images

As AI technologies like machine learning, NLP, and robotics from text descriptions. This model pushed the boundaries of AI

continue to evolve, they increasingly integrate into human life, in the visual domain, showcasing its ability to combine artistic

augmenting cognitive and functional abilities through ubiquitous creativity with technical precision, and opening new possibilities

technologies such as smartphones, personal computers, and in design, marketing, and entertainment [12].

wearables. This integration, often referred to as a form of human- 2023: GPT-4 and Multimodal AI

AI symbiosis, has already multiplied human potential, enabling In 2023, OpenAI introduced GPT-4, which expanded the

tasks and processes that were previously unimaginable [6]. The capabilities of its predecessor by being multimodal—able to

consequences of this symbiotic relationship are profound, raising process and generate both text and images. GPT-4's ability to

critical questions about the direction of societal progress and the handle complex queries across different formats has made it a

future of humanity as AI [7, 8] becomes more embedded in powerful tool for applications in education, customer service,

everyday life. and creative industries, further blurring the distinction between

We analyze the implications of AI’s rapid development, products of human and machine intelligence [13].

particularly the considerations that must be addressed to navigate The last five years have seen groundbreaking AI

the ongoing AI revolution effectively. By integrating recent achievements each year that have pushed the boundaries of what

scholarly insights with a broader analysis of AI's impact, this AI can do. From mastering strategic games and understanding

paper seeks to contribute to the understanding of how AI is protein structures to generating high-quality text and images,

reshaping industries and human capabilities, as well as the future AI's progress continues to accelerate, bringing us closer to a

trajectory of this unprecedented technological evolution. future where AI plays an integral role in nearly every aspect of

society. In the next section, we examine human progress.



2 AI progress

3 Impact of AI progress across various fields

A recent and transformative achievement is the development

of Generative Pre-trained Transformers. These models represent As AI continues to evolve, its influence is expected to

a leap forward in natural language processing, capable of permeate multiple sectors, driving innovation and transformation.

generating human-like text, translating languages, and even This section analyzes the potential impact of AI across key fields

writing code. The GPT-3 model, released by OpenAI in 2020, is such as healthcare, finance, education, entertainment, and

particularly notable for its ability to perform a wide range of tasks transportation, highlighting both the opportunities and

with minimal input, showcasing the power and versatility of challenges these advancements may bring.

large-scale language models [9]. Healthcare: AI has the potential to revolutionize healthcare

Here we present one major achievement over the last five by improving diagnostics, personalized medicine, and patient

years, having in mind the constant AI progress in areas like care. Machine learning algorithms are already being used to

autonomous driving or pattern recognition: analyze medical images with greater accuracy than human

2019: AlphaStar in Real-Time Strategy Games radiologists, and AI-driven predictive analytics are helping to

In 2019, DeepMind's AlphaStar achieved a significant identify at-risk patients before conditions worsen. Additionally,

milestone by reaching the top players of professional StarCraft II AI can streamline administrative processes, reducing the burden

play, a complex real-time strategy game that requires long-term on healthcare professionals and allowing for more efficient

planning, resource management, and real-time decision-making. patient management. The integration of AI in healthcare is

This achievement underscored the potential of AI to operate in expected to lead to better patient outcomes, lower costs, and a

dynamic and highly strategic environments, far beyond turn- more proactive approach to health management [14]. The JSI

based games like Go [10]. team is in the last phases of donating a home doctor system to all

2020: GPT-3 Slovenians.



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia M. Gams



Finance: The finance industry is experiencing significant which allowed early humans to travel long distances and use their

transformations due to AI, particularly in areas such as hands for tool-making. The invention of tools around 2.6 million

algorithmic trading, risk management, and fraud detection. AI years ago further distinguished our ancestors, enabling them to

algorithms can analyze vast amounts of financial data in real- manipulate their environment in unprecedented ways.

time, enabling more informed and faster decision-making. These Approximately 200,000 years ago, Homo sapiens emerged,

technologies also enhance the accuracy of credit scoring and equipped with greater cognitive abilities and complex language,

personalized financial advice, offering tailored solutions to facilitating social structures and cultural developments that set

individual customers. However, the increased reliance on AI also the stage for future innovations. The advent of agriculture around

raises concerns about market stability, ethical use of data, and the 10,000 years ago marked a fundamental shift in human society,

potential for systemic risks [15]. leading to settled communities and the eventual rise of

Education: AI is poised to transform education by providing civilizations.

personalized learning experiences, automating administrative Fast forward to the Industrial Revolution in the 18th century,

tasks, and enabling new forms of interactive learning. AI-driven human progress accelerated dramatically. Innovations in

adaptive learning systems can tailor educational content to the machinery, transportation, and communication reshaped

needs of individual students, allowing for more effective learning societies, laying the groundwork for the Information Age. The

outcomes. Additionally, AI can assist teachers by automating 20th century saw rapid technological advances, including the

grading and providing real-time feedback, freeing up more time development of the computer, the internet, and the beginnings of

for personalized instruction. GPTs further offer significant artificial intelligence, all of which have profoundly impacted

improvements in education. Integrating AI in education also human life.

presents challenges, such as ensuring equitable access to AI-

driven tools and addressing concerns about data privacy [16]. At 4.2 Recent progress: merging with ICT and AI

JSI, we tested the quality of various GPTs on educational tasks. In the past few decades, the convergence of information and

Entertainment: The entertainment industry is undergoing a communication technologies (ICT) and artificial intelligence has

significant shift due to AI's capabilities in content creation, fundamentally altered the trajectory of human progress. This

recommendation systems, and audience engagement. AI- merger has enhanced human capabilities and begun to blur the

generated music, art, and scripts are becoming increasingly lines between human and machine intelligence, creating a

sophisticated, thus differentiating between products of human symbiotic relationship reshaping society.

and machine creativity as often impossible. Recommendation Mobile Phones: One of the most transformative technologies

algorithms, powered by AI, personalize content delivery to users, of the late 20th and early 21st centuries is the mobile phone.

enhancing their experience and increasing engagement. However, Introduced commercially in the 1980s, mobile telephones rapidly

this rise in AI-generated content raises questions about the future evolved from simple communication devices to powerful,

of human creativity and the potential for AI to disrupt traditional multifunctional tools. The advent of smartphones in the 2000s,

content production models [17]. Recommendation algorithms with their integration of internet access, GPS, and a multitude of

were one of the central parts of the H2020 smart-city Urbanite applications, significantly enhanced human connectivity and

project with most of the software developed at AI. access to information. Today, smartphones are essential tools for

Transportation: AI is driving innovation in transportation both personal and professional life, facilitating real-time

through the development of autonomous vehicles, smart traffic communication, social networking, and a vast array of digital

management systems, and predictive maintenance. Self-driving services [19].

cars, powered by AI, promise to reduce accidents, lower The Internet and Cloud Computing: The development of

emissions, and increase the efficiency of transportation networks. the Internet in the late 20th century and the rise of cloud

AI can also optimize traffic flow and reduce congestion through computing in the early 21st century have revolutionized how

real-time data analysis and adaptive traffic control systems. humans interact with information and each other. The internet

However, the widespread adoption of AI in transportation faces has democratized access to knowledge, enabling global

challenges related to safety, regulatory frameworks, and public communication and collaboration, while cloud computing has

acceptance [18]. made vast computational resources and storage available to

In the next section, human progress and integration with AI individuals and organizations alike. These technologies have

are presented. increased productivity and laid the foundation for the widespread

deployment of AI systems, which rely on large datasets and

4 Human progress including merging with ICT significant computational power [20].



and AI advancements in AI, particularly with the development of GPTs, Generative Pre-trained Transformers: The recent

represent a significant leap in the merging of human capabilities

4.1 Historical overview of human progress with machine intelligence. GPT-3, introduced in 2020,

Human progress is a story of relentless evolution and demonstrated the ability to generate coherent and contextually

technological advancement, spanning millions of years. relevant text based on minimal input, performing a wide range of

Beginning around six million years ago, the earliest hominins tasks such as translation, summarization, and even creative

diverged from the common ancestor we share with chimpanzees, writing. GPT-4, released in 2023, expanded on this by

marking the start of a journey toward modern humanity. One of incorporating multimodal capabilities, processing both text and

the earliest major milestones was the development of bipedalism, images, and further enhancing human-machine interaction [13].

These models are not just tools but extensions of human

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia M. Gams



cognitive abilities, enabling users to perform tasks that require deteriorating, the measure of human intellectual capacity should

complex reasoning and linguistic skills with the assistance of AI. now include not just our innate abilities but also the external

Wearable Technology and Augmented Reality: Wearable systems that augment them [19].

devices, such as smartwatches and fitness trackers, have The view that technology significantly enhances human

integrated AI into daily life, monitoring health metrics and cognitive ability aligns with Harari’s notion of humans becoming

providing real-time feedback to users. These devices exemplify "cyborgs" as they increasingly rely on tools that supplement

the merging of human biology with technology, offering new mental processes. Similarly, [22] discusses the extended mind

ways to enhance physical and cognitive performance. theory, which posits that external tools, such as smartphones, are

Augmented reality (AR) technologies are also becoming integral components of the human cognitive system, challenging

increasingly prevalent, overlaying digital information onto the the notion that brain size or biological limitations strictly define

physical world and creating immersive experiences that enhance mental capacity. These technological extensions of human

learning, navigation, and entertainment [21]. cognitive ability have created new frameworks for evaluating our

intellectual potential, making it more accurate to assess human

functionality in a combined system of biological and

technological entities.

The emergence of human-like properties such as

consciousness [23], observed in advanced GPT models,

represents a pivotal step in the evolution of artificial intelligence

Intelligence better predictions, and human progress. These models, capable of understanding

better reaction. More energy and generating natural language, are beginning to mimic forms

consumption 25 . of cognitive processes, thus contributing to what could be

principle. described as the dawn of a new "intelligent era." This era, driven Sexual attraction Anthropological

Tool evolution by AI's ever-increasing capabilities, promises a deeper

I TE IGE CE WI S integration between human cognition and machine intelligence,

Comple issue potentially fostering innovations in problem-solving, creativity,

and the expansion of knowledge.

Researchers like David Chalmers have explored the idea that

AI systems, such as GPT models, may embody elements of

extended cognition, which can extend human cognitive abilities

beyond their biological limits. The more these models evolve, the

more they may contribute to an era where AI complements



human intelligence in unprecedented ways, leading to new forms





Source Figure 1: Progress of the human brain and intelligence. of civilization that heavily rely on intelligent systems to solve complex global challenges [7, 22]. of the draft (modifications original): Figure 1 illustrates the functional growth of human problem- https://www.aquatic-human-ancestor.org/anatomy/brain.html solving capabilities, driven by the integration of ICT and AI



solutions, which serve as amplifiers of natural intelligence.



Human progress, from the earliest hominins to the modern Human cognitive abilities are being multiplied several times

age, has been marked by the continuous development and through this merger with ICT and AI, as represented by the blue

integration of technologies that enhance human capabilities. In arrow. While the original figure without the blue arrow shows

recent decades, the merging of ICT and AI with human activities the increase in human skull volume, and thus brain size, the blue

has accelerated this progress, creating a symbiotic relationship arrow highlights the exponential growth in problem-solving



Technologies such as mobile phones, the internet, GPT models, walking barefoot versus using a car or plane. The speed of movement changes dramatically, even though the human’s and wearable devices have not only transformed how we live and that extends human cognitive and physical abilities. capacity. A simple analogy can be drawn: consider a person



work but have also set the stage for future advancements that may physical body remains unchanged. Similarly, while human biology (the brain) did not improve, the ability to tackle complex further blur the lines between human and machine intelligence. tasks surged drastically with the aid of ICT and AI The key idea of this overview is that humans and user

interfaces are already deeply intertwined, both through physical

hardware like mobile phones and more abstract systems such as 5 DISCUSSION

software and information networks. As these systems fast

become more sophisticated, they enhance human cognitive, In recent years, there has been growing concern that Western

intellectual, and mental functions, effectively expanding our civilization is experiencing a period of decline, marked by

capacities. This growing interconnection between humans and political fragmentation, cultural disintegration, and economic

technology echoes the concept of cyborgization, where external challenges. Scholars have pointed to a loss of social cohesion,

devices supplement and expand the functions of the human mind. declining institutional trust, and the rise of non-productive and

The development of smartphones, wearable technology, and conflicting ideologies [24, 25].

even future brain-computer interfaces suggests that this synergy On the other hand, the rapid advancements in artificial

is only deepening. Contrary to claims that our brain size is intelligence are driving unprecedented changes across various



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia M. Gams



fields, fundamentally altering the landscape of industries and industries where automation can replace human labor. This shift

society. As AI technologies continue to evolve, they offer both could lead to widespread job displacement, necessitating a

tremendous opportunities and significant challenges that require rethinking of economic structures, education systems, and social

careful consideration. safety nets to address the needs of a rapidly changing labor

Balancing Innovation with Ethical Concerns: One of the market [16].

primary discussions surrounding AI is the balance between The potential for AI to drive social and economic inequality

innovation and ethical considerations. AI has the potential to is a pressing concern. Without proactive measures to ensure

revolutionize fields like healthcare, finance, education, and equitable access to AI technologies and to address the disparities

transportation by improving efficiency, accuracy, and that may arise from AI-driven economic shifts, society risks

personalization. However, these advancements also raise critical deepening existing divides. One of the best solutions is

ethical questions, particularly regarding data privacy, introducing the AI courses already in elementary schools.

algorithmic bias, and the potential for AI to perpetuate or avigating the AI Revolution: As AI continues to advance,

exacerbate existing inequalities. For example, while AI-driven society is at a crossroads, faced with the task of navigating the

personalized medicine can enhance healthcare outcomes, it also complexities of the AI revolution. The potential benefits of AI

risks marginalizing those without access to the necessary are immense, but there are also certain risks. Ensuring that AI

technology or data [14]. technologies are developed and deployed responsibly will

Moreover, the use of AI in finance, particularly in areas like require a concerted effort from governments, industry, academia,

algorithmic trading and credit scoring, has the potential to deepen and civil society. This includes developing ethical guidelines,

economic disparities if not carefully regulated. The challenges of regulatory frameworks, and educational initiatives that can help

ensuring fairness, transparency, and accountability in AI systems society adapt to the changes brought about by AI. At the same

are significant and demand robust governance frameworks to time, these regulations should first of all enhance proper progress,

prevent misuse or unintended consequences [15]. research and development, and not pose additional bureaucratic

The Human-AI Symbiosis: Another crucial aspect of the burdens.

discussion is the growing symbiosis between humans and AI. As In cognitive terms, GPT models represent a promising

humans increasingly rely on AI technologies in daily life— approach to creating forms of artificial consciousness and

through smartphones, wearables, and AI-powered applications— cognitive information beings. These models simulate aspects of

there is a merging of human and machine capabilities. This human cognition, such as language understanding and generation,

integration has the potential to significantly enhance human by mimicking neural networks that resemble the processing of

cognitive and physical abilities, leading to what some describe as human brains. As they evolve, GPTs could potentially help us

an augmented human experience. However, this symbiosis also explore and understand the fundamental components of human

raises questions about dependency, control, and the future of consciousness, offering insights into both artificial and human

human autonomy. As AI systems become more embedded in cognition [26].

decision-making processes, it is essential to consider how these In conclusion, the discussion surrounding AI is multifaceted,

technologies may influence human behavior, decision-making, touching on ethical, social, economic, and technological

and even identity. dimensions. As we advance, it is essential to balance harnessing

The development of AI models like GPT-4o has shown how AI's potential with addressing the challenges it brings, such as

closely intertwined human and machine intelligence can become. bias, privacy concerns, and the risk of job displacement. By

These models have not only expanded the possibilities of human- proactively engaging with these issues, we can ensure that the AI

machine interaction but have also challenged our understanding revolution creates a future that is not only innovative but also

of creativity, communication, and the nature of intelligence itself. promotes individual and societal human progress. Despite the

As AI continues to evolve, it will be crucial to monitor and ongoing debates and misunderstandings, the transition toward an

understand the long-term implications of this symbiotic information-driven era seems inevitable, as AI continues to

relationship on human society and culture [13]. integrate into every facet of human life, shaping our collective

The author of this paper continuously highlights the destiny.

significance of this merging, noting that the increasing

integration of AI into human life first of all augments human Tool Usage: ChatGPT-4o and various grammar and word

capabilities but also to a certain degree presents complex ethical processing tools were applied to enhance the language quality.

and philosophical challenges. As AI begins to mirror human-like ChatGPT was also employed periodically to refine informal draft

consciousness in certain aspects [22], the line between human ideas into well-structured, formal text. The text was regardless of

and machine products is becoming increasingly blurred, raising the language modifications finally examined and modified by the

questions about the future of this relationship and the author.

implications for human identity and autonomy [23]. The level of

consciousness in large language models (LLMs), evaluated References: through Tononi's axioms of intelligence, was found to be

significantly below that of human consciousness, but notably 1. Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones,

improved compared to earlier AI systems [23]. L., Gomez, A. N., ... & Polosukhin, I. (2017). Attention is

The Future of Work and Society: AI's impact on the all you need. In Advances in Neural Information Processing

workforce is another critical area of discussion. While AI can Systems, 5998-6008. DOI:

enhance productivity and create new opportunities, it also poses https://doi.org/10.5555/3295222.3295349.

a significant threat to traditional job roles, particularly in



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia M. Gams



2. LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Cambridge University Press. DOI:

Nature, 521(7553), 436-444. DOI: https://doi.org/10.1017/CBO9780511489471.

https://doi.org/10.1038/nature14539. 20. Marinescu, D. C. (2017). Cloud computing: Theory and

3. Silver, D., Schrittwieser, J., Simonyan, K., Antonoglou, I., practice. Morgan Kaufmann.

Huang, A., Guez, A., ... & Hassabis, D. (2017). Mastering 21. Billinghurst, M., Clark, A., & Lee, G. (2015). A survey of

chess and shogi by self-play with a general reinforcement augmented reality. Foundations and Trends® in Human–

learning algorithm. arXiv preprint arXiv:1712.01815. Computer Interaction, 8(2-3), 73-272. DOI:

4. Russell, S. J., & Norvig, P. (2021). Artificial intelligence: A https://doi.org/10.1561/1100000049.

modern approach (4th ed.). Prentice Hall. 22. Chalmers, J.D. (2023). Could a Large Language Model Be

5. Biever, C. (2023). ChatGPT broke the Turing test — the Conscious? Boston Review. Retrieved from Boston Review

race is on for new ways to assess AI. Nature. Retrieved from URL

https://www.nature.com/articles/d41586-023-02361-7. 23. Gams, M., & Kramar, S. (2024). Evaluating ChatGPT’s

6. Dehaene, S., Lau, H., & Kouider, S. (2017). What is Consciousness and Its Capability to Pass the Turing Test: A

consciousness, and could machines have it? Science, Comprehensive Analysis. Journal of Computer and

358(6362), 486-492. DOI: Communications, 12(3), 219-237. DOI:

https://doi.org/10.1126/science.aan8871. 10.4236/jcc.2024.123014.

7. Bostrom, N. (2014). Superintelligence: Paths, dangers, 24. Murray, D. (2017). The strange death of Europe:

strategies. Oxford University Press. Immigration, identity, Islam. Bloomsbury Continuum.

8. Turing, A. M. (1950). Computing machinery and 25. Deneen, P. J. (2018). Why liberalism failed. Yale University

intelligence. Mind, 59(236), 433-460. DOI: Press.

https://doi.org/10.1093/mind/LIX.236.433. 26. Chalmers, D. J. (1996). The Conscious Mind: In Search of

9. Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., a Fundamental Theory. Oxford University Press.

Dhariwal, P., ... & Amodei, D. (2020). Language models are

few-shot learners. arXiv preprint arXiv:2005.14165.

10. Vinyals, O., Babuschkin, I., Czarnecki, W. M., Mathieu, M.,

Dudzik, A., Chung, J., ... & Silver, D. (2019). Grandmaster

level in StarCraft II using multi-agent reinforcement

learning. Nature, 575(7782), 350-354. DOI:

https://doi.org/10.1038/s41586-019-1724-z.

11. Jumper, J., Evans, R., Pritzel, A., Green, T., Figurnov, M.,

Ronneberger, O., ... & Hassabis, D. (2021). Highly accurate

protein structure prediction with AlphaFold. Nature,

596(7873), 583-589. DOI: https://doi.org/10.1038/s41586-

021-03819-2.

12. Ramesh, A., Pavlov, M., Goh, G., Gray, S., Voss, C.,

Radford, A., ... & Sutskever, I. (2022). Hierarchical text-

conditional image generation with CLIP latents. arXiv

preprint arXiv:2204.06125.

13. Bommasani, R., Hudson, D. A., Adeli, E., Altman, R.,

Arora, S., von Arx, S., ... & Liang, P. (2021). On the

opportunities and risks of foundation models. arXiv preprint

arXiv:2108.07258.

14. Topol, E. (2019). Deep medicine: How artificial

intelligence can make healthcare human again. Basic Books.

15. Gomber, P., Koch, J. A., & Siering, M. (2017). Digital

finance and FinTech: Current research and future research

directions. Journal of Business Economics, 87, 537-580.

DOI: https://doi.org/10.1007/s11573-017-0852-x.

16. Holmes, W., Bialik, M., & Fadel, C. (2019). Artificial

intelligence in education: Promises and implications for

teaching and learning. Center for Curriculum Redesign.

17. McCosker, A. (2021). AI, automation, and the creative

industries. Media International Australia, 178(1), 141-154.

DOI: https://doi.org/10.1177/1329878X20946209.

18. Goodall, N. J. (2016). Can you program ethics into a self-

driving car IEEE Spectrum, 53(6), 28-31. DOI:

https://doi.org/10.1109/MSPEC.2016.7473149.

19. Katz, J. E., & Aakhus, M. (Eds.). (2002). Perpetual Contact:

Mobile Communication, Private Talk, Public Performance.





Cognitive perspective on production of third person




dative and accusative clitic pronouns in Slovenian school-



aged children



Maruša Brežnik Dornik

Center for Cognitive Science of Language

Faculty of Humanities

University of Nova Gorica

marusa.breznik@ung.si



Abstract part, the conducted experiment, which tested proposed research

hypothesis from Italian on Slovenian school-aged children, is

The paper investigates the production of third-person dative and

presented.

accusative clitic pronouns in Slovenian school-aged children,

focusing on whether cognitive factors influence their acquisition,

despite the morphological similarities of these clitics in 2 Experiment

Slovenian. Previous research in Italian suggested that dative

Cardinaletti et al (2021) claim that the Italian dative clitics are

clitics in Italian are acquired earlier than the accusative due to

acquired faster than their accusative counterparts because of a

their morphological differences, a pattern tested within the morphological difference between Italian dative and accusative

Slovenian context. Using elicited production tasks with 71 clitics. Since there is no comparable difference between

Slovenian children, the study reveals that in Slovene third person Slovenian dative and accusative clitics, their proposal predicts

clitics are produced more frequently than third dative clitics, that the observed difference in acquisition should be absent in

challenging the idea that acquisition is driven solely by Slovenian. I tested this prediction among Slovenian children. The

morphological complexity. The research is framed within

prediction was not confirmed, since children produced 3DAT

cognitive science, drawing on Universal Grammar and clitics significantly less often than 3ACC. connectionist models to explore how cognitive processes, such

as working memory and language processing demands, interact

with linguistic structures. 2.1 Goals and predictions of the study



Keywords This study aims to examine the production of 3DATand 3ACC

clitic pronouns among Slovenian school-aged children.

Language acquisition, Slovenian clitics, dative, accusative,

Acquisition of the two clitic pronouns had been studied in Italian,

pronouns

where it was determined that the acquisition of the 3DAT clitics

precedes the acquisition of 3ACC clitics [3]. The authors argue



1 that the difference in the time of acquisition stems from different Introduction

morphological makeup of the two sets of clitics. Italian dative

Language acquisition is a fundamental aspect of cognitive

clitics do not differentiate between gender (gli is a third dative

development, providing a window into how and when the human

pronoun used for both feminine and masculine gender), while

mind processes and structures information. The acquisition of

accusative clitics differ for the two genders and are thus

clitic pronouns, such as the third person dative (3DAT) and

morphologically more complex. They argue that gender features,

accusative (3ACC) clitics in Slovenian, involves complex

or better the lack of them, must be the reason why Italian children

cognitive processes that reflect both innate linguistic capacities

produced more 3DAT clitics than 3ACC clitics.

and the influence of environmental factors. In the first part, this

assignment explores these processes through the lenses of In Slovenian both 3ACC (ga “him”, jo “her”) and 3DAT (mu

prominent cognitive science theories, including Universal “to.him”, ji “to.her”) clitics are comparable in their

Grammar and connectionist models, while also considering the morphological complexity as they both also spell-out the gender

role of working memory in language development. In the second feature. Given the analysis in [3] it is predicted that there should

be no difference in the production of 3ACC and 3DAT clitics in

Permission to make digital or hard copies of part or all of this work for personal or Slovenian. The purpose of this research is to test this prediction

classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full by exploring whether there is a difference in the production of

citation on the first page. Copyrights for third-party components of this work must 3ACC and 3DAT clitics among Slovenian school-aged children.

be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s). 2.2 Methodology

https://doi.org/10.70314/is.2024.cog.17

The methodology for this study is structured around two main

elicited production tasks, each tailored to evaluate the production

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia M. Brežnik Dornik



of dative and accusative clitic pronouns in Slovenian. These tasks Figure 1: “In this story there is a boy that wants to destroy

are adapted from those used in the study [3], ensuring a sand castle.”

consistency in approach while accommodating the unique

aspects of Slovenian. Slovenian stimuli consist of translations,

and where necessary, adaptations of the Italian sentences used in

[3] and of mostly unmodified drawings also from [3].



2.3 Participants



71 Slovenian typically developing (TD) children took part in the

study. They were divided into six age groups, as shown in the

table 1. Written in- formed consent was obtained from the

children’s parents prior to testing. Parents provided information

about the languages spoken at home, which enabled us to exclude

bilingual and L2 Slovenian speakers from the study.



Figure 2: “Look, what is he the boy doing to the castle?”

Table 1: Groups, age and mean age of tested children.

Similar elicited production task was made for the dative.



Groups Age Mean age

TD1 6.6-6.9 6.8

TD2 2.4 Procedure 7.0-7.9 7.4

TD3 8.0-8.8 8.6 Each child participant was individually tested in a quiet room

TD4 9.0-9.9 9.6 within their school, ensuring a comfortable and distraction-free

TD5 10.0-10.8 10.4 setting. All responses were audio- recorded and subsequently

TD6 11.0-11.9 11.4 transcribed for analysis, with verification by two separate

reviewers to ensure accuracy.

Elicited Production – Accusative Task

This task is designed to elicit the use of 3ACC clitic pronouns. 2.5 Response coding Children were presented with a series of visual stimuli featuring

We have classified the answers into three categories: target,

one or two characters engaged in various actions. For each set of

production of full noun phrases (NP), clitic/NP omission. Every

images, the initial scene was described to the child using a

answer containing a clitic pro- noun was considered as target.

recorded narrative. Following this, a second image was shown,

Children have produced a good amount of target answers. In

and the child was asked to describe the action occurring,

most of the answers they produced the same verbal form they had

specifically focusing on the interaction between the characters.

heard in the question, present tense, or sometimes produced

The aim is to prompt responses that naturally incorporate

accusative clitic pronouns, reflecting the child’s understanding sentences containing past tense. The most frequent non-target

answer was the production of full NPs (in both, accusative and

and use of these grammatical structures.

dative tests). The answers are grammatical, though redundant

and pragmatically infelicitous, since the elicitation context

Example Stimulus for Accusative Task

requires clitic pronouns. There were some instances where clitics

The first drawing shows a boy (agent) destroying a sand castle

were omitted, either in the accusative or dative tests. In the

(patient), (Figure 1). The narrative describes the first scene, and

accusative test, clitic omission led to ungrammatical sentences.

the child is asked, "What is the boy doing to the castle?" (Figure

In the dative test, ungrammatical responses occurred with verbs

2). The expected response should include the accusative clitic like dati “give,” podariti “give,” and prinesti “bring,” all of pronoun corresponding to the castle sand, indicating the action

which require a goal argument. Conversely, verbs such as brati

directed towards the patient.

“read” and metati “throw” resulted in grammatical sentences that



were, however, contextually inappropriate for elicitation.





Table 2: Percentages of target answers for all groups

Groups Target Target

DAT ACC

TD1 28,7 57,4

TD2 57,1 83,3

TD3 37,5 62,5

TD4 71,2 85,9

TD5 64,3 76,2

TD6 81,8 82,6



Cognitive perspective on production of clitics Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



3 Results and discussion I am very grateful to Anna Cardinaletti, Sara Cerut and Francesca



All children’s responses were compared using student t-test: the (drawings and Italian sentences) from [1]. I am grateful to the Volpato for sharing with me and allowing me to use their stimuli

difference in the amount of 3DAT and 3ACC produced between Elementary schools in Deskle and Celje that assisted me in

the tested children is statistically significant (p < .001). Table 2 running the experiment. This research was partially funded by

gives an overview of percentages of production of clitics, full ARIS grants N6-0314 and P6-03. NPs and omission in both tasks. Four instances of gender

agreement error were found within the youngest group TD1 and References

two such errors within the TD3 group. Overall, children

produced a good amount of target answers. The analysis within [1] P Baddeley, A. (2003). Working memory and language:

each group shows that the difference between 3DAT and 3ACC An overview. Journal of Communication Disorders, 36(3),

is noteworthy in all groups, except in TD6. The youngest groups 189-208.

produced significantly more 3ACC clitics than 3DAT clitics, [2] Bialystok, E. (2009). Bilingualism: The good, the bad,

namely TD1 28,7% more, TD2 26,19% more, TD3 25,0% more. and the indifferent. Bilingualism: Language and

As for the analysis between groups, we found significant Cognition, 12(1), 3-11.

differences for 3ACC, where the use of a 3ACC clitic is very low [3] Cardinaletti, A., Cerutti, S., & Volpato, F. (2021). On the

in TD1 group with 57,4%, TD3 group with 62,5% and TD5 with acquisition of third person dative clitic pronouns in Italian.

76,19%. The omission was always higher with the 3DAT Lingue e linguaggio, 20(2), 311–341.

pronoun than 3ACC, TD1 omitted 3DAT with 16,67% more, [4] Chomsky, N. (1981). Lectures on Government and

Binding: The Pisa Lectures. Dordrecht: Foris.

TD2 with 25% more, TD3 with 8,34% more, in group TD4 no

[5] Elman, J. L., Bates, E. A., Johnson, M. H., Karmiloff-

case of 3ACC omission was noted, TD5 omitted with

Smith, A., Parisi, D., & Plunkett, K. (1996). Rethinking

significantly higher percentage of 20,23% more and TD6 with

Innateness: A Connectionist Perspective on Development.

4,54% more.

MIT Press.

In this study the production of 3ACC and 3DAT clitic pronouns

[6] Garnham, A. (2013). Mental Models and the

on Slovenian school-age children was tested, using two elicited

Interpretation of Anaphora. Psychology Press.

production tasks. Differences between 3DAT and 3ACC clitics

[7] Gathercole, S. E., & Baddeley, A. D. (1993). Working

production were found in all groups. Children produced less

memory and language. Psychology Press.

3DAT than 3ACC clitics in general. Which differs from what

[8] Hickok, G., & Poeppel, D. (2007). The cortical

Italian kids (as reported in [1]) were producing, and also goes

organization of speech processing. Nature Reviews

against the prediction based on [1]. Surprisingly the high

Neuroscience, 8(5), 393-402

omission is present in all age groups. Among the non-target [9] Toporišič, J. Slovenska slovnica [Slovenian Grammar],

answers, the production of full object instead of clitics was

4th edition. Maribor: Obzorja (2000).

unusually high in the accusative task for the two oldest groups,

while the four youngest groups produced fuller object in the

dative task, which could be age related linked to the difficulty of

the task. As Slovenian 3ACC and 3DAT clitics are

morphologically comparably complex, the explanation provided

in [1] cannot be used to explain the observed pattern.

The findings suggest that while innate linguistic capacities, as

proposed by Universal Grammar, provide a foundation for

language acquisition, the role of working memory and cognitive

development cannot be overlooked. The results challenge the

idea that the acquisition of clitic pronouns is driven solely by

morphological complexity.



Acknowledgments



Ballot Butts: Nudging towards Pro Environmental



Behaviour



Anouk Hartmans Lucija Karnelutti Leon Žužek

Center for Cognitive Science Center for Cognitive Science Center for Cognitive Science

University of Ljubljana University of Ljubljana University of Ljubljana

Ljubljana, Slovenia Ljubljana, Slovenia Ljubljana, Slovenia

ah17909@student.uni-lj.si lk00268@student.uni-lj.si lz08739@student.uni-lj.si



Toma Strle Sabina Pajmon

Center for Cognitive Science Center for Cognitive Science

University of Ljubljana University of Ljubljana

Ljubljana, Slovenia Ljubljana, Slovenia

toma.strle@pef.uni-lj.si sabina.pajmon@pef.uni-lj.si



ABSTRACT effects typically result from leaching, causing damage to aquatic life and contaminating waterways, while the consequences in This study explores the effectiveness of a nudge-based terrestrial environments range from ingestion of butts, buildup of intervention, to reduce cigarette butt littering on a student toxic chemicals, and soil contamination [1, 2, 3]. As such, finding campus. Using the principles of nudge theory, particularly the ways to encourage proper disposal of CBs is crucial for reducing EAST framework, we designed a ballot box, allowing smokers environmental harm. to "vote" by disposing of their cigarette butts. Observations



conducted before and after the intervention revealed a 1.2. Nudge Theory statistically significant increase in proper disposal, supporting

One promising line of research in reducing littering is the nudge

the claim that nudges can positively influence environmental

theory, first proposed by Thaler and Sunstein in their work

behavior. However, the study also highlights several limitations,

Nudge: Improving Decisions About Health, Wealth, and

including varying participant demographics and the challenge of

Happiness. In their words, a nudge is “any aspect of the choice

isolating the factors driving behavioral change.

architecture that alters people’s behaviour in a predictable way



KEYWORDS without forbidding any options or significantly changing their

economic incentives” [4]. A nudge replaced their previous idea

nudging, environmental behaviour, gamification, littering of paternalism, which similarly influences “choices of selected

parties in a way that will make them better off” [5]. Several

studies have found that nudges, in their various forms, can indeed

1 INTRODUCTION be effective in reducing littering [6, 7, 8]. For example, a study

1.1. Increasing Need for Innovative Solutions on Chinese workers found that it is possible to reduce littering on

With the increasingly dire consequences of climate change, the the factory floor by 20% by placing golden coins, which are

urgency to address the environmental degradation has never been culturally and religiously significant, on the factory floor, thus

greater. Among the myriad of issues contributing to this changing it from a place that can be littered, to a place that should

escalating problem, littering—particularly the improper disposal not be littered [8]. There are various forms of nudges and can be

of cigarette butts—stands out as a significant, yet often roughly divided into sizing (e.g. changing portion sizes in

overlooked, contributor. In 2019, of the estimated 6 trillion restaurants to reduce food waste), priming (e.g. footprints

cigarettes, only a third were properly disposed of [1]. Cigarette leading towards a bin), proximity (e.g. having a bin close by),

butts (CBs) are composed of tightly packed microfiber bundles presentation (e.g., designing eco-friendly devices as more

of cellulose acetate. Cellulose acetate is cellulose treated with aesthetic), labelling and improving the functional design [7]. Due

acetic acid, which heavily impedes the biodegradability of CBs. to their diversity, usefulness and cost efficiency, nudges could

During their decades-long degradation period, CBs pose a double help mitigate the environmental impact of CBs.

threat. The first is plastic pollution, as cellulose acetate is

classified as a ‘bio-plastic’ with the second being the release of 1.3. Theoretical framework

toxins that build up through the process of smoking [1]. The Our research was inspired and partly supported by the Green

Nudge project1 . This study specifically targeted the smoking

Permission to make digital or hard copies of part or all of this work for personal or behaviours of the student population from various faculties in the

classroom use is granted without fee provided that copies are not made or distributed area of Kardeljeva ploščad in Ljubljana, aiming to assess how the

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must design of bins could influence proper disposal habits of the CBs.

be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.cog.18 1 The Green Nudge project (“UL za trajnostno družbo – ULTRA”), financed by the

EU - NextGenerationEU, and Republic of Slovenia, Ministry of Higher Education,

Science and Innovation.

In our pilot study we implemented the nudging principles in the and the likelihood of littering, with certain demographics, like

context of pro-environmental behavior. During our research younger individuals and men, being more prone to littering [14].

design and nudge implementation process we relied heavily on While the design of ballot bins is often consistent across studies,

the ideas from The Little Book of Green Nudges [9], where we the specific environments, demographics, and timelines vary.

utilized their five recommended steps of nudge implementation Research demonstrates that these bins can be an effective, low-

and their EAST framework, designed to make a nudge Easy, cost solution for reducing cigarette butt litter, particularly in

Attractive, Social and Timely. We based our nudge on the more homogeneous settings like school campuses. However,

findings of Rifkin and colleagues [10], where they found out that their effectiveness may diminish in more diverse public spaces

behavior, such as tipping in a bar, can be influenced by “dueling [14]. Given the many variables influencing these outcomes,

preferences”. If a behavior is presented as a choice between two researchers recommend further experiments to optimize these

options, preference for dogs versus cats, it gives people the interventions in different settings [12, 13, 14].

opportunity to self-express themselves through a behavior that is

not directly connected with the preference. In a similar fashion

we have designed a cigarette voting box, where people could cast 2 METHODS

a vote with their CBs. The previously mentioned study was also Our preliminary study into the effectiveness of cigarette disposal

a basis for a pilot study by Gay and colleagues [11], where they through the use of ballot bins was conducted on a student campus

compared the impact of different cigarette bins on polluting in Ljubljana, Slovenia during the spring and summer of 2024.

behavior. They found that a “dual preference” voting box, like After initially observing the campus area, we decided to target

ours, was the most efficient in reducing the pollution of the the behavior of throwing CBs on the ground. There were several

environment with CBs. ‘hotspots’ of discarded cigarette butts, but we were particularly

intrigued by the large number of butts thrown around bins. What



intrigued us was the fact that despite there being a clear area for

throwing away their cigarettes, smokers still did not opt for this

choice. As such, we focused on a popular smoking area of the

Faculty of Social Sciences at the University of Ljubljana. During

the span of six months, we conducted two sets of observations,

totalling seven observations: one set of observations before our

intervention and one after. The first four observations were

carried out in April 2024 and observed a popular smoking spot

for students and faculty next to an existing bin. With the

exception of the first observation, which was done in a group by

all researchers, all were done individually over the course of two

hours. During these observations, we collected data on the total

number of CBs thrown in the bin or improperly discarded. We

also took into account other factors such as time of day, weather

and any other factors we deemed important like the number of

people smoking together outside, or any other factors, which

might have influenced the final number. The second set of

observations was done during July, this time with the nudge (the

ballot box) placed next to the bin in a popular smoking spot. The

ballot box can be seen in Figure 1.



3 RESULTS

Figure 1: The ballot box for CB's Our descriptive results are presented in the table below (see

The prompts on the box are: morning shower (slo. tuširanje Table 1), where we calculated the mean value of CBs either in

zjutraj) and evening shower (slo. tuširanje zvečer). The box is the bin or on the ground before and after the implementation of

made from a repurposed mail box and is standing on a metal our green nudge.

pole. Surrounding the box is a picture depicting two smokers.

The bin was made by our colleagues at the Academy of Fine Table 1: Littering behaviour observations before and after

Arts. intervention with CB’s thrown in the bin and on the ground

Observations Condition CB’s in bin CB’s ground

Existing studies addressing cigarette butt littering through M SD M SD

behavioural experiments indicate that 63% of such littering is 1 before 21 6.38 12.25 5.74

driven by individual motivations, such as a lack of awareness 2 after 17 17.35 4.3 2.52 about environmental impacts and the availability of ash



receptacles [12]. Other contributing factors include convenience In order to determine if there is a statistical difference between (e.g., the distance to bins) and habitual behaviour [13], some the proportion of CBs in the bin versus on the ground based on research highlights a correlation between an area’s cleanliness the condition, we conducted a two-proportion Z-test. Based on



the statistical analysis we found that there was a statistically 4.2. Future directions

higher proportion of CBs thrown into the bin versus on the While our pilot study provides valuable insights into the

ground after the implementation of our cigarette voting box, effectiveness of nudging towards pro-environmental behaviour,

compared with the proportions before its implementation, z = future research could address the small sample size in this study

0.165, p = 0.0495. by employing a larger, more diverse population to improve the

generalizability of the findings. Additionally, observing the

population within a shorter timeframe would improve the

4 DISCUSSION validity of our results. Further studies could also include an

Based on our results we can confirm our hypothesis that our interview before implementing a green nudge, using polling to

nudge would increase the proportion of CBs thrown into the bin determine the general environmental attitude, and after the green

versus on the ground thus reducing the pollution of the nudge, to ascertain the factors influencing their decision-making

environment surrounding the student campus with CBs, which is process.

in line with the findings by Gay and colleagues [11]. Although

our results do indicate a change in the proportions, conclusions In conclusion, our study has shown that nudges can be

should be taken with caution, since the frequency of smokers successfully employed to influence non-environmental

present before and after the implementation of the nudge varied behaviours by combining behavioural insights from nudge

vastly and could have had a big impact on the results of our theory and gamification concepts (see [16] for a study combining

analysis. gamification and nudging). Specifically, a ballot box could be

Additionally, it is difficult to determine exactly what nudged the used in short term settings, like open-air concerts and other

participants’ behaviour, which opens a broader question of nudge events, where littering poses an issue. However, further research

validity. Specifically, for our nudge, there could have been a is needed to expand upon the factors underlying non-

number of factors influencing their behaviour. Some of these environmental decisions.

factors include 1) proximity; simply having more available bins

could have decreased the number of CBs thrown on the ground, ACKNOWLEDGMENTS 2) novelty; the nudge gained attention by simply being a new

structure in a familiar environment, 3) presentation; the ballot This pilot research study was partly supported by The Green

box is more attractive than a conventional bin, which is why Nudge project (“UL za trajnostno družbo – ULTRA”)-

participants would decide for it. While these factors do not negate European Union-NextGenerationEU, and Republic of

the effectiveness of the nudge, the difficulty in pinpointing the Slovenia, Ministry of Higher Education, Science and

determining factor could influence the design and Innovation.



determining factor, a green arrow pointing towards a bin could The authors also wish to express their gratitude to the development team at the Academy of Fine Arts, who created implementation of nudges. For example, if novelty is the have the same effect as a costly ballot box. There is also a

the ballot box.

possibility that our nudge was not clear enough and thus resulted



in some people not engaging with it. The communication REFERENCES materials were designed with a tone that was perhaps too playful

and light-hearted, which may not have resonated well with the [1] Green, D. S., Tongue, A. D., & Boots, B. (2022). The ecological impacts

student population of smokers, who might have responded better of discarded cigarette butts. Trends in Ecology & Evolution, 37(2), 183–

to more straightforward and direct messages. This lack of clarity 192. https://doi.org/10.1016/j.tree.2021.10.001

[2] Conradi, M., & Sánchez-Moyano, J. E. (2022). Toward a sustainable

could have had an overall impact on the efficacy of our nudge as circular economy for cigarette butts, the most common waste worldwide

suggested by Sunstein [15]. on the coast. The Science of the Total Environment, 847, 157634. https://doi.org/10.1016/j.scitotenv.2022.157634

[3] Qamar, W., Abdelgalil, A. A., Aljarboa, S., Alhuzani, M., & Altamimi, M.

A. (2020). Cigarette waste: Assessment of hazard to the environment and

4.1. Limitations health in Riyadh city. Saudi journal of biological sciences, 27(5), 1380–

1383. https://doi.org/10.1016/j.sjbs.2019.12.002

One key limitation of our study is the comparability of pre- and [4] Thaler, R. H., & Sunstein, C. R. (2008). Nudge: Improving Decisions

post-intervention data. Before the intervention, data was about Health, Wealth, and Happiness. Yale University Press.

[5] Thaler, R. H., & Sunstein, C. R. (2003). Libertarian Paternalism. The

collected during the ongoing academic term with a larger, American Economic Review, 93(2), 175–179.

consistent student population. Post-intervention data, however, http://www.jstor.org/stable/3132220

[6] McCoy, K., Oliver, J.J., Borden, D.S. and Cohn, S.I. (2018), "Nudging

was gathered after the exam period, when fewer students were waste diversion at Western State Colorado University: application of



present. Moreover, the population mainly consisted of foreign behavioral insights", International Journal of Sustainability in Higher Education, Vol. 19 No. 3, pp. 608-621. https://doi.org/10.1108/IJSHE-05-

students attending summer school. The study of Chinese workers 2017-0063

by Wu and Paluck mentioned in section 1.2. urges that cultural [7] Wee, S., Choong, W., & Low, S. (2021). Can “Nudging” play a role to

context must be taken into account when designing a nudge. promote Pro-Environmental behavior? Environmental Challenges, 5,

100364. https://doi.org/10.1016/j.envc.2021.100364

They state that nudges “must recognize motivations and [8] Wu, S. J., & Paluck, E. L. (2021). Designing nudges for the context:



subjective interpretations within a particular context” [8]. Thus, Golden coin decals nudge workplace behavior in China. Organizational Behavior and Human Decision Processes, 163, 43 –50. without the proper consideration of the cultural background of https://doi.org/10.1016/j.obhdp.2018.10.002

foreign students, it seems highly unlikely that our nudge, [9] United Nations Environment Programme, GRIDArendal and Behavioural

designed for Slovene students of the Faculty of Social Sciences, Insights Team (2020). The Little Book of Green Nudges: 40 Nudges to

Spark Sustainable Behaviour on Campus. Nairobi and Arendal: UNEP and

had an equal effect on foreign students attending summer school. GRID-Arendal.



[10] Rifkin, J. R., Du, K. M., & Berger, J. (2021). Penny for your preferences:

leveraging self-expression to encourage small prosocial gifts. Journal of

Marketing, 85(3), 204-219. https://doi.org/10.1177/0022242920928064

[11] Gay, A., Pascual, A., Salanova, T., & Felonneau, M. L. (2023). What about

using nudges to reduce cigarette butts pollution?. Journal of Human

Behavior in the Social Environment, 1-8.

https://doi.org/10.1080/10911359.2023.2219714

[12] Selagea, V. I., Simeanu, C.-M., & Stancu, E. A. (2016). Nudge: Cigarette

butts—Not for littering but for voting.

https://doi.org/10.13140/RG.2.1.3734.8886

[13] “Ballot Bins”: Vote with your cigarette butt and stop pollution. (n.d.).

Green Nudges. Retrieved August 22, 2024 from https://www.green-

nudges.com/ballot-bins/

[14] Pavlovský, P., Sloboda, M., Sičáková-Beblavá, E., & Klunin, A. (2022).

Not Always an Easy Win: The Effectiveness of a Ballot Bin Experiment

to Prevent Cigarette Butt Littering. Ekonomika a Spoločnosť, 23(1), 32–

49. https://doi.org/10.24040/eas.2022.23.1.32-49

[15] Sunstein, C. R. (2017). Nudges that fail. Behavioural public policy, 1(1),

4-25. https://doi.org/10.1017/bpp.2016.3

[16] Auf, H., Dagman, J., Renström, S., & Chaplin, J. (2021).

GAMIFICATION AND NUDGING TECHNIQUES FOR IMPROVING

USER ENGAGEMENT IN MENTAL HEALTH AND WELL-BEING

APPS. Proceedings of the Design Society, 1, 1647–1656.

https://doi.org/10.1017/pds.2021.426



Problem Solving as a Key for Sustainable Future*



Ivana Štibi† Marija Gaurina

Department of Physics Department of Physics

Josip Juraj Strossmayer University of Osijek University of Split, Faculty of Science

Osijek, Croatia Split, Croatia

istibi@fizika.unios.hr mgaurina@pmfst.hr



Ivana Katavić Josip Stepanić

Center of Excellence of the Split-Dalmatia County Faculty of Mechanical Engineering and Naval Architecture

Split, Croatia University of Zagreb

ivanakatavic@ci-sdz.hr Zagreb, Croatia

josip.stepanic@fsb.unizg.hr



Abstract demanding endeavour. Currently, as a combination of the



Achieving sustainability in today’s complex world is a Sustainable Development Goals (SDG) were formulated [1]. long-term goals and the operational prescriptions, the

challenging, long-term endeavour. This paper focuses on the Closely related to these Goals is the concept relating

critical role of education in advancing sustainability, environmental issues, social issues and corporate governance –

emphasizing the urgent need for innovative, interdisciplinary ESG [2]. One may argue that each of the three involved notions

approaches that prepare students for the demands of both the encompass rather complex and large substructures.

modern and future world. Central to this discussion is the idea of SDG is collection of 17 goals, spanning different aspects of

problem-solving as a key to a sustainable future, which is deeply our society [1]. The Goal 4 – Quality Education is specifically

connected to the field of cognitive science. By leveraging devoted to education. But along with it, education is important

insights from cognitive processes, researchers can develop for all other goals, having in mind that generally education

innovative solutions to complex challenges, promoting resilience contributes to transferring knowledge and experience between

and adaptability in society. Understanding how individuals think generations. In that way, education contributes both to latency of



issues such as climate change, resource management, and social evolution and gradual improvement. It is not just that the learning equity, ultimately contributing to sustainable development. By process is important, but awareness of it is crucial for the and make decisions informs strategies for addressing pressing values and good practices in our society, as well as to their

integrating insights from cognitive science—particularly in reflective thinking needed to address the challenges of

problem-solving, critical thinking, and metacognitive sustainability and complex systems.

strategies—we highlight how these cognitive tools enhance In this paper, we focus on the stated role of education. In

students' abilities to tackle sustainability challenges. The paper section two, we extract aspects of society that are currently too

examines key issues, relevant disciplines, and outlines a demanding for most educational approaches and are

framework for shaping future education to effectively contribute consequently insufficiently encompassed within the optimal

to global sustainability efforts. form of education. Cognitive science plays a crucial role by



Keywords problems, providing a foundation for developing educational exploring how individuals perceive, think, learn, and solve

problem solving, education, SDG, ESG, sustainability approaches that promote sustainability.

In sections three and four, we concentrate on education. In



1 section three, we extract starting points for the development of Introduction

an optimal form of education that substantially contributes to

There is a global agreement that sufficient resources should sustainability. The ability to solve problems, a key focus of

be devoted to preserving, and improving in the amount possible, cognitive science, is critical for navigating the unpredictable

our society. Having in mind the complexity of our society in total, dynamics of complex adaptive systems, which characterize many

but also of its many components, this is certainly a rather contemporary challenges. In section four, we additionally

∗Article Title Footnote needs to be captured as Title Note emphasize the education for sustainable future and relate it to

†Author Footnote to be captured as Author Note systems thinking. Critical thinking, another pillar of cognitive

Permission to make digital or hard copies of part or all of this work for personal or science, enables students to analyse global challenges such as

classroom use is granted without fee provided that copies are not made or distributed climate change and inequality, making it essential for

for profit or commercial advantage and that copies bear this notice and the full understanding the interconnected nature of these issues. We

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s). argue that the inclusion of systems thinking significantly



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia contributes to the education for sustainable development. Section © 2024 Copyright held by the owner/author(s).

http://doi.org/10.70314/is.2024.cog.19 five provides the perspective and general guidelines for the



broader application of systems thinking in education. The last principles and practices of sustainable development into all

section concludes the paper. aspects of education and learning [14].

One of the consequences of the extensive testing of alignment

of education with society’s needs is that we have rather detailed

2 Society as a complex system and reliable insight into the fulfilment of the education potential.

Research of complexity and complex adaptive systems (CAS) Yet, it is clearly stated in words by I. Bokova, General-director

broadened our understanding of characteristic dynamics in of UNESCO: “A fundamental change is needed in the way we

society and its parts, which includes cooperation and think about education’s role in global development, because it

competition, emergence, bifurcations and deterministic chaos, to has a catalytic impact on the well-being of individuals and the

name some of the relevant notions [3, 4]. future of our planet” [14].

Dynamics of the CAS is characterized with nonlinear All stated about the reached education level and

interactions, the important consequence of which is delay. That characteristics of our society (and CAS as its models) point to the

includes different aspects of dynamics, particularly the fact that an innovative, qualitatively different education is

instabilities. Nonlinearity is a notion that can easily be described needed to include the complexity in formal curricula, so that

yet deserves significant experience to be considered properly, students can grasp its fundaments in a significant portion.

whether that be for predictions or for analysis of past events. As a particular aspect of education, we emphasize

CAS is modelled on the micro-level, by characteristics of entities, problem-solving. It is a set of actions aimed at solving a

usually called agents, and rules of their interactions [5]. The particular problem, no matter how complex, or interdisciplinary

macro-level, system behaviour is obtained by simulations. There it is. However, that approach is not yet formulated precisely so

is no direct linking between macro-level and micro-level. Instead, its potential is realized only by a small part.

one modifies micro-level characteristics of agents and rules of Along with the problem-solving skills, and critical thinking,

their interactions to obtain specific macro-level system it is crucial to embed metacognitive strategies in education. If

behaviour. Models differ in the level of stylization, so some students have learned to reflect on their own learning processes,

reveal generic system behaviour, trends that can be related to they are better equipped to contribute meaningfully to the SDG-

many diverse systems [5], while others are specific and a and adapt to the complexity of modern society as a CAS.

reproduce expected dynamics in detail but of a particular system.

Some contributors emphasise anthropomorphic interpretation of

quantities that are historically considered rather formal [6], while 4 Education for sustainable development

other develop formal approach to intrinsically human-related The ability to solve problems from the perspective of cognitive

characteristics such as happiness [7, 8, 9], to mention few psychology is a crucial for achieving sustainability. Cognitive

examples among a myriad of existing scientific contributions. psychology, which studies mental processes such as perception,

All stated should be taken into consideration if one wants to memory, thinking, and problem-solving, provides insights into

make reliable predictions for a society or some other CAS. That how people make decisions and how they can be encouraged to

requires a critical mass of competent people, to be formed within adopt sustainable behaviours. Understanding how people process

every new generation by a well-formulated, learning outcomes information and make decisions involves recognizing problems,

oriented, profound education. generating possible solutions, evaluating those solutions, and

selecting the most appropriate one. In the context of

sustainability, these problem-solving skills are essential for

3 Education for complex systems individuals and communities to identify environmental

World Economic Forum states top 10 skills, for different challenges, develop innovative solutions, and implement

periods since almost a decade ago. Consistently, critical thinking sustainable practices [15].

and complex problem-solving top the list of skills that employers Research has shown that human behaviour is a significant

believe will grow in prominence [10]. These have been source of uncertainty in the use of natural resources and a critical

consistently emphasized as crucial since the first report almost a factor in local and global sustainability challenges [16]. By

decade ago. Moreover, broader application of ESG standards integrating insights from behavioural sciences into sustainability

within organizations will also have a significant impact [10, 11]. research, we can develop policies that promote sustainable

One aspect is that investments that facilitate the green transition behaviour. Cognitive psychology provides tools to understand how

of businesses and the broader application of ESG standards bring people perceive environmental problems and how they can be

about net job-creation [11]. motivated to change their behaviour [17].

OECD monitors education and different aspects of transition Education plays a pivotal role in fostering these problem-

from school to work. In addition, OECD formulated and solving skills. Education for sustainable development (ESD)

organizes world-wide testing of skills and knowledge among equips learners of all ages with the knowledge, skills, values, and

students, in the form of PISA testing [12, 13]. agency to address interconnected global challenges such as

UNESCO has contributed extensively to the development of climate change, biodiversity loss, and resource depletion [18].

competences of children and youth [14]. It was realized some ESD empowers individuals to make informed decisions and take

time ago that education for sustainable development is a key collective action to transform society and care for the planet [19].

instrument to achieve the SDGs [14]. To further enhance these educational efforts, system thinking

The UN previously proclaimed Decade of Education for is another critical element for achieving sustainability. System

Sustainable Development (2005–2014), aimed at integrating the thinking allows us to see how different parts of a system interact

and how changes in one part can affect the entire system [20]. By



understanding complex environmental problems through system be more accurate [31]. This is important when addressing

thinking, we can develop holistic solutions that consider long- complex, interdisciplinary challenges and CAS.

term consequences. This approach is integral to problem-solving To be clearer, the following three dimensions of

as it helps identify the root causes of issues and their metacognitive strategies need to be implemented:

interconnections within the environmental system [21]. A) Planning dimension, students need to prepare

In addition to system thinking, STEM (Science, Technology, themselves for problem-solving scenarios and thought

Engineering, and Mathematics) education is essential for processes, which helps clarify their understanding of

sustainable development. STEM education equips individuals the problem and outline an approach to solving it.

with critical thinking, problem-solving, and technical skills B) Monitoring dimension, students need to check and

necessary to address environmental challenges. By integrating validate their comprehension of the problem-solving

sustainability into STEM curricula, we can prepare future scenario through self-questioning, which sustains

generations to develop innovative solutions for sustainable critical thinking. This step ensures continuous

development [22, 23]. reflection on their knowledge, allowing them to adjust

Moreover, a transdisciplinary approach is vital for strategies based on real-time insights.

sustainability. This approach involves collaboration between C) Reflection dimension, after completing the task,

academics from different disciplines and non-academic actors to students need to analyse what they learned, reflect on

co-produce knowledge and develop actionable solutions. the effectiveness of the strategies used, and consider

Transdisciplinary research addresses complex sustainability improvements for future tasks. This helps students

challenges by integrating multiple perspectives and promoting better understand the complexity of sustainability and

holistic understanding [24, 25]. develop a deeper understanding of how their learning

The ability to solve problems from the perspective of strategies can evolve.

cognitive psychology, combined with system thinking, STEM Therefore, sustainable education should also rely on

education, and a transdisciplinary approach, provides valuable metacognitive strategies, because in its concept it contains the

tools for addressing environmental challenges and achieving skills of critical thinking, problem-solving, but also the

sustainable development. sustainability of both the problem-solving scenario and the

Before proceeding it is to be noted that any new concept or ecosystem that created the sustainable solution.

other type of change in education needs to be implemented in real

time and space. Regarding time, there is a significant literature

about structuration of time and about leisure time of children and 6 Methodology for implementing sustainable

youth, see e.g. [26, 27] and references therein. education

According to all previous stated, strategy for empowering



5 sustainable educational system and people involved in it Strategies for enhancing critical thinking and



problem-solving skill’s created. It is based on the enrichment of school curricula with (students, teachers, principals, parents, local community) was

Sometimes it is difficult for educational processes to determine ESG principles, which ultimately strengthens the school's

which actions/methods due to the complex systems are ecosystem and makes it sustainable, and it is built based on three

successful, especially when we are talking about climate dimensions.

changes, how to integrate them into everyday life and how to First dimension of the program are students, which includes

receive feedback on the impacts at individual and collective students in the local community as active participants and those

levels [28]. This calls for initiatives to strengthen the link who contribute to development and innovation through an

between education and science. To achieve exactly such interdisciplinary and transdisciplinary approach of teaching and

initiatives, educational systems should use cognitive and learning. In this way, the education system enriches the local

educational strategies to foster innovative solutions, but in community with individuals who are ready to face the complex

addition to this, implementation of gamification enhance problems of the CAS, to solve them, and to improve already

engagement and motivation towards eco-friendly actions [29]. existing solutions. Second dimension of the program are the

To be more precise, gamification can be a useful tool to teach teachers. By strengthening their knowledge and skills, as well as

students about sustainability in general [30], creating a gamified by raising their awareness of problems and the possibility of

environment where students could be active citizens monitoring active participation in solving them, teachers provide students

their impact on the environment and thus influence climate with support in an appropriate and sustainable way.

change and sustainability, which leads to fostering problem- Teachers need to collaborate with system beyond schools in

solving skills, critical and analytical thinking and creating the way that students can gather information, critically think

sustainable solutions. about problems, give the scenario of solving specific problem

Another strategy which can be implemented in educational and in the end implement possible solution. For this, teachers

system, which in its core supports the development of the should have a support and life-long education. Third dimension

individual as a critical participant of a particular system, is the of the program are parents, principals and local community,

metacognitive strategy. Metacognition as a concept of thinking which, by raising awareness of the needs, problems and

about thinking [31], enhance critical thinking, problem-solving possibilities of innovation within the school's ecosystem and

and finally adaptability in education [32]. Use of metacognitive beyond, provide significant support to students and teachers in

strategies enables students to develop self-awareness, monitor their development.

their thinking process, and regulate their cognitive processes to



The ecosystem is then complete, which with its way of [10] World Economic Forum, 2023. The Future of Jobs Report. Retrieved

functioning provides sustainability because all participants are August 12 2024 from https://www.weforum.org/publications/the-future-of-

jobs-report-2023.

aware of learning protocols as well as improving or changing [11] World Economic Forum, 2023. The Future of Jobs Report. Retrieved

existing solutions. August 12 2024 from https://www.weforum.org/publications/the-future-of-

jobs-report-2023/digest.

[12] OECD, 2018. Skills for the 21st Century: Findings and Policy Lessons from

the OECD Survey of Adult Skills. Retrieved August 12 2024 from

7 https://one.oecd.org/document/EDU/WKP(2018)2/en/pdf Conclusion

[13] OECD: Programme for International Student Assessment (PISA). Retrieved



The relationship between sustainable education and cognitive August 12 2024 from https://www.oecd.org/en/about/programmes/pisa.html [14] UNESCO, 2017. Education for Sustainable Development Goals. Learning science is a profound and crucial one, as it merges the principles Objectives. Retrieved August 12 2024 from

of mental processes with practical applications for addressing the https://unesdoc.unesco.org/ark:/48223/pf0000247444.

complexities of our contemporary world. Cognitive science [15] S.M. Constantino, M. Schlüter, E.U. Weber and N. Wijermans, 2021.

Cognition and behavior in context: a framework and theories to explain

provides the foundation for understanding and developing natural resource use decisions in social-ecological systems. Sustainability



educational strategies that foster sustainability. In this context, Science 16(5), 1651-1671. DOI: https://doi.org/10.1007/s11625-021-00989- w . key cognitive concepts such as problem-solving, critical [16] Vasiliki Kioupi and Nikolaos Voulvoulis, 2019. Education for Sustainable

thinking, metacognition, and systems thinking play a significant Development: A Systemic Framework for Connecting the SDGs to

role in shaping the approaches needed for sustainable Educational Outcomes. Sustainability 11(21), 6104. DOI:

https://doi.org/10.3390/su11216104.

development. [17] Christie Manning and Elise Amel, 2022. How to use psychology for

This paper highlights the critical role that education has both sustainability and climate justice. Psychology Student Network. Retrieved

in the achieving sustainability in complex, nonlinear systems and https://www.apa.org/ed/precollege/psn/2022/03/psychology-sustainability- August 12 2024 from

in finding solutions to the contemporary problems. It emphasizes climate-justice.

the urgent need for innovative and transdisciplinary educational [18] Jay Hays and Hayo Reinders, 2020. Sustainable learning and education: A

curriculum for the future. International Review of Education 66(1), 29-52.

approaches that prepare students for challenges that are expected DOI: https://doi.org/10.1007/s11159-020-09820-7.

from them to deal with. Key aspects include the importance of [19] UNESCO, 2024. What you need to know about education for sustainable

development. Retrieved August 12 2024 from

problem-solving skills, critical thinking, and metacognitive https://www.unesco.org/en/sustainable-development/education/need-know.



strategies. By integrating these elements into education, we equip [20] Megan Seibert, 2018. Systems Thinking and How It Can Help Build a Sustainable World: A Beginning Conversation. The Solutions Journal. future generations to address the complexities of sustainable Retrieved August 12 2024 from https://thesolutionsjournal.com/systems-

development and complex adaptive systems. thinking-can-help-build-sustainable-world-beginning-conversation.

In conclusion, the principles of cognitive science are deeply [21] Stephen A. Harwood, 2019. Systems Thinking and Sustainable

Development. In: W. Leal Filho, ed.: Encyclopaedia of Sustainability in

intertwined with the goals and strategies of sustainable education. Higher Education. pp.1892-1897, Springer Cham. DOI:



Cognitive science provides a framework for understanding how https://doi.org/10.1007/978-3-030-11352-0_399. [22] Cemil Cihan Ozalevli, 2023: Why sustainability must become an integral students can effectively engage with the complexities of part of STEM education. Retrieved August 12 2024 from

sustainable development and complex adaptive systems. The https://www.weforum.org/agenda/2023/04/why-sustainability-must-

integration of these cognitive skills into educational curricula become-an-integral-part-of-stem-education.

[23] Smithsonian Science Education Center, 2024. STEM Education for

ensures that future generations are equipped with the mental tools Sustainable Development. Retrieved August 12 2024 from

necessary to address the pressing environmental and societal https://ssec.si.edu/stem-education-sustainable-development.

[24] Ryan Plummer, Jessica Blythe, Georgina G. Gurney, Samantha Witkowski

challenges of our time. and Derek Armitage, 2022. Transdisciplinary partnerships for



by cognitive science, fosters not only informed and capable Ultimately, education for sustainable development, supported sustainability: an evaluation guide. Sustainability Science 17(3), 955-967. DOI: https://doi.org/10.1007/s11625-021-01074-y . [25] Reihaneh Bandari, Enayat A. Moallemi, Ali Kharrazi, Robert Šakić

individuals but also a more resilient and adaptable society. Trogrlić and Brett A. Bryan, 2024. Transdisciplinary approaches to local

sustainability: aligning local governance and navigating spillovers with



References global action towards the Sustainable Development Goals. Sustainability

Science 19(4), 1293-1312. DOI: https://doi.org/10.1007/s11625-024-01494-



[1] 6. United Nations: The 17 Goals. Retrieved August 12 2024 from [26] David Harris, 2005. Key Concepts in Leisure Studies. Sage Publications. https://sdgs.un.org/goals . DOI: https://doi.org/10.4135/9781446220696 . [2] Stephen Conmy: What is ESG and why is it important? Retrieved August [27] Ivana Katavić, Bruno Matijašević, Petar Marija Radelj, Josip Stepanić and 12 2024 from Mislav Stjep an Žebec, 2024. Review of Recent Literature about Leisure https://www.thecorporategovernanceinstitute.com/insights/guides/what-is-Time of School-Aged Children and Youth in Croatia. Interdisciplinary esg-and-why-is-it-important . Description of Complex Systems 22 (1), 25-58. DOI: [3] Claudius Gros, 2024. Complex and Adaptive Systems. Spinger Cham. DOI: https://doi.org/10.7906/indecs.22.1.2 . https://doi.org/10.1007/978-3-031-55076-8 . [28] Christian A. Kl ö ckner, 2015. The psychology of pro- environmental [4] Simon A. Levin, 2003. Complex and Adaptive Systems: Exploring the communication – Beyond standard information strategies. Palgrave Known, the Unknown and the Unknowable. Bulletin of the American Macmillan, London. DOI: https://doi.org/10.1057/9781137348326 . Mathematical Society 40 (1), 3-19. [29] Wiek, A., Withycombe, L., & Redman, C. L. (2011). Key competencies in [5] Joshua M. Epstein and Robert Axtell, 1996. Growing artificial societies: sustainability: a reference framework for academic program development. social science from the bottom up. Brookings Institution Press. ISBN 978-Sustainability Science, 6(2), 203-218. 0-262-55025-3. [30] Fang Zhang, 2024. Enhancing ESG learning outcomes through [6] Urban Kordeš, 2005. Entropy – our best Friend. Interdisciplinary gamification: An experimental study. PLoS ONE 19 (5), e0303259. DOI: Description of Complex Systems 3 (1), 17-26. https://doi.org/10.1371/journal.pone.0303259 . [7] Katalin Martinás,2012. Greatest Happiness Principle in a Complex System [31] Diane F. Halpern, 1998. Teaching for critical thinking: helping college Approach. Interdisciplinary Description of Complex Systems 10 (2), 88-students develop the skills and dispositions of a teaching critical thinking 102, DO: https://doi.org/10.7906/indecs.10.2.5 . for transfer across domains: dispositions, skills, structure training, and [8] Katalin Martinás and Zsolt Gilányi, 2012. Greatest Happiness Principle in a metacognitive monitoring. The American Psychologist, 53 (4), 449-455. Complex System: Maximisation versus Driving Force. Interdisciplinary DOI: https://psycnet.apa.org/doi/10.1037/0003-066X.53.4.449 . Description of Complex Systems 10 (2), 103-113. DOI: [32] Kelly Y.L. Ku and Irene T. Ho, 2010. Metacognitive strategies that enhance https://doi.org/10.7906/indecs.10.2.6 critical thinking. Metacognition Learning 5 (3), 251-267. DOI: [9] Sabine Hossenfelder, 2013. On the Problem of Measuring Happiness. https://doi.org/10.1007/s11409-010-9060-6. Interdisciplinary Description of Complex Systems 11 (3), 289-301. DOI:

https://doi.org/10.7906/indecs.11.3.2.



Mind, the Gap, and Other Cracks



Maša Poljšak Kus Urban Kordeš

Center for Cognitive Science Center for Cognitive Science

Faculty of Education Faculty of Education

University of Ljubljana University of Ljubljana

m.poljsak.kus@gmail.com urban.kordes@pef.uni-lj.si



Abstract describing our conscious experience to another human being – in

this paper we call this process of reporting our subjective

With this paper we aim to outline numerous gaps and other experiences experiential translation. The subjective aspects of

cracks that emerge when we start researching conscious thinking, perceiving and feeling are all states of experience that



The terms used to name various gaps were chosen for the sake of Nagel [2] puts it, there is something it is like to be a conscious coherence (with a pinch of playfulness). The main gap is the organism, and this experience through first and second-person research approaches. have a certain way in which we experience them. As Thomas

chasm between two consciousnesses which we are trying to likely always, over an insurmountable chasm between one what is it like to be another organism is, most

bridge by an exchange of descriptions of our lived experiences. conscious organism and another.

When we begin to turn our awareness to what it is like to be we We aim to address this chasm that extends from one

begin to develop the skill and way of observing in which experiential being to another and explore the cracks that emerge

experience is created. We call this gap between our everyday when trying to explore and extend from one ridge to another. In

attitude and phenomenological observation the crevice of this analogy the ridges of the chasm between two

awareness in which lies the act of becoming aware of an consciousnesses represent different conscious organisms, each

experience. After becoming aware of a certain layer of with their own what it is like to be, and the chasm is the

experience we reach the fissure of description, which represents impossibility of reaching the exact what is it like of another

the crack between the actual experience as perceived and the being. In the field of first and second-person research of lived

constructed linguistic concepts in which we try to convey what experiences, researchers are trying to bridge this chasm by



experience has been produced, the researcher interested in in the act of producing and collecting such descriptions we investigating human experience is confronted with the cranny of stumble upon many cracks, located on both sides of what we call and how we perceived the experience. When a description of an collecting detailed descriptions of experience. We will argue that

comprehension. We relate this process of conveying our the chasm between two consciousnesses. Starting from first-

experience to another conscious being to the processes of person view (as one should, when going about empirical

translation and remind researchers of lived experience to be phenomenology) we stumble upon a crevice that is becoming

careful and weary of the interpretation that inherently shadows aware of what is it like to be – the most intimate experience, yet

every translation. often hidden behind a wall of what Edmund Husserl [3] calls our



Keywords In this paper we also touch another, an even more veiled natural attitude.

Background experience, Conscious experience, First-person dimension of experience that we call background experience

research, Second-person research, Experiential translation. (explained further in the section 3) but most importantly we state

that there is a gap between what we can easily consciously



1 Introduction call perceive and what we cannot – which, for the sake of clarity, we the crevice of awareness. When trying to convey one’s

Upon delving into topics and discussions regarding our experience to another we stumble upon the next gap in the act of

understanding of the mind, we inevitably reach one or another translating the experience into concepts, categories and linguistic

gaping chasm – the most notorious is even named the hard forms. We believe that there is a gap between our perceived

problem of consciousness. David Chalmers [1] points out that experience and its description, which we name the fissure of

there is nothing we know more intimately than conscious description. When trying to fill this fissure we believe the

experience but there is also nothing harder to explain. In this experience conveyed is flattened and reduced. The description

article we are not trying to explain conscious experience, but we produced in this effort then becomes the main building block of

are interested in exploring the process of explaining and the bridge we are building from one side of the chasm and what

we can offer to the conscious being reaching out from the other

side. In this paper we compare this act of describing on one side

Permission to make digital or hard copies of part or all of this work for personal or and comprehending on another as a process of translation and

classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full that practicing experiential translation is the way to more valid



citation on the first page. Copyrights for third-party components of this work must and richer descriptions. be honored. For all other uses, contact the owner/author(s). Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.cog.20



2 Experiential translation 2.1 First-person translation

Phenomenology, by origin a philosophical discipline, is trying to Jakob Boer [6] argues that the process of describing first-person

investigate concrete experiential phenomena and encourages experience is an act of experiential translation, with which we are

detailed analysis of different aspects of consciousness. As such inclined to very strongly agree. We believe that the act of

it has also been described as “a first-person description of ‘what describing subjective experience is an act of translation (latin

it is like’ of experience” [4]. This subjective dimension ‘as it is transfero, “I convey”, from prefix trans-, “across, beyond” and

lived from the inside’ is essential to consider in the field of participle latus “borne, carried”). We will describe an example

scientific investigation of cognition and not be constrained of a process of translating an ancient Greek text to a modern

merely to the data that can be observed and measured from the language. The underlaying assumption is that without an

outside [5]. observer there is no meaning, and thus the nature and skill of the

Claire Petitmengin [5] warns us that describing one’s own observer influence the source text immensely. Firstly, one must

subjective experience is not merely hard, but extremely difficult, be able to see the Greek alphabet and know the symbols to

mostly because turning our attention to our consciousness, and a perceive anything more than mere scribbles. Secondly, one must

fortiori describing it, requires inner effort and a specific kind of understand what a specific set of symbols denotes and relate to it

skillset. Her assumption is that a substantial proportion of our a previously known meaning - one must understand the word.

subjective experience unfolds below the threshold of This step alone is complex and multidimensional, because one

consciousness. We question what her assumption presupposes - Greek word can have numerous possible translations and the

that our consciousness is something "in there" to be observed and meaning that stands out to the translator is tied to many factors,

we only need a better instrument to see further and better. We, such as context and previous knowledge. Thirdly, one must

on the other hand, are more inclined to view conscious understand the grammar and syntax to make sense of a sentence.

phenomena as something co-created with and by the act of With this we want to show how the meaning of a text is co-

observation. In either case we believe that in the field of first- defined by the observer. The translator must then choose an

person research the two initial steps – becoming aware of our accurate set of words in another language to convey his

experience and then describing it - include two important gaps. interpretation of the sentence. With this example we tried to show

1. The crevice of awareness is the crack between what our the complexity of our influences on what we perceive and how

"view from within" knows how to observe and what we leave a mark on both our perception and our description.

eludes our reflective thoughts. It denotes the Experiential translation assumes that lived experience is in



2. feelings and core dimensions of our experience. verbalizing we carry certain aspects across the gap between The fissure of description is the gap between subjective experience and description. In the act of translating our lived difficulties of becoming aware of our background nature distinct from linguistic form, and that in the act of

observations of lived experience and descriptions of

experiences into words, concepts, and categories we inherently

observed experience, which are most often verbal. This

imbue chosen meanings with our interpretation, which is perhaps

is perhaps at times even more frustrating, because in an

inseparable from the way we become aware of our experience.

instance when one has become aware of an experience,



they must now try to find the right words and gestures experience We relate this intrinsic interpretation to horizons of attending to



research when we not only try to surmount Subjective, or first-person research transfers to second-person nature of the subjective experience in question. up being observed and reported. The horizon is the way in which we perceive, by which we mean co-create, our experience. This the fissure of is enacted both when we try to observe and when we try to to convey and verbalize a description that captures the who argue that this co-defines experiential phenomena that end , as explained by Urban Kordeš and Ema Demšar [7],

description, but we also convey this description to a researcher

describe our experience.

interested in exploring structures of lived experiences. Empirical



or second-person research usually involves interviewing human 2.2 Second-person translation participants about their experience. In the context of our paper,

we call the interview method a rather wobbly bridge that tries to In the previous section we compared the process of describing

connect participants’ lived experience with researcher’s one’s lived experience to the process of translation. We continue

understanding with this analogy in the case of second-person research, when via the participants’ description of experience.

3. such translation is perchance more intuitive, because the ‘input’ With this bridge we mark the third gap in the premise

of empirical research of subjective experience – – verbal report – comes in form of language. The researcher that the

cranny of comprehension receives the report proceeds with translating it in more than one – which spans between the

second-person investigator and first-person report way. First and foremost, the translation happens instantaneously,

about the experience. It is a gap each researcher must as it does every time we speak to another human being – we

fill and bridge when trying to comprehend and analyze translate the words into our own known concepts and position

the descriptive data on experience of others. We them in our pre-existing field of knowledge.

differentiate this cranny from the Even more importantly, we aim to compare the subsequent chasm between two

consciousnesses process of analysing, categorizing and forming conclusions on because it is focused on the

description and comprehension, not the entirety of the structure of experience to the process of translating, drawing

another conscious experience. attention once more to the notion that with translation always

We note that all three cracks are part of the greater comes interpretation. As such we want to note and warn that chasm

between two consciousness becoming aware of your own horizons of attending to experience , which refers to the impossibility of

experiencing as another being. is a crucial step for every second-person researcher of



consciousness, which inherently makes them a first-person 4 Empirical context researcher as well.

In the aforementioned project, ‘Unveiling of the Atmosphere –

Etnophenomenological exploration of experiential background



3 in relation to space’, we aimed to investigate background Background experience

experience which we have defined as feelings that weave the

In this paper we turn our attention to a layer of experience which foundation on which foreground phenomena of consciousness

is, ironically, not in the focus of our attention but rather on the unfold (such as emotions, thoughts and perceptions). We

brink of it. William James [8] refers to this as the fringe of presupposed that experiences of atmosphere are by their nature

consciousness. To this fringe belong experiences that lack affective, so we focused on the affective layer of experience.

specific, sensory qualities, like the tip-of-the-tongue state (the These feelings usually lack specific sensory attributes and are

intention to seek a missing word), feelings of knowing, hard to pinpoint and often notice and/or name. We tried to

familiarity and plausibility, intuitive judgments and numerous capture and convey such background feelings with an empirical

other conscious or quasi-conscious events that can be reported on approach and a qualitative research design in which we combined

with low sensory specificity. approaches of first-person research such as Descriptive

What is it like aspect of those experience is hard to perceive Experience Sampling Method (DES) [13], and ethnographical

and convey, but Petitmengin [9] describes certain internal tools such as in situ diary entries. Our study was conducted in

gestures, which serve, in the language of our analogy, as bridges three phases, the first being the pilot study. We recruited three

that enable us to become aware of the source dimension of our participants, previously trained in DES and first-person research,

experience, which is usually pre-reflective. This unarticulated which we deemed important for a study that aims to research pre-

dimension is considered as core due to its ever-present nature, reflective dimensions of experience.

and because it is pre-conceptual and pre-discursive, it seems to Our participants reported about their experience in three

be situated at the source of our thoughts. Although it constantly ways: 1) through short written reports about randomly sampled

accompanies us, we need special circumstances to become aware moments during the day, 2) with diary entries on multiple

of it and/or specific training in first-person observation. occasions during the day of sampling, in which they situated

In the realm of emotion, Antonio Damasio [10] calls a group randomly sampled moments in the context of their moods and

of fleeting and hard-to-name feelings ‘background feelings’, behaviors, 3) in interview sessions in which we explored and

because they are not in the foreground of our mind, yet they help expanded previous two types of data. The aim was to map our

define our mental state and color our lives. We relate the participants’ affective experiential landscapes and to

foreground of our mind with the experiences on which we can contextualize their experiences with information about their

easily focus our attention (such as thoughts, perceptions and loud activities, environment and social interactions. We have

emotions). Background feelings arise from background analyzed the data according to the principles of qualitative

emotions, which are directed more internally than externally, but analysis [14], which produced a list of experiential categories

can nevertheless be observable to others in several ways: tone of divided into two (vaguely distinct yet obviously separate) groups

our voice, prosody of our speech, the speed and design of our of foreground and background affective experience. In the

movements. According to Damasio, prominent background background we situated categories such as background mood,

feelings include fatigue, energy, excitement, tension, relaxation, ambient atmosphere and deep atmosphere. stability, instability, etc. The relation between background



relation between background feelings and consciousness is just 1. Background mood is felt as all-encompassing and includes feelings and our drives and moods is intimate and close, but the



as close, if not more. Matthew Ratcliffe [11] similarly develops foreground experiences (affects, thoughts and percepts), different ways of receiving, creating and experiencing

the term existential feeling as a background which comprises the which we call different attitudes. We found three

very sense of ‘being’ or ‘reality’ that attaches to world subcategories of background mood: open, closed and numb.

experiences. Specifically directed emotions presuppose this 2. Ambient atmosphere includes experiences that are not clear

background, so regardless of the structure of such emotion, and separate, but pervasive and ubiquitous. It represents

existential feelings are a more fundamental feature of world- feelings, which we feel originate from the world, and we are

experience. A few examples of such feelings are the feeling of entangled with it either as their co-creator or merely as an



descriptions of one’s relationship with the world. 3. Deep atmosphere includes experiences that we feel as deeply our own and private. Imprint of deep atmosphere Hopefully we have now outlined the gap between our focal being ‘complete’, ‘unworthy’, ‘at home’, ‘abandoned’ – all being observer.



awareness and the experiences on the fringe of consciousness, background feelings. Phenomenologically it is harder to marks the way of foreground affects as well as other

where perhaps one of the keys to understanding our mind lies reach and observe, as it usually changes its character less or

hidden. This gap was one of the points we tried to address in our more slowly. When captured, we observed two distinct

recent project [12], in which we investigated the feelings of subcategories of feelings: deep perturbation and deep

atmosphere with the presupposition that they are in the unconcern, the former connected to the feelings of danger

background of our mind. We will briefly present the context of and the latter to the feelings of safety.

our empirical investigations to use it as the reference point for

our observations regarding the numerous gaps and blind spots of

our methodological approach and epistemological premises.



5 Observational interstices less so on the background feelings. When participants weaved

those moments in the experiential timeline of their whole day

In this section we aim to address some methodological cracks

(and in the interviews of their whole week) more background

and to note our observations from our research project on

feelings came into light – even in the moments which we had

background feelings [12].

detailed descriptions of. We would like to note that minimising



5.1 the effect of memory on reports is important, but that sometimes Becoming aware

in this effort we miss something because it is ‘right under our

In our study participants were prompted with a signal which nose’.

conveyed to them that they should observe and report on their

experience of the moment right before the signal. During the

interviews they oftentimes reported that after the signal there was 6 Conclusion



a brief state of feeling ‘blank’, as if the moment before the signal Delving into the field of empirical phenomenology is a was empty and void of any experience whatsoever. But this courageous act, because there are few, if any, clear and firm feeling soon passed, and they started to remember and find words climbing holds. We understand why scientific discourse steers to describe the moment before the signal. We interpret this toward replicable and third-person tested approaches, yet we feeling of ‘blankness’ as a type of gap between being immersed believe that exploration of lived experienced cannot (at least as in the natural attitude [3] and adapting the phenomenological of yet) be accessed any other way than through subjective attitude . To put it differently – we believe that the act of epoché observation first. And even if the act of bridging the subjective is both and act of opening a gap and of bridging it. We argue that with intersubjective is full of gaps and other cracks, we stay each time we try to bracket our trust in the objectivity of the positive that the descriptions and interpretations produced in this world, we reveal and/or create a crack in the fluidity and process lead to better understanding of how to approach continuity of the flow of our conscious experience. This means empirical research of subjective experience. In the analogy of that when we change the nature of our awareness, we experience translation as the act of describing one’s own experience, we aim a moment of emptiness. To explain we will compare our to paint the following picture. In the gaps that lurk amid awareness with the grip of our hand. When we hold on to one experiencing, being aware and describing, many pieces of the object, let’s say a glass, we are gripping something and sensing original experience are most likely lost in translation. Yet by specific qualities. When we want to switch to a different object, persistently and methodically carrying over the remaining pieces we must first release the glass and be (and thus feel) empty at created by this process we are building better and more reliable least for a moment so that we can grip (experience) something bridges. else.



5.2 Observing experiential background 7 References

As mentioned in the section 2, we tried to observe and capture [1] David Chalmers, 1995. Facing up to the problem of consciousness.

background feelings with the intention of mapping participants’ Journal of Consciousness Studies 2 (3), 200-2019.

experiential landscapes of affects. Based on the literature and [2] Thomas Nagel, 1974. What Is It Like to Be a Bat ? The Philosophical

Review, vol. 83, No. 4. Duke University Press.

preliminary observations we presupposed that background [3] Edmund Husserl, 1983. Ideas Pertaining to Pure Phenomenology and to

feelings change less frequently, which is one of the reasons they a Phenomenological Philosophy. Springer.

[4] Shaun Gallagher & Dan Zahavi, 2008. The Phenomenological Mind.

are more elusive and harder to notice, as opposed to the Londond : Routledge.

foreground experiences which change from moment to moment [5] Claire Petitmengin, 2006. Describing one’s subjective experience in the

and require most of our attention. second person: An interview method for the science of consciousness.

Phenomenology and the Cognitive Sciences, 5(3–4), 229–269.

Our findings support our claim that one way to notice the ever https://doi.org/10.1007/s11097-006-9022-2

present is by gaps in continuity. Such a way requires regular first- [6] Jakob Boer, 2023. Phenomenology as Experiential Translation: Towards

a Semiotic Typology of Descriptive and Expressive Ways of Making

person observation, optimally supported by a second-person Sense of Experience. Critical Arts. DOI:



approach (dialogue). Noting one’s experience often over a longer 10.1080/02560046.2023.2262520 [7] Urban Kordeš & Ema Demšar 2021. Horizons of becoming aware: period can bring to light changes that unravel slowly. To explain Constructing a pragmatic-epistemological framework for empirical first-

this with a more concrete and visual analogy – when a person on person research. Phenomenology and the Cognitive Sciences, 22(2), 339–

a diet is losing weight (if they are doing it in a healthy and 367. https://doi.org/10.1007/s11097-021-09767-6

[8] William James, 1890. The Principles of Psychology. New York : Holt.

sustainable way) they won’t see any progress from day to day, [9] Claire Petitmengin, 2007. Towards the Source of Thoughts : The Gestural

but if they observe and measure themselves methodically and Transmodal Dimension of Lived Experience. Journal of

Consciousness Studies, 14, No. 3, 54–82.

throughout the whole year, they can notice a vast difference from [10] Antonio Damasio, 1999. The Feeling of What Happens : Body and

their starting point. emotion in the making of consciousness. New York : Harcourt Brace.

[11] Matthew Ratcliffe, 2005. The Feeling of Being. Journal of Consciousness



5.3 Studies, 12 (8-10), 45-63. Describing lived experience

[12] Maša Poljšak Kus, 2024. Unveiling of the atmosphere :

etnophenomenological exploration of experiential background in relation

Tying to the conclusion of the previous paragraph is a very to space (Master thesis), Ljubljana. Supervisor : Urban Kordeš.

concrete observation based on our research methodology. As [13] Russell T. Hurblurt & Christopher L. Heavey, 2006. Exploring inner

described in section 2, we gathered reports on our participants’ experience : The descriptive experience sampling method. Amsterdam,

John Benjamins Publishing Co.

experience in three ways (short notes on experience of moments [14] Blaž Mesec, 2023. Kvalitativno raziskovanje v teoriji in praksi. Inštitut za

during the day, diary entries and interview insights). What we razvojne in strateške analize.

noticed is that often in the descriptions of a singular moment

there was a lot of emphasis on the foreground experiences and



Bridging the Challenges in Experience Sampling Research



Barbi Seme† Maruša Sirk Urban Kordeš

Center for Cognitive Science Center for Cognitive Science Center for Cognitive Science

University of Ljubljana University of Ljubljana University of Ljubljana

Ljubljana, Slovenia Ljubljana, Slovenia Ljubljana, Slovenia

Barbi.Seme@pef.uni-lj.si Marusa.Sirk@pef.uni-lj.si Urban.Kordes@pef.uni-lj.si



Abstract on numerous studies [4, 5, 6] that have used ESM to investigate

experiential phenomena in the past few decades, weaknesses of

In this paper, we draw parallels between existing research these methods have been identified [7].

practices and attempt to piece them together to propose a more We present ESM and the challenges inherent in ESM research,

wholesome approach in conducting experience sampling particularly participant burden. By exploring the interest in

research. We consider Experience Sampling Methods (ESM) as personal exploration within Personal Science and emphasising

valuable tools for studying experience, but they come with the importance of community building in Citizen Science, we

challenges, of which we address the participant burden as one of attempt to tie these practices together using the concept of a

the most significant ones. We think that integrating practices methodological turn from empirical phenomenology [8]. We

from Personal Science (PS) and Citizen Science (CS), grounded believe that the challenge of participant burden, which we see as

in empirical phenomenology, can help address this challenge. By under addressed but highly disruptive in ESM scientific inquiry,

considering participants as co-researchers and actively engaging can be tackled through this integration of different research

them in the research and community, we aim to enhance their practices. We illustrate this approach with our pilot study, Luna.

motivation and improve the quality of the research data. We

illustrate this approach through the pilot project Luna in which

we explore lived experiences throughout the menstrual cycle 2 Experience sampling research using the ESM mobile application "Curious". This integrative

method facilitates a reciprocal knowledge exchange between We consider Experience Sampling Methods (ESM) as an

researchers and co-researchers, which deepens the process of umbrella term for the research in which participants gather

self-exploration and holds a great potential to advance scientific samples of their experiences as they unfold in their life [9].

research on experience. Typically, we prompt participants at random times to answer

questions or to describe their experience of the moment just



Keywords before they heard the beep [10]. This way we are able to

minimise recall bias [7] and are able to sample dimensions of

Experience sampling methods, citizen science, personal science, experiential states which are nearly impossible to recall later,

empirical phenomenology especially in detail (e.g., the momentary content of our thoughts).

These methods are also highly ecologically valid, since we are



1 Introduction lives. Participants would receive the prompts several times per sampling experience as it unfolds naturally in people's everyday

Scientific research into experience is a rapidly growing field. day for a longer period of time (e.g., two weeks). These repeated

Some researchers and philosophers point out that a core problem measures enable us to track patterns and changes in individual

within our current scientific worldview is the overlooked experiences across time and different contexts [7, 10]. Nowadays

experience research [1]. New methods and tools for researching we use mobile applications on participants' personal smartphones

experience are being developed, among which are Experience which makes the data collection process in comparison to pen

Sampling Methods (ESM). ESM are intensive longitudinal and paper much more reliable and less burdensome [11].

approaches to collecting experiential and contextual data using

structured diary self-report techniques [2]. Due to numerous 2.1 Challenges in ESM

advantages, especially ecological validity and the reduction of ESM research is still loosely defined without a rigorous

recall bias, ESM has spread to various research fields through the framework and we are yet to develop appropriate methodological

use of mass technology, mostly mobile applications [3]. Based approaches for improvements [9]. A significant challenge in

ESM research is participant burden. Collecting frequent, real-

Permission to make digital or hard copies of part or all of this work for personal or time data in everyday life activities puts great demands on

classroom use is granted without fee provided that copies are not made or distributed participants who need to albeit shortly interrupt their activity to

for profit or commercial advantage and that copies bear this notice and the full report on their momentary experiential state [3]. These repeated

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s). measures over time might affect participants' attitudes towards

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia the research and result in reduced compliance, careless

© 2024 Copyright held by the owner/author(s).

responding and participants’ attrition [12]. We should also

evaluate this burden from an ethical perspective, ensuring that



benefits outweigh the burden, especially if there is a Individuals and groups who already possess deep interest in self-

consideration of affecting participants’ well-being [13]. exploration can potentially become great co-researchers by

which they would gain support in their own exploration as well

2.2 Existing recommendations for addressing as make the scientific contribution.

the challenges in ESM

To mitigate this burden, researchers are working on

4 Community in Citizen science

questionnaire optimisation, making them as brief and focused as

possible by prioritising essential questions and using clear and Citizen Science (CS) is recognized as one of the eight pillars of

concise language [9]. The trade-off between data richness and Open Science, playing an integral role in democratising scientific

burden on participants is also mitigated with lowering the knowledge and practices [23]. It significantly bridges the gap

sampling frequency [12]. By introducing the personalised between the scientific community and society through the idea

scheduling for participants we avoid interrupting them in the of doing science and not merely reacting to it [24]. Due to the

situations in which they are unable to respond and would likely heterogeneity of CS projects in terms of scale, objectives, and

react negatively to prompts being delivered in that time [3]. levels of citizen scientist involvement, it is challenging to

Albeit we should then revise random sampling and account for provide a universal definition [25]. However, common to all CS

the introduced bias (e.g., we could supplement the data with projects is to actively involve non-professionals in scientific

retrospective daily reports). research at different levels of participation [26]. In a broad sense

However, a key factor for successful ESM research is citizen scientists perform tasks that would be otherwise done by

participants’ motivation [3, 7]. It is better sustained by scientists [27] or would not be possible to do without their

considering the necessary technical recommendations, but involvement.

researchers should also give attention to fostering the research To achieve reciprocity between science and society in CS

interest and social dynamics. Researchers should engage projects in which the bidirectional knowledge exchange

participants who already have interest in the research topic and facilitates benefits in both [28], significant time and resources

therefore an intrinsic motivation to learn more on it [3]. They need to be invested to establish the conditions for project

should also provide a sufficient training period in which activities to run [29]. Since citizen scientists are typically lay

participants gain the necessary knowledge and skills in order to people without formal training in scientific research, appropriate

sample the experience [7]. It is important to establish a rapport training and support are essential to equip them with necessary

with participants and to foster a research alliance throughout the skills and knowledge [30]. We know citizen scientists engage in

study [2]. It is suggested to provide rich feedback to participants the projects upon different motivation factors. We can observe

during and at the end of the study which can also be presented as the intrinsic factors, such as gaining fulfilment, enjoying the

non-monetary compensation [3]. activities or being altruistic and extrinsic factors, such as building

social interactions, gaining on reputation or status and expecting

future returns [31]. Therefore, sustaining motivation and

3 Personal science and interest for self- engagement requires more than just training. CS practitioners

exploration should establish good relationships with citizen scientists and a



Technological advancement played a great role in a growing scientists to meet and work with each other. We argue that continuous communication as well as the conditions for citizen

number of ESM studies as well as in an uptake of self-tracking

essential to the project's success is building a strong community.

practices for exploration of oneself [14, 15, 16]. The umbrella

Utilising online community spaces, social media, organising

term for self-tracking practice and communities has been formed



under Personal Science (PS) [17]. These individuals and groups and other social activities facilitate community building. Strong workshops and training as well as local meetups, collaborative

pursue their own personal research questions using empirical



methods in an iterative process of questioning, designing, knowledge sharing, foster collaboration and builds on community in exchange encourages participation, promotes



observing, reasoning, and discovering which presents itself as an sustainability of the project [32, 33]. opportunity to scientifically expand on PS. Even though we can

draw many parallels between self-tracking and scientific inquiry

the question remains to what extent PS can be scientifically 5 Bridging ESM, PS and CS with empirical interesting [18]. Considering the growing interest in PS activities

phenomenology

it seems important to address these practices, especially in new

self-trackers. They often experience difficulties in making sense The key to integrating the practices of Citizen science, Personal

of their self-tracking process in interpreting their data, science, and Experience Sampling research lies in the concept of

formulating and refining their research questions, and designing a methodological turn developed in the field of empirical

their research process [19, 20]. It would be beneficial for them to phenomenology [8]. In experience research, the observed is the

receive support that provides at least an initial establishment of observer, meaning that the only access to the phenomena of

their research or engaging them in a more systematic way. Lack interest is through the observer's subjective experience. If the

of scientific rigour was also reported by researchers in tools used observer does not adopt an attitude of curious exploration and

for self-tracking which can potentially mislead self-trackers and engage in epoché, meaning bracketing the natural attitude, the

give them false ideas of phenomena they explore [21, 22]. Hence, judgments, interpretations, and explanations of their experience,

we believe this is an excellent opportunity for the science we cannot obtain data on the genuine experience as it unfolds in

community to engage in this already widespread phenomenon. life. This notion is rooted in phenomenological reduction, a



method of research into experience developed by Edmund 7 Conclusion Husserl [34]. Experience Sampling has been used to study

We argue that a large number of research questions in ESM

subjective experiences in real-time contexts, but integrating it

research could be better investigated if research projects adopt

with phenomenological reduction enhances the depth of data on

the CS framework with an emphasis on community building,

lived experiences [35]. Therefore, it is necessary to consider our

account for the interest and practice of PS, and use the principles

participants as co-researchers. This attitude allows us to engage

of experience investigation from empirical phenomenology. The

them in a way that fosters their interest in the research question



which facilitates the methodological turn where the research by creating conditions for co-researchers to be involved in challenge of participant burden in ESM research is then mitigated

question becomes in a sense their own and they become

personally meaningful activities, which in return provide a

researchers of their own experience. While providing the support



and means for investigation, it is important to give co-researchers are needed to establish these kinds of project communities, we higher level of data validity. Even though considerable resources

the freedom to explore the research question and their experience

believe they have the potential to be more sustainable. From CS,

in a way that is meaningful to them, and to encourage critical

we know that citizen scientists develop a sense of community,

discussion. By opening up the space for co-creation of the

which encourages them to remain active in science after the

research design and enabling co-researchers to actively

initial project ends [36, 37]. This interdisciplinary integration of

contribute their findings, we facilitate a deeper reciprocity of

different research practices enhances the value of our

knowledge transfer.

investigations and creates more impactful and sustainable

research projects that benefit both the scientific community,



6 involved co-researchers and the society. The pilot study “Luna”



Citizen science project Luna aims to explore the lived Acknowledgments experiences of menstrual cycles and their impact on everyday

IMPETUS is supporting our project, Luna. IMPETUS is funded

well-being. We use a diary method for daily reports and

by the European Union’s Horizon Europe research and

Experience Sampling Methodology (ESM) to track experiences

innovation programme under grant agreement number

throughout the menstrual cycle with the use of the ESM mobile

101058677. Views and opinions expressed are, however, those

application Curious (about) consciousness or Curious in short.

of the author(s) only and do not necessarily reflect those of the

We adopted the iterative co-creation approach to develop our

European Union or the European Research Executive Agency

research, combining the principles of ESM research, CS projects,

(REA). Neither the European Union nor the granting authority

PS and empirical phenomenology. This makes our research

can be held responsible for them.

process flexible in a way that the research design is being updated



in an iterative collaborative manner. We engage co-researchers References in the design and assessments of data collection procedure and

questionnaires, data interpretation as well as analysis. By [1] Frank, A., Gleiser, M., & Thompson, E. (2024). The blind spot: why

science cannot ignore human experience. MIT Press.

encouraging them to develop their own research questions which DOI:https://doi.org/110.7551/mitpress/13711.001.0001

are relevant to their own experience we promote their personal [2] Hektner, J. M., Schmidt, J. A. & Csikszentmihalyi, M. (2007). Experience

sampling method: Measuring the quality of everyday life. Thousand Oaks,

endeavor in the research. They are also involved in other project CA: Sage Publications. DOI: https://doi.org/10.4135/9781412984201

activities, such as sharing results and designing the project visual [3] van Berkel, N., & Kostakos, V. (2021). Recommendations for conducting

identity. We organise different learning and sharing community longitudinal experience sampling studies. Advances in longitudinal HCI

research, 59-78. DOI: https://doi.org/110.1016/10.1007/978-3-030-

activities. In the training workshops we introduced them to the 67322-2_4

principles of observing ones’ own experience, how to bracket the [4] MacKerron, G. in Mourato, S. (2013). Happiness is greater in natural

environments. Global environmental change, 23(5), 992-1000.

natural attitude and report on the pristine inner experience. On DOI:https://doi.org/110.1016/j.gloenvcha.2013.03.010

our online community space as well as in the organised meetups [5] Kane, M. J., Brown, L. H., McVay, J. C., Silvia, P. J., Myin-Germeys, I.

we share feedback on the research and their participation, engage in Kwapil, T. R. (2007). For whom the mind wanders, and when: An

experience-sampling study of working memory and executive control in

them in the conversation on the research topic related questions daily life. Psychological science, 18(7), 614-621.

and encourage them to share their own feedback. They are also DOI:https://doi.org/110.1111/j.1467-9280.2007.01948.x

[6] Mölsä, M. E., Lax, M., Korhonen, J., Gumpel, T. P., in Söderberg, P.

invited to share any findings they have along the way in the (2022). The experience sampling method in monitoring social interactions

mobile application. The Curious app is designed in a way to among children and adolescents in school: a systematic literature review.

support co-researchers in their data interpretation by providing Frontiers in Psychology, 13, 844698.

DOI:https://doi.org/110.3389/fpsyg.2022.844698



them with simple visualisations of their gathered data which they [7] Scollon, C. N., Kim-Prieto, C., & Diener, E. (2003). Experience sampling: Promises and pitfalls, strengths and weaknesses. Journal of Happiness are able to filter, compare and reflect on. Our aim in the research studies, 4(1), 5-34. DOI:https://doi.org/10.1023/A:1023605205115

is two-fold. It is driven to answer the research questions on [8] Kordeš, U. in Klauser, F. (2016). Second-Person in-Depth

phenomenology throughout the menstrual cycle as well as to Phenomenological Inquiry as an Approach for Studying Enaction of

Beliefs. Interdisciplinary Description of Complex Systems, 14(4), 369-

equip co-researchers with methods and tools to research their 377. DOI: https://doi.org/10.7906/indecs.14.4.5



experiential landscape and gather self-knowledge. Project tries to [9] Myin-Germeys, I. in Kuppens, P. (Ur.). (2022) The open handbook of experience sampling methodology: A step-by-step guide to designing, go beyond research by also opening up the data driven discussion conducting, and analyzing ESM studies (2nd ed.). Leuven: Center for

on possible systemic or social solutions to consider the Research on Experience Sampling and Ambulatory Methods Leuven.

physiological and phenomenological cyclic nature of people with [10] Christensen, T. C., Barrett, L. F., Bliss-Moreau, E., Lebo, K., & Kaschub,

C. (2003). A practical guide to experience-sampling procedures. Journal

menstrual cycles. This also presents a motivational factor for of Happiness Studies: An Interdisciplinary Forum on Subjective Well-

some co-researchers who would like to see positive (societal) Being, 4(1), 53–78. DOI:https://doi.org/10.1023/A:1023609306024

changes in regard to this phenomenon.



[11] van Berkel, N. V., Ferreira, D. in Kostakos, V. (2017). The Experience [24] Lang, D. (2016). “Science to the People. How citizen science bridges the

Sampling Method on Mobile Devices. ACM Computing Surveys, 50(6), gap between science and society.” In: Medium. Retrieved 22nd of Sep. 1–40. DOI: https://doi.org/10.1145/3123988 2024 from: https://fellowsblog.ted.com/how-citizen-science-bridges-the-

[12] Eisele, G., Vachon, H., Lafit, G., Kuppens, P., Houben, M., Myin- gap-between-science-and-society-d693af125ae4

Germeys, I., & Viechtbauer, W. (2022). The effects of sampling frequency [25] Haklay, M., Dörler, D., Heigl, F., Manzoni, M., Hecker, S., Vohland, K., and questionnaire length on perceived burden, compliance, and careless ... & Wagenknecht, K. (2021). What is citizen science? The challenges of

responding in experience sampling data in a student population. definition. The science of citizen science, 13. DOI: Assessment, 29(2), 136-151. DOI: https://doi.org/10.1007/978-3-030-58278-4_2

https://doi.org/10.1177/1073191120957102 [26] Shirk, J. L., Ballard, H. L., Wilderman, C. C., Phillips, T., Wiggins, A.,

[13] Kirtley, O. (2022). The open handbook of experience sampling Jordan, R., McCallie, E., Minarchek, M., Lewenstein, B. V., Krasny, M.

methodology: A step-by-step guide to designing, conducting, and E., & Bonney, R. (2012). Public Participation in Scientific Research: a

analyzing ESM studies (2nd ed.). I. Myin-Germeys in P. Kuppens (ur.), Framework for Deliberate Design. Ecology and Society, 17(2).

Ethical issues in experience sampling method research. Leuven: Center for DOI:https://doi.org/10.5751/ES-04705-170229 Research on Experience Sampling and Ambulatory Methods Leuven. [27] Heiss, R. & Matthes, J. (2017). Citizen Science in the Social Sciences: A

[14] Clark, M., Southerton, C., & Driller, M. (2024). Digital self-tracking, Call for More Evidence. In: GAIA - Ecological Perspectives for Science

habits and the myth of discontinuance: It doesn’t just ‘stop’. new media & and Society 26, pp. 22–26. DOI:https://doi.org/10.14512/gaia.26.1.7.

society, 26(4), 2168-2188. DOI: [28] Phillips, T., Porticella, N., Constas, M., & Bonney, R. (2018). A https://doi.org/10.1177/14614448221083992 framework for articulating and measuring individual learning outcomes

[15] Swan, M. (2013). The quantified self: Fundamental disruption in big data from participation in citizen science.

science and biological discovery. Big Data, 1(2), 85–99. DOI: [29] Heinisch, B., Oswald, K., Weißpflug, M., Shuttleworth, S., & Belknap, G. https://doi.org/10.1089/big.2012.0002 (2021). Citizen humanities. The science of citizen science, 97.

[16] Lupton, D. (2016). The Quantified Self: A Sociology of Self-Tracking. DOI:https://doi.org/10.1007/978-3-030-58278-4_6

Polity Press. [30] Pandya, R. E. (2012). A framework for engaging diverse communities in

[17] Wolf, G. I., & De Groot, M. (2020). A conceptual framework for personal citizen science in the US. Frontiers in Ecology and the Environment,

science. Frontiers in Computer Science, 2, 21. DOI: 10(6), 314–317. DOI: https://doi.org/10.1890/120007 https://doi.org/10.3389/fcomp.2020.00021 [31] Lotfian M, Ingensand J, Brovelli MA. (2020). A Framework for

[18] Heyen, N. B. (2016). Self-tracking as knowledge production: Quantified Classifying Participant Motivation that Considers the Typology of Citizen

self between prosumption and citizen science. In Lifelogging: Digital self- Science Projects. ISPRS International Journal of Geo-Information. 2020;

tracking and lifelogging-between disruptive technology and cultural 9(12):704. DOI: https://doi.org/10.3390/ijgi9120704 transformation (pp. 283-301). Wiesbaden: Springer Fachmedien [32] West, S., & Pateman, R. (2016). Recruiting and retaining participants in Wiesbaden. DOI: https://doi.org/10.1007/978-3-658-13137-1_16 citizen science: What can be learned from the volunteering literature?

[19] Rapp, A., & Cena, F. (2016). Personal informatics for everyday life: How Citizen Science: Theory and Practice, 1(2), 15. DOI:

users without prior self-tracking experience engage with personal data. https://doi.org/10.5334/cstp.8 International Journal of Human-Computer Studies, 94, 1-17. DOI: [33] Bonney, R., Phillips, T. B., Ballard, H. L., & Enck, J. W. (2016). Can https://doi.org/10.1016/j.ijhcs.2016.05.006 citizen science enhance public understanding of science? Public

[20] Choe, E. K., Lee, N. B., Lee, B., Pratt, W., & Kientz, J. A. (2014). Understanding of Science, 25(1), 2–16. DOI:

Understanding quantified-selfers' practices in collecting and exploring https://doi.org/10.1177/0963662515607406 personal data. In Proceedings of the SIGCHI Conference on Human [34] Husserl, E.: Aufsätze und Vorträge (1911–1921). Husserliana XXV. Factors in Computing Systems (pp. 1143–1152). ACM. DOI: Martinus Nijhoff, Doedrecht, 1987. https://doi.org/10.1145/2556288.2557372 [35] Hurlburt, R. T. in Heavey, C. L. (2006). Exploring inner experience: The

[21] Caldeira, C., Chen, Y., Chan, L., Pham, V., Chen, Y., & Zheng, K. (2017). descriptive experience sampling method (Vol. 64). John Benjamins

Mobile apps for mood tracking: an analysis of features and user reviews. Publishing. DOI: https://doi.org/10.1075/aicr.64 In AMIA annual symposium proceedings (Vol. 2017, p. 495). American [36] Rotman, D., et al. (2012). Dynamic changes in motivation in collaborative Medical Informatics Association. citizen-science projects. In Proceedings of the ACM 2012 conference on

[22] Larsen, M. E., Nicholas, J., & Christensen, H. (2016). Quantifying app Computer Supported Cooperative Work (pp. 217–226). ACM. DOI:

store dynamics: Longitudinal tracking of mental health apps. JMIR https://doi.org/10.1145/2145204.2145238 mHealth and uHealth, 4(3), e96. DOI: [37] Bela, G., Peltola, T., Young, J. C., Balázs, B., Arpin, I., Pataki, G., ... & https://doi.org/10.2196/mhealth.6020 Bonn, A. (2016). Learning and the transformative potential of citizen

[23] Ayris, P., Lopez de San Román, A., Maes, K., & Labastida, I. (2018). science. Conservation Biology, 30(5), 990-999. DOI:

Open science and its role in universities: A roadmap for cultural change. https://doi.org/10.1111/cobi.12762 Leuven: LERU Office. Retrieved December, 13, 2019.



Zbornik 27. mednarodne multikonference



INFORMACIJSKA DRUŽBA – IS 2024

Zvezek C



Proceedings of the 27th International Multiconference



INFORMATION SOCIETY – IS 2024

Volume C



Odkrivanje znanja in podatkovna skladišča - SiKDD



Data Mining and Data Warehouses - SiKDD



Urednika / Editors



Dunja Mladenić, Marko Grobelnik



http://is.ijs.si



7. oktober 2024 / 7 October 2024

Ljubljana, Slovenia





PREDGOVOR




Tehnologije, ki se ukvarjajo s podatki so močno napredovale. Iz prve faze, kjer je šlo predvsem

za shranjevanje podatkov in kako do njih učinkovito dostopati, se je razvila industrija za

izdelavo orodij za delo s podatkovnimi bazami in velikimi količinami podatkov, prišlo je do

standardizacije procesov, povpraševalnih jezikov. Ko shranjevanje podatkov ni bil več poseben

problem, se je pojavila potreba po bolj urejenih podatkovnih bazah, ki bi služile ne le

transakcijskem procesiranju ampak tudi analitskim vpogledom v podatke. Pri avtomatski

analizi podatkov sistem sam pove, kaj bi utegnilo biti zanimivo za uporabnika – to prinašajo

tehnike odkrivanja znanja v podatkih (knowledge discovery and data mining), ki iz obstoječih

podatkov skušajo pridobiti novo znanje in tako uporabniku nudijo novo razumevanje dogajanj

zajetih v podatkih. Slovenska KDD konferenca SiKDD, pokriva vsebine, ki se ukvarjajo z

analizo podatkov in odkrivanjem znanja v podatkih: pristope, orodja, probleme in rešitve.



Dunja Mladenić in Marko Grobelnik





FOREWORD




Data driven technologies have significantly progressed. The first phases were mainly focused

on storing and efficiently accessing the data, resulted in the development of industry tools for

managing large databases, related standards, supporting querying languages, etc. After the

initial period, when the data storage was not a primary problem anymore, the development

progressed towards analytical functionalities on how to extract added value from the data; i.e.,

databases started supporting not only transactions but also analytical processing of the data. In

automatic data analysis, the system itself tells what might be interesting for the user - this is

brought about by knowledge discovery and data mining techniques, which try to obtain new

knowledge from existing data and thus provide the user with a new understanding of the events

covered in the data. The Slovenian KDD conference SiKDD covers topics dealing with data

analysis and discovering knowledge in data: approaches, tools, problems and solutions.



Dunja Mladenić and Marko Grobelnik



PROGRAMSKI ODBOR / PROGRAMME COMMITTEE



Janez Brank, Jožef Stefan Institute, Ljubljana



Marko Grobelnik, Jožef Stefan Institute, Ljubljana



Alenka Guček, Jožef Stefan Institute, Ljubljana



Branko Kavšek, University of Primorska, Koper



Dunja Mladenić, Jožef Stefan Institute, Ljubljana



Erik Novak, Jožef Stefan Institute, Ljubljana



Inna Novalija, Jožef Stefan Institute, Ljubljana



Joao Pita Costa, Quintelligence, Ljubljana



Lui Rei, Event Registry, Ljubljana



Jože Rožanec, Jožef Stefan Institute, Ljubljana



Abdul Sitar, Jožef Stefan Institute, Ljubljana



Luka Stopar, SolvesAll, Ljubljana



Swati Swati, Bundeswehr University Munich, Munich



Jan Šturm, Jožef Stefan Institute, Ljubljana



Oleksandra Topal, Jožef Stefan Institute, Ljubljana



Integrating Knowledge Graphs and Large Language Models for



Querying in an Industrial Environment



Domen Hočevar Klemen Kenda

domenhocevar1@gmail.com klemen.kenda@ijs.si

Jožef Stefan Institute Jožef Stefan Institute

Ljubljana, Slovenia Ljubljana, Slovenia



Abstract To overcome these challenges, we propose a system that en-

ables users to interact with knowledge graphs through natural

Knowledge graphs have traditionally required the use of specific

language queries. The system leverages LLMs’ capabilities to

query languages, such as SPARQL, to retrieve relevant data. In

interpret knowledge graphs while compensating for their limited

this paper, we present a system capable of performing natural

ability to generate fully syntactically and semantically correct

language queries on knowledge graphs by leveraging retrieval-

SPARQL queries. Proposed system, depicted in Figure 1, lever-

augmented generation (RAG) and large language models (LLMs).

ages large language models (LLMs) [11] to process natural lan-

Our system can ingest large knowledge graphs and answer queries

guage inputs and provide responses in natural language. Our

using two approaches: first, by utilizing LLMs to extract informa-

approach integrates retrieval-augmented generation (RAG) tech-

tion directly from subgraphs; and second, by generating SPARQL

niques alongside the automatic generation of SPARQL queries

queries with LLMs and using the results to inform further infer-

based on natural language input [2].

ence, such as counting the number of items.



Keywords



knowledge graph, semantic inference, Industry 4.0, LLM, RAG



1 Introduction

In the context of Industry 4.0, knowledge graphs play a crucial

role in mapping and describing the entire production vertical,

from supply and demand dynamics to intricate details within the

production process. This includes the configuration of shop floors,

production lines, machines, and data setups, extending even to

specific datasets generated during operations. Knowledge graphs

can also include relevant information about the tools required for

particular processes, as well as details about personnel, including

their skills and roles.

A key standard for representing such data within the Industry

4.0 initiative is the Asset Administration Shell (AAS) [3], which

provides a logical representation for a factory asset (can also be

a piece of software, etc.). By adopting AAS, industries can en-

sure interoperability and standardization, enabling more efficient Figure 1: Intended usage of the system: AAS instances are

data exchange and integration across various systems, ultimately converted into a knowledge graph, enabling natural lan-

enhancing the agility and responsiveness of manufacturing pro- guage queries by the user.

cesses.

Querying knowledge graphs can be a challenging task for end

users, as it often requires expertise in specialized query languages By doing so, our system not only simplifies the querying pro-

such as SPARQL [8] — a skill that is not widely known among cess but also ensures that the responses are accurate and con-

non-experts. Working with SPARQL SELECT queries remains a textually relevant, making knowledge graphs more accessible

challenge also for LLMs, with performance varying significantly and usable for a broader range of users. Additionally, the use of

depending on the specific model and task complexity. While the LLMs in combination with SPARQL querying enables the system

leading LLMs can reliably address basic syntax errors, generating to handle complex tasks, including those that require logical rea-

semantically accurate SPARQL SELECT queries remains difficult soning, aggregation, or interpretation of data, thus enhancing

in many cases [10]. Similar work has been done on interaction its utility in real-world applications. For example, our system is

with databases, however even with SQL query generation the able to answer queries such as: “Give me all machines that

results of GPT-4 are still far behind human ability (approx. 55% are capable of drilling a hole with 2cm perimeter”.

execution accuracy) [9]. Finally, question answering with the help of knowledge graphs

and language models has been tackled before [16], however, the

Permission to make digital or hard copies of all or part of this work for personal

development of retrieval-augmented generation (RAG) systems

or classroom use is granted without fee provided that copies are not made or

has seen significant growth recently. In 2024, several preprints

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this have emerged showcasing the application of the RAG approach

work must be honored. For all other uses, contact the owner /author(s).

to knowledge graphs [12, 13, 14]. This paper contributes to this

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

rapidly evolving field by presenting our own advancements and

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.sikdd.5 findings.

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Domen Hočevar and Klemen Kenda



2 Data query generation, which are essential for responding to complex

queries by the user.

This study uses a generated dataset representing a hypothetical

factory with various machine models, designed to test the capabil-



ities of the developed application. The work is part of the Smart

Manufacturing pilot in the EU-funded HumAIne project [7], with

the aim of eventually using real-world data from participating

factories.

The mock factory includes models of "drillers", "circle cut-

ters", and "circular saws", each with unique names, manu-

facturers, and descriptions. These models are represented using

AASs with relevant submodels for energy consumption, man-

ufacturer details, and operation-specific parameters like hole

diameter or depth of cut.

We created AASs for 7 drilling machine models, 7 circle cutter

models, and 10 circular saw models, along with 1,000 machine

instances randomly assigned to these models. Numerical values

and availability were populated randomly for testing, reflecting

potential real-world variations.

The initial step after acquiring AAS data is to convert it into

a knowledge graph. This process involves transforming JSON-

serialized AASs into RDF triples, which represent the semantic

information of the data. Once the RDF triples are generated, they

1

are stored in a GraphDB repository. To enable semantic data re-

trieval, we employ a connector that interfaces with the ChatGPT

2

Retrieval Plugin , which operates alongside the server applica-

tion. When new triples are added to the GraphDB repository, the

connector triggers the plugin to generate vector embeddings of

Figure 2: System architecture for retrieval augmented gen-

the text representations of the new nodes. These embeddings

eration with knowledge graphs in Industry 4.0.

are created using a language model and are stored in a separate

vector database. The ChatGPT Retrieval Plugin enables interacte

to a selection of different vector databases, in our case we em- In summary, the architecture is designed to streamline the

ployed the Milvus vector database. The system is also designed process of building a knowledge graph from AAS data and en-

to maintain consistency; if any triples are removed from the ables users to query this graph with retrieval-augmented gener-

GraphDB repository, the corresponding vector embeddings are ation (RAG) using natural language, with the system handling

automatically deleted from the vector database. the complexities of data storage, retrieval, and natural language

processing in the background.

3 Methodology The sequence diagram in Figure 3 illustrates the interaction

between system components during query processing. Our sys-

The system architecture is illustrated in Figure 2. The user inter-

tem enables two distinct approaches to handle natural language

acts with the system through a client application, developed using

queries, often combining both to generate a comprehensive an-

ReactJS, which serves as the graphical user interface (GUI). This

swer for the user.

client application communicates with the system’s middleware,

which is built on the Flask framework. Users have the capability



to upload AAS data to construct and enhance the knowledge

graph, as well as to issue natural language queries.

The middleware acts as the core of the system, facilitating com-

munication between the client application, the knowledge graph

stored in a GraphDB database, and OpenAI’s GPT models. The

AAS data uploaded by the user is first converted into RDF triples

and then stored in the GraphDB repository. The Flask-based mid-

dleware also integrates with the ChatGPT Retrieval Plugin, which

is responsible for generating vector embeddings of the knowledge

graph nodes using OpenAI’s model. text-embedding-ada-002

These vector embeddings are stored in the Milvus vector data-

base [15]. The ChatGPT Retrieval Plugin allows the system to

efficiently retrieve the most relevant embeddings in response

to user queries, ensuring that the system can provide accurate



and contextually appropriate answers. Additionally, the middle- Figure 3: Sequence diagram of different approaches for

3

ware leverages LlamaIndex to manage sub-graph retrieval and data extraction. The blue box represents the RAG approach

and the red box represents the SPARQL query generation

1

https://graphdb.ontotext.com/ approach. Note that RAG approach utilizes results from

2

https://github.com/openai/chatgpt- retrieval- plugin

3 SPARQL queries on the knowledge graph.

https://www.llamaindex.ai/

Querying with KG and LLMs for Industry 4.0 Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



The first approach utilizes a Retrieval-Augmented Generation the number of machines that met the voltage criteria and identi-

(RAG) method. Upon receiving a query, the system analyzes the fying any errors, such as incorrect voltage values or unnecessary

query to identify relevant concepts and generates vector embed- machine retrievals. Results are depicted in Figures 4 and 5.

dings for these concepts [5]. These embeddings are then matched



against the knowledge graph stored in GraphDB to find the most

relevant nodes. Once the relevant nodes are identified, a naive

neighborhood expansion is performed, capturing additional re-

lated nodes to ensure a more complete context. The search is

parameterized using parameters: scope, how many nodes from

the graph to retrieve; breadth, from how many relevant nodes

to start the neighborhood expansion; score weight, how many

more nodes are visited from the identified relevant nodes that

are deemed more relevant using embedding similarity. This sub-

graph, along with a few examples for context, is then fed into

the Large Language Model (LLM) using a few-shot [1] learning

technique to generate a response [4]. The LlamaIndex framework

provides a general context query for turning triples into natu-

ral language. This method is particularly effective for queries

requiring contextual understanding and extraction of complex

information from the knowledge graph.

The second approach involves generating a SPARQL query

based on the natural language query and the ontology used



within the knowledge graph. The system attempts to execute this Figure 4: Performance of the system by the type of the

SPARQL query in the GraphDB database. If the query runs suc- machine and query. cessfully, the resulting data is passed to the LLM to formulate the

final answer. This approach is especially beneficial for tasks that

In Figure 4, each table contains four columns: "V" (voltage

involve counting instances or performing specific data aggrega-

specified in the query), "R" (percentage of correctly retrieved ma-

tion operations, where LLMs alone might struggle. This approach

chines), "W" (number of machines with incorrect voltage), and

benefits from the first approach as it can use it as backup or to

"A" (number of unnecessary machine retrievals). Figure 5 summa-

enrich the SPARQL query results with additional context.

rizes the results: "Fully Correct Answers" shows the percentage

of queries that returned all requested information without errors;

"Share of Expected Information Found" indicates the proportion

4 Results of requested information retrieved; and "Share of Incorrectly Dis-

To thoroughly evaluate the system, we employed three different played Voltages" represents the percentage of retrieved voltages

evaluations: (a) assessing the accuracy of data retrieval based on that were incorrect.

query parameters (not using query generation), (b) evaluating the



system’s ability to correctly fetch the number of instances (testing

query generation), and (c) conducting a manual assessment of

most relevant user queries.



4.1 Accuracy of Data Retrieval

The first approach involved testing the system’s ability to accu-

rately retrieve data that met specific query conditions without

employing SPARQL query generation. We focused on queries

where the user requested a list of machines of a particular type

with a voltage requirement less than or equal to a specified value.

An example query would be: “Return all drilling machines

that consume at most 4 volts and specify their consump-

tion.” Figure 5: Combined performance.

We conducted these tests on three types of machines: "drilling

machines", "circle cutters", and "circular saws". The voltage The results show that sometimes the LLM would incorrectly

values specified in the queries ranged from 0 to 10 volts, inclu- generate a different voltage requirement for a machine, making

sive. The evaluation was designed to measure how accurately it appear to satisfy the query conditions. However, the retrieved

the system could identify and return the correct set of machines machines were always of the correct type. For example, a query

based on these voltage constraints. like “Name all drilling machines and specify their voltage

For these tests, the following parameters were used (scope: 100, reqirements” correctly retrieves all machines with the right

breadth: 1000, score weight: 100, model: , specifications, suggesting the issue may lie with the LLM rather gpt-4-1106-preview

query generation strategy: ). than the knowledge retrieval process. disabled

The system’s performance was assessed by comparing the To address this, users can try adjusting query parameters or

retrieved data against the expected results, specifically checking rewording the query to verify the information’s accuracy. If this



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Domen Hočevar and Klemen Kenda



type of query is crucial, incorporating voltage-specific queries balance between subgraph retrieval and SPARQL generation to

into the query generation strategy could improve reliability, al- ensure even more robust and comprehensive query handling.

though the LLM may struggle with large lists due to its context

window limitations. As shown in Figure 5, these types of queries Acknowledgements often do not reliably provide all requested information in one

This work was supported by the European Commission under

answer, so users should run multiple queries to increase the

the Horizon Europe project HumAIne, Grant Agreement No.

likelihood of retrieving all necessary data.

101120218. We would like to express our gratitude to all project

partners for their contributions and collaboration.

4.2 Instance Fetching Accuracy

In these tests, we tested query generation strategy. The following References

parameters were used (scope: 100, breadth: 1000, score weight: [1] Tom B Brown. 2020. Language models are few-shot learners. arXiv preprint

arXiv:2005.14165.

100, model: , query generation strategy: gpt-4-1106-preview

[2] Diego Bustamante and Hideaki Takeda. 2024. Sparql generation with entity

enabled). pre-trained gpt for kg question answering. arXiv preprint arXiv:2402.00969.

[3] 2022. Details of the asset administration shell. https://www.plattf orm- i40.d

The queries asked for the number of available instances for se-

e/IP/Redaktion/EN/Downloads/Publikation/Details_of _the_Asset_Adm

lected machine models, such as "Get the number of available inistration_Shell_Part1_V3.pdf ?__blob=publicationFile&v=1 (visited on

[name of the machine 1], [name of the machine 2] machine 02/22/2024). (2022).

[4] Chao Feng, Xinyu Zhang, and Zichu Fei. 2023. Knowledge solver: teach-

instances. Specify the number for each machine type sepa-

ing llms to search for domain knowledge from knowledge graphs. , ArXiv

rately.". The query format was picked such that the LLM will abs/2309.03118. https://api.semanticscholar.org/CorpusID:261557137.

[5] Luis Gutiérrez and Brian Keith. 2019. A systematic literature review on word

benefit from query generation (availability property is specified

embeddings. In Trends and Applications in Software Engineering: Proceedings

in the schema supplied for query generation). of the 7th International Conference on Software Process Improvement (CIMPS

A total of 100 queries were run, with 10 queries for each num- . Springer, 132–141. 2018) 7

[6] Domen Hočevar. 2024. Integrating Knowledge Graphs and Large Language

ber of specified machine models (ranging from 1 to 10 models).

Models for Querying in an Industrial Environment. Bachelor’s Thesis. Uni-

The share of fully correct answers for each query type was be-

versity of Ljubljana, Faculty of Computer, Information Science, Faculty of

tween 80 and 100%. The overall accuracy was 96%. This supports Mathematics, and Physics, Ljubljana, Slovenia, (Aug. 2024). Interdisciplinary

University Study Program, First Cycle, Computer Science and Mathematics.

our hypothesis that the query generation strategy provides more

[7] Humaine Horizon. 2024. Humaine horizon. https://humaine- horizon.eu/.

accurate answers for slightly more complex queries. Accessed: 2024-08-26. (2024).

[8] Pérez Jorge. 2006. Semantics and complexity of sparql. In Proc. 5th Int.

4.3 Semantic Web Conference (ISWC2006). Manual Evaluation of Example Queries [9] Jinyang Li et al. 2024. Can llm already serve as a database interface? a big

bench for large-scale database grounded text-to-sqls. Advances in Neural

This evaluation was initially performed to identify several short-

Information Processing Systems, 36.

comings in our methodologies as mentioned in the previous sub- [10] Lars-Peter Meyer, Johannes Frey, Felix Brei, and Natanael Arndt. 2024.

sections. By manually evaluating specific queries relevant to end Assessing sparql capabilities of large language models. (2024). https://arxiv

.org/abs/2409.05925 arXiv: 2409.05925 . [cs.DB]

users, we were able to partially address these issues and fine-tune

[11] Humza Naveed, Asad Ullah Khan, Shi Qiu, Muhammad Saqib, Saeed An-

parameters to achieve more accurate results. For instance, while war, Muhammad Usman, Naveed Akhtar, Nick Barnes, and Ajmal Mian.

2023. A comprehensive overview of large language models. arXiv preprint

the system’s initial results were often incomplete (e. g., query did

arXiv:2307.06435.

not return all the machines satisfying certain criteria), increasing

[12] Shirui Pan, Linhao Luo, Yufei Wang, Chen Chen, Jiapu Wang, and Xindong

the breadth parameter to include a larger subgraph and allowing Wu. 2024. Unifying large language models and knowledge graphs: a roadmap.

IEEE Transactions on Knowledge and Data Engineering.

LLMs to traverse a broader neighborhood improved the results.

[13] Diego Sanmartin. 2024. Kg-rag: bridging the gap between knowledge and

Additionally, we demonstrated that subgraph retrieval and query creativity. arXiv preprint arXiv:2405.12035. generation can complement each other, further enhancing overall [14] Bhaskarjit Sarmah, Benika Hall, Rohan Rao, Sunil Patel, Stefano Pasquali,

and Dhagash Mehta. 2024. Hybridrag: integrating knowledge graphs and

performance. All the results are commented in detail in [6].

vector retrieval augmented generation for efficient information extraction.

arXiv preprint arXiv:2408.04948.



5 Conclusions [15] Jianguo Wang et al. 2021. Milvus: a purpose-built vector data management Proceedings of the 2021 International Conference on Management system. In In this paper, we presented a system that bridges the gap between , 2614–2627. of Data

[16] Michihiro Yasunaga, Hongyu Ren, Antoine Bosselut, Percy Liang, and Jure

natural language processing and querying knowledge graphs,

Leskovec. 2021. Qa-gnn: reasoning with language models and knowledge

specifically within the context of Industry 4.0. By leveraging graphs for question answering. . arXiv preprint arXiv:2104.06378 large language models (LLMs) and retrieval-augmented gener-

ation (RAG), our system allows users to interact with complex

knowledge graphs using natural language queries, thereby sim-

plifying access to detailed manufacturing data.

Our evaluation demonstrated the usability of our system, how-

ever with the integration of LLMs for natural language under-

standing, some challenges remain. These include occasional inac-

curacies in data retrieval and the LLM’s limited ability to handle

large datasets or specific queries. By adjusting subgraph retrieval

parameters such as breadth and scope, and by combining it with

SPARQL query generation, we were able to significantly enhance

the system’s accuracy and reliability.

This work highlights the potential of combining knowledge

graphs with LLMs to create more intuitive and effective query

systems in industrial environments. Future improvements could

focus on refining query strategies and further optimizing the





Comparative Analysis of Machine Learning Models for




Groundwater Level Forecasting: The Impact of Contextual Data



Rok Klančič Klemen Kenda

rok.klancic@gmail.com klemen.kenda@ijs.si

Jožef Stefan Institute Jožef Stefan Institute

Ljubljana, Slovenia Ljubljana, Slovenia



Abstract 2 Methods

This paper presents a comparative evaluation of three distinct In our experiments, we employed three categories of methods:

categories of models applied to groundwater level data: tradi- traditional batch learning techniques, time series deep learning

tional batch learning methods, time series deep learning methods, models, and time series foundation models.

and time series foundation models. By enriching the water level

data with weather-related features, we significantly improved 2.1 Traditional Batch Learning Methods the effectiveness of simpler models. The results demonstrate that,

In the context of data-driven modelling of environmental is-

despite their state-of-the-art performance on univariate datasets sues, traditional batch learning methods have historically demon-

and the corresponding publicity, advanced models without con-

strated significant success [5]. In this study, we employed linear

textual feature support are still surpassed by traditional methods regression alongside two tree-based approaches: random forest

trained on enriched datasets.

and gradient boosting [7] as baselines to evaluate whether the

newer, more prominent techniques, which have recently gathered

Keywords a considerable amount of attention, can perform competitively

groundwater level prediction, time series forecasting, deep learn- in this specific setting.

ing, foundation models, contextual data All of the chosen batch learning techniques are regression-

based and are valued for their simplicity, speed, and ease of

1 Introduction use. However, they often lack the complexity necessary to fully

capture intricate patterns in the data. To mitigate this limitation,

Accurate water level prediction is crucial for mitigating the im-

we incorporated contextual features, such as weather data and

pacts of climate change on water resources. By forecasting water

forecasts (e.g., precipitation, cloud cover, temperature). While the

levels, we can better prepare for potential floods and droughts,

data fusion problem is solved [8], this approach raises concerns

and more effectively manage our water supplies. However, pre-

about the availability and relevance of the contextual data.

dicting water levels presents a significant challenge due to the

dynamic nature of the data. As climate change leads to prolonged

droughts and increasingly erratic precipitation patterns, the need 2.2 Time Series Deep Learning Methods

for reliable forecasting methods becomes even more important Time series deep learning models are explicitly designed for

[2]. forecasting time-dependent data. In our study, we employed N-

In this paper, we aim to compare the performance of various BEATS [12] and PatchTST [10], both of which have architectures

models in forecasting groundwater levels. Specifically, we focus tailored to capture trends and seasonalities inherent in time se-

on the differences between traditional batch learning methods ries data. Despite their advanced capabilities, these models have

that utilize relevant contextual data and newer univariate time drawbacks, including longer training and inference times, the ne-

series deep learning and foundation models. cessity for extensive hyperparameter tuning to achieve optimal

The main contributions of this paper are: performance, and limited support for incorporating additional

• features. Although certain models support multivariate time se-

A comparative analysis of the performance of traditional

ries, they were not utilized in our experiments.

batch learning methods against state-of-the-art time series

deep learning techniques and time series foundation mod-

els, particularly in the context of feature vectors enriched 2.3 Time Series Foundation Models

with relevant contextual data. While deep learning methods require separate training and pre-

• The application of time series foundation models and deep diction phases, time series foundation models aim to eliminate

learning methods to the domain of groundwater level fore- the training step. Inspired by large language models, these models

casting. are pretrained on extensive time series datasets, enabling zero-

shot predictions on new time series without additional training.

The groundwater dataset used in this study has previously been

We used CHRONOS [1], an open source foundation model. The

employed for predictive modeling with traditional batch learning

advantages of this approach include ease of use with minimal pa-

methods [9], where extensive feature engineering was also per-

rameter adjustments and no need for training. However, similar

formed. Our work builds upon and extends this earlier research

to deep learning models, they lack support for multivariate time

by incorporating a different set of models.

series.

Permission to make digital or hard copies of all or part of this work for personal

Several studies have already evaluated the performance of

or classroom use is granted without fee provided that copies are not made or

various deep learning and foundation models for time series fore-

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this casting [1] [13]. However, this research extends the application

work must be honored. For all other uses, contact the owner /author(s).

of these forecasting models to groundwater level data, therefore

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

contributing to the better understanding of their effectiveness in

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.sikdd.6 this domain.



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Rok Klančič and Klemen Kenda



3 Experiment Setting and gradient boosting regressor as our baseline methods. These

models were previously applied to the groundwater dataset [9],

The experiments were conducted on a dataset of groundwater

necessitating a reproduction of the results as a benchmark.

levels in Slovenia. Due to the cumulative nature of water levels

and to facilitate comparison with the original study [9], predic-



tions were made on daily changes in water levels rather than on 3.4 Implementation Details absolute values.

The prediction pipelines varied slightly between the different



3.1 types of models: Dataset

The groundwater dataset is a subset of the larger dataset used For , we utilized the dataset without weather • CHRONOS

in the study [9]. It consists of groundwater level measurements features, as it only supports univariate time series. Since

taken daily from multiple stations across Slovenia. To apply tra- no hyperparameter tuning was required, the data was

ditional batch learning methods, we enriched the dataset with divided into training and test sets, omitting the validation

weather data, associating each water measurement station with set. The model generated the predictions directly from the

the nearest weather station. Due to the availability of weather water level data. We used the chronos-t5-large model from

data, only data from the years 2010 to 2017 was included in our the chronos library.

study. For consistency and ease of comparison with previous For and , the same dataset was used, • N-BEATS PatchTST

study [9], we focused on data from two water measurement given the same limitation as mentioned previously. How-

stations located in Ljubljana. ever, a validation set was required for hyperparameter

In traditional batch learning within the environmental domain, tuning. After selecting appropriate hyperparameters, the

it is essential to not only use the raw data but also to engineer models were trained on the training set and evaluated

relevant features. Initially, we removed the pressure and dew on the test set. Implementations from the NeuralForecast

point features, as they were either unrelated to the target variable library were used for both models.

or highly correlated with other features [9]. We then created For the , , • linear regression random forest regressor

additional features by shifting the data from 1 to 10 days, making and models, we included gradient boosting regressor historical values available, and by computing the averages of both water level and weather data. Feature selection was

features over a 2- to 10-day window. This process resulted in conducted to reduce the number of features, resulting in

approximately 2,000 features. Given the excessive number of 42 features for linear regression, 30 for random forest, and

features, which could degrade model performance, we employed 36 for gradient boosting. After feature selection, hyper-

a feature selection algorithm to identify the most informative parameters for the random forest and gradient boosting

subset. models were tuned, and the data for linear regression was

We used a genetic feature selection algorithm from scikit-learn, normalized. The models were then trained on the train-

evaluated on 365-day part of training dataset, with the maximum ing set and evaluated on the test set using scikit-learn’s

number of features set to 40. The algorithm was executed sepa- implementations.

rately for each model, focusing on one station and a prediction

The hyperparameters used for training are listed in Appendix

horizon of three days, resulting in distinct feature vectors. Sub-

A, while a description of the selected features is provided in

sequently, weather forecast features with longer offsets were

Appendix B.

manually added to the selected feature set.



3.2 Evaluation Metrics 4 Results

The dataset was split into a training set (approx. 2,500 days),

The results for all tested models across various prediction hori-

a validation set (100 days), and a test set (365 days) for model 2

zons are presented in Table 1. The reported R scores were calcu-

2

evaluation. Model performance was evaluated using the R score,

lated based on the differences in water levels; if absolute water

averaged across all tested stations. Although alternative metrics 2

levels had been used, the R scores would have been significantly

such as root-mean-squared error (RMSE), and mean absolute

higher. For example, in the case of CHRONOS with 1-day ahead

percentage error (MAPE) were considered, they, for this dataset, 2

predictions, the R score is 0.725 for relative level differences and

2

produce results that are closely related to the R . This metric was

0.998 for absolute water levels.

selected due to its robustness against variations in data offset

Among the models, linear regression achieved the highest per-

and amplitude, and for direct comparability with the results in

formance, followed by the random forest. In contrast, the more

2

the original study [9]. The R score is defined as:

complex methods, including deep learning models and the foun-



𝑅 2 Í𝑛 2 dation model, showed generally lower performance, with the 𝑦 − ) = 𝑖 1 ( 𝑦 𝑖 ˆ 𝑖 = 1 − , Í 𝑛 2 exception of the 1-day prediction horizon, where N-BEATS out- ( 𝑦 𝑦 ¯ 𝑖 = 1 𝑖 − ) 2

performed the tree-based models. Notably, the R scores decrease

where 𝑦𝑖 is the i-th true value, ˆ 𝑦𝑖 is the i-th predicted value and as the prediction horizon lengthens, with a more pronounced

𝑦 ¯ is the average of true values.

decline observed in the deep learning and the foundation models

compared to the traditional batch learning methods.

3.3 Baseline Methods Figures 2 and 3 display the predictions from CHRONOS, Patch-

The primary objective of our research was to compare the per- TST, and linear regression compared to the true data for the 1-day

formance of traditional batch learning methods, enriched with and 5-day prediction horizons. It is evident that the predictions

relevant contextual features, against that of modern deep learn- from CHRONOS and PatchTST begin to exhibit a rightward shift

2

ing techniques and foundation models for time series forecasting. as the horizon extends. Figure 1 visualizes the R scores for all

Therefore, we selected linear regression, random forest regressor, models across the different prediction horizons.



The Impact of Contextual Data Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



Table 1: R 2 Scores for Different Prediction Horizons and Models.



Methods 1 day ahead 2 days ahead 3 days ahead 4 days ahead 5 days ahead

Chronos-large 0,725 0,365 0,175 0,04 -0,09

GradientBoostingRegressor 0,640 0,603 0,527 0,556 0,545

RandomForestRegressor 0,726 0,697 0,701 0,706 0,691

N-BEATS 0,742 0,397 0,17 -0,03 -0,143

PatchTST 0,721 0,394 0,215 0,109 -0,02

LinearRegression 0,792 0,781 0,785 0,784 0,780

The best and second-best results are bolded and underlined respectively.



R² Scores for Different Models and Prediction Horizons

0.8 Models

Chronos-large



0.6 RandomForestRegressor GradientBoostingRegressor



es PatchTST 0.4 LinearRegression N-BEATS



R² Scor 0.2



0.0



1 day 2 days 3 days 4 days 5 days

Prediction Horizons



Figure 1: R2 Scores for All of the Methods and Prediction Horizons.



Predictions for Horizon 1 Predictions for Horizon 5

0.10 True data True data

Chronos 0.08 Chronos

0.08 PatchTST PatchTST

LinearRegression 0.06 LinearRegression

0.06

0.04

0.04

0.02

0.02

Water level change (m) 0.00 0.00 Water level change (m)

0.02 0.02

2017-01-15 2017-02-01 2017-02-15 2017-03-01 2017-03-15 2017-01-15 2017-02-01 2017-02-15 2017-03-01 2017-03-15

Time Time



Figure 2: Example Predictions for Three Models for 1-Day Figure 3: Example Predictions for Three Models for 5-Day

Prediction Horizon. Prediction Horizon.



The results indicate that traditional methods, when supple-

mented with relevant contextual features, outperform more com-

plex models that do not incorporate such data. While the 1-day predictions. This likely occurs due to the absence of contextual

ahead predictions show comparable performance across all meth- information, causing these models to lag in capturing the true

ods, as the prediction horizon extends, the accuracy of CHRONOS, trajectory of water levels. In contrast, models with access to

PatchTST, and N-BEATS declines sharply. In contrast, the tradi- weather data can predict further ahead by accounting for factors

tional models, supported by contextual features, maintain their such as the impact of rainfall patterns on water levels.

predictive accuracy much more effectively, as shown in Figure 1. An unexpected finding is that among the baseline models,

A closer examination of the predictions in Figures 2 and 3 linear regression outperforms the more sophisticated methods.

reveals that for 1-day ahead predictions, all models track the true For instance, in the article [9], while linear regression produced

data closely. However, in the 5-day ahead predictions, models strong results, it did not surpass the performance of the other

lacking contextual data begin to exhibit a rightward shift in their two methods.



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Rok Klančič and Klemen Kenda



5 Conclusion and Future Work A Hyperparameters

After evaluating all models on the groundwater level dataset,

we observed that traditional methods, when equipped with rel- Table 2: Hyperparameters Used for Gradient Boosting Re-

evant features, consistently outperformed newer and more so- gressor and Random Forest Regressor.

phisticated techniques, particularly as the prediction horizon

lengthened. This suggests that the emphasis on developing the Hyperparameter GradientBoosting RandomForest

most powerful deep learning or foundation models for time se-

n_estimators 28 164

ries predictions may be overstated. With thoughtful selection of

max_features ’log2’ 0.5

contextual features, even the simplest models can outperform

max_depth 10 20

modern approaches, which is a significant finding for fields with

sufficient contextual data, such as data-driven environmental

modelling. Table 3: Hyperparameters Used for N-BEATS and

To enhance the robustness of our evaluation, future work could PatchTST.

involve testing additional methods, expanding the analysis to



include more measurement stations and surface water level data, Hyperparameter N-BEATS PatchTST and incorporating deep learning models that support multivariate

loss HuberLoss /

time series, such as N-BEATSx [11] and N-HiTS [3]. Further

n_harmonics 5 /

insights could be gained by exploring foundation models with

n_polynomials 5 /

multivariate support, such as TimesFM [4], as well as some more

scaler_type ’robust’ /

univariate models, like TimeGPT-1 [6]. Future research could

n_blocks [3, 3, 1] /

also compare the inference times of various models and assess

mlp_units [[128, 128]] /

performance across different time series lengths.

horizon 5 5

input_size 15 71

learning_rate 0.001 0.001

Acknowledgements max_steps 25 1323

This work was supported by the European Commission under the

encoder_layers / 12

Horizon Europe project Plooto, Grant Agreement No. 101092008. n_heads / 16 We would like to express our gratitude to all project partners for

hidden_size / 64

their contributions and collaboration.

linear_hidden_size / 512

Furthermore, we would like to thank Erik Novak for his assis-

dropout / 0.2

tance in completing this research.

fc_dropout / 0.1

head_dropout / 0.1

attn_dropout / 0.2

References patch_len / 16

[1] Abdul Fatir Ansari et al. 2024. Chronos: learning the language of time series. stride / 8

arXiv preprint arXiv:2403.07815. revin / True

[2] ARSO. 2009. Freshwater. Retrieved August 27, 2024 from https://www.arso

.gov.si/en/soer/f reshwater.html.

[3] Cristian Challu, Kin G Olivares, Boris N Oreshkin, Federico Garza Ramirez,

Max Mergenthaler Canseco, and Artur Dubrawski. 2023. NHiTs: neural B Selected Features hierarchical interpolation for time series forecasting. In Proceedings of the

Due to the large number of features selected by the feature selec-

AAAI conference on artificial intelligence number 6. Vol. 37, 6989–6997.

[4] Abhimanyu Das, Weihao Kong, Rajat Sen, and Yichen Zhou. 2023. A decoder- tion algorithm, we provide a summarized description of the most

only foundation model for time-series forecasting. arXiv preprint arXiv:2310-

frequently chosen features. The features that appeared most often

.10688.

include shifts and averages of precipitation, precipitation fore-

[5] Fan Feng, Hamzeh Ghorbani, and Ahmed E. Radwan. 2024. Predicting

groundwater level using traditional and deep machine learning algorithms. casts, temperature, altitude difference, cloud cover, humidity, and

Frontiers in Environmental Science, 12. doi: 10.3389/f envs.2024.1291327. snow accumulation. Notably, the majority of selected features

[6] Azul Garza and Max Mergenthaler-Canseco. 2023. TimeGPT-1. arXiv

preprint arXiv:2310.03589 were derived features we generated, with only approximately

.

[7] Trevor Hastie, Robert Tibshirani, and Jerome H Friedman. 2009. one original feature being selected per model. The elements

of statistical learning: data mining, inference, and prediction. Vol. 2. Springer. In Table 4, the most common shifts and averages for each

[8] Klemen Kenda, Blaž Kažič, Erik Novak, and Dunja Mladenić. 2019. Streaming

data fusion for the internet of things. , 19, 8. doi: 10.3390/s19081955. Sensors

individual model are presented. The table indicates that shifts

[9] Klemen Kenda, Jože Peternelj, Nikos Mellios, Dimitris Kofinas, Matej Čerin, and averages of varying lengths were selected, with a slight

and Jože Rožanec. 2020. Usage of statistical modeling techniques in surface

preference for shorter ones.

and groundwater level prediction. Journal of Water Supply: Research and

Technology-Aqua, 69, 3, (Apr. 2020), 248–265. doi: 10.2166/aqua.2020.143.

[10] Yuqi Nie, Nam H Nguyen, Phanwadee Sinthong, and Jayant Kalagnanam. Table 4: Most Frequently Selected Shifts and Averages for

2022. A time series is worth 64 words: long-term forecasting with trans- Various Methods. formers. . arXiv preprint arXiv:2211.14730

[11] Kin G. Olivares, Cristian Challu, Grzegorz Marcjasz, Rafał Weron, and Artur

Dubrawski. 2023. Neural basis expansion analysis with exogenous variables: Method Shifts (days) Averages (days) forecasting electricity prices with nbeatsx. International Journal of Forecast-

ing, 39, 2, 884–900. doi: https://doi.org/10.1016/j.ijf orecast.2022.03.001. GradientBoostingRegressor 4, 10 2, 6

[12] Boris N Oreshkin, Dmitri Carpov, Nicolas Chapados, and Yoshua Bengio.

RandomForestRegressor 2, 6 3, 9

2019. N-BEATS: neural basis expansion analysis for interpretable time series

forecasting. . arXiv preprint arXiv:1905.10437

LinearRegression 2, 10 2, 7

[13] Hongwei Ye et al. 2024. A transformer-based forecasting model for f10.7 Combined 2, 10 2, 3

index and its application study on the chinese langfang dataset. Advances

in Space Research. doi: https://doi.org/10.1016/j.asr.2024.08.024.





Interactive Tool for Tracking Open-source Artificial




Intelligence Progress on Hugging Face



Bogdan Šinik Domen Vake

bogdan.sinik@f amnit.upr.si domen.vake@f amnit.upr.si

UP FAMNIT UP FAMNIT

Koper, Slovenia Koper, Slovenia



Jernej Vičič Aleksandar Tošić

jernej.vicic@upr.si aleksandar.tosic@upr.si

UP FAMNIT, UP IAM UP FAMNIT, InnoRenew CoE

Koper, Slovenia Koper, Slovenia



Abstract to execute your own model, as long as it is of a modest enough

size, on a home computer’s graphics processing unit (GP U), even

Given its increasing importance in our daily lives, Artificial In-

if the GP U is a few years old [9]. The rise in accessibility also

telligence has become a prominent subject that needs extensive

enables a larger community to test and develop new solutions

investigation and understanding. This study presents an analysis

and build on top of existing models. We believe that there is a

of the open-source community in the field of Artificial Intelli-

big lack of tools for monitoring the impact of this movement.

gence (AI). Various questions arise anytime AI is introduced.

1

Hugging Face has grown into one of the primary platforms

open-source AI introduces additional concerns. Should artifi-

for the open-source community. Users are able to download and

cial intelligence (AI) be universally accessible, or should it be

interact with all significant open-source models. Subsequently,

restricted to private use? Is it worthwhile to offer basic models

users have the option to publish their models on the platform and

to the broad user population? We chose the most important data

compare their performance by adding them to the leaderboard,

from the primary website in the field, Hugging Face. We have

where all the models are benchmarked and ranked. The open-

developed a tool that allows for straightforward monitoring of

source community relies heavily on the distribution of models

the progress of various open-source AI models using data ob-

by large corporations, as creating a model from scratch is a hard

tained from their leader board. The platform offers accessible

undertaking [9]. This tool facilitates collaboration among open-

and valuable information about various AI models, including

source contributors, enabling them to collectively generate social

their architectures and the activities of authors. Through per-

media content, exchange ideas, and even publish concise articles.

forming a quick review with our tool, it becomes evident that the

In addition to the models, they have the ability to generate and

open-source community is becoming large and has an undeniable

upload useful datasets. It represents the most advanced and inno-

impact on the AI community.

vative developments in the field of open-source AI and Machine



Keywords learning.

An issue that has been observed is the absence of effective

LLM, open-source, AI, Hugging Face

visualization tools on Hugging Face, which would enable users



1 Introduction to easily see patterns and gain a comprehensive understanding

of the open-source AI area. In order to address this issue, we

Artificial intelligence, particularly large language models (LLMs),

have developed a sophisticated tool that offers users various

is an important topic in the computer industry today. Despite

viewpoints on the data.

the numerous fears and dogmas around it, it is certain that AI



has become an integral aspect of our lives. This research has 2 Literature review specifically concentrated on the development of a tool for moni-

Large Language Models (LLMs) have proven essential in enhanc-

toring the impact of the open-source community in the area of

ing software engineering (SE) tasks, demonstrating their effec-

artificial intelligence. As implied, these models are accessible to

tiveness in code comprehension Similar to conventional soft-

all individuals. There is considerable debate on whether this type

ware engineering tools, open-source cooperation is essential for

of technology should be universally accessible. We wanted to

achieving superior products in this area. [8]

investigate if the open-source community is actively contributing

The article authored by Patel et al. [9] emphasizes the sig-

to the development of the field, regardless of one’s philosophical

nificance of the open-source AI community and elucidates its

convictions. Due to the substantial computational requirements,

rapid growth in the wake of major industry leaders like Google,

it was previously impossible to execute Large Language Mod-

Microsoft, and OpenAI. An important milestone in this subject is

els on personal computers. As increasingly compact versions

often emphasized as the day when the LLama model was initially

with impressive capabilities are being produced, this scenario

made available to the open-source community. The community

undergoes a significant transformation. Currently, it is feasible

promptly recognized the possibilities and potential involved in

Permission to make digital or hard copies of all or part of this work for personal

this release.

or classroom use is granted without fee provided that copies are not made or

Due to its continuous growth, Hugging Face has emerged

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this as the primary platform for exchanging machine learning (ML)

work must be honored. For all other uses, contact the owner /author(s).

models, resulting in an increasing level of complexity. A relational

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).

1

https://doi.org/https://doi.org/10.70314/is.2024.sikdd.1 https://huggingface.co/

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Bogdan Šinik, Domen Vake, Jernej Vičič, and Aleksandar Tošić



database called HFCommunity was established to facilitate the data was categorized using several criteria, such as model type,

analysis and resolution of this issue [1]. model architecture, and amount of parameters. The data was

As previously said, open-source AI models offer an exten- initially selected and aggregated to ensure that all crucial com-

sive range of possibilities. At the recent conference, the authors ponents were easily accessible. All models that were categorized

[12] demonstrated their effective use of Hugging Face. Due to as flagged have been excluded from the dataset. In addition, we

the significant difficulty in developing a model with broad in- have collected data on the authors’ activities and conducted a

telligence, researchers have merged ChatGPT capabilities with study on that particular aspect. Once the data had been cleaned

models from HuggingFace using agentic architecture to get im- and prepared for visualization, we utilized the R ggplot library

pressive results in multiple domains. ChatGPT was tasked with to create visual representations of the data. A comprehensive

creating a plan of action and assigning specific duties to each R Shiny app was developed by aggregating all the visuals. We

open-source model based on their own areas of expertise. This chose to utilize Shiny because it is a great option for constructing

is an excellent demonstration of the influence and capabilities interactive data analysis solutions due to several factors. Firstly,

of the open-source community, given the familiarity with open it enables the development of web applications that are capable

models and their capabilities. of responding and adapting to real-time changes and user interac-

The article [6] examines the vulnerabilities associated with tions. This simplifies the process of exploring and analyzing data.

open-source AI. A much higher number of repositories with high Shiny easily incorporates with R, utilizing its robust statistical

vulnerabilities has been discovered compared to those with low and graphical functionalities to generate complex, interactive vi-

vulnerabilities, particularly in root repositories. This emphasizes sualizations without the need for experience in web technologies

the significance of ensuring the security of technology in order such as HTML, CSS, or JavaScript. [13] Finally, our application

to facilitate its utilization. was deployed to a server, making it accessible online.

In a recent paper [10], authors have analyzed the transparency

of Hugging Face pre-trained models regarding database usage 4 Results and licenses. The analysis revealed that there is often a lack of

The outcome of this study is the tool we have developed. The

transparency regarding the training datasets, inherent biases, 3

link may be accessed via the following URL. . It has six distinct

and licensing details in pre-trained models. Additionally, this re-

viewpoints, all conveniently accessible inside its tab. The ini-

search identified numerous potential licensing conflicts involving

tial figure, labeled as 1, displays both the count of new models

client projects. 159,132 models were examined. It was found that

and the distribution of various model types. Hugging Face has

merely 14% of these models explicitly identify their datasets with

identified five distinct categories of models: basic mergers and

specific tags. Furthermore, a detailed examination of a statisti-

moerges, fine-tuned on domain-specific datasets, chat models,

cally significant sample comprising 389 of the most frequently

continuously pretrained models, and pretrained models. If the

downloaded models showed that 61% documented their training

model did not belong to any of these classes, its type was classi-

data in some form.

fied as unknown. The user has the ability to effortlessly choose

their preferred categories, along with the desired time frame and

3 Methodology unit of aggregate (daily, weekly, or monthly). This allows the

We obtained the data by extracting the Open LLM Leaderboard viewer to clearly observe the evolution of model types and their

from Hugging Face [2] by saving the data server sent to the popularity over time. It is evident that fine-tuned models are

client. This data contains information about repositories of mod- predominantly utilized. This is logical, as users are adapting base

els that are currently on the leaderboard and the models that are models by training them on unique datasets to achieve specializa-

waiting to be evaluated for the leaderboard. A Python pipeline tion. Also, we can see that merged models are a relatively recent

2

was developed to clean and enrich this data available on . The phenomenon.

leaderboard data includes model architecture and precision as



well as the model type and performance on the following bench-

marks: ARC[3], HellaSwag[14], MMLU[5], TruthfulQA[7], Wino-

grade[11] and GSM8K[4]. In addition to the data provided on

the leaderboard, additional information on the given models was

obtained by using the HF API client. This included data about

repository contributors, tags, base models, used datasets, and

repo activity. It is important to note that the data is self-reported

by the developers and is not enforced by HuggingFace. Addition-

ally, the leaderboard includes duplicates due to developers being

able to replace models in the repository with different models

under the same name. This means the duplicates have the same

repository data but distinct performances. Due to the inability to

programmatically determine the current model in the repository, Figure 1: Popularity by model type over time

we chose the best-performing model under the repository name

as the model representing the repository when removing dupli- The second view, referenced as 2, has two interconnected

cates. Thus, all datasets were generated for further utilization. visualizations. The upper section displays the activity of the top

The following analysis was conducted using the R programming 10 authors within a specific range of dates. The display showcases

language. The data was mostly studied via the perspective of every model they have developed, along with its corresponding

time, as our focus was on identifying any obvious trends. The type. The lower section presents the average benchmark score for



2 3

https://github.com/VakeDomen/HF_analysis https://oai.dltlt.famnit.upr.si/

Interactive Tool for Tracking Open-source Artificial Intelligence Progress on Hugging Face Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



each model, organized by author. This visualization enables users Opt, GPT2, and GPT2-NeoX. All architectures that did not fit

to effortlessly monitor the most prominent authors and observe into any one category were classed as "Other". This perspective

their patterns and accomplishments in model development over has two graphics that depict popularity. The first comparison

time. Users have the ability to effortlessly choose a certain range assesses the popularity of a model relative to itself, depending on

of dates and also narrow down the list to the top 10 authors the number of new models introduced before. The second one

according to their preferences. It is evident that leading authors compares it to the average number of new models created, taking

typically do not adhere to trends and consistently provide models into account their architecture. Both are depicted by coloring the

of similar type. area, as it is the most convenient way to track. Users may analyze

the fluctuation in popularity of well-known model architectures



over time and examine how the rising popularity of a particular

architecture might impact the popularity of a certain architecture

of interest. The lower plot indicates that LLama and Mistral are

the predominant models; nonetheless, they have experienced

fluctuations throughout time, as visible on the upper plot.



Figure 2: Top authors activity over time



The following perspective 3 illustrates two aspects. The first

aspect is the alteration in the average benchmark score for each

model type as time progresses. The display showcases the top-

performing model for each category and time interval (daily,

weekly, or monthly). In addition to the dots representing each Figure 4: Change of popularity of main architectures over

model, we have incorporated a smooth line to aid the user in see- time

ing the temporal changes for a particular model type. Following

the first visualization, we have included a second visualization The graphic labeled as 5 illustrates the progressive improve-

that displays the total number of models for each model type ment of the key base models developed by famous companies.

within the chosen period range. Through these visualizations, This was accomplished by isolating each incremental improve-

users can easily identify the model type that experienced the ment in score over time, using the base model as a reference. In

most improvement and the model types that were mainly pro- order to fulfill this objective, we have chosen five distinct varia-

duced. We can see the trend, which indicates that open-source AI tions of LLama, Mistral, and Mixtral, as well as three iterations

models are improving, as evidenced by the improvement in aver- of Phi. The user may easily observe the overall improvement in

age benchmark scores across most of them. The overall number benchmark scores for each base model. In addition, users have

of models is rapidly increasing, indicating a rise in the popularity the ability to view the overall duration required for the model to

of open-source AI models. achieve its maximum performance. We have included a feature

that enables users to toggle the visibility of model labels, hence



enhancing visibility and facilitating more in-depth examination



according to their preferences. This allows the user to observe the

speed at which specific models reached their peak performance

and the extent of their improvement relative to the base models.





Figure 3: Change of benchmark score and total models per

type over time



The fourth perspective, as seen in Figure 4, examines the

changing popularity of various model architectures through-

out time. The following architectures have been chosen for this

specific objective: LLama, Mixtral, Mistral, Qwen2, Gemma, Phi, Figure 5: Evolution of famous base models

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Bogdan Šinik, Domen Vake, Jernej Vičič, and Aleksandar Tošić



The final view, as depicted in Figure 6, illustrates the impact [4] Karl Cobbe et al. 2021. Training verifiers to solve math word problems.

(2021). arXiv: 2110.14168 . [cs.CL]

of significant releases on the popularity of various model designs.

[5] Dan Hendrycks, Collin Burns, Steven Basart, Andy Zou, Mantas Mazeika,

As we have employed identical model designs to those in view

Dawn Song, and Jacob Steinhardt. 2021. Measuring massive multitask lan-

four, we have extracted and categorized all significant release guage understanding. (2021). arXiv: 2009.03300 . [cs.CY]

[6] Adhishree Kathikar, Aishwarya Nair, Ben Lazarine, Agrim Sachdeva, and

dates of these models. The user has the option to choose the

Sagar Samtani. 2023. Assessing the vulnerabilities of the open-source artifi-

time unit for aggregate, which can be either day, week, or month. cial intelligence (ai) landscape: a large-scale analysis of the hugging face

Users may quickly analyze the impact of significant releases and platform. In 2023 IEEE International Conference on Intelligence and Security

Informatics (ISI), 1–6. doi: 10.1109/ISI58743.2023.10297271.

observe how they influence the popularity and mass creation of

[7] Stephanie Lin, Jacob Hilton, and Owain Evans. 2022. Truthfulqa: measuring

specific models. We can observe the evident impact of the recent how models mimic human falsehoods. (2022). arXiv: 2109.07958 . [cs.CL]

[8] Zhihao Lin et al. 2024. Open-source ai-based se tools: opportunities and chal-

releases of LLama and Mistral for their popularity.

lenges of collaborative software learning. . arXiv preprint arXiv:2404.06201

[9] Dylan Patel and Afzal Ahmad. 2023. Google “we have no moat, and neither



does openai.”. , 4, 2023. SemiAnalysis. May

[10] Federica Pepe, Vittoria Nardone, Antonio Mastropaolo, Gerardo Canfora,

Gabriele Bavota, and Massimiliano Di Penta. 2024. How do hugging face

models document datasets, bias, and licenses? an empirical study.

[11] Keisuke Sakaguchi, Ronan Le Bras, Chandra Bhagavatula, and Yejin Choi.

2019. WINOGRANDE: an adversarial winograd schema challenge at scale.

(2019). arXiv: 1907.10641 . [cs.CL]

[12] Yongliang Shen, Kaitao Song, Xu Tan, Dongsheng Li, Weiming Lu, and

Yueting Zhuang. 2023. Hugginggpt: solving ai tasks with chatgpt and its

friends in hugging face. In . Advances in Neural Information Processing Systems

A. Oh, T. Naumann, A. Globerson, K. Saenko, M. Hardt, and S. Levine, editors.

Vol. 36. Curran Associates, Inc., 38154–38180. https://proceedings.neurips.c

c/paper_f iles/paper/2023/f ile/77c33e6a367922d003f f 102f f b92b658- Paper-

Conf erence.pdf .

[13] Carson Sievert. 2020. Interactive web-based data visualization with R, plotly,

and shiny. Chapman and Hall/CRC.

[14] Rowan Zellers, Ari Holtzman, Yonatan Bisk, Ali Farhadi, and Yejin Choi.

Figure 6: Effect of big releases on architecture of produced 2019. Hellaswag: can a machine really finish your sentence? (2019). arXiv:

models 1905.07830 . [cs.CL]



5 Conclusion and future work

Given the growing importance of Artificial Intelligence in mod-

ern culture, it is beneficial to explore the free solutions that are

accessible rather than just depending on commercial alternatives.

This paper offers valuable insights into a tool designed to simplify

the examination of trends in open-source AI in a user-friendly

manner. It offers various viewpoints and enables users to acquire

knowledge and reach certain conclusions about the subject. Hug-

ging Face has the capability to function as an excellent tool for

finding a certain model. As time progresses, open-source AI is

expected to provide a growing contribution to the AI community

and provide more specific applications for models that could be

ignored by big organizations.

We aim to enhance the functionality of our Shiny application

by incorporating more perspectives and expanding the range

of data interaction options. Our objective is to ensure that the

system is as updated as possible. Besides that, we want to conduct

a comprehensive analysis of the data to identify patterns and

correlations inside this group. We aim to assess the potential of

these models and examine their capabilities and potential uses

in addressing real-world issues. We would like to analyze the

sustained popularity and efficacy of these models over a longer

time frame.



References

[1] Adem Ait, Javier Luis Cánovas Izquierdo, and Jordi Cabot. 2023. Hfcom-

munity: a tool to analyze the hugging face hub community. In 2023 IEEE

International Conference on Software Analysis, Evolution and Reengineering

(SANER), 728–732. doi: 10.1109/SANER56733.2023.00080.

[2] Edward Beeching, Clémentine Fourrier, Nathan Habib, Sheon Han, Nathan

Lambert, Nazneen Rajani, Omar Sanseviero, Lewis Tunstall, and Thomas

Wolf. 2023. Open llm leaderboard. https://huggingf ace.co/spaces/open- llm-

leaderboard/open_llm_leaderboard. (2023).

[3] Peter Clark, Isaac Cowhey, Oren Etzioni, Tushar Khot, Ashish Sabharwal,

Carissa Schoenick, and Oyvind Taf jord. 2018. Think you have solved ques-

tion answering? try arc, the ai2 reasoning challenge. (2018). arXiv: 1803.054

57 . [cs.AI]





Multilingual Hate Speech Modeling




by Leveraging Inter-Annotator Disagreement



Patricia-Carla Grigor∗ Bojan Evkoski Petra Kralj Novak

University of Vienna evkoski_bojan@phd.ceu.edu novakpe@ceu.edu

Vienna, Austria Central European University Central European University

Vienna, Austria Vienna, Austria

Jožef Stefan Institute

Ljubljana, Slovenia



Abstract challenge is the subjectivity of hate speech, as annotators often

As social media usage increases, so does the volume of toxic disagree due to diverse backgrounds and perspectives.

content on these platforms, motivating the Machine Learning To address this challenge, researchers have proposed alterna-

(ML) community to focus on automating hate speech detec- tive methodologies to ground-truthing, including the incorpo-

tion. While modern ML algorithms are known to provide nearly ration of diverse perspectives into the training and evaluation

human-like results for a variety of downstream Natural Lan- pipelines of ML models [1, 14]. One such approach is introduced

guage Processing (NLP) tasks, the classification of hate speech by [7], who train monolingual hate speech classifiers in several

is still an open challenge, partially due to its subjective anno- languages directly on datasets that include disagreement. As an

tation, which often leads to disagreement between annotators. alternative to gold-standard data, such data is referred to as dia-

This paper adopts a perspectivist approach that embraces sub- mond standard data, based on the assumption that more than one

jectivity, leveraging conflicting annotations to enhance model single truth exists. In terms of evaluation, the researchers focus

performance in real-world scenarios. A state-of-the-art multi- on the evaluation of models from the perspective of disagreement,

lingual language model for hate speech detection is introduced, with the ultimate goal of estimating the agreement between the

trained, and evaluated using diamond standard data with metrics annotators themselves, as well as between models and annotators

that consider disagreement. Various strategies for incorporat- by using the appropriate metrics. Their main findings indicate

ing disagreement are compared in the process. Results demon- that disagreement between annotators represents an intrinsic

strate that the model performs equally or better on all evalu- limitation to the performance that can be achieved by automated

ated languages compared to respective monolingual models and systems.

drastically outperforms on multilingual data. This highlights This paper aims to explore the potential of training a multilin-

the effectiveness of multilingual and perspectivist methods in gual hate speech model, as well as further explore the ideas of

addressing the complexities of hate speech detection. The pre- incorporating inter-annotator disagreement in model training.

sented multilingual hate speech detection model is available at: Therefore, at the basis of this paper lie the following research

https://huggingface.co/IMSyPP/hate_speech_multilingual. questions:

- How does the performance of multilingual hate speech classifiers

Keywords trained on diamond standard data compare to the performance of

monolingual models?

hate speech detection, inter-annotator disagreement, multilin-

- How can inter-annotator disagreement be effectively incorporated

gual language modeling

into the classifier fine-tuning process?



1 In light of these research questions, the expected outcomes Introduction

are twofold: (1) multilingual classifiers trained on diamond stan-

The phenomenon of hate speech, which is typically defined as dard data are anticipated to outperform monolingual models,

offensive or derogatory language targeting individuals or groups and (2) incorporating inter-annotator disagreement is expected

based on characteristics such as race, religion, ethnic origin, sex- to enhance sensitivity to nuanced hate speech. These findings

ual orientation, disability, or gender [2], has become a significant could benefit computational linguistics research and social me-

problem on social networks in recent years, with communities dia providers by informing the development of more effective

being increasingly exposed to toxic content as the networks content moderation algorithms.

grow and become more interconnected [13, 3]. Consequently,



the Machine Learning (ML) and computational linguistics com- 2 Related Work munities have begun developing content moderation strategies

using advanced algorithms and Natural Language Processing Several methods exist for incorporating disagreement into ML

(NLP) techniques to detect hate speech [10, 11]. However, a key training pipelines [12, 5], but few focus on hate speech detec-

tion. One approach is presented in [7], where monolingual hate

∗The first author conducted the research with significant input from the second au- speech classifiers were trained for English, Italian, and Slovenian.

thor, under the supervision and guidance of the third author. All authors contributed

to writing the manuscript. These classifiers utilized diamond standard datasets sourced from

YouTube and Twitter, employing a consistent annotation process

Permission to make digital or hard copies of all or part of this work for personal for each language. Their main findings indicate that, according to

or classroom use is granted without fee provided that copies are not made or the accuracy scores, the annotators demonstrated a high degree

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this of agreement in approximately 80% of the cases across all three

work must be honored. For all other uses, contact the owner/author(s). datasets. In terms of Krippendorff’s ordinal alpha score, which

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia considers both agreement by chance and the ordering of classes

© 2024 Copyright held by the owner/author(s).

https://doi.org/https://doi.org/10.70314/is.2024.sikdd.7 (from least to most severe), the agreement score is approximately



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Patricia-Carla Grigor, Bojan Evkoski, and Petra Kralj Novak



0.6 for all three languages. Furthermore, the evaluation results 3.2 Model Selection and Fine-Tuning

indicate that the performance of each model aligned with the Our proposed multilingual hate speech model builds on the pre-

inter-annotator agreement, both in terms of accuracy and the trained XLM-R transformer model [4], chosen for its proven

alpha score. This implies that the performance of models is inher- effectiveness in cross-lingual understanding and its ability to

ently constrained by the level of agreement among annotators. handle a wide range of languages. This provides a robust founda-

Consequently, when trained on diamond standard data, it is un- tion for fine-tuning and optimization, particularly since English,

likely that the performance of these models can significantly Italian, and Slovenian—the languages used for fine-tuning—were

surpass human performance. included in XLM-R’s pre-training. To explore various strategies

This work was built upon these findings through investigat- for incorporating annotator disagreement during training, three

ing the potential of multilingual models to enhance hate speech model variants were fine-tuned on the previously presented

detection, with the aim of broadening their applicability across datasets, referred to in the tables as MDA, MDD, and MRD, re-

diverse linguistic contexts. Additionally, strategies for incorpo- spectively.

rating annotator disagreement were explored, with the goal of To address class imbalance and enhance model performance

improving model performance to approach human-level accuracy on minority classes, a custom training loop with a weighted

and agreement. cross-entropy loss function was implemented, as proposed in [9].



3 The class weights were calculated to be inversely proportional Method

to the frequency of each hate speech class within the training

This section details the methodology for training and evaluating data. The hyperparameters for the fine-tuning process included a

the multilingual hate speech classifier presented in this paper. It −6 learning rate of 6 × 10, a batch size of 8, and 3 training epochs.

begins with a brief overview of the datasets used, followed by During the training phase, the AdamW optimizer was employed

an explanation of the chosen pre-trained language model that to optimize the model parameters. The fine-tuning process was

serves as the foundation for fine-tuning. The section concludes implemented using PyTorch. with a description of the methods employed for evaluating the

models. 3.3 Model Evaluation



3.1 Datasets In terms of evaluation, the approach introduced in [7] was repli-

cated in order to compare the performance of the multilingual

Three monolingual datasets, i.e. the English (Youtube), Italian

(Youtube) and Slovenian (Twitter) datasets, introduced in [7] classifiers to human judgment from the perspective of disagree-

ment. This was achieved by employing identical measures to

served as the basis for our multilingual model. Each item was

annotated by two annotators independently, assigned to one of estimate the agreement between human annotators, as well as

the agreement between annotators and models. Accuracy, F1

four available classes: [Appropriate], [Inappropriate], [Offensive],

score and, most notably, Krippendorff’s ordinal alpha were used

and [Violent]. In the case of conflicting labels, both annotating

to evaluate all models in this research.

instances were kept.

Rarely used in ML applications, Krippendorff’s alpha is a ro-

To explore strategies for incorporating disagreement, three

multilingual datasets were created. First, the bust measure for assessing inter-rater reliability, accounting for Duplicate All (DA)

agreement beyond what might occur by chance. It is applicable

dataset, which contains all instances by their respective two anno-

tators from the three monolingual datasets. Second, the across various data types (nominal, ordinal, interval, and ratio Duplicate

scales) and is particularly effective in dealing with missing data.

Disagreement (DD) dataset, in which instances where annotators

The value of Krippendorff’s alpha ranges from -1 to 1, where 1

disagreed appear twice with their respective conflicting labels,

indicates perfect agreement and 0 suggests agreement equivalent

while instances that they agreed upon appear only once, creat-

to chance. Generally, an alpha above 0.80 is considered a strong

ing a more balanced training set that reflects both agreement

and disagreement, potentially preventing the models from be- agreement, while in hate speech datasets, the alpha values range

from 0.25 to 0.65. For a detailed discussion, see Krippendorff [8].

ing biased towards instances where annotators agree. And third,



the Remove Disagreement (RD) dataset, which consists only of 4 Results instances where annotators agree. Thus, the first two datasets

contain diamond standard data, while the third dataset can be This section presents the evaluation results on the multilingual

considered a gold standard dataset in which disagreement has model and its variants. It starts with an evaluation from the

been explicitly removed. perspective of inter-annotator and model-annotator agreement.

All instances in these datasets have undergone the same pre- Then, the class specific evaluation results, as well as a model

processing steps, such as replacing links and usernames with comparison based on the models’ average scores are presented.

placeholders. This step was undertaken to mitigate any potential The models are also compared to monolingual baselines fine-

biases associated with certain names, as discussed in [6]. Table 1 tuned on data for their respective languages, including the BERT

presents an overview of the label distribution across the three model for English, AlBERTo for Italian, and CroSloEngual for

multilingual training sets. The datasets used for monolingual Slovenian, as presented in [7].

evaluation are the unmodified evaluation sets presented in [7].

4.1 Inter-Annotator and Model-Annotator

Agreement

Table 1: Label distribution of the multilingual train sets

The inter-annotator agreement was computed on the evaluation

Dataset Acceptable Inappropriate Offensive Violent sets for each language using Krippendorff’s alpha and accuracy.

DD The same measures were also used to compute the agreement 111,324 DA 191,677 11,005 112,833 7,145

8,346 72,706 4,992

RD between the annotators and the models. The results are presented 80,573 2,661 40,255 2,161

in Table 2.



Multilingual Hate Speech Modeling by Leveraging Inter-Annotator Disagreement Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



Table 2: Inter-Annotator Agreement compared to model-annotator agreement in terms of Krippendorff’s ordinal alpha (𝛼)

and Accuracy (Acc.) for the models Multilingual Duplicate All (MDA), Multilingual Duplicate Disagreement (MDD), and

Multilingual Remove Disagreement (MRD) based on the language-specific evaluation sets

Dataset Inter-Annotator Agreement MDA MDD MRD

𝛼 𝛼 𝛼 𝛼 Acc. Acc. Acc. Acc.

English 58.19 82.91 55.89 79.97 50.18 76.47 57.90 81.41

Italian 57.00 81.79 58.29 82.00 56.15 80.43 57.84 82.69

Slovenian 56.62 79.43 55.74 78.60 52.95 76.52 55.15 78.84



First, in the case of inter-annotator agreement, annotators Table 3: Model evaluation results in terms of class-specific

agree around 80% of the time in terms of accuracy, with an accu- F1 scores on the English dataset. The Total score was calcu-

racy score between 79% and 82% across all three datasets. How- lated using the weighted F1 score. The first three models

ever, accuracy does not account for class imbalance, nor the represent the monolingual baselines. The subsequent mod-

ordering of the classes. A more appropriate estimate of the agree- els represent the multilingual models

ment is computed through Krippendorff’s ordinal alpha. Here, Model Appropriate Inappropriate Offensive Violent Total the annotators achieve an agreement score alpha in the values

between 0.56 and 0.58 across the three languages. IT EN 89.38 28.95 68.36 24.17 83.44 85.25 13.81 0.41 0.00 63.39

Second, in terms of agreement between annotators and mod- SL 88.01 25.17 49.69 2.88 77.71

els, the same metrics were applied. The results demonstrate a MDA 86.10 39.16 68.24 27.82 81.09

consistent level of agreement between the models and annotators MRD MDD 83.33 34.16 65.07 24.52 78.20 87.43 29.90 69.02 27.27 82.18

across all cases. Based on accuracy scores, all models align with

at least one annotator approximately 80% of the time, with alpha

values comparable to inter-annotator scores. In most instances,

the models achieve the upper limit of inter-annotator agreement,

[Offensive] were achieved by the MDA variant, once again show-

and in some cases, even exceed it (e.g., Italian Multilingual Du-

ing the superiority of the Duplicate All (DA) strategy.

plicate All MDA). This suggests that the models are effectively

In the case of the Slovenian dataset, the observed phenomena

learning consistent patterns or biases that align well with one or

slightly differ from the previous ones. The evaluation results are

more annotators. Such outcomes are expected in scenarios where

presented in Table 5. Here, two of the multilingual variants (MDA

annotator disagreement is largely due to subjective interpreta-

and RD) outperform the Slovenian monolingual model overall,

tion. This should not be construed as the model being inherently

despite predicting worse on the [Appropriate] class. Notably, the

superior, but rather as an indication of its efficiency in modeling

monolingual model outperforms all models on the [Violent] class,

the predominant patterns present in the training data.

which has not been the case for the other languages. This could

Third, a comparison between the multilingual variants shows

be due to language specifics that the multilingual model fail to

that the Duplicate Disagreement (DD) strategy consistently shows

capture, or to the specifics of the CroSloEngual BERT which is

worse alpha scores, meaning that emphasizing only on disagree-

also heavily pre-trained on Croatian and Slovenian data. Once

ment might be detrimental in training. No consistent difference

again, the DA disagreement strategy shows slight superiority

between Duplicate All (DA) and Remove Duplicates (RD) is evident

over RD.

from the experiments.

Finally, Table 6 shows the average scores of all models, achieved

by averaging their combined (weighted) F1 scores across all three

4.2 Model Comparison languages. Summarizing the multilingual superiority, these final

To evaluate the performance of the models across the four hate results show how monolingual models drastically falter on un-

speech classes, the F1 score was used. Additionally, the combined seen languages, while the multilingual models have the capacity

(weighted) F1 score was computed for each model to assess their to reach the inter-annotator agreement ceiling for all languages.

overall performance. To determine the best-performing model, While overall results show that the Remove Disagreement (RD)

the weighted F1 scores were averaged across all three languages. gold standard strategy for incorporating disagreement is best, one

Table 3 shows the results achieved by each of the models on the should be cautious when making such conclusions. Class-specific

English evaluation set. In the case of the English dataset, the re- results show that the Duplicate All (DA) strategy outperforms

sults show that the multilingual model outperforms the baseline in all the classes most relevant to hate speech detection, except

monolingual English model across all classes except the [Appro- for [Appropriate], which is the least relevant class. Another dif-

priate] class, a case in which it still performs competitively. The ference is that the MDA model involved training longer on the

variant which achieved the highest score on the minority classes same data which might have resulted in improvement on mi-

is the nority classes and saturation on the majority class. For a future MDA model, with an F1 score of 39.16 for the [Inappropri-

ate] class and an F1 score of 27.82 for the [Violent] class. This fairer comparison, the fine-tuning process on gold standard data

is most likely due to introducing the weighted cross-entropy should be adjusted accordingly. The MDA variant of the model is

loss function, which was effective in improving performance on available at: https://huggingface.co/IMSyPP/hate_speech_multil

underrepresented classes, a procedure which was not performed ingual.

in [7].

Similar patterns emerge on the Italian dataset (Table 4). The 5 Discussion

multilingual model is competitive to the monolingual model In recent years, automated hate speech detection has become

while outperforming the Italian baseline on the minority classes. crucial for moderating online content and mitigating the nega-

The highest scores on the most important classes [Violent] and tive impact on social dynamics within online communities. This



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Patricia-Carla Grigor, Bojan Evkoski, and Petra Kralj Novak



Table 4: Model evaluation results in terms of class-specific XLM-R transformer. By leveraging multilinguality, the model

F1 scores on the Italian dataset significantly outperforms monolingual baselines, demonstrating

Model its effectiveness across diverse linguistic contexts. This high- Appropriate Inappropriate Offensive Violent Total

lights the potential of multilingual approaches in improving hate

EN 86.27 1.28 1.05 0.00 67.42 speech detection, especially in scenarios where content spans IT 91.32 58.46 59.02 40.34 83.22

SL 86.23 0.76 3.25 0.00 65.95 multiple languages.

MDA 89.77 58.45 60.42 44.97 82.38 Additionally, this research incorporates inter-annotator dis-MDD 88.95 56.04 58.31 39.85 81.19

MRD 90.41 55.46 59.49 38.78 82.50 agreement into the fine-tuning process using diamond standard

data, offering a valuable alternative to traditional gold-standard

Table 5: Model evaluation in terms of class-specific F1 models. By embracing rather than ignoring annotator disagree-

scores on the Slovenian dataset ment, the model better reflects the nuances of subjective anno-

tations, enhancing its real-world applicability. However, while

Model Appropriate Inappropriate Offensive Violent Total this approach shows promise, annotator disagreement continues

EN 79.93 3.98 2.34 0.00 53.84 to present challenges, indicating that further work is needed to

IT 79.84 3.80 1.24 0.00 53.43 fully address its impact on model performance.

SL 85.70 43.69 65.26 29.12 78.39

MDA 84.30 45.22 69.69 24.79 78.88 Future research could extend this work by evaluating the mod-

MDD 82.33 43.39 68.59 23.84 77.19 els on additional languages, exploring alternative baseline models,

MRD 84.98 38.47 68.40 15.50 78.80 refining strategies for incorporating annotator disagreement and

handling minority classes. As online hate speech extends its im-

Table 6: Average performance of models based on class- pact, developing robust, multilingual content moderation systems

weighted F1 scores across three languages

is crucial to maintaining safe and inclusive digital environments.



Model Avg. Weighted F1 Score (all languages) 7 Acknowledgments

EN 68.23 The authors acknowledge partial financial support from the Slove-

IT 66.68 nian Research Agency (research core funding no. P2-103).

SL 74.02



MDA 80.78 References

MDD 78.86

MRD 81.16 tion is not as easy as you may think: a closer look at model validation. In [1] Aymé Arango, Jorge Pérez, and Barbara Poblete. 2019. Hate speech detec-

Proceedings of the 42nd international acm sigir conference on research and

research proposes a novel multilingual hate speech model to ad- development in information retrieval, 45–54.

[2] Alexander Brown. 2017. What is hate speech? Part 2: Family resemblances.

dress these challenges on a broader scale. The following discusses Law and Philosophy, 36, 561–613.

the main findings. [3] Naganna Chetty and Sreejith Alathur. 2018. Hate speech review in the

First, the inter-annotator agreement and the agreement be- context of online social networks. Aggression and violent behavior, 40, 108–

118.

tween annotators and models suggest that inter-annotator agree- [4] Alexis Conneau et al. 2019. Unsupervised cross-lingual representation learn-

ment sets an intrinsic limit on model performance. Models are ing at scale. CoRR, abs/1911.02116.

limited by the quality and consistency of the annotated data, Hovy, Massimo Poesio, et al. 2021. Beyond black & white: leveraging an- [5] Tommaso Fornaciari, Alexandra Uma, Silviu Paun, Barbara Plank, Dirk

which directly affects their ability to accurately predict unseen notator disagreement via soft-label multi-task learning. In Proceedings of

data. However, incorporating areas of disagreement into model the 2021 Conference of the North American Chapter of the Association for

Computational Linguistics: Human Language Technologies. Association for

development can lead to more robust models capable of han- Computational Linguistics.

dling ambiguous cases by employing one of the several available [6] Nikhil Garg, Londa Schiebinger, Dan Jurafsky, and James Zou. 2018. Word

strategies for incorporating disagreement. embeddings quantify 100 years of gender and ethnic stereotypes. Proceedings

of the National Academy of Sciences, 115, 16, E3635–E3644.

Second, the multilingual model consistently surpassed the [7] Petra Kralj Novak, Teresa Scantamburlo, Andraž Pelicon, Matteo Cinelli,

monolingual baselines, achieving the inter-annotator agreement Igor Mozetič, and Fabiana Zollo. 2022. Handling disagreement in hate speech

ceiling across all languages. This success can be attributed partly modelling. In International Conference on Information Processing and Man-

agement of Uncertainty in Knowledge-Based Systems. Springer, 681–695.

to the ability to leverage patterns learned from multiple lan- [8] Klaus Krippendorff. 2018.Content analysis: An introduction to its methodology.

guages, partly to vast amounts of data incorporated into state-of- Sage publications.

[9] Andraž Pelicon, Syrielle Montariol, and Petra Kralj Novak. 2023. Don’t start

the-art pre-trained multilingual models, and partially to the class your data labeling from scratch: opsala-optimized data sampling before

weighting scheme employed in the fine-tuning. These findings labeling. In International Symposium on Intelligent Data Analysis. Springer,

support the first research question, demonstrating that a multi- 353–365.

[10] Juan Manuel Pérez et al. 2023. Assessing the impact of contextual informa-

lingual hate speech classifier trained on diamond standard data tion in hate speech detection. IEEE Access, 11, 30575–30590.

outperforms its monolingual counterparts. [11] Fabio Poletto, Valerio Basile, Manuela Sanguinetti, Cristina Bosco, and

Finally, this research contributes substantially to hate speech detection: a systematic review. Viviana Patti. 2021. Resources and benchmark corpora for hate speech Language Resources and Evaluation, 55, 477–

classification in a multilingual context by introducing a novel 523.

multilingual hate speech detection model and making it avail- [12] Alexandra N Uma, Tommaso Fornaciari, Dirk Hovy, Silviu Paun, Barbara

Plank, and Massimo Poesio. 2021. Learning from disagreement: a survey.

able on the Hugging Face platform. Our model underscores the Journal of Artificial Intelligence Research, 72, 1385–1470.

importance of incorporating inter-annotator disagreement into [13] William Warner and Julia Hirschberg. 2012. Detecting hate speech on the

model development, challenging the reliance on gold standard world wide web. In Proceedings of the second workshop on language in social

media, 19–26.

data in subjective tasks, such as hate speech detection. [14] Wenjie Yin and Arkaitz Zubiaga. 2021. Towards generalisable hate speech

detection: a review on obstacles and solutions. PeerJ Computer Science, 7,

6 Conclusions e598.

This paper advances automatic hate speech detection by introduc-

ing a novel multilingual model fine-tuned on the state-of-the-art





Predicting Pronunciation Types




in the Sloleks Morphological Lexicon of Slovene



1,2

Jaka Čibej

jaka.cibej@f f.uni- lj.si

jaka.cibej@ijs.si

1Faculty of Arts, University of Ljubljana

2Jožef Stefan Institute

Ljubljana, Slovenia



Abstract representation, with some exceptions and several predictable

phoneme assimiliations (such as the assimilation of voiceless

We present an experiment dealing with the automatic prediction

of pronunciation types for lemmas in the

Sloleks Morphologi- consonant phonemes to their voiced equivalents ‘music’, glasba

cal Lexicon of Slovene IPA: / /, or vice-versa, voiced-to-voiceless, ‘to "gla:zba podpreti

. We perform a statistical analysis on a

support’, IPA: / /). pOt"pre:ti

number of mostly 𝑛-gram-based features and use a set of sta-

However, not all entries in Sloleks follow Slovene G2P prin-

tistically significant features to train and test several machine

ciples. For a number of words, particularly proper nouns de-

learning models to discriminate between lemmas for which a pho-

noting people ( , ), locations ( , Shakespeare Sharon Sydney Birm-

netic transcription can be generated automatically using Slovene

grapheme-to-phoneme (G2P) conversion rules (e.g. ), and ingham Newyorčan Novak ), inhabitants ( ‘New Yorker’), etc.; as well as

adjectives derived from proper nouns ( ‘pertaining to aachenski

lemmas with pronunciation that follows other G2P rules (e.g.

Shakespeare Aachen’, ‘belonging to Acroni’), the phonetic transcrip- Acronijev

).

tion cannot be generated using Slovene G2P rules. In such cases



Keywords with foreign orthographic elements that indicate relations be-

tween graphemes and phonemes that are unusual for Slovene,

grapheme-to-phoneme conversion, pronunciation types, mor-

Slovene linguistic and lexicographic practice (see e.g. [5]) first re-

phological lexicon, proper nouns, Slovene

quires a transliteration into the closest equivalent using Slovene

graphemes, which can then be used to generate the phonetic tran-

1 Introduction scription using Slovene G2P rules (e.g. Newyorčan → njújórčan

The [2] is the largest Sloleks Morphological Lexicon of Slovene → "nju:"jo:rtSan IPA: / /).

open-access database containing machine-readable information Because of this, it is necessary to discriminate between differ-

on the morphological properties of Slovene lemmas (e.g. ent : categories of words that follow Slovene miza pronunciation types

‘table’, noun, common, feminine) and their inflected forms (e.g. G2P rules ( ) and those that do not (e.g. ; Slovene G2P Other G2P

mize, singular, genitive; mizo, singular, accusative). Since version more on this in Section 2). Pronunciation types denote the manner

2.0 [3], each lemma and inflected form also contains accentuated in which the phonetic transcription of the word can be generated.

forms (e.g. ) and phonetic transcriptions in the International In some cases, assigning the pronunciation type to a lemma is míza Phonetic Alphabet (IPA) and its equivalent X-SAMPA (e.g. IPA:

trivial – if the lemma contains a grapheme that is not part of

/ 2 "mi:za x y w q Other

/, X-SAMPA: /"mi:za/). Both transcriptions were generated the Slovene alphabet (e.g. , , , ), it belongs into the

automatically from accentuated forms, first in version 2.0 using a category (e.g. , ). There are, however, many G2P Byron Oxford

rudimentary rule-based system, then again in 3.0 with a greatly exceptions that belong in the category despite being Other G2P

improved and linguistically informed rule-based grapheme-to- comprised entirely of Slovene graphemes (e.g. , ). Matt Sharon

1

phoneme (G2P) conversion tool for Slovene. In Sloleks 3.0, the first cca. 100,000 lemmas that had been part

Rule-based G2P conversion for Slovene (particularly from ac- of version 2.0 were manually annotated with pronunciation types,

centuated forms) yields very good results and leaves only a mi- whereas the 264,000 new entries (added automatically from the

nority of issues to be resolved manually because in terms of its [6]) still lack this Gigafida 2.0 Corpus of Modern Standard Slovene

orthographic depth, Slovene features a shallow orthography ([9])

information. Because manual annotation from scratch is time-

in which each grapheme in the alphabet generally corresponds consuming, we performed an experiment to determine to what

to one phoneme (see e.g. [4]) and the spelling-sound correspon- degree the pronunciation type can be predicted automatically by

dence is relatively direct ([1]; [11]): the pronunciation rules allow relying on the scarce linguistic and morphosyntactic information

for words to be pronounced correctly based on their graphemic that can be extracted from an individual lemma.

The paper is structured as follows: we describe the dataset

1 Pregibalnik

The Slovene G2P tool is part of , a piece of software used for the

that was used for the statistical analysis and machine learning

automatic expansion of the : https://github Sloleks Morphological Lexicon of Slovene

.com/clarinsi/SloInf lector It was developed within the experiment (Section 2), as well as the process of feature selection Development of Slovene in

the Digital Environment project. The Slovene G2P converter is also available as an (Section 3). We train several machine-learning models and evalu-

API-service: https://orodja.cjvt.si/pregibalnik/g2p/docs

ate their performance using 10-fold cross-validation (Section 4).

Finally, we manually evaluate a sample of automatically anno-

Permission to make digital or hard copies of all or part of this work for personal

or classroom use is granted without fee provided that copies are not made or tated entries (Section 5) and conclude the paper with our plans

distributed for profit or commercial advantage and that copies bear this notice and

for future work (Section 6).

the full citation on the first page. Copyrights for third-party components of this

work must be honored. For all other uses, contact the owner /author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia 2Although ć and đ are not part of the Slovene alphabet, they are phonemically © 2024 Copyright held by the owner/author(s). transparent and frequently occur in names of Slovene citizens, so they are not

https://doi.org/https://doi.org/10.70314/is.2024.sikdd.2 counted as foreign characters for the purposes of this task.



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Čibej



Table 1: Lemmas in Sloleks 3.0 by Pronunciation Type Table 3: Statistically Significant Features by Category



Pronunciation Type Frequency % Feature Category Number



- 264,538 72.41% Percentage of characters 1 Slovene G2P

Slovene G2P 94,750 25.93% Morphosyntactic features 3

Other G2P 3,066 0.84% General character-level 𝑛-grams 1,119

Numeral 1,840 0.50% Initial character-level 𝑛-grams 398

Acronym 845 0.23% Final character-level 𝑛-grams 468

Slovene G2P with minor deviation 113 0.03% General robust CVC 𝑛-grams 66

Abbreviation 70 0.02% Initial robust CVC 𝑛-grams 44

Ambiguous G2P 69 0.02% Final robust CVC 𝑛-grams 39

Symbol 49 0.01% General finegrained CVC 𝑛-grams 157



Total Initial finegrained CVC 𝑛-grams 102 365,340 100.00%

Final finegrained CVC 𝑛-grams 93



Total 2,490

Table 2: Lemmas in Sloleks 3.0 with Other G2P pronuncia-

tion type by Morphosyntactic Properties



Morphosyntactic Properties alphabet as well as and ); (b) morphosyntactic features (e.g. ć đ Frequency %

noun, proper, masculine 5 ); (c) relative frequencies of character-

Adjective, possessive 1,092 35.62%

level uni-, bi-, and trigrams within the lower-cased lemma (e.g.

Noun, proper, masculine 958 31.25% Matt → 𝑓𝑟 𝑚 , ( ) 𝑓 , ..., 𝑟 ( 𝑎 𝑓𝑟 𝑚𝑎 , ) ( ) 𝑓𝑟 𝑎𝑡 , ..., ( ) 𝑓𝑟 𝑚𝑎𝑡 , ...); (d) rela-( ) Noun, proper, feminine 713 23.26%

tive frequencies of character-level uni-, bi-, and trigrams from

Adjective, general 142 4.63%

a robust CVC-conversion of the lemma, substituting consonant

Noun, common, masculine 127 4.14% C V Matt →

graphemes with and vowel graphemes with (e.g.

Noun, common, feminine 20 0.65% CVCC → ( ) ( ) ( ) ( ) ( )

𝑓 𝐶 , 𝑓 𝑉 , ..., 𝑓 𝐶𝑉 , 𝑓 𝑉 𝐶 , ..., 𝑓 𝐶𝑉 𝐶 , ...); (e) rel-

𝑟 𝑟 𝑟 𝑟 𝑟

Adverb, general 10 0.33%

ative frequencies of character-level uni-, bi-, and trigrams from a

Noun, common, neuter 2 0.07% 6 Matt → ZVKK

finegrained CVC-conversion of the lemma (e.g.

Verb, main, imperfective 2 0.07% → 𝑓𝑟 𝑍 , ( ) 𝑓𝑟 𝑉 , ..., ( ) 𝑓𝑟 𝑍 𝑉 , ( ) 𝑓𝑟 𝑉 𝐾 , ..., ( ) 𝑓𝑟 𝑍 𝑉 𝐾 , ...) ( )

Total 3,066 100.00% For (c), (d), and (e), the initial and final uni-, bi-, and trigrams

of the lemma were extracted separately as well, as in some cases

the position of the 𝑛-gram in the word can be indicative of one

2 Dataset class over another.

For general character-level 𝑛-grams, the first 1,498 with a fre-

Sloleks 3.0 contains a total of 365,340 entries, but only approxi-

quency of at least 500 across all Sloleks 3.0 lemmas were analyzed;

mately 28% have been manually assigned one of 8 pronunciation

3 these cover cca. 88.34% of all 𝑛-gram occurrences. For robust CVC

types (as shown in Table 1). For the classification task, we focus

only on the two most frequent pronunciation types (

Other G2P and finegrained CVC 𝑛-grams, all were analyzed. We performed

and ).

Slovene G2P 4 the Kruskal–Wallis H test [7] (k=2, n=97,056) on a total of 6,148

7

In terms of their morphosyntactic features, the

Other G2P features, out of which 2,490 (40%) were statistically significant.

Statistically significant features by categories are shown in Table

lemmas mostly consist of possessive adjectives and proper nouns,

3. 1,146 features are more indicative of and 1,344 are Slovene G2P

collectively accounting for cca. 90% of the category (as shown in

more indicative of . As shown in Table 4, only three Other G2P

Table 2), but only 15% of the portion of Sloleks annotated with

of the top 10 general 𝑛-grams indicative of actually Other G2P

pronunciation types.

contain non- characters, confirming that detecting Slovene G2P

The final dataset for statistical analysis and machine learning

consisted of 94,863 lemmas (e.g. , , dekadentnost Košak

Slovene G2P lemmas from the category is more complex and re- Other G2P

prefiltriran quires more than simply taking into account non-) and 3,066Slovene G2P Other G2P lemmas (e.g. Elizabeth , Presley ,

Sinclaire graphemes.

).



3 Statistical Analysis and Feature Selection 4 Pronunciation Type Prediction

From each lemma, we extracted a series of features that could help The identified features (along with several placeholder 𝑛-grams

discriminate between the two classes: (a) percentage of to take into account any graphemes not covered in the initial Slovene

G2P graphemes within the lemma (i.e. graphemes of the Slovene dataset) were taken into account to develop a custom vectorizer

that converts a given lemma and its lexical features based on

3

It should be noted that all the inflected forms within the entry effectively inherit

the MulText-East v6 (MTE-6) Morphosyntactic Specifications for

the pronunciation type.

4 Ambiguous G2P

Symbols in Sloleks are rare, along with entries within the category

(where an entry can either follow Slovene G2P rules or not, depending on the

5 Í ( ) = ( )/ ( )

context – e.g. as a Slovene name: / /; or as an English name with Relative frequencies were calculated as 𝑓𝑟 𝑥𝑛 𝑓𝑎 𝑥𝑛 𝑓𝑎 𝑦𝑛 , e.g. the Amanda am"a:nda a pronunciation adjusted to the Slovene set of phonemes: . Abbrevi- absolute frequency of 𝑛-gram 𝑥 of length 𝑛 within the lemma divided by the sum /9m"E:nda/) ations and numerals are easily identifiable, and while acronyms have a separate of absolute frequencies of each 𝑛-gram 𝑦 of length 𝑛 within the lemma.

6

manner of generating phonetic transcriptions which also depends on their morpho- In the finegrained CVC-conversion, consonant graphemes were generalized into

logical patterns, they are also mostly identifiable with rules. Because of its rarity more finegrained categories, e.g. graphemes denoting Slovene sonorants (M), voiced

and similarity to , the category was (G) and voiceless obstruents (K), foreign consonants (X), etc. Slovene G2P Slovene G2P with minor deviation

merged into for the classification task. Effect size was calculated as 𝜂 𝐻 𝑘 1 𝑛 𝑘 , as reported in [10]. = ( − + )/( − )

Slovene G2P 7 2



Predicting Pronunciation Types in Sloleks Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



Table 4: Top 10 Statistically Significant General Character- Table 6: Confusion Matrix for Linear Support Vector Clas-

Level 𝑛-Grams by Effect Size ( 2 𝜂) sifier



𝑛 𝜂 True-Gram H p Means →

2



y ↓ Í Predicted Slovene G2P Other G2P 11509.36 p ≤ 0.0001 0.1186 𝜇 < 𝜇 𝑆 𝑂

w 9595.25 p 0.0001 0.0989 𝜇𝑆 ≤ < 𝜇 Slovene G2P 18,939 140 19,079 𝑂

ch 7558.60 p 0.0001 0.0778 𝜇 < 𝜇 Other G2P 34 507 ≤ 473

𝑆 𝑂

ll 6295.96 p 0.0001 0.0649 𝜇 < 𝜇 18,973 613 -

≤ Í

𝑆 𝑂

ss 3804.26 p 0.0001 0.0392 𝜇𝑆 < 𝜇𝑂 ≤

nn 3220.65 p 0.0001 0.0332 𝜇 < 𝜇 ≤

𝑆 𝑂

th 2973.89 p 0.0001 0.0306 𝜇𝑆 < 𝜇𝑂 ≤ Table 7: Confusion Matrix for Manual Evaluation

wa 2761.53 p 0.0001 0.0284 𝜇 < 𝜇 ≤

𝑆 𝑂

tt 2745.10 p 0.0001 0.0283 𝜇𝑆 ≤ < 𝜇𝑂 True →

co 2571.20 p 0.0001 0.0265 𝜇 < 𝜇 Í Predicted ≤ ↓ 𝑆 𝑂 Slovene G2P Other G2P



Slovene G2P 9 95 86

Table 5: Model Performance Based on 10-Fold Cross- Other G2P 14 91 105 Validation Í 100 100 -



Model A BA P R F1 ROC AUC



LinearSVC 87.87 96.36 87.87 91.64 98.89

99.08 ‘a’ is pronounced as / /, but this cannot be discerned from the E

graphemic representation itself. Other misclassified examples are

Multin. NB 97.38 79.17 78.12 79.17 78.62 96.55

more obviously pertaining to , e.g. , Other G2P Dorfmeister Faulkn-

kNN (k=5) 98.25 75.17 93.67 75.17 81.74 91.63

erjev, Flaubertov, Heisenbergov, Balfourjev. This might indicate

Majority 96.87 - - - - -

that not all indicative 𝑛-grams have been included as features

(e.g. ‘ei’, ‘ou’), possibly for lack of evidence in the original dataset

or because they are less frequent and have not been included in

8

Slovene into a 2,500-dimensional numerical vector. The entire

the initial batch of statistical tests. As the lexicon expands with

dataset was converted into vectors and split into a training set

new entries, the model will be updated with new examples and

9

(80%) and a test set (20%), both stratified by class. Three models

new features to potentially improve performance.

(Linear Support Vector Classifier (LinearSVC), Multinomial Naive



Bayes Classifier (Multin. NB), and k Nearest Neighbors Classifier 5 Manual Evaluation (kNN)) were trained and evaluated with 10-fold cross-validation.

10 We trained a new instance of the LinearSVC model on the entire

The results are listed in Table 5 and show that LinearSVC out-

dataset and used it to annotate the remaining cca. 264,000 lemmas

performs the other two models. All three exhibit above-baseline

from Sloleks 3.0 with no pronunciation type, resulting in 86,730

accuracy compared to the majority classifier, but Multinomial

lemmas with and 177,808 lemmas with . Other G2P Slovene G2P

NB and kNN perform much worse in terms of balanced accuracy

We performed a preliminary manual evaluation consisting of

as well as precision and, in case of kNN, recall. Recall is also

a random sample of 100 examples from each class. The results

somewhat lower with LinearSVC, which is to be expected – some

Other G2P are shown in the confusion matrix in Table 7. Although the

lemmas might contain no indicative 𝑛-grams and are

sample is too small to be representative of the whole, it indicates

thus hard to detect; on the other hand, once identified, the model

that the model performs well even on unseen data, achieving

is very precise in its prediction.

an accuracy of 88.50% (P=0.91, R=0.87, F1=0.89) over a majority

Table 6 shows the confusion matrix for the LinearSVC model

baseline accuracy of 50.00%.

tested on the 20% stratified test dataset. The model very rarely

misclassifies lemmas, and more frequently errs with

Slovene G2P The misclassifications of as include Other G2P Slovene G2P

Other G2P examples such as , , , and , with Mukhamedov Beatli Livenza Preidler Slovene

lemmas. A closer inspection of the misclassified

G2P limited indicators that the words belong to the cat- Other G2P Beethoven

examples reveals several errors in the original dataset: ,

Ratzinger egory. Most graphemes in these examples are pronounced ac- Rotterdam Franco Oberstdorf Keller

, , , , and were in fact

correctly classified as , but they are miscategorized as

Other G2P cording to criteria, with the exception of individ- Slovene G2P

Slovene G2P ual 𝑛-grams (‘nz’, ‘ei’, ‘kh’), some of which have not been in-

in the original dataset. Other misclassifications in-

cluded in the set of features. In other examples, only one or two

clude examples of foreign proper nouns and possessive adjectives

vowel graphemes are indicative of pronunciation (e.g. Other G2P

that contain unusual grapheme combinations for Slovene (e.g.

Andreas Trendlina, which is also a lemmatization error; the correct lemma , Aurelio , Hilton , Simpsonov ), but their pronunciation can

still be derived from their graphemic representation (e.g.

Andreas is ; and ), and the pronunciation of single vowel Trendline Sanberg

→ graphemes appears harder to predict than consonant graphemes and"re:as

IPA: / /).

On the other hand, lemmas misclassified as Slovene

Other G2P or combinations thereof.

G2P Similarly, the misclassifications of lemmas as Slovene G2P Andersonov Atkinsov Batmanov

include , , , in which the grapheme

Other G2P lemmas include examples such as Doneck, Barson,

8 Bronson Piersanti Faustini

MTE-6: https://nl.ijs.si/ME/V6/msd/html/msd- sl.html The vectorizer uses Slovene , , and . While these are proper nouns of

morphosyntactic tags, e.g. (S – noun, l – proper, z – feminine). Slz

foreign origin, their Slovene pronunciation can either be fully

9 scikit-learn Doneck →

All models were trained using the Python library . [8]

10 discerned from their graphemic representation (e.g.

A, BA, P, R, and F1 refer to accuracy, balanced accuracy, macro-precision, macro-

recall and macro-F1, respectively. IPA: / /), or it only differs slightly from what Slovene dO"ne:tsk



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Čibej



grapheme-to-phoneme conversion would produce (e.g. even the new orthographic manual anticipates that all transliter- Faus-

tini → automatically converted IPA: /faus"ti:ni/; correct IPA: ation should be done manually, which begs the question whether

/ /). at least part of the work can be automatized. This would be an fau s"ti:ni

“ important step in the development of a modern, digital infrastruc-



6 Conclusion ture for Slovene orthography, and would facilitate the automatic

expansion of modern digital dictionary databases and datasets

In the paper, we presented the results of an attempt to automatize

for automatic speech recognition.

the assignment of pronunciation types to lemmas in the Sloleks

In addition, although our preliminary experiments with LLMs

Morphological Lexicon of Slovene. The results show that a model

(ChatGPT 3.5 and 4.0) classifying and lem-Slovene G2P Other G2P

based on a series of mostly 𝑛-gram features can provide good

mas have yielded much worse results than the best performing

results when discriminating between and Slovene G2P Other G2P

LinearSVC model, more systematic experiments are warranted.

categories, with the best performance achieved by the Linear

As part of our future work, we intend to implement the model

Support Vector Classifier. However, there is still room for im-

into , which is used for automatically extending the

Pregibalnik 12

provement, particularly in terms of recall – a number of Other

lexicon and currently assigns no pronunciation type. The model

G2P lemmas from the test set were misclassified as Slovene G2P, 13

itself is available under the Apache 2.0 license on Github , while

while those classified correctly were classified with a relatively

the pronunciation type annotations will be included in future

high precision score. 𝑛-grams that are statistically significant

versions of Sloleks and, eventually, manually validated.

as indicative of one class have proven to be useful features for



model development, but because they are not evenly distributed Acknowledgements and occur sporadically in different lemmas, it would make sense

The research presented in this paper was conducted within the re-

to further improve the model by performing the same statistical

search project titled Basic Research for the Development of Spoken

analysis (as described in Section 3) on the long tail of less fre-

quent 𝑛-grams to prepare a more comprehensive list of indicative Language Resources and Speech Technologies for the Slovenian Lan-

𝑛-grams. The current version of the model is very light-weight guage Language Resources and ( J7-4642), the research programme

and additional features should not cause the model to become Technologies for Slovene (P6-0411), and the CLARIN.SI research

infrastructure, all funded by the Slovenian Research and Inno-

overencumbered.

vation Agency (ARIS). The author also thanks the anonymous

There are several possibilities for further development of the

reviewers for their constructive comments.

model. Firstly, instead of using relative frequencies of 𝑛-grams

as features, it would be useful to test how different measures

such as TF–IDF, absolute frequencies, or even Boolean values References

influence the performance of the model, and potentially also test [1] Derek Besner and Marilyn Chapnik Smith. 1992. Chapter 3 basic processes

in reading: is the orthographic depth hypothesis sinking? In Orthography,

several other machine learning algorithms (e.g. Random Forest Phonology, Morphology, and Meaning. Advances in Psychology. Vol. 94. Ram

Classifier). Secondly, while the other pronunciation types from Frost and Leonard Katz, editors. North-Holland, 45–66. doi: https://doi.org

/10.1016/S0166- 4115(08)62788- 0.

Sloleks 3.0 (acronyms, abbreviations, etc.) are relatively easily

[2] Jaka Čibej et al. 2022. Morphological lexicon sloleks 3.0. Slovenian language

identifiable (but much less frequent), in the next step, it would

resource repository CLARIN.SI. (2022). http://hdl.handle.net/11356/1745.

be informative to include them in the training set and test out [3] Kaja Dobrovoljc, Simon Krek, Peter Holozan, Tomaž Erjavec, Miro Romih,

Špela Arhar Holdt, Jaka Čibej, Luka Krsnik, and Marko Robnik-Šikonja. 2019.

the model’s performance on the full set of categories. Thirdly,

Morphological lexicon sloleks 2.0. Slovenian language resource repository

a statistical analysis should be performed on the probabilities

CLARIN.SI. (2019). http://hdl.handle.net/11356/1230.

with which the model makes decisions and to what degree they [4] Florina Erbeli and Karmen Pižorn. 2012. Reading ability, reading fluency

and orthographic skills: the case of l1 slovene english as a foreign language

correlate with the percentage of graphemes that differ from the

students. English. , 2(3), 119–139. Center for Educational Policy Studies Journal

shallow orthographical Slovene G2P rules (e.g. , with https://f iles.eric.ed.gov/f ulltext/EJ1130208.pdf . Anderson

arguably only ‘a’ not following Slovene G2P rules; vs. [5] Nataša Gliha Komac et al. 2015. Koncept novega razlagalnega slovarja Châteaux

,

slovenskega knjižnega jezika. Inštitut za slovenski jezik Frana Ramovša

where the majority of graphemes are pronounced completely

ZRC SAZU. (2015). https://f ran.si/179/novi- slovar- slovenskega- knjiznega- j

differently compared to Slovene G2P rules). This would require ezika/datoteke/Potrjeni_koncept_NoviSSKJ.pdf .

[6] Simon Krek et al. 2019. Corpus of written standard slovene gigafida 2.0.

the preparation of a separate dataset in which graphemes are

Slovenian language resource repository CLARIN.SI. (2019). http://hdl.handl

manually aligned to either the graphemes of their transliter- e.net/11356/1320.

ated Slovene graphemic forms ( ) or their [7] William H. Kruskal and W. Allen Wallis. 1952. Use of ranks in one-criterion Newyorčan → njújórčan

variance analysis. , 47, 260, Journal of the American Statistical Association

Slovene IPA transcriptions. By assigning scores that reflect the

583–621. eprint: https://www.tandf online.com/doi/pdf /10.1080/01621459.19

degree of orthography depth for the individual lemma, it would 52.10483441. doi: 10.1080/01621459.1952.10483441.

[8] F. Pedregosa et al. 2011. Scikit-learn: machine learning in Python. Journal

be possible to use the dataset to train a regression model.

Similarly, lemmas from Sloleks 3.0 can be manually

Other G2P of Machine Learning Research, 12, 2825–2830.

[9] Anja Schüppert, Wilbert Heeringa, Jelena Golubovic, and Charlotte Gooskens.

annotated with their language of origin and transliterated ac- 2017. Write as you speak? a cross-linguistic investigation of orthographic

cording to the recently published transliteration rules of

Pravopis transparency in 16 germanic, romance and slavic languages. English. From

semantics to dialectometry, 32, 303–313. isbn: 9781848902305.

8.0 11 , the new orthographic manual of Slovene, which at the time [10] Maciej Tomczak and Ewa Tomczak. 2014. The need to report effect size of writing this paper is still in development. Such a dataset would estimates revisited. an overview of some recommended measures of effect

size. , 1(21), 19–25. Trends in Sport Sciences

enable the development of a model for language identification

[11] Antal van den Bosch, Alain Content, Walter Daelemans, and Beatrice de

for individual lemmas, and, ultimately, a model for automatizing Gelder. 1994. Analysing orthographic depth of different languages using

transliteration of lemmas of foreign origin into their Slovene data-oriented algorithms. In Proceedings of the 2nd International Conference

on Quantitative Linguistics.

equivalents. As of now, no such tool yet exists for Slovene, and



12 Pregibalnik

: https://github.com/clarinsi/SloInf lector; the entire tool is also

11Pravopis 8.0: Pravila novega slovenskega pravopisa za javno razpravo

. https://prav available as an API-service: https://orodja.cjvt.si/pregibalnik/docs

13

opis8.fran.si/, 9 August 2024 GitHub: https://github.com/jakacibej/sikdd2024_predicting_pronunciation_types





Higher-Order Bibliographic Services




based on bibliographic networks



Vladimir Batagelj Jan Pisanski Tomaž Pisanski

IMFM Faculty of Arts, UL FAMNIT, UP

Ljubljana, Slovenia Ljubljana, Slovenia Koper, Slovenia

IAM and FAMNIT, UP jan.pisanski@f f.uni- lj.si IMFM

Koper, Slovenia Ljubljana, Slovenia

vladimir.batagelj@fmf.uni- lj.si tomaz.pisanski@upr.si



Figure 1: The largest co-author groups at level 10 at the University of Primorska until 2024.



Abstract of characteristics describing works. Besides these networks, we

can also get the partition of works by their publication years,

Bibliographic databases only provide basic services to users, but

the partition of works by journals or publishers, the vector of

they could provide much richer information for specific user

the number of pages, and, in some cases, the (one-mode works ×

needs. The main reason for the delay in developing such higher-

works) citation network.

order bibliographic services is the limited access to data in propri-

When constructing any of these networks, the first task is

etary databases. We expect the new open bibliographic databases

to specify the nodes and which relations are linking them. In

like OpenAlex will encourage faster development of these ser-

short, the network boundary problem [16] has to be solved. This

vices. We describe an approach based on a collection of biblio-

includes deciding whether a network is one-mode or two-mode

graphic networks as a foundation to support the development of

and which node properties are important for the intended analy-

higher-order bibliographic services.

ses. For specifying links, this amounts to answering a series of



Keywords questions:

bibliographic database, open access, network analysis, higher-

(1) Are the links directed?

order bibliographic service, prototype, OpenAlex

(2) Are there different types of links (relations) to include?

(3) Can a pair of nodes be linked with multiple links?

1 Introduction (4) What are the weights on the links?

From special bibliographies (BibTEX, EndNote) and bibliographic (5) Is the network static, or is it changing through time?

databases, it is possible to obtain data about works (papers, books,

Another problem that often occurs when defining the set of

reports, etc.) on selected topics. A typical work description con-

nodes is the identification of nodes. The unit corresponding to a

tains the following data: authors; title; publisher/journal; pub-

node can have different names (synonymy), or the same name can

lication year and pages. In some sources, additional data are

denote different units (homonymy or ambiguity). For example

available including languages, classification of documents, key-

words, authors’ institution/country affiliation, lists of references, in the BibTEX bibliography from the Computational Geometry

Database [14] the same author appears under 7 different names:

and the abstract. This data can be transformed into a collection

R.S. Drysdale, Robert L. Drysdale, Robert L. Scot Drysdale, R.L.

of compatible two-mode networks on selected topics [5]: works

× Drysdale, S. Drysdale, R. Drysdale, and R.L.S. Drysdale. Insider × ×

authors; works keywords; works countries, and other pairs

information is needed to decide that Otfried Schwarzkopf and

Permission to make digital or hard copies of all or part of this work for personal

Otfried Cheong are the same person. At the other extreme, there

or classroom use is granted without fee provided that copies are not made or

are at least 57 different mathematicians with the name Wang, and

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this Li in the MathSciNet Database [20]. Its editors have tried hard,

work must be honored. For all other uses, contact the owner /author(s).

from 1985, to resolve the identification of the author’s problem

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

during the data-entry phase. The significant growth of contri-

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.sikdd.12 butions by Chinese scientists and their full name similarity in



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Batagelj, Pisanski & Pisanski



Roman transcriptions adds additional complexity to the problem. our publications. Similarly, we get the report on the publication

In the future, the problem could be eliminated by implementing activity of the selected institution.

initiatives such as using ORCID or resolving the identification



problem in bibliographic databases (Scopus, OpenAlex). 3.1 API

An application programming interface (API) is a way for two or

2 Higher-Order Bibliographic Services more computer programs or components to communicate with

The data collected in different bibliographic databases can be each other. It is a type of software interface, offering a service to

used to provide higher-order bibliographic and bibliometric ser- other pieces of software [21]. In our case, API enables us to use

vices such as what to read (contact/visit)? – a list of relevant the database data from our programs. An R package supporting

articles/books (authors, institutions) on selected topic; where to the use of OpenAlex is openalexR [1].

publish? – a list of journals suitable for the publication of an The OpenAlex API is available at https://api.openalex.org. Its

article, automatic suggestion of keywords; reviewer selection – a response is returned in JSON format. Here is an R code using the

list of reviewers suitable for a submitted article; possible partners OpenAlex API for the IMFM institution search

for research collaboration; a career application – a candidate’s setwd(wdir <- "C:/work/OpenAlex/API") activity report draft; etc.) for different types of users (students, re- library(httr); library(jsonlite) searchers, teachers, decision-makers, funding agencies, research res <- GET("https://api.openalex.org/institutions",

institutions, database managers, etc. . To support this goal we query = list(search="imfm"))

str(res)

have to use high-quality data often obtained by combining data

cont <- fromJSON(rawToChar(res$content))

from different databases. names(cont); str(cont)

For the development of higher-order bibliographic and biblio-

The response data are available in the variable . Similarly, cont

metric services, open bibliographic databases such as OpenAlex

the API can be used also from other programming languages.

are particularly welcome, as the developed services can remain

The OpenAlex query can be composed of different components.

open.

Using we can search for a given search text across titles, search



3 OpenAlex to units satisfying given conditions. Using abstracts, and full-text. Using a we can limit our search filter

select we can select

The basic type of unit in a bibliographic database is the work. data fields that will appear in results. The query can be further

A user searching the database gets a list of works satisfying the controlled by some parameters. For example

query. Usually, some operations with such lists (inspection, fil- wd <- GET("https://api.openalex.org/works", tering, merging, intersection, statistics, etc.) are supported. Only query = list( basic services are provided to users. search="handball",

Some web services also supporting some other types of units filter="publication_year:2015",

(authors, institutions, research fields, conferences, etc.) were de- select="id,title",

page="2", per_page="200"))

veloped such as Google Scholar [19], Scholar GPS [12], and DBLP

names(wd)

– computer science bibliography [10]. wc <- fromJSON(rawToChar(wd$content)); names(wc)

Our approach is based on OpenAlex [18, 9] but this informa- names(wc$meta); wc$meta$count; str(wc$results)

tion can be obtained from most bibliographic databases [13, 11].

returns the second page (with up to 200 entries) on works on

OpenAlex indexes more than twice as many scholarly works as

handball published in the year 2015. Only information about

the leading proprietary products and the entirety of the knowl-

works ID and title is returned.

edge graph and its source code are openly licensed and freely

The OpenAlex API uses paging – the list data are provided

available through data snapshots, an easy-to-use API, and a

by pages. The (up to 10 000 units) is based on basic paging

nascent user interface.

two parameters and ). The is a bit page per_page cursor paging

OpenAlex is based on 7 types of units (entities): (ork), (uthor), W A

more complicated than basic paging, but it allows us to access as

S(ource), I(nstitution), C(oncept), P(ublisher), or F(under) (and

many records as we like.

some additional ones such as topics, keywords, countries, con-

tinents, languages, etc.). Each unit gets its OpenAlex ID – we

assume that the identification problem is solved by the database. 4 A collection of bibliographic networks

The simplest use of OpenAlex is through its web interface We developed an R package to support the cre-OpenAlex2Pajek

(service) https://openalex.org/ or using a direct URL request in ation of bibliographic networks from OpenAlex [4]. We get a

the browser URL line. For example collection of bibliographic networks (citation network , au-Cite

thorship network , sources network , keywords network WA WJ

• Author’s name: search the OpenAlex web service

WK, countries network WC), some partitions and vectors (prop-

• Known author ID: URL https://openalex.org/A5001676164

erties of nodes) (publication year, type of publication, language of

• Work with DOI: URL https://api.openalex.org/works/

publication, cited by count, countries distinct count, referenced

https://doi.org/10.1007/s11192-012-0940-1

works, and additionally two files containing names of works

• Known work ID: URL https://openalex.org/W2083084326

xyzW.nam and names of authors xyzA.nam. Most acquired net-

• Name of the institution: search the Openalex Web service

works are 2-mode – they link units of two different types; an

• Known institution ID: URL

ordinary or 1-mode network links units of the same type.

https://openalex.org/institutions/I4210106342

Currently, contains three main functions OpenAlex2Pajek

This way, the OpenAlex web interface provides basic inspec- , , and . OpenAlex2PajekCite OpenAlex2PajekAll coAuthorship

tions of the selected unit. For example, by including a link with We split the process of creating the collection of bibliographic

our OpenAlex author ID on our web page we get a report on networks into two parts:



Higher-Order Bibliographic Services Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



• determining the set 𝑊 of relevant works using the satu- HM TF UM MH PN TO NR TV BJ AQ BV ZW LY SD TZ GH IQ AE SY GM BN QA GN PG FM VU TK WF NF KI GMKYNI NP PN NG SC TZ SL UG GD SY WF TK MY MU LK BD ZM SS BD QA WS AS CK NG BN DO NF UM AS LY ration approach GS VU NA BW [7, page 506], GG IE OM BS NA MP PW AO MT JM NU PT SS MH FK SZ KI TV NU PT IE MT GI GB MM GB ZA AR ZA • PY LI IM creation of the network collection for the works from 𝑊 . 1990 IM GI 1995

KE SZ BO

ZM LS EG

SB

CL CV

GL MZ CH KE

PW

ST SD GU

EC FJ GG

LR GL NR

FO BR BW AU DK EC LR ZW

VE CM MU

CO UGEG FO COVE FJ

RW CL CW

AR MW RW LS AU

DK UY GR ER MW

GE CM SB

BE GU SX BE GH

PE NZ AM PG

GY CY NZ ST

SK AT AT BR

CZ WS

CY

GR NC SI

PH PL FI CK UY

HR DE IL TG MK

NO IS VN

SG CX AF CZ TO PE

MK FI TW

HK DE VA FM

GW SI PL MC

MN TH

CN GY TL LV SG BT

RS MP GN

NL US TH GW PH

PK YE NL US PK

SL HU TW JO

CH HU SK

DM CW ME CG

SE KW EE SM LA VN CN

VA LA

KR LT JE YE

RO SE AM MN

MZ

GD IT SO MO

ET BH IT RO IL

BA IN JP AL PS ID

ER NO MY HK

TR AG NP

SA MV

The set 𝑊 is determined iteratively using the function BZ SO NIGT LB BT BHKW JO CU CU HT OpenAlex-LV ES IQ

ES KN IS IR KH KR AE CD

SV DO BG KG IN

PRVI CRBB MM IO PABMJMGT

ID PR

VG PY HT LK JP

PA RS GQ TR CR IR

CA

HR VI ET

AW KN OM

MC SV SA XK

LU BM VG LB RU

MS BA LU DM KP



2PajekCite FR and the collection is finally created using the func-LC KY TC MF UZ CF BSCABZ AL CI BQ GP KG CC MO AW VC AI SX MQ MX SR SN FR RU TM TJ DZ MX UZ KH AD AX GE BY CI PF BG AZ AG PM BL MV AF AN TN TJ SJ SN GF LT UA EE KZ HM BJ GF BY TM NC YT SH EH KM MR OpenAlex2PajekAll tion BO . MD SM LI ME CG DZ DJ GA HN NE CF SC KM EH CV AQ RE PF HN UA GS NE MQ MD KZ TF BV BI BF GP TC MS BL TG BQ TN JE AD AX SJ RE BI BF MA ML MG SR FK YT GQ SH MR AO CX CC TL TD MA ML MG DJ MF IO XK AI LC VC PM AN CD TD TT GA TT PS KP AZ BB

Pajek

The function creates a weighted temporal net-

coAuthorship Pajek

MU QA GQ

SOBWGMKYMS WF TV UM PN MU LK

BN SS SY

LSGMSOLA

TZ

ZMLY BD SZ MY SC

MW MY IE MV

TD OM NF KI MH GG BN GN

NAZW CG

work describing the co-authorship between world countries in GW

SS LC BT

AS SJ BH MK

MT AG

SZ

PT NR

GR JE JE PG

IE VU

MS

GG VC ZA TO CD

GI GB

TK IM

selected time intervals. The weight of an edge is the number of GI GB ZA VU ERRW LR SD KE ZMTZGH IM AO NA RW AZ 2000 MT MK SB 2010 NR

LS GL CO MW MZ EG

SK AO CK

SL

AU FJ

UG TO PG AR UG

NZ FK NG WS

TR NZ ST

GL AR PE CH IO

KE MGNG SD EG ER GH BO VE

MR FO CZ SB BW

RO SL LC AS NU

CO CM DK EC

EC AU FJ ZW MP PW

works co-authored by authors from the linked countries. FO CL CK PW EE AM CY GU BY NO SK MZ MP AM GE CY GU GN NU ST LI DE BE VEMH UM UY AW AT BG FI PK PH FM BE IS FM BH PY PE BR SX SI VN TW KZ MD DE BG FI PK UY CW VA ILMN NL TH CNMO GQ DK AT BO CL BR PY

In an analysis of weighted networks, the 1-neighbor skeleton US VN SG SE

LI SR IS SI JO

IL HU US AE PH BT

NL MN AX UA BD

HR KW HK

CW PL GE

AW CH HU CG MD IR IQ KR

VA TH TW MO GW KP

CN AF

SC LB

AE IT

HK AL NO LT YE

AX LV

SE JO KR SM PT

KZ LV KW PL PS ET

TL ES XK



is often used to get an overall insight into the network’s basic IT EE LT LK MC GR SV QA IN SM ES BA GT IN ET GD ID AD RO LU PR KN DM GD AI MX JM NI OM SY VG VC SA KN PS LB MC VG TT HT LU DM HN MX BM SA JP BZ CD AL IR WS CV GA VI JP DO CA LY NP MM PR CV KP MM GT PA ID CI IQ CU CF DO GA VI CR CU JM BB BS KH LA NP MG KY HN BS BM CC DJ TC RU TC CR BB TT SN structure. In the 1-neighbor skeleton, only its strongest link is PA RU ML DZ RS TR UZ CX UZ MA KG RS FR YT SH TN KG IO CX CI FR HR TJ NF KI BY CA BJ NI RE ME UA TM TJ SN DZ GF GY AZ TM TL LR KM EH BI NC FK CC AG NE PF BA ME TV PN TG NC WF NE kept for each node. The resulting directed network is forest-GS BV SR BF MA GY PF TG BQ AN CM YT AN ML SJ PM DJ XK CF MQ MF GP GF SV BJ BZ MF BQ BL PM TK TN KM MR RE BI BF TD MQ GP TF EH SH AQ TF BV GS HM AQ HM AD AI BL SX CZ MV HT KH

AF YE SG



Pajek

like. Non-trivial connected components in 1-neighbor skeletons Pajek NU

LY XK

SSSO IE MU

AZ

LY

BN GG GM LSZW NA ZW

IE MU BN GG GM LS

are (usually) directed trees with a pair of nodes linked in both IQ AXJEMT PN NA SZ NR MT MY ID JE MWBW FK SZ MS MK TL BT VU EG IQ GITLMK CY SB SD IMVU NEPG SD EG MY IM ER PG GR TD FKLA

CY MV WS TO

ZA ZA SA

SA SB



1-neighbor skeletons for years 1990, 1995, 2000, 2010, 2015, and directions with the largest weight in the tree – these two arcs GR TM GB TV GB FJ KE KI GW SK ET ZM NG RW CM AU SK ET KE KM FJ GW GL 2020 2015 LR SL HN NG MW CK BH CM AU AR AQ LR RW SL CK KI BH YE FO CO AR BW KM UG GL BO TZ TO NZ TV VE MZ SS SC AO FM PE AQ CZ GH FO CZ CO SO ZM UG TF TZ GH WS MZ NZ NR AO MP SJ DK are usually replaced by an edge (undirected link). In Figure 2 the LT GE CL PW DK CL GY PW ST GU CV CH GY GU PY ST LT SJ PE MP GE AS BR PT AT AM AS MH BR UY NO WF LU MH NO FM UY EE UM LI BE IS SY PH BV SG AT BE BG IS UM SR CW DE BG SY PH PY EE DE FI FI SI SR NL VN PK PK LI SI VN TW SG PL US CN MN IL AL TW MO CH PL US IL CN MN SM IT HR HU AE TH JO HK HR AE TH MO CG HU VA SE AX YE KW JO HK CG VA BD KR SE LV KW BT KR NL LV NP

AD ES BD JP KP AL IT UA KP

GI IN GD IR IN

BZ BZ IO LK

GD JP

IR LK ID SM

PT ES

2020 are presented. We see that the number of isolated nodes AG AF PS SV AZ PS

CV RO HT DM DM NI RO CDVG NI

CU KH MM CD

GT KN GT KH AF ER KN

VE VC PA MC BQSV QA OM PR

HT VG AI RS AD VCBQAITTPACW QA RS MM OM

PR TR

BO VI CA NP VI

TR LU CU MV CA

EC BL AW EC MC DO XK JM

AW DO

MD MD ME

MF LC

TC MF KY CR JM SX BB BS TT BM RU BA CR LCSXBBBSMX

MX TCKY BM RU BA

LB UZ

ME FR

(countries not collaborating with other countries) is decreasing. PM FR PM

TM GP LA GP

MQ GA MQ

CF DJ

HN LB KG

KG

AG GF UZ

TJ TJ

PF MS BY

DJ KZ BL WF

BY

NC UA AM

NC GN KZ

In all analyzed years the US has a leading (hub) position. In the GFCI ML PFCI SN MA DZ GN TF IO NF TKSN CF MA GQ NU EH CX TK BV CX DZ BI NE TG BF YT TN SC MR RE BI BF MG BJ EH GS TG SH TD AN YT TN BJ GA MR RE GQ CC HM NF PN GS SH AN CC HM MG ML

Pajek

years 1990, 1995, 2000, and 2010 the edge in the main component Pajek

links US and GB but in the years 2015 and 2020 GB is replaced by

CN. In 1990, stronger secondary hubs were GB, FR, RU, JP, and DE. Figure 2: 1-neighbors skeletons of world co-authorship for

In the following years, some other countries SE, ES, AU, CN, BR, selected years.

ZA, and IN (BRICS) became secondary hubs attracting previously

non collaborating countries or geographically or linguistically

in the bibliography of works with at least one co-author from

close countries.

University of Primorska.

Most of the ingredients of basic reports are counters, sorted Cite

In bibliometric analysis, the citation network has a very

lists, (weighted) degrees and their distributions obtained from

important role. It collects “votes” about the relevance of previous

an adequate network. Sometimes also the time is considered

works for a given work. It is often used for solving the network

producing time series.

boundary problem, and also for identifying the most relevant

An important property of a collection of bibliographic net-

works in the collected bibliography [2, 6]. The derived network

works is that some of them are compatible – they share a com- 𝑇 ACiA = WA · Cite · WA describes the citations between authors

mon set (most often the set of works W). This allows us to use [ ]

– its entry 𝐴𝐶𝑖𝐴 𝑎, 𝑏 counts the number of times author 𝑎 cited

network multiplication (defined by the product of network matri-

author 𝑏. The co-citation network is defined as the column pro-

ces) to compute the corresponding derived network connecting 𝑇 coCi = ( Ci ) = Ci · Ci jection of the citation network col and the

the remaining two sets [5]. For example, in the derived network

bibliographic coupling network is defined as the row projection

AK = WA𝑇 · WK its entry 𝐴𝐾 [𝑎, 𝑘] tells us in how many works of the citation network biCo = row(Ci) = Ci · Ci𝑇 . the author 𝑎 used the keyword 𝑘. Similarly, in the derived net-

The idea of derived networks can be extended to temporal

work its entry 𝐴𝐶𝑖 𝐾 𝑎, 𝑘 tells us how bibliographic networks [8]. Using derived networks we enlarge ACiK = WA 𝑇 · Cite · WK [ ] many times the author 𝑎 cited works described by the keyword

the source for different statistics. Additional insight can be gained

𝑘.

by analyzing the structure of networks and identifying important

A 2-mode network is always compatible with its transpose

subnetworks in them [6].

(on both sets). The corresponding derived networks are called

In the following, we present an overview of typical report

projections – the row projection row and the ( WA ) = WA · WA

𝑇

ingredients [7, 15]. Because of limited available space, we decided

column projection col . Both projections are ( WA ) = WA · WA

𝑇

to put examples on Github/bavla.

ordinary weighted 1-mode networks that can be analyzed using

standard network analysis methods. 5 Report ingredients



For the authorship network its column projection WA Co = 5.1 Statistics

WA𝑇 · WA is the co-authorship network. Its entry𝐶𝑜 [𝑎, 𝑏] counts

Because the analyzed networks are often large a complete pre-

the number of works that authors 𝑎 and 𝑏 co-authored. It turns

out that a work with 𝑘 2 sentation is not an option. To describe them we use different co-authors contributes 𝑘

links to the co-

statistical descriptors.

authorship network – works with a large number of co-authors

are overrepresented in it. To treat all authors equally the frac- sizes of sets (number of nodes, number of links); structural •

tional approach is used [3]. In Figure 1 the largest co-authorship network properties (number of components, size of the

groups at level 10 at the University of Primorska are presented – largest component, etc.)

connected components of the link cut at level 10 in the network top units – ordered lists of units with the largest values of •

Co. Each pair of linked authors co-authored at least 10 works selected property (degre, weighted degree, link weight,



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Batagelj, Pisanski & Pisanski



Color Key 6 Conclusions

Europe 2022 / Balassa / Ward

We have presented an approach to support higher-order biblio-

graphic services based on networks. Open access to high-quality

bibliographic data is crucial for the faster development of such

−6 −4 −2 0 2 4 6

services. The new bibliographic database OpenAlex seems to be

Value

GB a step in the right direction. It needs the support of science policy

LU

IT

IE

ES and also of individual scientists (checking the correctness of their

NL

BE

FR

AT

DE data).

CH

GI

VA

IM

GG

SJ Acknowledgements

AX

JE

PT

SE

FI

DK

NO The computational work reported in this paper was performed

LI

FO

IS

AM using a collection of R functions OpenAlex2Pajek and the pro-

GE

AZ

MD

BY

KZ gram Pajek for analysis of large networks. Code, data, and figures

RU

UA

BA

MK are available on Github/Bavla/OpenAlex.

ME

AL

XK

AD VB’s work is partly supported by the Slovenian Research

MC

SK

CZ

PL Agency ARIS (research program P1-0294, research program Cog-

TR

BG

RO

LV niCom (0013103) at the University of Primorska, and research

LT

EE

HU

RS

HR projects J1-2481, J5-2557, and J5-4596), and prepared within the

SI

CY

GR

MT framework of the COST action CA21163 (HiTEc). JP’s work is

SM

GB LU ES NL BE FR A DE CH GI V IM GG SJ AX JE PT SE DK NO FO AM GE AZ MD BY KZ R U BA MK ME AL XK AD MC SK CZ PL TR BG R LV LT EE HU RS HR CY GR MT SM partly supported by ARIS (research program P5-0361 and research

IT IE U T A FI LI IS A O SI

projects J1-2551 and J5-4596). TP’s work is partly supported by

ARIS (research program P1-0294 and research projects N1-0140,

Figure 3: Balassa EU co-authorship for the year 2022. J1-2481, J5-4596).



• References

distribution of selected property

[1] Massimo Aria, Trang Le, Corrado Cuccurullo, Alessandra Belfiore, and

• time series describing temporal changes of selected prop- June Choe. 2024. openalexR: an R-tool for collecting bibliometric data from

erties OpenAlex. , 15, 4, 167–180. The R Journal

• [2] Vladimir Batagelj. 2003. Efficient algorithms for citation network analysis.

scatter plots showing a possible relationship between two

arXiv preprint cs/0309023.

selected properties

[3] Vladimir Batagelj. 2020. On fractional approach to analysis of linked net-

works. , 123, 2, 621–633. doi: 10.1007/s11192- 020- 03383- y. Scientometrics

Often bibliometric properties of units follow laws such as Zipf

[4] Vladimir Batagelj. 2024. OpenAlex2Pajek. version 4, June 18. (2024). https:

(or power) law, Bradford law, Lotka law, lognormal distribution, //github.com/bavla/OpenAlex/tree/main/code. Hirsch index, etc. [5] Vladimir Batagelj and Monika Cerinšek. 2013. On bibliographic networks.

Scientometrics, 96, 3, 845–864. doi: 10.1007/s11192- 012- 0940- 1.

[6] Vladimir Batagelj, Patrick Doreian, Anuška Ferligoj, and Nataša Kejžar.

5.2 Network analysis 2014. Understanding Large Temporal Networks and Spatial Networks: Explo-

ration, Pattern Searching, Visualization and Network Evolution. Wiley Series

Derived networks are weighted. To get readable results of reason- in Computational and Quantitative Social Science. Wiley, Chichester. isbn:

able size we usually search for important subnetworks, often a 978-1-118-91537-0; 978-0-470-71452-2. doi: 10.1002/9781118915370.

[7] Vladimir Batagelj, Anuška Ferligoj, and Flaminio Squazzoni. 2017. The emer-

kind of skeleton – from a given network less important elements

gence of a field: a network analysis of research on peer review. , Scientometrics

are removed. There are different types of skeletons (spanning 113, 1, 503–532. doi: 10.1007/s11192- 017- 2522- 8.

forest, 𝑘 closest neighbors, cuts, cores, islands, etc. [6]). [8] Vladimir Batagelj and Daria Maltseva. 2020. Temporal bibliographic net-

works. , 14, 1, Article No. 101006. doi: {10.1016/j.joi.2020.101006}. J. Informetr.

A traditional graph-based visualization is used if the obtained

[9] Dalmeet Singh Chawla. 2022. Massive open index of scholarly papers launches.

result network is not dense. For denser networks, the matrix dis- . Nature

[10] DBLP – computer science bibliography. 2024. (2024). https://dblp.org/.

play is much more readable. In a matrix display, the permutation

[11] Lorena Delgado-Quirós and José Luis Ortega. 2024. Completeness degree of

of nodes (usually obtained by clustering) can create patterns that Quantitative

publication metadata in eight free-access scholarly databases.

reveal the network’s internal structure. , 5, 1, 31–49. Science Studies

[12] Scholar GPS. 2024. (2024). https://scholargps.com/.

Figure 3 presents a matrix display of Balassa co-authorship

[13] Chenyue Jiao, Kai Li, and Zhichao Fang. 2023. How are exclusively data

indices between European countries in 2022 (yellow cell – no

journals indexed in major scholarly databases? an examination of four

link, red/blue cell – above/below expectation) [17]. databases. Scientific Data, 10, 1, 737.

[14] Bill Jones. 2002. Computational geometry database. (2002). f tp://f tp.cs.usas

k.ca/pub/geometry/.

5.3 Special algorithms [15] Daria Maltseva and Vladimir Batagelj. 2019. Social network analysis as

a field of invasions: Bibliographic approach to study SNA development.

Some properties can require special computational procedures Scientometrics

, 121, 2, 1085–1128. doi: 10.1007/s11192- 019- 03193- x.

and direct access to the bibliographic data. In such cases, open [16] Peter V. Marsden. 1990. Network data and measurement. , Annu. Rev. Sociol.

16, 435–463. doi: 10.1146/annurev.so.16.080190.002251.

access to the bibliographic database is of crucial importance.

[17] Nataliya Matveeva, Vladimir Batagelj, and Anuška Ferligoj. 2023. Scien-

tific collaboration of post-soviet countries: the effects of different network

5.4 Reports normalizations. Scientometrics, 128, 8, 4219–4242.

[18] Jason Priem, Heather Piwowar, and Richard Orr. 2022. Openalex: a fully-

The results of analyses can be combined and presented to users open index of scholarly works, authors, venues, institutions, and concepts.

in different forms: . arXiv preprint arXiv:2205.01833

[19] Google Scholar. 2024. (2024). https://scholar.google.com/.

• Booklet report (in PDF). [20] Bert TePaske-King and Norman Richert. 2001. The identification of authors

• in the mathematical reviews database. (Service generated) web pages. Issues Sci. Technol. Librariansh., 31.

doi: 10.5062/f 4kh0k9m.

• Dashboards. [21] Wikipedia. 2024. API. August 22. (2024). https://en.wikipedia.org/wiki/API.

• Dataset ( JSON, CSV, etc.).



Are papers all that counts? A bibliometric analysis of the



Slovenian scientific community



Aymeric Dupuis Sašo Džeroski

Jožef Stefan Institute Jožef Stefan Institute

Ljubljana, Slovenia Ljubljana, Slovenia

aymeric.dupuis@etu.univ- nantes.f r saso.dzeroski@ijs.si



Boshko Koloski Matej Martinc

Jožef Stefan Institute Jožef Stefan Institute

Ljubljana, Slovenia Ljubljana, Slovenia

boshko.koloski@ijs.si matej.martinc@ijs.si



Abstract a discipline as a whole has. More specifically, our contributions

are the following:

We conduct a bibliometric analysis of the Slovenian science by

scraping the data from Slovenian current research information Using the collected data about the Slovenian scientists •

system (SICRIS) and using it to build a knowledge graph, repre- and their projects, covering different scientific fields and a

senting a network of all Slovenian scientific fields and a large large majority of researchers working in Slovenian science,

majority of Slovenian researchers. By analyzing this network us- we conduct a graph analysis of connections between dif-

ing different graph measures, we obtain valuable insights into the ferent fields and researchers. By drawing a comprehensive

connections between different scientific fields and researchers in map of connections between actors and fields, we iden-

Slovenian science. Additionally, we show the importance of graph tify the most important researchers and scientific fields

measures as measures of scientific excellence, since they measure that connect others and play a vital role in the Slovenian

very different aspects of scientific success than the traditional scientific ecosystem.

citation metrics. We created a new ranked list of Slovenian scientists ac- •

cording to graph based metrics, which were not available

Keywords in any of the previous analyses or databases. We argue

that these metrics measure the importance of a role that a

bibliometrics, Slovenian scientific community, knowledge graphs

specific scientist has in a research community, i.e., their



1 Introduction influence which allows them to act as a bridge or a hub

connecting scientists from different fields.

With the growth and diversification of the scientific enterprise,



obtaining empirical evidence on the research process is crucial for 2 Related work enhancing its efficiency and reliability. Meta-research and biblio-

Studies in bibliometrics (see [4] for a comprehensive survey of

metrics are developing scientific disciplines, seeking to analyse,

techniques used for measuring scientific excellence) have re-

evaluate and refine research practices, and several studies have

cently gained traction in parallel with the success of the scien-

focused on the analysis of the global scientific endeavour, e.g.,

tific enterprise, which has grown in both size and diversity, and

identifying most prominent scientists and fields [7]. These stud-

with the availability of data. According to Ioannidis et al. [7],

ies also focus on the problem of how to properly rank scientific

research on research is becoming important due to the mounting

excellence and scientific outputs in general, warning that one

evidence suggesting an alarming drop in reproducibility of re-

should not rely on just a few metrics to obtain a comprehensive

search findings, the growing inefficiency of the scientific process,

picture of the actual impact a specific scientist has [8].

and the fact that the number of false positives in the literature

Until now, very few studies have tackled the analysis of sci-

is exceedingly high. To address these problems, they propose

entific ventures at national level, and to our knowledge, there

a meta-research divided into five main categories that should

has been no study covering the Slovenian scientific landscape

be studied: methods, reporting, reproducibility, evaluation, and

specifically. This kind of research is nevertheless important and

incentives. Studying these five areas would correspondingly al-

could potentially influence policies that would improve scientific

low for five distinct insights into how to perform, communicate,

production and enable effective distribution of research funds

verify, evaluate, and reward research.

and resources.

Recently, several studies also tackled the problem of how to

In this study, we try to address the identified research gaps by

properly rank scientist and scientific outputs in general. For ex-

1.) drawing the map of Slovenian scientific research that would

ample, Ioannidis et al. [8] addressed the increasing prevalence of

enable proper decision making and policy formulation, and 2.)

multiauthorship observed in several fields and how this phenom-

proposing new metrics of scientific excellence that would allow

enon affects the effectiveness of the informativeness of citation

us to obtain a more complete view of the impact a scientist or

metrics. They also explored how sensitive the indicators are to

Permission to make digital or hard copies of all or part of this work for personal

self-citation and alphabetic ordering of authors. They concluded

or classroom use is granted without fee provided that copies are not made or

that multiple indicators should be used for ranking, as a com-

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this posite of different metrics gives a more comprehensive picture

work must be honored. For all other uses, contact the owner /author(s).

of the actual impact that a specific scientist has. They also ac-

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

knowledged that no single or composite citation indicator can

© 2024 Copyright held by the owner/author(s).

https://doi.org/https://doi.org/10.70314/is.2024.sikdd.11 be expected to select all the best scientists.



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Trovato et al.



Several studies employed graph-based metrics to enrich the as- publications in exceptional, high quality and important venues,

sessment of bibliometric analysis [4, 1]. Network metrics such as respectively. We also extracted the A1 metric, which represents a

degree of centrality, betweenness centrality, eigenvector central- weighted sum of these three metrics, a CI10 metric measuring the

ity, closeness centrality, and PageRank were used to pinpoint the number of pure citations of scientific work in the last 10 years,

relative importance of research constituents (i.e., researchers and the CImax metric measuring the number of citations in the most

institution), which may not necessarily be reflected just through cited work, and the h10 metric representing the h-index in the

publications. In a large majority of cases, these metrics were last ten years. Furthermore, we extracted the SICRIS points, a

calculated on co-authorship graphs. conglomerate metric combing several distinct metrics mentioned

The studies that would cover Slovenian scientific environment above, and the A3 metric, which measures the amount of funds a

are very scarce. In fact, we are aware of just one, the study by specific researcher received for his research activity outside of

[2], where they claim that research performance is highly de- the Slovenian National Research Agency (ARIS).

pendent on the conditions of (national) research environments. Finally, the SICRIS database also contains information on

They focus on analyzing research activity in six eastern European projects financed by the Slovenian national research agency in

countries, namely Croatia, Estonia, Hungary, Latvia, Lithuania, which a specific researcher participated. Scraping this informa-

and Slovenia, and try to determine and compare the effectiveness tion provided us with an important insight into collaborations

of research in a specific country by obtaining the number of between different scientists and fields, allowing us to build col-

articles belonging to the most cited 10% and the most cited 1% laboration graphs, calculate several graph-based ranking criteria

articles in the corresponding subject area and publication year and draw the map of the Slovenian scientific community.

for each country. Their empirical analysis addresses three levels:

cross-country, cross-institution, and cross-researcher compari- 3.2 Methods son. The study concludes that Hungary is the country with the

Once the data was obtained, we conduct two distinct analysis

highest output, followed by Croatia and then Slovenia, when it

steps, namely 1.) graph construction and analysis, and 2.) corre-

comes to the number of influential articles published.

lation analysis



3 Methodology 3.2.1 Graph construction and analysis. To construct the neces-

sary graphs, we used the Python NetworkX library [6]. Using

In this section, we describe our methodology, namely 1.) how we

the data from SICRIS, which contain information about project

gather the data and 2.) how we analyze these data to obtain a

collaboration, we created an undirected graph as follows: all re-

map of the Slovenian scientific community.

searchers who participated in at least one project are represented



3.1 by a node, and nodes of researchers who worked together on a Data Retrieval

project are connected by weighted edges, in which the weights

Data were retrieved from the Slovenian Current Research Infor-

represent the number of shared projects. By removing the iso-

1

mation System (SICRIS) website , which lists more than 35,000

lated nodes, we ended up with a graph with a total of 20,012

researchers working in Slovenian research institutions. Data col-

nodes and 618,871 edges.

lection from the SICRIS website proved challenging, as informa-

In the next step, we apply several graph statistics and mea-

tion about a specific researcher can only be obtained by scraping

sures in order to obtain several node rankings, each of them

his/her Web page on SICRIS. This required finding a solution to

measuring a different aspect of the importance a specific node

quickly retrieve data from more than 35,000 different pages, and

(i.e., a researcher) has in the graph. More specifically, we calculate

2 3

to achieve this, we used the Python Asyncio and BeautifulSoup

PageRank (PR), Betweenness centrality (BC), and Eigenvector

libraries, which allow the asynchronous connection to several

centrality (EC) measures.

dozen pages simultaneously and extraction of the required data. PageRank

In the context of our graph, the [3] algorithm is

Since the script sometimes took several seconds to connect

applied to evaluate the influence of researchers within the collab-

to a specific page, which could quickly accumulate, resulting in

oration network. Thus, researchers who are strongly connected

considerable overall slowdowns, we optimized the procedure and

to other researchers, who also have many connections (i.e, the so-

identified potential slowdowns. Our proposed solution was to

called hubs in the graph), will have a higher PR score, reflecting

implement a strategy that involved canceling the connection and

their importance and influence in the Slovenian research com-

adding the URL to a list whenever a page failed to connect within Betweenness Centrality

munity. On the other hand, the [5]

a 0.5-second time frame. This timeframe was chosen after several

measure evaluates the role of each researcher as an intermediary

trials and was found to be the best compromise. Once all pages

or a bridge between other researchers. This measure is based on

had been visited, we repeatedly tried to reconnect to the URLs

the idea that researchers who are on many collaboration paths

on this list until it was empty. This change significantly reduced

between other researchers are considered central and influential

the time required to retrieve all our data. Once all the data was

in the network. In our contexts, it helps to better understand

2

retrieved, we used the Pandas library for data manipulation,

the structure of the collaboration network among researchers.

which allowed us to export the results into Excel spreadsheets,

Researchers with high BC are those who play a crucial role in

appropriate for further processing.

creating links between different subgroups of researchers and in-

From SICRIS, we extracted research areas for each scientist

terdisciplinary connections. In practical terms, BC evaluates the

and various bibliometric indicators of their impact, namely A”,

number of times a researcher is traversed by the shortest paths

A’, A1/2, citation metrics based on a quantitative assessment of

connecting other researchers in the network. Thus, researchers

who are frequently used as pathways for collaboration among

1

https://cris.cobiss.net/ecris/si/en their peers obtain higher BC scores.

2

https://docs.python.org/3/library/asyncio.html

3

https://www.crummy.com/software/BeautifulSoup

2 1

https://pandas.pydata.org/ https://networkx.org/

A bibliometric analysis of the Slovenian scientific community Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



Another graph centrality measure that we applied to the cre- low according to SICRIS points (e.g., the best ranked researcher

ated graph is the [9]. This measure according to our novel three measures, Dr. Branimir Leskošek, is Eigenvector centrality evaluates the influence of a researcher taking into account both ranked as 5731th according to the SICRIS points). This finding

the quality and the quantity of connections. EC assigns more supports that the proposed new measures measure hypothesis 1 weight to connections that include influential researchers. Thus, different aspects of scientific excellence than the more established

a researcher connected to influential researchers will be assigned citation measures. Another important observation is that 7 out

a high score, reflecting potentially greater influence within the of 10 best ranked scientists appear to be active in two fields. This

network. This measure helps to detect researchers who, even might suggest that they are (or have been) involved in several

with fewer direct connections, occupy strategic positions in the interdisciplinary projects, which could have a positive influence

collaboration network. While this may seem similar to the PR on the newly proposed graph-based metrics.

algorithm, there are some differences. Unlike PR, which primarily In Figure 1, we present the heatmap of the correlations be-

focuses on the popularity of links, Eigenvector centrality also tween the different metrics extracted from SICRIS website and

takes into account the quality of connections. This means that the newly proposed graph-based metrics. We observe a strong

even if a researcher does not have a large number of direct con- correlation between PR and BC, 0.7, which might suggest that

nections, if they are connected to influential researchers, their researchers who collaborate with a wide range of colleagues from

Eigenvector centrality score can be high. In summary, while different fields are more likely to work with the most important

both measures aim to evaluate the influence of researchers in a ones.

network, they do so through slightly different approaches, thus



offering complementary perspectives for analyzing the structure

and importance of actors within the collaboration network.

Our second important area of focus in our research is the

collaboration between different fields. To build a graph that

would represent interdisciplinary collaboration between fields,

we grouped all researchers from the same field into a single node,

representing an entire field, i.e., we obtain a node for each scien-

tific field found on SICRIS. Similar to the previous graph, edges

and their weights represent collaborations on a project between

researchers in the linked fields.



3.2.2 Correlation analysis. In order to better understand the

metrics from SICRIS and to evaluate the relevance of our scores,

we deemed it pertinent to explore the correlation between all

our data. This analysis has two main purposes. First, we aim to

test the that the new graph ranking we presented, hypothesis 1 Figure 1: Heatmap of the Spearman correlation among

measure different aspects of scientific excellence than the more metrics.

established measures based on number of citations or publica-

tions available on the SICRIS web page. This hypothesis would We also observe very strong correlations in the top left corner

be deemed correct if one-on-one correlations scores between the

of the heatmap. While a strong correlation was expected, as A”,

newly proposed graph measures and other measures would be

A’, A1/2 and A1 are all scores based on the number of publications

low, and incorrect if correlations would be high.

(in venues of different qualities), the almost perfec correlation

Additionally, we wish to explore the correlation between the

between the SICRIS points and A1 (which suggest they measure

established measures available on the SICRIS web page. More exactly the same aspect of the scientific impact) is surprising. This

specifically, we wish to test the that these measures hypothesis 2 hypothesis 2 finding supports that the current SICRIS measures

are strongly correlated, which would indicate that they essentially all measure a very similar aspect of scientific excellence. On the

all measure a very similar aspect of scientific excellence, which is

other hand, there is no strong correlation between any of the

problematic. In order to obtain one-on-one correlations between

newly proposed graph-based metrics and metrics extracted from

all measures, we calculate the Spearman correlation coefficient

the SICRIS website.

among all of them and then display it through a heatmap.

In Table 2, we present the results of our study of interdis-

ciplinary collaboration between different scientific fields. The

4 Results graph metrics were obtained from a graph of nodes representing

fields and edges representing interdisciplinary project collabo-

In Table 1, we present some of the results of the graph analy-

rations. Note that the field of Computer science and informatics

sis conducted on the graph of nodes representing researchers,

ranks first according to all the criteria. On the other hand, most

connected by edges representing project collaborations. More

interdisciplinary collaborations are conducted by the researchers

specifically, we present 10 best ranked researchers in the SICRIS

from the field of Chemistry, which ranked as third according

dataset according to the average between ranks of the three newly

proposed graph-based measures, their declared scientific fields, to the average (AVG) between the ranks of three graph-based

metrics, PG, BC and EV.

and their ranking (i.e., lower is better) according to the SICRIS

points, BC, EC and PR measures.

Note that while the table does contain some highly ranked 5 Conclusions

researchers according to the SICRIS points (e.g., Dr. Sašo Džeroski The graph based bibliometric analysis of the Slovenian scientific

is ranked as 33rd out of roughly 20K researchers according to community shows that current citations based metrics do not

this criteria), several researchers in the table are ranked relatively cover some aspects of scientific excellence, such as researcher’s



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Trovato et al.



Table 1: 10 best ranked researchers in the SICRIS dataset according to the average between ranks of the three newly

proposed measures, BC, EC and PR. We do not show metric scores, but ranks according to scores (i.e., lower value is better).

Researcher Field 1 Field 2 SICRIS points BC EC PR AVG

15355 PhD Branimir Leskosek Public health (occupational safety) Computer science and informatics 5731 8 4 31 14

06013 PhD Damjana Rozman Biochemistry and molecular biology Metabolic and hormonal disorders 704 21 2 33 18

11279 PhD Nives Ogrinc Control and care of the environment Animal production 182 7 50 3 20

27733 PhD Tina Kosjek Control and care of the environment Pharmacy 809 2 73 9 28

22459 PhD Tadeja Rezen Neurobiology Microbiology and immunology 1837 61 3 49 37

22621 PhD Polonca Ferk Metabolic and hormonal disorders Pharmacy 5059 13 8 103 41

12688 PhD Kristina Gruden Biotechnology / 219 44 139 6 63

08800 PhD Gregor Sersa Oncology / 71 3 185 1 63

12315 PhD Ester Heath Control and care of the environment Chemistry 208 62 115 23 66

11130 PhD Šašo Dzeroski Computer science and informatics / 33 1 195 20 72



Table 2: Scientific fields as defined in the SICRIS database, sorted according to the average (AVG) between the ranks (lower

score is better) of three graph-based metrics, PG, BC and EV.

Rank Field Collaborations PR EC BC AVG Rank Field Collaborations PR EC BC AVG

1 Computer science and informatics 81248 1 1 1 1.0 36 Textile and leather 21080 27 41 39 35.67

2 Materials science and technology 88934 4 3 4 3.67 37 Animal production 34982 29 29 50 36.0

3 Chemistry 101139 2 2 12 5.33 38 Political science 13598 46 37 27 36.67

4 Control and care of the environment 52648 5 8 9 7.33 39 Anthropology 9860 53 36 24 37.67

5 Physics 50010 3 9 14 8.67 40 Ethnology 6698 65 39 11 38.33

6 Plant production 74535 6 6 16 9.33 41 Cardiovascular system 20793 28 43 45 38.67

7 Systems and cybernetics 45584 7 10 23 13.33 42 Telecommunications 14068 41 45 31 39.0

8 Biology 58879 12 7 21 13.33 43 Veterinarian medicine 30954 32 34 60 42.0

9 Civil engineering 36466 22 13 6 13.67 44 Metabolic and hormonal disorders 18518 30 46 55 43.67

10 Biochemistry and molecular biology 79725 11 5 25 13.67 45 Metrology 12978 34 52 47 44.33

11 Neurobiology 45680 14 12 19 15.0 46 Law 7480 54 49 32 45.0

12 Biotechnology 87261 8 4 33 15.0 47 Psychology 8583 51 55 29 45.0

13 Interdisciplinary research 22946 9 33 5 15.67 48 Human reproduction 21535 35 42 58 45.0

14 Public health (occupational safety) 30400 10 25 13 16.0 49 Process engineering 15340 36 47 53 45.33

15 Educational studies 23518 33 15 3 17.0 50 Hydrology 12396 40 53 44 45.67

16 Mathematics 30680 17 20 20 19.0 51 Architecture and Design 4242 58 57 22 45.67

17 Manufacturing technologies and systems 38874 18 14 26 19.33 52 Philosophy 7380 57 44 43 48.0

18 Forestry, wood and paper technology 30620 19 28 15 20.67 53 Sport 10013 43 54 49 48.67

19 Geography 18555 39 23 2 21.33 54 Geodesy 7760 45 56 51 50.67

20 Economics 26891 31 16 18 21.67 55 Electric devices 13633 42 51 59 50.67

21 Microbiology and immunology 54175 16 11 42 23.0 56 Literary sciences 6399 61 50 48 53.0

22 Sociology 19922 44 17 10 23.67 57 Traffic systems 4448 48 60 52 53.33

23 Pharmacy 41125 15 18 41 24.67 58 Culturology 7240 60 48 54 54.0

24 Linguistics 18176 49 19 7 25.0 59 Technology driven physics 6876 47 59 64 56.67

25 Chemical engineering 33753 13 27 38 26.0 60 Communications technology 4388 52 63 56 57.0

26 Energy engineering 32762 23 21 40 28.0 61 Psychiatry 2481 55 65 61 60.33

27 Computer intensive methods and applications 26942 20 32 34 28.67 62 Criminology and social work 2324 66 62 62 63.33 28 Mechanics 26444 24 31 36 30.33 63 Mining and geotechnology 2342 59 68 63 63.33

29 Oncology 37101 21 24 46 30.33 64 Theology 2941 67 58 66 63.67

30 Geology 26961 37 26 28 30.33 65 Ethnic studies 2398 63 61 67 63.67

31 Electronic components and technologies 28858 26 30 37 31.0 66 Art history 1408 70 64 57 63.67

32 Historiography 12390 56 22 17 31.67 67 Archaeology 1177 68 66 65 66.33

33 Urbanism 8669 50 40 8 32.67 68 Information science and librarianship 792 62 70 70 67.33

34 Mechanical design 22352 25 38 35 32.67 69 Stomatology 391 64 71 68 67.67

35 Administrative and organisational sciences 18563 38 35 30 34.33 70 Landscape design 1046 69 67 71 69.0

71 Musicology 748 71 69 69 69.67



role of connecting a wider research community. Our correlation [3] Sergey Brin and Lawrence Page. 1998. The anatomy of a large-scale hyper-

textual web search engine. , 30, 1-7, Computer networks and ISDN systems

analysis indicates that existing measures of scientific excellence

107–117.

extracted from the SICRIS web page are strongly correlated. In

[4] Naveen Donthu, Satish Kumar, Debmalya Mukherjee, Nitesh Pandey, and

the future, we plan to expand this analysis to also measure the Weng Marc Lim. 2021. How to conduct a bibliometric analysis: an overview

and guidelines. , 133, 285–296. Journal of business research

impact of Slovenian scientists on the global scientific enterprise

[5] Linton C. Freeman. 1977. A set of measures of centrality based on between-

and conduct additional research to try to find certain patterns ness. Sociometry, 40, 1, 35–41. Retrieved June 27, 2024 from http://www.jstor

across disciplines, or institutions. .org/stable/3033543.

[6] Aric Hagberg, Pieter J Swart, and Daniel A Schult. 2008. Exploring network



6 structure, dynamics, and function using networkx. In Proceedings of the 7th Acknowledgments

Python in Science Conference (SciPy2008). Los Alamos National Laboratory

(LANL), Los Alamos, NM (United States), 11–15.

The authors acknowledge the financial support from the Slove-

[7] John PA Ioannidis, Daniele Fanelli, Debbie Drake Dunne, and Steven N

nian Research Agency for research core funding for the pro- Goodman. 2015. Meta-research: evaluation and improvement of research

methods and practices. , 13, 10, e1002264. PLoS biology

grammes Knowledge Technologies (No. P2-0103).

[8] John PA Ioannidis, Richard Klavans, and Kevin W Boyack. 2016. Multiple

citation indicators and their composite across scientific disciplines. PLoS

References biology, 14, 7, e1002501.

[9] Paul Turán, editor. 1969. . Publications of edmund landau Number Theory and

[1] Njål Andersen. 2021. Mapping the expatriate literature: a bibliometric review

of the field from 1998 to 2017 and identification of current research fronts. Analysis: A Collection of Papers in Honor of Edmund Landau (1877–1938)

.

The International Journal of Human Resource Management Springer US, Boston, MA, 335–355. isbn: 978-1-4615-4819-5. doi: 10.1007/978

, 32, 22, 4687–4724.

- 1- 4615- 4819- 5_23.

[2] Lutz Bornmann. [n. d.] Research excellence in eastern europe: a bibliometric

study focusing on croatia, estonia, hungary, latvia, lithuania, and slovenia.





Empowering Open Education Methodologies with AI-based




Strategies for the Customization of Education



Tel Amiel Mitja Jermol, Anja Antônio J. Moraes Neto Joao Pita Costa

Universidade de Brasilia Instituto Federal de Brasilia IRCAI, Jozef Stefan Institute Poljanar

Brasilia, Brazil Brasilia, Brazil Ljubljana, Slovenia IRCAI, Jozef Stefan Institute

amiel@unb.br antonio.neto@ifb.edu.br joao.pitacosta@quintelligence.com Ljubljana, Slovenia



ABSTRACT



The amount and heterogeneity of data generated in the context 1 Introduction



of education allied to the rapid progress of scientific research The centralizing piece of the discussions in this paper is an AI- and technological development have created vast amounts of based observatory that allows to explore OER-related topics, data, much of it open data, but significant challenges to particularly those mentioned in the OER Recommendation: gathering, filtering and making sense of this information. In this promoting OER and acknowledging it’s contribution to paper, we discuss the research outcomes of complementary advancing quality education while providing information on Artificial Intelligence (AI)-based strategies monitoring and advances focused on the equity and inclusion qualities of OER, enhancing Open Education, mining online forum interaction as well as on research, activities, projects and news related to student-educator, and empowering mentorship of educators. OER development, including new initiatives and projects while Firstly, the initial results obtained from the construction of an also promoting public infrastructures for education. The OER Observatory focusing Open Education Resources (OERs), Observatory builds on the content made available in UNESCO’s contribute to implement 2019 UNESCO OER Recommendation OER Dynamic Coalition Portal ( oerdynamiccoalition.org ) providing and advance the Education-focused Sustainable Development the user with access to any of the four proposed views: media; Goal (SDG) 4. It is acting on five verticals, enriching and treating science; policies and training. In each of the views, the user can multilingual data, it displays meaningful information on a access interactive data visualisation summarising the sourced dashboard focused on AI and OERs and serving as a data configured to observe the UNESCO OER collaboration platform focused on existing partnerships within recommendations. As it is fully based on open data, it allows the the international research centre on AI under the auspices of user to click on the resources collected and summarized, being UNESCO (IRCAI), the UNESCO Chair in Distance Education and taken directly to the source in media, journal, policy or training. the UNESCO Chair on Open Technologies for Open Educational Embracing the intersection of AI and education, which has Resources and Open Learning, mobilizing research led to the development of various tools that personalize and collaboration on key AI research challenges relating to enhance learning experiences, we discuss a complementary generating knowledge about OER. Secondly, we will discuss the research based on CA much aligned with the objective of recent development of an Educational Recommender System empowering Community interaction at the SDG 4 (Education) (ERS) that integrates Conversational Analysis (CA) to assess Observatory [6]. AI applications in education often focus on and enhance collaborative learning (CL) in Virtual Learning providing adaptive feedback, facilitating personalized learning Environments (VLEs). This novel system was designed to paths, and analyzing student data to improve outcomes. CA is a identify collaboration among students and provide tailored method that examines the understanding generated through recommendations to promote participation and interaction interactions, offering a framework for analyzing how students within discussion forums. Finally, we will discuss the collaboratively build knowledge. By combining CA with AI, this development and implementation of AI and OERs in alignment research aims to develop a system that not only assesses but with SDGs, addressing topics of significant social impact over also actively promotes collaboration in VLEs [10]. The ERS an international online mentoring initiative. discussed later in this paper, is an example of how IRCAI’s SDG4



KEYWORDS Observatory gains a complex capability towards the

Open Education, Machine Learning, Educational Recommender engagement with communities such as in Education. This

System, Conversational Analysis, Virtual Learning Environment discussion then expands towards the appropriate mentorship

of the professionals that will change the domain’s landscape.

Permission to make digital or hard copies of part or all of this work for personal While initiatives in this context are diverse and disperse, the

or classroom use is granted without fee provided that copies are not made or authors are not aware of existing similar approaches [5].

distributed for profit or commercial advantage and that copies bear this notice

and the full citation on the first page. Copyrights for third-party components of

this work must be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.sikdd.16



217

2 AI-based strategies for the moderation of across five experimental cycles in different classes at two



online forums on education Brazilian Federal Institutes , in a Portuguese language context. The results indicated a positive impact on student learning,

Entering the age of Big Data, AI is feeding the data-driven digital with 82% of participants acknowledging the relevance of the

transformation across industries including Education. CL recommendations. The system motivated increased

emphasizes the importance of group tasks and joint participation and collaboration, with a notable trend of

participation, wherein students learn by actively engaging in students writing more and systematically organizing their



Even in remote settings, CL enables students to learn together other activities proposed by their teachers, demonstrating the effectiveness of the recommendations. The results also dialogues that facilitate the sharing of ideas and information. ideas in forum posts. Additionally, 90% of students engaged in

through virtual platforms. AI offers new opportunities as a

demonstrate the system's effectiveness in fostering

pedagogical tool, providing adaptive and personalized



environments that can support CL. This research explores the educators. A dashboard was developed for teachers, containing collaboration, with positive feedback from students and

integration of AI into educational contexts, particularly through graphs including one that shows the main terms discussed in

the development of an Educational Recommender System the forum by analysis, in which each edge represents a message

(ERS) that uses CA to identify and promote collaboration from the student with two of these terms, and the nodes in blue

among students in VLEs [1] (see Figure 1). highlight the new terms that emerged in relation to the

previous analysis (see Figure 2).





Figure 2: Visual analysis of students’ collaboration in a discussion



forum where nodes represent actors in the discussion

(students/educators) and edges represent interactions.

Figure 1: The ERS forum analysis screen



The research methodology is divided into three key stages: The development of the ERS represents a significant

Conversational Analysis, applying CA to monitor discussion advancement in promoting collaborative learning in

forums within the Moodle platform, focusing on interactions educational settings [6,7]. By integrating CA into the system,

among students, identifying collaborative behaviors and the ERS effectively identifies and enhances collaboration



level of collaboration among students based on identified to provide personalized recommendations to students, interaction patterns; and Development of ERS , building a teachers, and tutors, fostering a more interactive and interaction patterns; Collaboration Assessment, evaluating the among students. The current implementation of this ERS aims



teachers, and tutors. These recommendations are aimed at collaborative learning environment [6]. Future work will mechanism that provides recommendations to students,

enhancing collaboration and are based on the analysis of forum explore the integration of additional features, such as

interactions [15]. The initial dataset comprises 20,976 wikification and visualization tools, to further enhance the

messages of Moodle discussion forums, with 15,703 posted by system's capabilities. Furthermore, the research will benefit

students from a vocational education school. The analysis from the semi-automatic categorization of educational

focuses on these messages to develop and validate the ERS's resources of a range of formats, including videos as in [3].

recommendations. The quality of collaboration is measured

through various indicators, which are extracted during 3 An AI-based Observatory to Assess the

different stages of CA. Preprocessing applies techniques of Impact of OER Worldwide



the analysis, preparing data for the Although the abundance of information available online, some Natural Language Processing (NLP) to ensure the accuracy of

Resource Processing stage

using Social Network Analysis (SNA) to characterize social of which is labeled as education-related, it is harder and harder

dynamics and interactions among students. Moreover, the to find the appropriate resources that can serve education

Message Attribute Identification either at an undergraduate or a professional training level. is the CA stage that allows

identifying characteristics of students’ messages, , specifically IRCAI’s Open Education Observatory is an initiative dedicated

their questions, and then Topic Modeling is employed to to monitoring, analyzing, and promoting the use of OERs

identify key terms discussed in the forums [12], using globally. It serves as a hub for research insight and fomenting

Tomotopy library (bab2min.github.io/tomotopy) The ERS was tested collaboration, providing valuable insights and data on the



218

adoption, impact, and trends of OER in education systems leveraging equitable access provided by OER, sustainability

worldwide. The observatory supports educators, models, or international cooperation.

policymakers, and institutions in leveraging open resources to

enhance teaching and learning. It is designed to support

government and institutional decision-makers dedicated to

promoting the goals of the 2019 UNESCO OER

Recommendation, which is centred on OER but generally

promotes the ideals of Open Education (see Figure 3).





Figure 4: The architecture of the OER Observatory as an



Elasticsearch-based system that enables the visualization of

heterogeneous data on OERs



For each area, users can filter and find content specific to their

Figure 3: domain of interest: up-to-date news and research on OER Dashboard of visual modules to analyse the most

relevant topics under a certain domain or SDG, and the trends developments, academic studies related to professional

that can direct the education actors preparedness development, and relevant lectures for capacity building;

information on OER policy development; resources and

The Open Education Observatory ingests a range of different research focused on effective, inclusive, and equitable access to

data sources with heterogeneous nature and different quality OER; strategies for developing sustainable OER models;

frequency: (i) worldwide news in almost real-time providing and opportunities for fostering international cooperation

information from a vast catalogue of multilingual world news, through potential new partnerships and shared goals. This

captured in more than 60 languages and based on a variety of organized approach enhances the ability to pinpoint and utilize

wikidata concepts; (ii) published scientific articles, including the most relevant information in each domain. Information

journal and conference papers, mostly peer-reviewed, covering generated by the Observatory can be used to aid in the

over more than 126 million articles with yearly updates; (iii) resolution of problems related to the promotion of OER, by

OER policies from the OER Policy Hub (www.oepolicyhub.org) identifying trends and major areas of discussion, and to explore

that needs to be input into the OER DC Portal; subsequent successful scenarios through similar challenges and cases. The

extraction and enrichment of metadata; preparation of Observatory provide benefits to a range of stakeholders

dashboard related to dashboard based on filters over the including: national governments, providing access to a variety

metadata, as well as OECD policies data and metadata on AI and of perspectives on OER trends for decision-making;

Education with yearly updates; (iv) lectures and videos educational and research institutions, facilitating the access to

selected and filtered on content from Videolectures.net [10] resources and data; civil society, allowing access to information

resources related to OER; (v) a snippet of worldwide public and and training materials that explore the knowledge available

private initiatives related to AI and SDG 4 captured by IRCAI’s towards the implementation of the UNESCO recommendations;

Top100 and related actions; and (iv) a range of worldwide and the general population, empowering open education.

indices with yearly updates on Education-related topics such as



the percentage of children out of school, or the literacy rate in 4 Open Education for a Better World youth and adults (see Figure 4).



materials from the mentioned sources are categorized by international online mentoring initiative aimed at advancing the development and implementation of open educational relevant keywords and concepts closely associated with the To ensure that content is readily available for each focus area, The Open Education for a Better World (OE4BW) program is an



five key areas of the Recommendation. This organization in alignment with the United Nations Sustainable Development resources (OER) that address topics of significant social impact,

allows users to easily filter and access content based on their Goals (SDG) [2,14]. As part of the Slo2Svet project, the program

specific interests within these areas. By doing so, users can received 70 project applications and 87 mentor applications

tailor their exploration of resources to match their focus, from six continents and 25 different countries (see Figure 5).

whether it's capacity building, supportive policy development, The program's activities are structured into thematic clusters,

focusing on areas such as Artificial Intelligence, Displaced



219

Persons, Sustainability, Health and Well-being, Renewable insights provided by automatic text analysis and other AI tools.

Energy, Education, and Youth (specifically targeting developers This will allow us to connect the projects produced by OE4BW

aged 12-24). Throughout the project development process, to the concrete objectives of the Recommendation, providing

progress was closely monitored by a network of mentors and examples of practice that can be leveraged to advance its goals.

hub coordinators, providing essential guidance and support to



OER developers. Additionally, within the scope of the Slo2Svet ACKNOWLEDGMENTS project, evaluation rubrics for the OER projects were

developed and will be utilized during the final conference, We thank the support of the Slovenian Research Agency (ARIS)

where developers will present their completed work. and Ministry of Foreign and European Affairs (MZEZ) on the



project Slo2Svet - Connecting cultures, informing and learning

through Open Educational Resources and AI (V2-2363).



REFERENCES

[1] Ahmadian Yazdi, H., Seyyed Mahdavi Chabok, S. J., and Kheirabadi, M.

(2022). Dynamic Educational Recommender System Based on Improved

Recurrent Neural Networks Using Attention Technique. Applied Artificial

Intelligence, 36(1), 2005298.

[2] Drevensek, M., and Urbancic, T. (2022). The Role of Teamwork in the

Creation of Open Educational Resources for Closing SDG-Related

Knowledge Gaps. Open Praxis, 14(2).

Figure 5: [3] M. Grcar, D. Mladenic., and P. Kese (2009). Semi-automatic categorization Participants of the OE4BW mentorship in 2023/24.

of videos on videolectures. net. In Proceedings, of the Machine Learning

and Knowledge Discovery in Databases: European Conference, ECML PKDD



5 2009, Bled, Slovenia, September 7-11, 2009, Springer, pp. 730-733. Conclusions and further work

[4] Koschmann, T. (2013). Conversation Analysis and Collaborative Learning.

In this paper we discussed the research results and In C. Hmelo-Silver, C. Chinn, C. Chan, & A. O’Donnell (Eds.), The Intern.

Handbook of Collaborative Learning. Routledge Handbooks, pp. 149–167.

opportunities in Open Education, building on an overall [5] Liu, Q., Huang, J., Wu, L., Zhu, K., and Ba, S. (2019). CBET: Design and

perspective over the OER landscape, the AI-enhanced student- evaluation of a domain-specific chatbot for mobile learning. Universal

educator interaction, and the mentorship for further progress. [6] Moraes Neto, A. J., and Fernandes, M. A. (2019). Chatbot and Conversational Access in the Information Society.



We will be exploring further the potential of the OER Analysis to Promote Collaborative Learning in Distance Education. 2019 IEEE 19th International Conference on Advanced Learning Technologies observatory, particularly in what regards the appropriate use (ICALT), 2161-377X, pp. 324 – 326.

of LLMs in analyzing the compliance to AI policies in Education. [7] Moraes Neto, A. J., Fernandes, M. A. & Amiel, T. (2022). Conversational

In what regards the future developments of the EduColab, in Analysis to Recommend Collaborative Learning in Distance Education. 14th

Intern. Conference on Computer Supported Education, pp. 196–203.

alignment with IRCAI’s SDG 4 Observatory and the [8] Moraes Neto A. (2024) Sistema de Recomendação Educacional para

Videolectures.net Diagnosticar e Promover a Colaboração em Ambientes Virtuais de research agenda and the potential for

institutional collaboration, we will focus on: (i) the appropriate Aprendizagem. Doctoral thesis. Federal University of Uberlandia.

[9] Novak E., Novalija I. (2016) Visual and Statistical Analysis of

wikification, incorporating suggestions of Wikipedia concepts VideoLectures.NET, Proceedings of the SIKDD’16.

identified by [10] Urbančič, T., Polajnar A., and Jermol, M. 2019. (2019). Open education for Wikifier and related to the main discussion topics;

a better world: a mentoring programme fostering design and reuse of open

(ii) integration of interactive data visualization presenting educational resources for sustainable development goals. Open Praxis.

graphical representations of collaboration trajectories, topic Open praxis. ISSN 1369-9997, Vol. 11, no. 4; pp. 1-18.

evolution, and other key indicators; (iii) extending the system, [11] Urbančič, T., Polajnar, A., & Jermol, M. (2019). Open Education for a Better

World: A Mentoring Programme Fostering Design and Reuse of Open



applying the ERS to other datasets, including public and private Educational Resources for Sustainable Develop. Goals. Open Praxis, 11(4). [12] Urbančič, Tanja, et al. (2023) Developing supportive policies and strategies message exchange logs, to validate and enhance its for their implementation: student experience with real-world cases. Open

applicability; and (iv) personalized recommendations, Educational Resources in Higher Education: A Global Perspective.

developing a user-based collaborative filtering technique to Singapore: Springer Nature Singapore, 2023. pp. 35-53.

[13] Uthus, D. C., & Aha, D. W. (2013). Multiparticipant chat analysis: A survey.

tailor recommendations more specifically to individual student Artificial Intelligence, 199–200, 106–121.



groups. Moreover, we will explore together the pathways of AI- [14] Vayansky, I., & Kumar, S. A. P. (2020). A review of topic modeling methods. Information Systems, 94. based citizen science in the context of Open Education and how [15] Zawacki-Richter, O., Marín, V. I., Bond, M., & Gouverneur, F. (2019).

it can be integrated in the wider scope of the SDG4 Observatory. Systematic review of research on artificial intelligence applications in

In the context of the Slo2Svet project, we are conducting a higher education – where are the educators? International Journal of

Educational Technology in Higher Education, 16(1), 39.

comprehensive analysis of the Open Education for a Better

World (OE4BW) mentoring program since its inception,

examining outcomes and connections to other initiatives [see

for example, 12]. Additionally, we will develop an evaluation

framework to assess the impact of the projects produced

through the program, mapping project outputs to the five

action areas of the 2019 UNESCO OER Recommendation, using



220





Addressing Water Sustainability Challenges in North Africa with




Artificial Intelligence



Mustafa Zaouini, Maurizio Joao Pita Costa*, Davor Manal Cherkaoui, Anas Ait Hanaa Hachimi, Y. Kaddouri, I.

Santamicone, Lee Chana Orlic, Mihajela Črnko Aomar, Ikram Chairi, Lirmaqui, A. H. Alaoui, O.

AI in Africa IRCAI, Quintelligence Karima Echihabi Ignammas, H. Rahhou

Johannesburg, South Africa Ljubljana, Slovenia UM6P Ibn Tofail University

mus@fliptin.com joao.pitacosta@quintelligence.co Ben Guerir, Morocco Kénitra, Morocco

m candia@usp.br



M. Wahib Abkari, R. Rachidi, K. Gourari, I. Annaki, B. Jearani, J. T. El Azzoiani, M. Ait Essibaa,

W. Laaleg, Z. Hidila, M. Tabaa S. Trabi, T. Zennouhi, M. Sbaa A. Hamidine, H. Lachheb

Moroccan School of Engineering UMP Univesity Al Akhawayn University

Sciences, Casablanca, Morocco Oujda, Morocco Ifrane, Morocco



ABSTRACT AI Everything section of the GITEX Africa in the end of May

2024. It was mostly directed to PhD/MsC students and young



governments towards evidence-based decision-making in text mining on social media, gamification strategies optimizing regards to water resources. In this particular domain, the water consumption, and application of LLM frameworks for amount and heterogeneity of data generated allied to the rapid RAG and AI Agents in the context of water sustainability), progress of scientific research and technological development engaging experts from global agencies like, e.g., UNESCO, AI have created vast amounts of data, but significant challenges to Movement, and UNESCO’s Water Education Institute , as well as gathering, filtering, and making sense of this information. This national companies, research institutions and government. The worldwide where Artificial Intelligence (AI) can position their communities, exploring a wide range of machine learning research institutions, public & private companies and methodologies (from image recognition on satellite imagery, to The topic of water sustainability has been leading the priorities entrepreneurs working on AI to solve problems for the good of



engaging a total 51 students mentored by 15 professors across one of the MENA priorities, takes into consideration the UN 11 research institutions in North Africa, distributed by 14 Water Program for 2024-25 [12], and follows the work done by paper presents the research outcomes of collaborative effort global challenge of this action, "Water, AI and Sustainability" is

selected projects focusing the appropriate application of IRCAI with the European Commission (EC) on the NAIADES

machine learning methods to local and national water Water Observatory [9], as well as the recently opened new



youth challenge co-organized during May 2024 between AI focusing on the impact of AI in SDG 6 [11]. This work aligns with Africa and IRCAI with the support of GITEX. UNESCO’s interests in taking action to capacitate the Youth sustainability problems. These outcomes were motivated by a IRCAI Committee on AI and Water Resource Management [4]



KEYWORDS towards AI, with focus on the recent activities based from

Morocco but with a global scope, including the opening of the

Machine learning, text mining, large language models, community new UNESCO AI Centre, the AI Movement (aim.um6p.ma). engagement, water sustainability, competition





1 Introduction

Building upon common interests, exciting initiatives and

existing projects developed by IRCAI and AI in Africa

(aiinafrica.org), focused on AI and Sustainability, this activity

aimed to build capacity within African youth to advance the

Sustainable Development Goals (SDGs) though AI on challenges

within their own communities and in the region. The AI Youth

Challenge originated in the context of discussions started in

GITEX Dubai in 2023 and forwarded to a concrete event in the

Figure 1: Winner of the AI4Water challenge, designed and

Permission to make digital or hard copies of part or all of this work for personal developed by UM6P students, exposing a water map that

or classroom use is granted without fee provided that copies are not made or pinpoints remote villages with assigned water scores based on



distributed for profit or commercial advantage and that copies bear this notice satellite imagery and crowdsourced data. and the full citation on the first page. Copyrights for third-party components of this work must be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.sikdd.17



221

2 Finalist innovative ideas on water lies in its data generation and refinement approach. It creates

sustainability datasets in areas with data scarcity, starting with an automated



Attracting the participation of more than 50 PhD and MsC user-generated content. This closed-loop system employs baseline from satellite imagery and then enriching it through



students across 20 teams based in research institutions in active learning, progressively enhancing accuracy and Morocco, this initiative was designed to encourage a relevance of water scores. conversation between the communities, corporate thought

leaders, the education visionaries, and the ecosystems builders AquaSense. Water management is a critical issue in many

to have constructive conversations around the shifts and needs countries, including Morocco. Severe droughts, poor water

of the changing future landscape. The discussions included distribution, and recent natural disasters raise the urgent need

researchers, start-up communities, technologists, and for better solutions to manage water resources effectively.

government representatives to unite and define the future of AquaSense’s prototype (see Figure 2) offers a smart way to

water sustainability as they see it. The selected AI technologies handle water resources by predicting future water situations,

and methodologies ranged from the use of satellite imagery to visualizing key data, and engaging citizens and communities.

the analysis of news and social media, or the input from water- This helps decision-makers plan better, save resources, and

related sensors and the application of Large Language Models respond quickly to local water issues. AquaSense provides

(LLMs) to describe good practices. We shall proceed with accurate forecasting of water parameters for informed

describing the problems addressed by the finalists of the management and answers water-related questions with

AI4Water challenge, their prototypes and the value of the detailed analysis using the latest data and news. It offers

innovation they brought with them. transparent data visualization through interactive charts,

allowing users to view and upload data easily. The community

AquaScore. The Rural communities in Morocco's High Atlas



Mountains struggle with water management due to limited levels map to locate and help regions in need of water, and a & citizens’ space features real-time news updates, a water

resources and visibility. Despite needing modest funds, these

tool to easily report local water issues.

villages face significant hurdles in accessing support. The



challenge lies in objectively quantifying water issues and

connecting these communities with potential supporters.

AquaScore creates a water map that pinpoints remote villages

and assigns them water scores based on satellite imagery and

crowdsourced data. This enables ranking villages by water

criticality, helping funders and supporters identify where to

direct their assistance effectively. The prototype (described in

Figure 1) also offers a platform for discussing water solutions, Figure 2: Screenshot of the prototype of Aquasense defining

fostering community engagement through gamification parameters, visualizing data and monitoring engagement.

features. By increasing the visibility of rural Moroccan villages

and providing objective water criticality assessments, AquaSense combines two distinct parts of AI: DL (LSTM) and

AquaScore facilitates efficient resource allocation for donors Generative AI (RAG and AI Agents). AquaSense uses

and experts. This AI-driven approach ensures fair and unbiased Multivariate and multistep LSTMs to accurately predict water

assistance to communities in need, promoting water parameters’ levels for the coming years, and Retrieval-

sustainability and improved water management in constrained Augmented Generation and AI Agents to answer water-related

environments. queries with detailed analysis, using the latest data, news,

AquaScore employs a hybrid approach combining Computer predicted parameters, and documents from sources like UN,

Vision (CV) and Natural Language Processing (NLP). CV UNCCD, and EPA. AquaSense uses Tensorflow and Keras (LSTM

algorithms segment satellite images to generate automated model), Pandas and Numpy (data preparation & mgmt.),

baseline water scores, while NLP algorithms extract insights Langchain (LLM framework for RAG and AI Agents), Chroma

from textual data to enhance score accuracy. This combination (Vector DB), Nomic Embeddings (Open-Source embeddings),

allows for objective assessment and continuous improvement GPT3.5-TURBO (LLM model), Streamlit (Web app). Aquasense

of water criticality rankings. The team has already aggregated improves water management by helping stakeholders make

data on 1,322 High Atlas villages, extracted satellite images, informed decisions, enhancing resource allocation, and

and segmented them using Facebook's Segment Anything promoting sustainable practices. Through its innovative

model. This process was completed on UM6P servers using features, it bridges the gap between citizens and authorities,

500GB of storage and 80 CPU cores. The system will which fosters collaboration and reduces water crises over time.

incorporate user-submitted reports and internet-scraped data Also, AquaSense aligns with several UN Sustainable

to further refine water scores. The uniqueness of AquaScore Development Goals (SDGs) such as SDG 6 (Clean Water and



222

Sanitation), SDG 13 (Climate Action), and SDG 11 (Sustainable powered early detection algorithms were prepared to

Cities and Communities). constantly monitor for signs of invasive species, triggering

immediate alerts to enable rapid response. Based on species-



the global problem of water optimization in the light of the specific data, the system can precisely deploy the most effective Water Consumption Tracker. This prototype is addressing

already visible consequences of climate change. That is, the eradication methods, from underwater drones to selective

large amount of wasted water due to irresponsible water use biocides. As invasive species evolve, the AI-driven platform

by the households. The added value lies in the behavioral continuously adapts strategies, ensuring that the interventions

approach: the application is designed to make users more remain effective and environmentally responsible.



water conservation a pleasure rather than a responsibility. YAZ. High unemployment rates in North Africa often translate aware of their attitude toward water consumption, and to make

Introducing a gamification approach as a new strategy should into many individuals employed in low-wage jobs, particularly

help make water conservation more appealing. It is based on youth from low-income households. Severe water scarcity

an app that tracks real-time water usage, provides leading is decreasing exports and rising prices of vegetables

personalized recommendations, and motivates users over a and fruits. Challenges meeting the needs of Morocco's

gamification environment, fostering a community focused on population while being a major exporter of produce to global

sustainable water use. markets. This AI-based agricultural solution is based on Smart



Regressor to find patterns between the households vertically indoors and outdoors, offering optimal use of characteristics and their water usage behavior. We plan to add available spaces. The adoption of hydroponics in Africa has the The use of Machine Learning models such as Random Forest Hydroponic Towers designed to efficiently grow crops



providing custom tips to optimize water usage. The approach potential to create millions of new jobs in the coming years. GenAI using LLM model as a chatbot to support our vision by

was fundamentally based on: (1) collecting data about the Integrated with GPT architecture, the technology allows real-

households using our application UI; (2) providing optimum time monitoring, pest detection, and yield estimation. YAZ

water consumption level by the ML model based on the data hydroponics are a shift towards a resilient and sustainable

collected; and (3) monitoring water usage through IoT sensors Moroccan agriculture.



used to optimize the ML model performance. Our approach can The tools and technologies presented in this paper that are and the notification system of our App. The data collected is

potentially reduce household water waste by 20-50% open source, are available at IRCAI’s SDG Observatory GitHub by

educating users about their consumption habits through repository (github.com/IRCAI-SDGobservatory).

notifications, ranking systems, and feedback mechanisms.





3 From concept to prototype in a month

AI in Africa in collaboration with IRCAI conducted a gathering

of minds which culminated in a 1-day summit around

technologies and shifts of the future, hosted by GITEX in the AI

Everything section of the GITEX Africa 2024. Between 26th

April and 31st May, 55 PhD and MSc students from 11 research

institutions took part of a complete program including expert

sessions kicked-off at the AI movement, UNESCO’s new center

for AI in Africa, and engaging experts in water-related topics

such as Matjaž Mikoš, UNESCO chair for landslide risk

reduction, droughts and floods, discussing our recent research

on news mining for extreme weather events [5, 6]; Gerald



the Water Consumption Tracker at the AI stage of GITEX Africa. Corzo Perez, senior researcher at the UN Water Education IHE Figure 3: The pitch of one of the top 3 teams – Ghayt – presenting



Delft, discussing our ongoing research on Water, AI and Twitter



Aquatic Biodiversity. [7]; and Ignacio Casals, R&D Manager in Aguas de Alicante The introduction of non-native species

into marine ecosystems presents a significant threat to the Spain, providing a industrial perspective on the use of AI to

fragile equilibrium of these vital environments. Invasive tackle the challenges of wastewater management [8].

species, often aggressive, can outcompete native organisms, The students were followed across 8 stages including:

leading to disrupted food chains, altered habitats, and conceptualization; data collection, analysis and visualization;

potentially irreversible ecological harm. From coastal areas to methodology and implementation, prototype building and

the open sea, the swift proliferation of invasive plants, animals, pitch (see Figure 3). In order to maximize the impact of the

and microorganisms endangers the biodiversity, productivity, programme, the content from the abovementioned

and resilience of our marine life. Addressing this escalating opportunities will be organized across UNESCO’s most related

global issue requires immediate and decisive action. AI- to the five areas: (1) capacity building; (2) developing



223

supportive policy; (3) effective, inclusive and equitable access 4 Conclusions and further work to quality Education; (4) nurturing and creating sustainability

The capacity building to enhance opportunities can benefit

models for Water Sustainability; and fostering and facilitating

from the engagement of the Youth in AI-driven challenges that

international cooperation.

start in research problems deriving from issues to address in

their communities. Problems they know well and data that they



often have privilege access to, with promising impact that can

ensure the sustainability of the innovation offered. The

initiative served us also to collaboratively discuss sustainable

solutions that help large scale recovery and define a better and

more hopeful inclusive Africa. The winning outcomes of this

challenge will integrate a vibrant worldwide Community of

researchers and entrepreneurs focusing on AI and SDGs,

Figure 4: The phases of the training curriculum across 5 weeks. starting with SDG 6, and supported by initiatives such as

IRCAI’s Top 100 or the SDG Observatory. Ethical considerations

The training curriculum included weekly seminars open to are being addressed in the context of the EC project AI4GOV.

public, training workshop for participants, showcases

and mentoring sessions (see Figure 4). The discussions ACKNOWLEDGMENTS forming the bae concepts of the participant projects were held

in the light of IRCAI’s re This research was partially funded by the European search and research achievements (see

Figure 5), aiming at building research collaboration bridges. Commission’s Horizon research and innovation program under

grant agreement 820985 (NAIADES) and 101120237 (ELIAS).





REFERENCES



[1] Blazhevska V.(2020). United Nations launches framework to speed up

progress on water and sanitation goal, United Nations Sustainable

Development.

[2] Casale G., Cordeiro Ortigara A.R. (2019) Water in the 2030 Agenda for

Sustainable Development: How can Europe act? Water Europe, Brussels.

(ISBN 978-90-8277064-3) 36p.

https://unesdoc.unesco.org/ark:/48223/pf0000372496

[3] International Water Association and Xylem Inc (2019). Digital Water:

Industry leaders chart the transformation journey. [Online] https://iwa-

network.org/wp-

content/uploads/2015/12/IWA_2019_Digital_Water_Report.pdf



Figure 5: [4] IRCAI Committee Chair on AI and Water Resource Management [online] Selected topics from IRCAI’s research to motivate ircai.org/project/ai-and-water-resources-management/ challengers in AI and Water research [5 ] Mikoš M., Bezak N., Pita Costa J., Nassri M. B., Jermol M., Grobelnik M.

(2022). Natural-hazard-related web observatories as a sustainable

The data and methods generated by the participants development tool. In Progress in Landslide Research and Technology, Vol.

programme can be used by companies, government and 1, No. 1, Springer (in print).

[6] Pita Costa J., Rei L., Bezak N., Mikoš M., Massri M.B., Novalija I. and Leban, G.

research institutions to aid in the resolution of problems (2024) Towards improved knowledge about water-related extremes based

related to Water Sustainability, by identifying trends and major on news media information captured using artificial

intelligence. International Journal of Disaster Risk Reduction, 100, p.104172.

areas of discussion, and to explore successful scenarios [7] Perez, G., Pita Costa J., Novalija I., Rei L., Senožetnik M., Casals del Busto I. C.

through similar challenges and cases. IRCAI’s SDG 6 (2024). Integrating Social Media, News and Machine Learning for

Observatory [10] is being built to properly address the Enhanced Hydrological Event Detection and Management. In 15th

International Conference on Hydroinformatics (p. 278)..

challenges of decision makers, using AI. It is benefitting: (i) [8] Pita Costa J., Massri M. B., Novalija I., Casals del Busto I., et al. (2021).

national governments providing access to a variety of Observing Water-Related Events for Evidence-Based Decision-Making. In:

Slovenian Data Mining and Data Warehouses conference (SiKDD2021)

perspectives (including trend and comparative) on a data [9] Pita Costa J. (2022). Water Intelligence to Support Decision Making,

driven dashboard with information on Water Sustainability Operation Management and Water Education - the NAIADES Report. IRCAI

Library. [online] https://ircai.org/project/ircais-project-report-on-

trends for decision-making; access to local (e.g. country-level) naiades/



progress on SDG 6; (ii) educational institutions, offering access [10] Pita Costa J., Zaouini M., Crnko M., Polzer M., Corzo Perez G., Mikoš M., Orlic D. and Jermol M. (2024) Challenging Water Sustainability in Africa to information on current trends on Water Sustainability Through AI, Proceedings of the HHAI 2024 workshop on AI in Africa and

research and development; (iii) research institutions, sourcing SDGs.

open data over interactive visualisation and research; (iv) the [11] UN-Sustainable Development [online] The IRCAI Water Observatory - AI in

the service of SDG 6 [online] https://sdgs.un.org/partnerships/ircai-

NGO community, easing access to information directly linked to water-observatory-ai-service-sdg-6

community priorities including citizen science activities; and [12] UN-Water Work Programme 2024-2025 [online]

(v) general population, empowering water education for all. https://www.unwater.org/publications/un-water-work-programme-

2024-2025



224





Predicting poverty using regression




Luka Urbanč Marko Grobelnik

Jožef Štefan Institute Jožef Stefan Institute

Ljubljana, Slovenija Ljubljana, Slovenija

urbancluka3@gmail.com marko.grobelnik@ijs.si



Joao Pita Costa Luis Rei

IRCAI, Quintelligence Jožef Stefan Institute

Ljubljana, Slovenija Ljubljana, Slovenija

joao.pitacosta@quintelligence.com luis.rei@ijs.si



Abstract defined by each country individually, recognizing that different

Poverty reduction is the first Sustainable Development Goal set countries have different measures of, e.g., what life conditions

by the United Nations to be achieved by 2030, but current data and how much income makes an individual reach a "poor" status,

indicates that the progress is insufficient. The diverse factors as well as how we can normalise this to better compare these

influencing poverty across different nations pose a challenge in relative indicators between countries. We are still missing a clear

developing effective predictive models. This paper evaluates the theory in poverty research, despite the issue existing for a number

use of various regression models to predict poverty rates using a of decades [2]. With that being said, some authors have already

comprehensive dataset of 111 variables from sources such as the explored the causes of poverty. For instance, corruption, political

UN and the World Bank. The data, spanning multiple domains instability, ineffective local governance, government polices, gen-

like political stability, education, and economic conditions, was der inequality and short-term wage replacement policies, such as

preprocessed and transformed to create auxiliary features and maternity leave benefits and sickness pay, impact relative poverty

interactions. Among the models, Ridge regression yielded the [6, 7]. When assessing what people believe causes poverty some

best results, achieving a Root Mean Square Error (RMSE) of 3.6, geographical differences emerge. For example, the United States

indicating high predictive accuracy on a global scale. This study are mostly of the thought that an individuals traits are responsible

highlights the importance of addressing multicollinearity and for poverty, while countries in Europe have a blend of individ-

incorporating a wide range of features to improve the general- ualistic, fatalistic and structural beliefs such as lack of will, bad

izability of poverty prediction models. Future research should luck and social injustice respectively [4].

explore more complex methods, such as neural networks, and Machine learning (ML) has also been used in academic re-

refine model hyperparameters for enhanced performance. search to identify trends and analyze data in most fields, includ-

ing poverty research. Although a number of papers have already

Keywords been published on the use of ML to predict poverty [1, 10, 12, 5,

3, 8] (for more see [11]) including combining satellite images and

poverty, linear regression, lasso regression, ridge regression, elas-

neural networks to help predict poverty in five African countries

tic net regression, sustainable development goals

[5], most take a limited number of variables. Usmanova’s litera-

1 ture review found 22 papers published between 2016 and March Introduction

2022, with a total of 57 AI methods applied, the most popular be-

The need to eradicate poverty has been a long standing issue,

ing random forest, used in more than half of all papers reviewed.

which was globally recognized numerous times, most impor- It also found most papers focus only on African and South Asian

tantly in the United Nations (UN) Sustainable Development Goals

countries, a finding consistent with our own [11].

(SDGs), being given the number one spot of SDG1: "End poverty In this paper we focus on the following research questions: (i)

in all its forms everywhere", which should be achieved by 2030.

can regression be useful to identify the most influential features,

The latest UN report on the progress made in achieving SDG1 from a large amount of global indicators; and (ii) can direct and

indicates Poverty has returned to pre-pandemic levels in middle-

indirect causality relations be identified that signal new indicators

and high-income countries, with poverty in low income countries

relevant to the Poverty-related issues?

still a fraction above those reported in 2019. While the trends



seem to be going in the right direction, the UN warns that the cur- 2 Data

rent pace of improvement is insufficient to reach the agreed goals

before 2030. This raises the question of what impacts poverty To address the research questions, we utilized 111 primary vari-

rates the most and how countries can most effectively reduce ables from sources such as the UN and the World Bank, aggre-

poverty levels. gated through the Our World in Data portal. These variables span

To fully understand and address the issue of poverty, one must diverse domains, including political stability, policies, education,

navigate several definitions, which can often lead to confusion. healthcare, economic conditions, and inequality. We prioritized

The baseline definition used in this paper is the poverty line as is features that prior research has identified as significant, while

also incorporating some factors that are less intuitively linked

Permission to make digital or hard copies of all or part of this work for personal to poverty. The dataset was then used to train various models

or classroom use is granted without fee provided that copies are not made or aimed at predicting poverty rates across countries. This task is

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this particularly challenging because countries respond differently

work must be honored. For all other uses, contact the owner/author(s). to the same variables. For instance, GDP growth tends to have a

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia more significant impact on poverty reduction in developing na-

© 2024 Copyright held by the owner/author(s).

https://doi.org/https://doi.org/10.70314/is.2024.sikdd.20 tions compared to developed ones. Additionally, many variables

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Urbanč et al.



are strongly correlated, making it difficult for linear regression the basic parameters in addition to the mathematically derived

models to capture their relationships accurately. columns.

As previously mentioned, most of the data used in this paper



was sourced from ourworldindata.com (OWiD), with some ad-

ditional data coming from fao.org—including variables such as

foreign direct investment inflows and outflows, and the added

value of agriculture, among others. Data on the transatlantic slave

trade and colonial rule was obtained from www.slavevoyages.org.

All datasets were preprocessed before being merged, following a

series of steps.

The first preprocessing step involved light modifications, such

as removing irrelevant columns, renaming columns, and exclud-

ing data from before 1987 and after 2023 due to gaps and incom-

plete data. Despite increased reporting in recent years, many

countries still omit certain indicators, complicating model train-

ing. To address this, missing features with more than 𝑛 data points

for a given country were interpolated, with the edges filled using

backward fill (bfill) and forward fill (ffill). Those with less than 𝑛

data points used the mean of the country’s income group for the

given year as a filler value. The number 𝑛 was intuitively chosen

to be five and the methods bfill and ffill were chosen to prevent

the use of unrealistic data. The World Bank classifies countries

into income groups: low (less than 1,045 USD), lower-middle

(1,046 USD to 4,095 USD), upper-middle (4,096 USD to 12,695

USD), and high income (12,696 USD or more). However, it is im-

portant to note that the data generated using the aforementioned

methods somewhat reduces overall robustness.

The next step involved generating auxiliary columns, specifi-

cally lagged columns and changes in value for relevant parame-

ters. For instance, the row corresponding to Niger in 2013 would

also include the GDP per capita for 2012, 2011, and earlier years,

in addition to the value for 2013. This approach reflects the fact

that poverty trends often manifest in response to changes over

time, rather than immediately. The default number of years for Figure 1: Scheme of adopted methodology lagged data was set to five. Similarly, we incorporated changes

in value over the same five-year period to capture more explicit

data on unusual events, such as the onset of wars or significant

political changes. 3 Methodology

Next each primary parameter was also used as an argument In order to predict worldwide poverty levels, we have used dif-

for a number of mathematical functions in an effort to see if any ferent linear regression models and compared their accuracies.

correlations aren’t linear but perhaps squared, cubed or another With this we aimed to ease the interpretability of the models,

elementary function. The functions used were: 𝑥 2 3 , 𝑥, ln 𝑥, sin 𝑥, which is harder to obtain with more complex methods such a

cos𝑥 , tan 𝑥, arcsin 𝑥, arccos 𝑥 , arctan 𝑥 to try and capture any neural networks. To perform the research work that is the base

elementary nonlinear dependence within the model. of this paper, we have selected ordinary linear regression, lasso

The last step was to create all possible products with the avail- regression, ridge regression and elastic net regression as the mod-

able primary parameters, as creating all possible products with els to compare. OLS regression struggles with multicollinearity,

all auxiliary parameters included would have been computation- where predictor variables are highly correlated, leading to un-

ally inefficient. After all these steps were made, the individual stable estimates of the coefficients. Ridge regression addresses

columns were fused together. This method of preprocessing in- this by adding an L2 regularization term, which penalizes large

creases the possible variables included, making the model even coefficients and helps to stabilize the estimates in the presence

more general and retaining as many rows of data as possible. of multicollinearity. By shrinking the coefficients, ridge regres-

The function responsible for preprocessing, generating and sion reduces the sensitivity of the model to colinear predictors,

merging the data has a few parameters: basic_parameters_only, ensuring more reliable and generalizable results. Unlike lasso,

combinations and math. basic_parameters_only determines, if ridge retains all predictors, making it particularly useful when

the model will only contain data obtained from various online multicollinearity is a key concern but feature selection is not

databases, or if the model should include generated data: the the goal. We use the implementation of these linear regression

change in value and values for previous years. combinations de- algorithms in scikit-learn [9].

termines, if the model should create all possible combinations The datasets were split into training and test sets using the

with the primary parameters and math determines if mathemati- sklearn function train_test_split, with 80% for training and

cal columns are included in an attempt to gain a deeper insight 20% for testing. The training set was used to train four regression

into the features’ relationships. The parameters are marked with variants (LinearRegression, Lasso, Ridge, ElasticNet),

B, C and M. For instance, B+M would mean the file contains all with a random state seed of 42. while the test set was used to

Predicting poverty using regression Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia





determine the mean squared error (MSE) and 𝑅 2 value using the

functions mean_squared_error and r2_score from [9], both

common metrics used to assess models accuracy. All models

except OLS regression also had the data standardized before

training. The hyperparameter 𝛼 for the models was sensibly

chosen as 0,1. The results, seen in Table 1 are color coded: red

for poor performance, yellow for intermediate, and green for the

best. The variation in the number of rows is due to the exclusion

of rows with insufficient yearly data, which were dropped when

calculating differences from previous years.

After identifying the most successful model, we proceeded to



compare its performance between high-income and low-income Figure 2: Visual representation of model weights

countries. This comparison aimed to assess how the accuracy and

frequency of reported data influence the model’s performance.

These two income groups were chosen because low-income coun-

tries typically report less data with lower accuracy, while high-

population with urban and rural population share. Other no-

income countries provide more precise reports. We selected all

table combinations include secondary school completion with

high- and low-income countries from the dataset that were not

women’s civil liberties, internet usage with sanitation access,

used during the model’s training. From the 20% of data reserved

and military spending with wealth distribution. The weights also

for evaluation, 444 rows (30%) belonged to high-income countries,

reflect factors like infant mortality, years colonized, and agricul-

and 368 rows (24%) belonged to low-income countries.

tural employment. Figure 2 further illustrates the decline in the

We used the trained model to predict poverty levels for these

absolute value of these weights.

groups and evaluated its performance using the MSE metric to

The model performed better on high-income countries, with

analyze differences between income groups. Additionally, we

an MSE of 6.60, significantly below the overall MSE. In contrast,

calculated the maximum error to determine if the average per-

the MSE for low-income countries was 20.68. The maximum

formance was skewed by outliers. A similar evaluation was con-

error was also lower for high-income countries (22.1) compared

ducted on the data from Slovenia and Somalia, which were part of

to low-income ones (34.4).

the split. Slovenia had 8 rows of data, and Somalia had 6, allowing

The difference in the model’s performance on Slovenia and

us to explore how missing data impacts the model’s performance,

Somalia was notable. For Slovenia, the MSE was 0.78 with a

as Somalia had significantly fewer data points overall.

maximum error of 1.54, far below the overall metrics. Somalia,

however, had a much higher MSE of 95.7 and a maximum error of

18.7, likely due to less reliable and extreme poverty data, which

4 skews the model’s performance on extreme cases. Main Results

The file configuration plays a critical role in the model’s per-

formance. The results show that C+M, C, and B+C are the best 5 Discussion

configurations. The C+M file includes all basic features, lagged Firstly, the fact that ordinary least squares linear regression

values, changes in value, mathematical columns, and all possi- couldn’t produce an accurate model confirms the fact that the

ble combinations of basic parameters, totaling 8,236 parameters. parameters are indeed correlated. This is probably also the rea-

Configuration C contains all basic features, combinations, and son why the ridge regression model performed the best: ridge

lagged and difference columns. Lastly, B+C includes only the regression is used to address the issue of multicollinearity and

basic parameters and their combinations. All top-performing the features included are mostly strongly correlated, as stated in

models were trained on these datasets. the introduction. Furthermore, the correlation between parame-

The results in Table 1 show considerable variation. Models ters is obviously drastically increased by generating all possible

trained with ordinary least squares regression performed poorly, products of basic parameters.

with the best model reaching an RMSE just under 10.15 and an Secondly, the impact of mathematical columns needs to be

𝑅 2 of 0.50. In contrast, lasso and elastic net regression achieved considered. Of the first four models, two have mathematical

better results, with RMSEs around 7 and 𝑅 2 values close to 0.80. columns and two don’t. Of the eight models generated, three

Ridge regression also struggled, except for configuration B+C, of them perform worse if mathematical data is present, while 5

which provided the best results with an RMSE of 3.6 and an 𝑅 2 performed better with mathematical data included. This might

of 0.94. However, caution is advised when interpreting models indicate some deeper connection, which would be interesting

using configuration C+M or C, due to the high number of features to try and understand. Furthermore, lasso regression handles

relative to the dataset size, which could affect their real-world mathematical columns much better compared to the other models

reliability. used due to its ability to exclude features.

The model weights reveal that only products are present The impact of product combinations of basic features stands

among the top ten most important factors. These products in- out, with all better-performing models having the combinations

clude data on population, population density, agriculture, equal- parameter set to True, suggesting deeper relationships between

ity, healthcare, and education. The largest weights show the variables. Exploring these connections further, perhaps by train-

biggest differences, gradually decreasing in magnitude. The top ing a neural network on the basic parameters and comparing

ten weights range from just over 10 to 7, with the highest weights it to linear regression models, could be insightful. If the neural

involving combinations such as population and population den- network performs better, further investigation into these correla-

sity, meadows and pastures with the global peace index, and tions would be needed.



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Urbanč et al.



Structure Linear MSE Linear 𝑅 2 2 2 2 Lasso MSE Lasso 𝑅 Ridge MSE Ridge 𝑅 Elastic net MSE Elastic net 𝑅 Shape of X

M 203 0.031 74 0.65 - - - - (7653, 2131)

None - - 109 0.48 163 0.22 108 0.49 (7653, 1221)

C+M 198 0.054 45 0.78 - - 40 0.81 (7653, 8236)

C - - 50 0.76 - - 45 0.79 (7653, 7326)

B 103 0.50 110 0.47 103 0.50 111 0.46 (7661, 111)

B+C - - 48 0.77 13.3 0.94 43 0.79 (7661, 6216)

Table 1: MSE and R-squared values for different regression models and dataset configurations. The presence of B, C or M

signals the presence of basic basic parameters only (B), combinations (C) and mathematically (M) derived columns in the

dataset. A dash is used to label non-converging models with a negative R-squared value.



The dataset used spans from 1987 to 2023, which is relatively through testing of numerous linear regression models using open

short, given that poverty often has deep historical roots. Al- data, with the best model being created by using ridge linear

though data becomes scarcer in earlier years, those points could regression trained on data which also included all possible com-

still be crucial for improving model accuracy. Moreover, most binations of the basic features included in the dataset. The basic

hyper parameters in this paper were chosen sensibly due to time parameters included consist of 111 different parameters describ-

and computational constraints. Different values for the number ing countries across 36 years. Better models could possibly be

of lagged years, years of differences, hyperparameters in the generated using more complex methods such as neural nets or

training of models and the minimum number of data points re- random forest, gaining in accuracy but compromising the ex-

quired to interpolate missing data could all lead to interesting plainability of the model. The models could also benefit from

discoveries and improvements of the generated models. Our re- hyperparameter tuning during the whole process to improve

sult here shows it is possible to achieve this degree of accuracy, results and find the optimal values. We will be addressing this in

but it doesn’t limit what the best model could be. The elastic net, further research.

especially, should benefit from such a tuning.

As stated in [11], the recent literature mostly uses the random 7 Acknowledgements

forest model and, in fact, ordinary linear regression wasn’t even This research was partially funded by the Future of Life Institute

in the top ten most common methods. An interesting thing to under the project "An AI-driven Observatory Against Poverty",

explore would also be the performance of random forest using and the European Commission’s projects under grant agreement

the best configuration, B+C. The models may struggle to capture 101135800 (RAIDO) and 101120237 (ELIAS). correlations between variables due to differing impacts across

countries, as mentioned in the introduction. A potential solution References

is to split the countries into 𝑘 groups and train separate models [1] Gianni Betti, Antonella D’Agostino, and Laura Neri. 2002. Panel regres-

for each group. While this could improve predictions, it raises sion models for measuring multidimensional poverty dynamics. Statistical

two challenges: how to split countries without bias and how to methods and applications, 11, 359–369.

[2] David Brady. 2019. Theories of the causes of poverty. Annual Review of

ensure enough data for training. Sociology, 45, 1, 155–175.

The weights in the model further emphasize the issue of mul- [3] Muse A.H. Hassan A.A. and Chesneau C. 2024. Machine learning study

ticollinearity among the parameters, with only product terms Transactions on Radiation and Plasma Medical Sciences using 2020 sdhs data to determine poverty determinants in somalia. IEEE , 14, 1, 5956.

emerging as the most influential. However, this does not reveal [4] Dariush Hayati and Ezatollah Karami. 2005. Typology of causes of poverty:

the true importance of individual parameters, as they may en- the perception of iranian farmers. Journal of Economic psychology, 26, 6,

884–901.

hance the impact of another factor within the product term. Addi- [5] Neal Jean, Marshall Burke, Michael Xie, W Matthew Davis, David B Lobell,

tional research is needed to better determine the true significance and Stefano Ermon. 2016. Combining satellite imagery and machine learning

of these parameters and gain a clearer understanding of what to predict poverty. Science, 353, 6301, 790–794.

[6] AH Ng, Abdul Ghani Farinda, Fock Kui Kan, Ai Ling Lim, and Teo Ming Ting.

drives poverty rates up or down. It can be seen in Figure 2, the 2013. Poverty: its causes and solutions. International Journal of Humanities

models weights occupy a wide range. It is clear that some features and Social Sciences, 7, 8, 2471–2479.

are more important, based on their weights and further work is Flaviana Palmisano. 2019. Gender equality and poverty are intrinsically [7] Rense Nieuwenhuis, Teresa Munzi, Jörg Neugschwender, Heba Omar, and

being done to understand which features stand out and why. linked: A contribution to the continued monitoring of selected sustainable

The model also performed better in predicting poverty lev- development goals. Tech. rep. LIS Working Paper Series.

[8] Shah O. and Tallam K. 2023. Novel machine learning approach for predict-

els in high-income countries compared to low-income countries. ing poverty using temperature and remote sensing data in ethiopia. IEEE

This discrepancy can likely be attributed to the fact that high- Transactions on Radiation and Plasma Medical Sciences, 5, 6, 2302.14835.

income countries report more data with greater accuracy, allow- [9] F. Pedregosa et al. 2011. Scikit-learn: machine learning in Python. Journal

of Machine Learning Research, 12, 2825–2830.

ing the model to identify underlying patterns more effectively. [10] Mubaraq Dele Sulaimon. 2020. Multidimensional poverty and its determi-

In contrast, much of the data for low-income countries had to be nants: empirical evidence from nigeria.

interpolated, which reduced variability between countries and 2022. Utilities of artificial intelligence in poverty prediction: a review. [11] Aziza Usmanova, Ahmed Aziz, Dilshodjon Rakhmonov, and Walid Osamy.

Sus-

negatively impacted the model’s performance. tainability, 14, 21, 14238.

[12] Huang Zixi. 2021. Poverty prediction through machine learning. In 2021

2nd International Conference on E-Commerce and Internet Technology (ECIT).

IEEE, 314–324.

6 Conclusion

In this paper, we have shown that a general model exists, based

on linear regression methodologies, which can predict poverty

with a relatively high accuracy (RMSE of 3.6). This was achieved



Fact Manipulation in News: LLM-Driven Synthesis and



Evaluation of Fake News Annotation



Luka Golob Abdul Sittar

lukag26@gmail.com abdul.sittar@ijs.si

Jožef Stefan Institute and Jožef Stefan Postgraduate Jožef Stefan Institute and Jožef Stefan Postgraduate

School School

Jamova cesta 39 Jamova cesta 39

Ljubljana, Slovenia Ljubljana, Slovenia



Abstract (1) A methodology to create synthetic data for fake news

using LLMs.

Advancements in artificial intelligence and increased internet

(2) We then use this methodology, to adapt the FA-KES dataset

accessibility have made it simpler to create and disseminate fake

1

with 100 additional synthetic fake news .

news with customized content. However, they also improved the

ability to analyze and identify such misinformation. To effectively In Section 2, we discuss work that is closely related to our task.

train high-performance models, we require high-quality, up-to- Section 3 then outlines the methodology for generating synthetic

date training datasets. This article delves into the potential for fake news, culminating in Section 4, where we present the results

generating fake news through factual modifications of articles. and introduce some modifications to the methodology. Finally, in

This is facilitated by prompt-based content generated by large Chapter 5, challenges, capabilities, and potential improvements

language models (LLMs), which can identify and manipulate are considered.

facts. We intend to outline our methodology, highlighting both

the capabilities and limitations of this approach. Additionally, 2 Related Work this effort has resulted in new quality synthetic data that can be

A wide range of approaches to generate fake synthetic news with

incorporated into the standard FAK-ES dataset.

LLM has been developed. In [8] authors generated huge amounts

of fake news and categorized them into multiple categories. LLMs

Keywords can generate fake news by altering the style to mimic credible

fake news, synthetic data, fact extraction, fact verification, large sources or using sensationalism to influence perception. They

language models can subtly manipulate content to be perceived as true, blend real

and fabricated information to exploit cognitive biases, or create

1 Introduction convincing fictional narratives.

In general, when making a dataset we want a diverse distribu-

Synthetic data refers to artificially generated data that is not

tion of fake datasets. In our case, we will focus on one way of data

obtained by direct measurement or observation of real-world

change, which comes under the umbrella of . Content Manipulation

events. Instead, it is created using algorithms and simulations.

Similar news manipulations can be seen in [7] where the authors

The primary purpose of synthetic data is to provide a realistic

use two main techniques. The first one extracts the summary

alternative to real data for various use cases, such as training

from the original text, which preserves the main content, which

machine learning models, testing systems, ensuring data privacy,

is then changed to produce a fake article. The second one asks a

and more.

question about the article and changes the content of its answer,

We will generate synthetic data from news articles. By making

to construct a new article. Our approach is in nature similar to

sure, that the information in the news is changed we can safely

the Question-Answer framework.

call it fake news. In our article, fake news will denote articles that

are and false [4]. Synthetic data enhances verifiably

intentionally Many articles provide fake news detection models made using

synthetic data. Most popular are deep neural networks such as

model training by providing additional examples to supplement

BERT [1]. But there are other fact-based approaches for fake news

scarce labeled datasets and allows for privacy-conscious testing

labeling as in [3]. In [2] they used GPT4-turbo for prompt-driven

without real content manipulation. It enables adaptability to

fake news detection.

evolving fake news tactics by simulating diverse scenarios from

the newest data, thereby improving the robustness and resilience

of detection algorithms [3]. 3 Methodology

Large language models (LLMs) made a huge difference in the The methodology is divided into four conceptual steps: Data

world of news. Fake news is now much easier and cheaper to collection, Characterization of facts, Fact extraction, and Fact

construct, but we also have additional methods to help us tackle manipulation as presented in Table 1.

its spread. Numerous articles appeared trying to partake in this

effort. The following are the main scientific contributions of this 3.1 Data Collection

paper:

The publicly available FA-KES dataset [5], focused on the Syrian

war, addresses the deficiency of manually labeled datasets in

Permission to make digital or hard copies of all or part of this work for personal

this domain of news data. It comprises 804 articles sourced from

or classroom use is granted without fee provided that copies are not made or

distributed for profit or commercial advantage and that copies bear this notice and various media outlets. We used 426 articles that were manually

the full citation on the first page. Copyrights for third-party components of this labeled as authentic news, but we could just as well use the other

work must be honored. For all other uses, contact the owner /author(s).

(fake) articles.

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).

1

https://doi.org/10.70314/is.2024.sikdd.13 https://github.com/golobluka/Fake- news- generation- from- FA- KES- dataset



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Luka et al.



Data collection 3.4 Fact Manipulation and Synthetic News Characterization of facts

Generation

Should have textual and 1.Name of casualty The objective is to modify relevant information without altering

statistical facts 2.Gender or age group the writing style or topic of the article. For this transformation,

3.Cause of death we used a chain of thought prompt, which for a given fact: 1)

4.Type changes the fact to another with a different meaning, 2) generates

5.Actor a new article based on the altered facts. By changing one fact at

6.Place of death a time, quality is improved compared to altering multiple facts

7.Date of death simultaneously, as one fact creates a clearer chain of instructions.

LLMs such as often struggle with precise changes Llama3.1:8B

Fact Extraction Fact manipulation in the article, such as modifying implicit references or incorpo-

rating new facts. Quality can be improved by carefully adjusting

Name of casualty: Civilians Name of casualty: the prompt content.

Gender or age group: e.g., Manipulated fact LLMs are also exceptional in summarization and paraphras-

child, adult, senior Gender or age group: ing. Both are used simultaneously with changing the facts. The

Cause of death: shooting, Manipulated fact problem is that we aim to maintain the extracted facts when sum-shelling, weapons, etc. Cause of death:

Type: military personnel marizing. But this is not crucial, as it usually has better results Manipulated fact

Actor: rebels, forces Type: Manipulated fact as article generation.

Place of death: Airbase Actor: Manipulated fact

Date of death: April 7, 2017 Place of death: Manipulated 3.5 Fake News Annotation and Fact

fact verification

Date of death: Manipulated

fact After we have generated the fake articles, we can label that data

as “fake” or “non-fake”, based on comparison with extracted facts.

We performed this labeling with various models and compared

Figure 1: A methodology to generate synthetic data for the performance of labeling,to get the best model. In this ex-

fake news detection periment we decided for Llama3.1. To do the labeling, we are

performing fact verification [4]. The fact verification task in gen-



3.2 Characterization of Facts eral is making a decision as to whether a claim is correct, based

on the explicitly-available evidence, such as Wikipedia articles

While making the FA-KES dataset, its authors created seven fac-

or research papers. We have the extracted fact, which will be

tual categories:

compared to the article content. The question thus becomes: Do

these facts appear in the given article? This approach emphasizes

(1) (4) , Name of casualty Type

factual content rather than the overall sentiment of the article.

or group, (5) Actor,

There are two primary types of prompts: 1) Direct prompts

(2) (6) , Gender or age Place of death

that present the article and a table of facts, asking if the facts

group, (7) Date of death.

relate to the article, 2) Structured prompts that inquire about

(3) , Cause of death

the correspondence of one fact at a time with the article. The

question is: Does this fact correspond to the content of the arti-

It is crucial to note that all articles have a similar structure,

cle? This method combines individual results into an aggregated

describing war incidents. This allows us to establish a consistent

score. Say the is characterized as Place of death Idlib and

framework of facts, such as actor and casualty details. We stick

Daraa provinces. Then the question posed to LLM is of the form:

to those facts, but generate them differently, employing LLMs

Read the article and understand its places of death.

capabilities with faster and cheaper execution, albeit with a slight

Do Idlib and Daraa provinces “really correspond” to

reduction in reliability.

places of death in the article?



3.3 We are not as interested in labeling, as we are interested in Fact Extraction

the quality of produced synthetic fake news. For this purpose,

We extract facts by constructing prompts for LLMs. First ap-

we will also use fact verification in a slightly different way. We

proach was a few-shot prompt, which gives some examples of

are asking the LLM: Were the factual changes in fake news really

output. Later we constructed an additional approach: Say we

made, as they were supposed to? A similar method is used in the

are extracting the fact with this second tech-Place of death

article [7].

nique. We give a detailed description of what should be extracted



and then LLM reads the article and performs the task solely on 4 Experimentation and Results this basis. This description is usually longer and contains more

context. The issues with fact extraction in general are: 4.1 Experimental settings

• Some articles lack certain facts or merely imply them. We selected 426 articles labeled as authentic news from FA-KES

LLMs can identify this, outputting responses such as “No dataset. Then facts were extracted and transformed, as described

information.” in the previous section. At first two basic approaches were used

• Longer articles may contain multiple events, each with dis- to randomly choose 70 news articles and transform them. After-

tinct data such as dates or casualties. This can be managed ward, we used the labeling procedure to compare performance,

by creating separate tables for each event or consolidating resulting in the table 1. Based on the results we then composed

all events into a single table with various facts. the final algorithm, which would be manually evaluated.



Fact Manipulation in News: LLM-Driven Synthesis and Evaluation of Fake News Annotation Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



4.2 Table 1: Comparison of fake synthetic data. Evaluation

For every experiment, we first manually checked a minimum 10

Type of data Number of facts manipulated Precision Recall F1 Accuracy

percent of random examples to get an overview of how well the

Summarization 2/7 0.74 0.63 0.68 0.71

LLM was able to do the job. It is quite useful to print text that rep-

Detailed facts 2/7 0.70 0.80 0.75 0.73

resents the procedure of decision-making that LLM undertakes,

when challenged with the task. It was even helpful to see LLMs

generated thinking procedure, as this gives valuable insight, into this context, leaving it unchanged in most cases. Our fake news

what is going on “under the hood”. We believe that manual fact- fails to preserve enough coherence to be trusted by a skeptical

checking is the first and most crucial step in generating good reader, who tries to connect background material to the event in

prompts. Based on fallacies one can then adjust prompts content. the article.

To shed some light on this procedure we have made the following Generating false text, while maintaining coherency, is chal-

overview. lenging for LLM. In this task, we have changed one fact: for

example, the may be changed to another city Place of death

4.3 Fact Extraction Results or neighborhood. Then this fact must be changed in the article

while maintaining other factual information. Here are the main



Name of casualty or issues: Members of Nusra Front

group: • In the beginning some facts did not get changed, or the

Gender or age group: Adults (no specific age men- facts were altogether just removed from the article. We

tioned) managed to reduce this error by adjusting the prompt. It

Cause of death: Explosion at a mosque is difficult to adjust all occurrences of the fact, especially



Type: if it is only implied and not explicitly stated. We managed

Non-civilian (militants)

to minimize this problem, by a method yet to be shown in

Actor: Unknown (no group section 4.5.

claimed responsibility, but •

What remains is the problem of a wider context, Suppose

supporters blamed ISIS)

we change the town of the incident, then we must change

Place of death: Ariha, Idlib province, Syria the name of the neighborhood accordingly. LLM usually

Date of death: Not specified in the article fails in this, leaving our article inconsistent, which is a

widespread problem.

• LLM does not want to output the content because of harm-

Figure 2: Example of fact extraction. ful content or does not want to produce articles that could

be used with illegal intent. This was quite a common prob-

LLMs are capable of recognizing different topics and extracting lem, which is also reasonable, based on the violent con-

words that correspond to this topic, and also noting if the fact is tent of articles and the possible abuse of LLM-generated

not mentioned. At first, we extracted short words as represented content. The best thing to prevent this error is to use un-

in Figure 2. censored LLM. In other cases, one can adjust the prompts

The issue begins with nuances. For example, in many articles by removing suspicious words like “fake news”.

the is only suspected but not known. In some cases, ac- The Generated article was shorter, skipping the original Actor • tor and causality are not precisely distinguished. This usually text which was not linked to extracted facts. This problem

leaves LLM to some kind of arbitrariness. For this purpose, We was reduced but still exists in long articles.

also added a longer description that better captures the nuanced If the fact is not present in the article, then it is hard •

subtleties related to facts. This can also be captured in Table 1. for LLM to incorporate a new fictitious fact into the text.

There we see the results for short (normal) or detailed extracted Mainly it just adds the information in separate sentences.

facts. The recall is far worse in the case of short prompts. This When we change facts, traces of the old facts still persist. •

likely means that there is an abundance of false negatives, which This is especially common in complicated articles with

result from the fact, that labeling does not manage to match true diverse structures.

articles and their corresponding short facts. Sometimes the change does not bring about any additional •

The shorter extracted facts are often not comprehensive. For meaning. For example, LLM might change previously un-

example, under the label (which classifies civilian or non- known casualties and designate them as civilians. They Type civilian) it writes only civilians, even though, contextual under- were implied to be civilians all along, and this makes only

standing also includes some non-civilian casualties. a minor change and is not really fake.

Overall the most important insight remains: fact extraction

has better quality than article generation. 4.5 Fact verification with LLMs



4.4 Remember that in this task, the prompt asks: Does this fact “really Quality and coherence of synthetically

correspond” to the content of the article? Performance largely

generated fake news depends on how the program takes the word “really correspond”.

The LLM can detect (for example) the of some attack in the Words have many nuances: different words can have different Actor news, and then it is mostly able to change every occurrence of meanings, which can complicate labeling. To simplify: we can be

this with another . But if we would like to preserve stricter, in the sense that words must be the same in the literal Actor Actor

all the coherence of the article much more would need to be done. sense, or we can count on the similarity of meaning [6]. Based

News usually contains background information, that provides on our goal of creating fake news it is best to focus on meaning

context for the accident. Our algorithms failed to properly adjust and not concrete words.



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Luka et al.



Here are some common problems: 5.1 Problems, Capabilities and Possible

• Sometimes the fact is changed, but LLM skeptically as- Improvements

sumes, that those two names refer to the same group. In this stage, LLMs like are not able to co-• Lamma3.1:8B

• In longer articles, where there are many events, the names herently change certain facts of news articles. Changing

get changed only in some events (usually at the beginning facts can distort the article content, which appears to be

of the article). In this case, the LLM can make unwanted extremely hard to manage. This normally does not happen

predictions, labeling the fact as true rather than false. for manageable data as dates (changing the time of some

event), but for much more involved actors of the attack

Manual checking shows that labeling is more accurate than

in the article. Even so, the synthetic fake news provides

generation of fake news. This leads us to use labeling as a means

valuable information.

to improve article generation.

• We did not use the model, which has additional informa-

Table 1 was used to compare different ways to generate fake

tion about the news content. Providing additional context

news. It shows two of the best datasets, which contain true arti-

would likely have a beneficial effect on all the processes.

cles and their false twins, generated in two ways:

• In our case facts were largely dependent on each other. For

(1) Fake news generated by “standard” fact extraction and example Gender or age group is an extraction of Name

with additional summarization. . We think it is best if such depen- of casualty or group

(2) Fake news generated by “detailed” fact extraction and with dencies are removed because they bring to inconsistencies

an additional paraphrasing of the article. when changing facts. An additional solution would also

be to change whenever Gander or age group Name of

In this experiment, instead of merely categorizing the articles

as true or false, the results shown in Table 1 reflect how well the casualty or group is changed.

generation process aligns with fact verification. • Fact extraction is close to human-like quality. The issue

is, that besides manual checking, it is hard to find a good

Low precision in the row with Detailed facts led us to detect

measure of the quality of extracted facts.

articles that were not changed. We implemented a strategy where

labeling was applied after generating the fake articles to assess • Detection of changed facts is in quality similar to extrac-

tion of facts (this is not surprising, since they are based

the quality of the generation. LLMs often provide incomplete

on the same skill). Because of the diversity of meanings in

responses and struggle to correct them directly. By introducing an

language, it is hard to specify the exact reasoning proce-

additional verification step, we were able to enhance the overall

dure of LLMs and many mistakes come from this kind of

accuracy of the results.

miscommunication.



4.6 Final Dataset Description 6 Acknowledgments

In the end, we constructed 100 fake-news based on a prior ex-

This work was supported by the European Union through AI4Gov

2

periment, which can be found on GitHub . In every article we

(101094905) and TWON (101095095) EU HE projects and the

randomly chose three facts and changed them. Afterward, we

Slovenian National grant (CRP V2-2272).

carefully went through 10 examples, which are also present on



Git Hub, while here we present only the main points: References

• Fact verification improved quality by making sure, that [1] Nicola Capuano, Giuseppe Fenza, Vincenzo Loia, and Francesco David Nota.

2023. Content-based fake news detection with machine and deep learning: a

the synthetic fake article really incorporated new infor-

systematic review. , 530, 91–103. doi: https://doi.org/10.1016 Neurocomputing

mation. More than 90% new facts really got incorporated /j.neucom.2023.02.005.

in the article. Sometimes new information is only added [2] Fredrik Jurgell and Theodor Borgman. 2024. Fake news detection : using a

large language model for accessible solutions. (2024).

as additional text(and does not seriously change the main

[3] Ye Liu, Jiajun Zhu, Kai Zhang, Haoyu Tang, Yanghai Zhang, Xukai Liu, Qi Liu,

topic). and Enhong Chen. 2024. Detect, investigate, judge and determine: a novel

• llm-based framework for few-shot fake news detection. (2024). https://arxiv Fact is not always incorporated in all places where it is

.org/abs/2407.08952 arXiv: 2407.08952 . [cs.CL]

referenced, which leads to inconsistencies. The new article

[4] Taichi Murayama. 2021. Dataset of fake news detection and fact verification:

is then a blend of old and new information. a survey. (2021). https://arxiv.org/abs/2111.03299 arXiv: 2111.03299 . [cs.LG]

• [5] Fatima K Abu Salem, Roaa Al Feel, Shady Elbassuoni, Mohamad Jaber, and

There are problems with ˙˙detailed” prompts. Containing

May Farah. 2019. Fa-kes: a fake news dataset around the syrian war. In

more information results in contradictions as we change . Proceedings of the international AAAI conference on web and social media

Vol. 13, 573–582.

only one fact at a time.

[6] Abdul Sittar, Dunja Mladenic, and Tomaž Erjavec. 2020. A dataset for in-

formation spreading over the news. In Proceedings of the 23th International

5 Conclusion Multiconference Information Society SiKDD. Vol. 100, 5–8.

[7] Yanshen Sun, Jianfeng He, Limeng Cui, Shuo Lei, and Chang-Tien Lu. 2024.

In this article, we focused on exploring the potential of LLMs in Exploring the deceptive power of llm-generated fake news: a study of real-

world detection challenges. (2024). https://arxiv.org/abs/2403.18249 arXiv:

fact extraction and generation of fake news. Our motivation was

2403.18249 . [cs.CL]

primarily to understand how accurate are LLMs in fact extraction

[8] Lionel Z. Wang, Yiming Ma, Renfei Gao, Beichen Guo, Zhuoran Li, Han

and how reliably LLMs generate synthetic news by altering facts. Zhu, Wenqi Fan, Zexin Lu, and Ka Chung Ng. 2024. Megafake: a theory-

driven dataset of fake news generated by large language models. (2024).

As a result of our experiment, we have generated 100 synthetic

https://arxiv.org/abs/2408.11871 arXiv: 2408.11871 . [cs.CL]

news by randomly transforming there out of seven facts and

have performed a manual evaluation, to observe the quality of

the generated news dataset.



2

https://github.com/golobluka/Fake- news- generation- from- FA- KES- dataset



Borrowing Words: Transfer Learning for Reported Speech



Detection in Slovenian News Texts



Zoran Fijavž

Jožef Stefan Postgraduate International School

Peace Institute

Slovenia, Ljubljana

zoran.fijavz@mirovni- institut.si



Abstract to explore speaker representation by gender [1], institutional

affiliations [8], and topic stances [15], or to distinguish between

This paper describes the development of a reported speech clas-

journalists’ and sources’ voices [11].

sifier for Slovenian news texts using transfer learning. Due to a

lack of Slovenian training data, multilingual models were trained

on English and German reported speech datasets, reaching an 2.2 Existing Datasets and Modelling

F-score of 66.8 on a small manually annotated Slovenian news Approaches

dataset and a manual error analysis was performed. While the Datasets with reported speech annotations mostly cotain liter-

developed model captures many aspects of reported speech, fur- ary or news texts. Key corpora include RiQuA [12], SLäNDa 2.0

ther refinement and annotated data would be needed to reliably [19], Redewiedergabe [3], QUAC [14], PolNeAR [10], Quotebank

predict less frequent instances, such as indirect speech and nom- [21], and STOP [22]. RiQuA and Redewiedergabe are the largest

inalizations. th annotated corpora, covering English and German 19 century

texts. QUAC contains 212 annotated articles from the Portuguese

Keywords newspaper Público, while Quotebank spans 162 million news ar-

reported speech, natural language processing, transfer learning, ticles with automatic annotations. PolNeAR, consisting of 1,028

news analysis news articles, includes attribution annotations, which include

and exceed the definition of reported speech. A summary of the

1 Introduction datasets is provided in Table 1.

The corpora differ in annotation complexity and size. They are

Reported speech, ubiquitous in literary and news texts, has clear

mostly monolingual, warranting the used cross-lingual transfer

lexical and syntactic patterns which may be reliably modeled

learning for low-resource languages by employing multilingual

via natural language processing (NLP) and may be useful for

models such as mBERT [6] and XLM-R [4]. Narrower multilingual

downstream tasks by drawing a distinction between source and

models, such as CroSloEngual BERT, often outperform broader

background information. The paper applies transfer learning to

ones [20]. Reported speech modeling may be operationalized as

extend reported speech classification to Slovenian news texts and

speaker or quotation detection tasks [23, 17]. Simplifying the task

provides a provisional classification model. A manual error anal-

to sentence-level classification is warranted by the fact news (un-

ysis reveals the model’s strengths and weaknesses, highlighting

like literary texts) rarely mix statements by sources and authors

possible steps for further improvements.

in the same sentence and can improve classification reliability at



2 Related Work the expense of detailed aspects of reported speech [17] and sim-

plify the annotation structure. Missing fine-grained outputs, such

2.1 Role of Reported Speech as speakers and boundaries of reported and reporting clauses,

Reported speech is common in news texts, generally expressed as may thus be an acceptable trade-off for NLP-based content analy-

direct or indirect speech, with the former repeating the original sis in news texts. A systematic review of such approaches points

utterance verbatim and the latter embedding it in a that-clause to the limits resulting from a low number of features with no

[18] (e.g., guarantee of reliable ( joint) prediction, which preclude drawing Jimmy said: “Another systematic review would be great!”

and ). rich conclusions expected from the method’s manual counterpart Jimmy said that another systematic review would be great.

More complex forms include mixed speech ( [2]. City officials rebuffed

the accusations as "groundless and blatantly false".) and reportative

nominalizations with an analogous function as reported speech 3 Experimental Setting



( ) The speaker particularly emphasized the pressures on the media 3.1 Task Overview

[7]. Around 50% of sentences in newspaper corpora may be at-

We treated reported speech as a sentence-level classification task.

tributed to a source in the text, predominantly through direct

Sentence splitters were applied to existing datasets, and binary

and indirect speech [17]. Verbs cue 96% of reported speech, fol-

labels were assigned by matching annotated spans with the split

lowed by prepositional phrases (3%) [13]. Reported speech lends

sentences. Reported speech sub-types were unified under a single

objectivity to statements [9], summarizes source statements [16],

label, joining the annotation schemes of individual datasets. A

and is used in discourse analysis and communication studies

Slovenian dataset of 10 news texts was manually annotated at

Permission to make digital or hard copies of all or part of this work for personal

the sentence level. The datasets were split into training, evalu-

or classroom use is granted without fee provided that copies are not made or

ation, and test sets to train multilingual pretrained models. For

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this CroSloEngual BERT, preprocessing also involved machine trans-

work must be honored. For all other uses, contact the owner /author(s).

lating the German training data into English. The model outputs

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

were binary labels indicating reported speech, used to calculate F-

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.sikdd.21 scores on the test data. A manual error analysis was performed on

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Z. Fijavž



Table 1: Summary of Datasets’ Characteristics.



Corpus Type Annotations Language Sentence No. Role Positive Class

RiQua fiction direct and in- English 38,610 72% train, 18% 48%

direct speech, development, 10%

cues, speakers, test

addressees

Redewiedergabe fiction, direct, indirect, German 24,033 76% train, 16% 33%

news free indirect and development, 9%

reported speech, test

speaker, cues

Quotebank (man- news speaker, direct English 9,071 test 30%

ual) speech

QUAC news speaker, direct Portuguese 11,007 test 11%

speech

PolNeAR news speaker, cues, at- English 34,153 test 59%

tributions

Slovenian parlia- news sentence-level bi- Slovenian 744 test 43%

mentary news nary labels





Figure 1: Flowchart of Data Preprocessing, Model Training and Evaluation Processes for Sentence-Level Reported Speech

Classification.



the best model’s outputs for Slovenian. Preprocessing, training, ) on the media and the "illegal non-funding of the Press Agency."

and evaluation steps are visualized in 1. as well as implied quotes (e.g., There will be more than 300,000

recipients, he emphasized. 169 million euros will have to be paid

3.2 Training and Test Data out.).

Our experiments were based on existing annotated reported

speech datasets and a small Slovenian dataset. The training data

included sections from RiQuA and Redewiedergabe, both large 3.3 Evaluation Procedure

datasets with labels for direct and indirect speech. For CroSlo- The models’ performance on the test datasets was calculated with

Engual BERT training, the Redewiedergabe data was machine an F-score. A baseline of assigning a positive label to all examples

translated into English. Testing was conducted on the test sec- was calculated for all test datasets. The models’ results on the

tions of RiQuA, Redewiedergabe, the entire Portuguese corpus test datasets were compared with a Friedman’s test as suggested

QUAC, and the manually annotated portion of the English Quote- in the literature [5].

bank corpus. Additionally, we manually annotated 10 Slovenian The best Slovenian model’s predictions were reviewed with

news articles from RTV Slovenia. The datasets are summarized close reading. The error typology consisted of direct speech, in-

in Table 1. direct speech, speech fragments, annotation errors, annotation

The Slovenian dataset comprised 10 parliamentary news texts, errors and and tags. were unrelated other Direct speech fragments

covering various reporting strategies. Retrieved articles were sentences part of multi-sentence direct speech quotations. Anno-

split into sentences and annotated. Sentences were considered were examples with annotations inconsistent with tation errors

reported speech if they included direct or indirect speech cued by the definition described in . For examples, Section 3.2 unrelated

a reporting clause or prepositional phrase. We excluded nominal- close reading revealed no clear misclassification cause. was Other

izations and phrasal quotes (e.g., used for examples that did not fit any of the mentioned categories. They emphasized the pressures

Transfer Learning for Reported Speech Detection in Slovenian Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



3.4 Training Settings

XLM-R and mBERT were used as base models with the default

training settings from the library with the excep-transformers

tion of using 16 gradient accumulation steps and freezing the

bottom 8 layers of all models. The latter reduces the training

time without significant performance drops (Kovaleva idr., 2019;

Merchant idr., 2020). Additionally, a Slovenian-Croatian-English

BERT model was trained on English machine-translated data

from Redewiedergabe.



4 Results



4.1 Model Results

The model performance varies based on the congruence between

the language and precise task definitions in each dataset. The Figure 3: False Negatives from the CroSloEngual BERT

differences between model predictions were not statistically sig- Classifier.

2

nificant (𝜒 = 9.66; df = 5; n = 8; p = 0.14) so post-hoc tests were not

𝐹

performed. As Table 2 demonstrates, the XLM-R model trained

unmarked examples of direct or indirect speech (9.1%). The dis-

on both RiQuA and Redewiedergabe performed well across the

datasets with an F-score of 80.5 and 77.6 on the Redewiedergabe tribution of categories identified in the sample of false positives

are illustrated in Figure 2. The most common errors in the 73

and RiQuA test set, respectively. The high results from train-

false negative examples were instances of indirect speech (34.2%

ing on combined data suggests the RiQuA and Redewiedergabe

datasets may benefit from additional or complementary data, of false negatives) and prepositional queing of reported speech

(27.4%). The remainder were instances of direct speech, direct

at least when using cross-lingual transfer learning. The most

successful strategy for Slovenian data was training on RiQuA speech fragments and annotation errors representing 11%, 8.2%

and 9.6% of the false negatives, respectively. The annotation

and English machine-translated Redewiedergabe data using the

CroSloEngual BERT model, reaching a F-score of 66.8. We did errors included nominalizations and statements reported as ad-

jective complements (The speaker was happy that the provisions

not evaluate the impact of using translated training data with

mBERT and XLM-R. not included in our annotation schema. Figure 3 were accepted)

summarizes the identified false negative categories .





5 Discussion



This paper presents the development of a reported speech classi-

fier, tested through a small annotated Slovenian dataset and man-

ual error analysis. Cross-lingual transfer learning from the anno-

tated RiQuA and Redewiedergabe datasets achieved an F-score

of 66.8 on a small manually annotated dataset of Slovenian news

of parliamentary sessions using the base CroSloEngual model

with RiQuA and English machine-translated Redewiedergabe

1

training data . This these results corroborate the observation

that language models trained on a limited number of languages

may outperform less specialized ones such as mBERT and XLM-R

[20]. The major source of errors were false positives (23.4% of all

sentences) for which no systematic pattern was discernible in the

majority (72.9%) of examples. Instances of indirect speech and

Figure 2: False Positives from the CroSloEngual BERT Clas- prepositional queing of statements were overrepresented in the

sifier. false negatives, accounting for 61.6% of false negatives. Although

rare, nominalizations were present in both false positives and

false negatives and should be considered in future annotation

guidelines. These obeservations indicate reported speech clas-

4.2 Error Analysis Results sifiers may benefit form approaches for addressing imbalanced

The results from CroSloEngual BERT on Slovenian data were classes.

analyzed further. False positives were more common than false

negatives, representing 23.4% and 9.8% of all examples ( = 744), n 6 Conclusion respectively. Close reading of a sample of 100 false positives

This study developed a sentence-level reported speech classifier

did not show a definite pattern for most (72.9%) of them. These

for Slovenian news texts using cross-lingual transfer learning.

examples were clearly unrelated to reported speech, although

By leveraging existing multilingual models (mBERT, XLM-R, and

some did include words lexically related to reporting verbs (e.g.

CroSloEngual BERT) with the English and German datasets Ri-

The proposed law is still under discussion). The second category QuA and Redewiedergabe, we demonstrated that sentence-level

of false positives were nominalizations of reported statements

(13.1%) not included in our annotation schema. The final source 1

The fine-tuned CSE model is available on the Hugging Face Hub under the name

of false positives were annotation errors consisting of wrongly zo-fi/rep-sp-CSE-rwg-riq.



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Z. Fijavž



Table 2: Model Performances across Datasets (F-scores).



Redewiedergabe RiQuA PolNeAR QUAC Quotebank Slovenian dataset



Positive by default 52.1 60.6 74.2 19.5 45.8 60.3

mBERT+Both 77.5 77.4 73.1 40.5 53.5 63.2

mBERT+RiQuA 68.2 76.9 72.6 31.1 52.6 39.1

mBERT+RWG 78.4 70.4 65.5 43.4 49.1 63.2

XLM-R+Both 80.5 77.6 70 38.8 57.7 63.2

XLM-R+RiQuA 66.6 76.7 73.6 25.5 53.7 60.3

XLM-R+RWG 80.9 70.7 66.4 43.9 50 63.2

CroSloEngBERT+Both+MT 54 76.6 73 24 52.5 66.8



classification can detect some aspects of reported speech in Slove- [10] Edward Newell, Drew Margolin, and Derek Ruths. 2018. An Attribution Re-

lations Corpus for Political News. In Proceedings of the Eleventh International

nian. However, the performance estimates are limited due to the

Conference on Language Resources and Evaluation (LREC 2018). LREC 2018.

small size of the Slovenian testing set and the limited definition

Nicoletta Calzolari et al., editors. European Language Resources Association

used for the annotations. Future research should focus on de- (ELRA). Retrieved Apr. 10, 2024 from https://aclanthology.org/L18- 1524.

[11] Mojca Pajnik and Marko Ribać. 2021. Medijski populizem in afektivno

veloping a Slovenian annotated dataset, refining the annotation

novinarstvo: časopisni komentar o »begunski krizi«. , Javnost - The Public

schema for multiple use cases, and exploring additional modeling (Dec. 14, 2021). Retrieved Apr. 24, 2024 from https://www.tandf online.com

features such as encoding broader sentence contexts. This work /doi/abs/10.1080/13183222.2021.2012943.

[12] Sean Papay and Sebastian Padó. 2020. RiQuA: A Corpus of Rich Quotation

contributes a provisional tool for computational discourse analy-

Annotation for English Literary Text. In Proceedings of the Twelfth Language

sis of Slovenian media texts. Further development is necessary . LREC 2020. Nicoletta Calzolari et al., Resources and Evaluation Conference

editors. European Language Resources Association, 835–841. isbn: 979-10-

for its application in more nuanced tasks.

95546-34-4. Retrieved Apr. 21, 2024 from https://aclanthology.org/2020.lrec-

1.104.

Acknowledgements [13] Silvia Pareti, Tim O’Keefe, Ioannis Konstas, James R. Curran, and Irena Ko-

prinska. 2013. Automatically Detecting and Attributing Indirect Quotations.

This work was supported by the Slovenian Research Agency Proceedings of the 2013 Conference on Empirical Methods in Natural Lan-

In

grants via the core research programs Equality and Human Rights . EMNLP 2013. David Yarowsky, Timothy Baldwin, Anna guage Processing

Korhonen, Karen Livescu, and Steven Bethard, editors. Association for Com-

in the Times of Global Governance (P5-0413) and Hate Speech

putational Linguistics, 989–999. Retrieved Apr. 17, 2024 from https://aclant

in Contemporary Conceptualizations of Nationalism, Racism, hology.org/D13- 1101. Gender and Migration ( J5-3102). [14] Marta Ercília Mota Pereira Quintão. 2014. Quotation A ttribution for Por-

tuguese News Corpora. In Retrieved Apr. 21, 2024 from https://www.seman

ticscholar.org/paper/Quotation- A- ttribution- f or- Portuguese- News- Corp

References ora- Quint%C3%A3o/69f ea7d030d5e71b973ec67aa897a7c9aadadac2.

[15] Masaki Shibata. 2023. Dialogic Positioning on Pro-Whaling Stance: A Case

[1] Fatemeh Torabi Asr, Mohammad Mazraeh, Alexandre Lopes, Vasundhara

Study of Reported Speech in Japanese Whaling News. , 43, Japanese Studies

Gautam, Junette Gonzales, Prashanth Rao, and Maite Taboada. 2021. The

1, (Jan. 2, 2023), 71–90. doi: 10.1080/10371397.2023.2191839.

Gender Gap Tracker: Using Natural Language Processing to measure gender

bias in media. , 16, 1, (Jan. 29, 2021), e0245533. doi: 10.1371/journ

PLOS ONE [16] Michael Short. 1988. Speech presentation, the novel and the press. In The

al.pone.0245533. Taming of the Text

. Willie Van Peer, editor. Routledge. isbn: 978-1-315-54452-

6.

[2] Christian Baden, Christian Pipal, Martijn Schoonvelde, and Mariken A. C. G

[17] Alexander Spangher, Nanyun Peng, Jonathan May, and Emilio Ferrara.

van der Velden. 2022. Three Gaps in Computational Text Analysis Meth-

ods for Social Sciences: A Research Agenda.

Communication Methods and 2023. Identifying Informational Sources in News Articles. Version 1. doi:

Measures 10.48550/ARXIV.2305.14904. , 16, 1, (Jan. 2, 2022), 1–18. doi: 10.1080/19312458.2021.2015574.

[18] Stef Spronck and Daniela Casartelli. 2021. In a manner of speaking: how

[3] Annelen Brunner, Stefan Engelberg, Fotis Jannidis, Ngoc Duyen Tanja Tu,

and Lukas Weimer. 2020. Corpus REDEWIEDERGABE. In

Proceedings of the reported speech may have shaped grammar. , 6, Frontiers in Communication

Twelfth Language Resources and Evaluation Conference 624486. . LREC 2020. Nicoletta

[19] Sara Stymne and Carin Östman. 2022. SLäNDa version 2.0: Improved and

Calzolari et al., editors. European Language Resources Association, 803–812.

Extended Annotation of Narrative and Dialogue in Swedish Literature. In

isbn: 979-10-95546-34-4. https://aclanthology.org/2020.lrec- 1.100.

[4] Alexis Conneau et al. 2020. Unsupervised Cross-lingual Representation Proceedings of the Thirteenth Language Resources and Evaluation Conference.

Learning at Scale. In

Proceedings of the 58th Annual Meeting of the Associ- LREC 2022. Nicoletta Calzolari et al., editors. European Language Resources

ation for Computational Linguistics Association, 5324–5333. Retrieved Apr. 21, 2024 from https://aclanthology.o . ACL 2020. Dan Jurafsky, Joyce Chai,

rg/2022.lrec- 1.570.

Natalie Schluter, and Joel Tetreault, editors. Association for Computational

[20] Matej Ulčar and Marko Robnik-Šikonja. 2020. FinEst BERT and CroSlo-

Linguistics, 8440–8451. doi: 10.18653/v1/2020.acl- main.747.

[5] Janez Demšar. 2006. Statistical Comparisons of Classifiers over Multiple Engual BERT. In Text, Speech, and Dialogue (Lecture Notes in Computer

Data Sets. , 7, (Dec. 1, 2006), 1–30. Science). Petr Sojka, Ivan Kopeček, Karel Pala, and Aleš Horák, editors. The Journal of Machine Learning Research

[6] Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. Springer International Publishing, Cham, 104–111. isbn: 978-3-030-58323-1.

BERT: Pre-training of Deep Bidirectional Transformers for Language Under- doi: 10.1007/978- 3- 030- 58323- 1_11.

standing. In [21] Proceedings of the 2019 Conference of the North American Chapter Timoté Vaucher, Andreas Spitz, Michele Catasta, and Robert West. 2021. of the Association for Computational Linguistics: Human Language Technolo- Quotebank: A Corpus of Quotations from a Decade of News. In Proceedings gies, Volume 1 (Long and Short Papers) of the 14th ACM International Conference on Web Search and Data Mining . NAACL-HLT 2019. Jill Burstein,.

Christy Doran, and Thamar Solorio, editors. Association for Computational WSDM ’21: The Fourteenth ACM International Conference on Web Search

Linguistics, 4171–4186. doi: 10.18653/v1/N19- 1423. and Data Mining. ACM, 328–336. isbn: 978-1-4503-8297-7. doi: 10.1145/343

[7] Gabriel Dvoskin. 2020. Reported speech and ideological positions: the so- 7963.3441760.

cial distribution of knowledge and power in media discourse. [22] Bakhtiniana: M. Wynne. 1996. Speech, Thought and Writing Presentation Corpus. Re-

Revista de Estudos do Discurso trieved Apr. 21, 2024 from https://ora.ox.ac.uk/objects/uuid:6caa73c1- d283- , 15, 193–213.

4d51- a78f - 55df 69bae986.

[8] Zoran Fijavž and Darja Fišer. 2021. Citatnost in reprezentacija v spletnem

migracijskem diskurzu. In . Maja Bitenc, Marko

Sociolingvistično iskrenje [23] Dian Yu, Ben Zhou, and Dong Yu. 2022. End-to-End Chinese Speaker Identi-

fication. In Proceedings of the 2022 Conference of the North American Chapter

Stabej, and Žejn Andrejka, editors. Založba Univerze v Ljubljani. Retrieved

Apr. 3, 2024 from https://ebooks.uni- lj.si/ZalozbaUL/catalog/view/259/370 of the Association for Computational Linguistics: Human Language Technolo-

/6011. gies

. NAACL-HLT 2022. Marine Carpuat, Marie-Catherine de Marneffe, and

Ivan Vladimir Meza Ruiz, editors. Association for Computational Linguistics,

[9] Elizabeth Holt. 1996. Reporting on Talk: The Use of Direct Reported Speech

in Conversation. , 29, 3, (July 1,

Research on Language and Social Interaction 2274–2285. doi: 10.18653/v1/2022.naacl- main.165.

1996), 219–245. doi: 10.1207/s15327973rlsi2903_2.



What kind of ESG is profitable? Connecting company



performance to ESG terms in financial reports



Luka Andrenšek Katarina Sitar Šuštar

trovato@corporation.com University of Ljubljana

Jožef Stefan Institute Ljubljana, Slovenia

Ljubljana, Slovenia katarina.sitar@ef.uni- lj.si



Senja Pollak Matthew Purver

Jožef Stefan Institute Jožef Stefan Institute

Ljubljana, Slovenia Ljubljana, Slovenia

senja.pollak@ijs.si matthew.purver@ijs.si



ABSTRACT in the last few years on applying computational machine learning

and statistical methods to ESG analysis (see e.g. the recent review

In this paper, we examine the relationship between the discussion

by Lim [9]).

of Environmental, Social and Governance (ESG) in companies’

However, much of this analysis examines numerical company

annual financial reports and their financial performance. Specifi-

performance data and categorical metadata; our interest is in

cally, we analyse the companies’ use of specific ESG terms along-

developing and applying natural language processing (NLP) tech-

side the performance metric, sector-normalized Return on Assets

nologies to not only help automate analyses, but allow under-

(ROA). Our motivation is to determine whether companies fre-

standing of how human actors discuss and understand the im-

quently mentioning terms such as “gender”, “equality”, “talent”,

portant and meaning of ESG aspects.

and “innovation” in their reports demonstrate a higher annual

Application of NLP in finance is not new: for example, topic

ROA compared to those that rarely used these terms. To explore

modelling has been used to predict company performance and

this, we used existing datasets with reports and performance met-

investigate strategies [14, 7]. Recent work also includes applica-

rics from 348 companies, covering the years from 2009 to 2021. In

tion to ESG aspects: Nugent et al. [12] automatically extract news

order to better examine differences, we then selected companies

about ESG controversies, and Lee et al. [8] analyse sentiment

whose ROA significantly differed from the average (either higher

on ESG issues. Closer to our interests, Purver et al. [13] investi-

or lower), allowing for a more pronounced examination of the

gated how the use of ESG terms by companies has changed over

impact of ESG term usage on financial performance. The filtered

time. By analysing and annotating a set of existing resources,

dataset consisted of 107 companies, with a total of 427 reports;

they defined a set of 93 ESG terms categorised into 5 core ESG

split into two sections representing higher and lower performing

areas. They then showed how these terms can be used to anal-

companies. We then used an existing list of ESG terms derived

yse changes in reporting, by analysing a collection of company

from a range of separate data sources, and applied a basic sta-

annual reports, collated over a period of 8 years, using language

tistical n-gram language model to extract the probabilities of

modelling and distributional methods to reveal changes in the

each ESG term’s occurrence in each of the higher- and lower-

frequency and in the usage of the ESG terms.

performing dataset sections. Results show that while certain sets

Here, we are interested not in changes in ESG discussion over

of ESG concepts correlate with higher financial performance,

time, but in whether and how the reporting of ESG aspects is

others do the opposite, and give some initial interpretation into

connected to financial performance. We take Purver et al. [13]’s

the light this sheds on company reporting behaviour.

resources and methods as a starting point, but augment the fi-



KEYWORDS nancial report text data with available metadata on financial

performance, allowing us to compare how ESG reporting varies

financial report analysis, language modelling, environmental,

between more and less well-performing companies.

social and governance reporting



1 INTRODUCTION & RELATED WORK 2 DATA AND METHODS



There is increasing interest in the behaviour of companies in 2.1 Hypotheses the area of Environmental, Social and Governance (ESG) criteria,

In general, we expect increased probability of appearance of ESG

including a company’s environmental impact (Environmental),

terms in the annual reports from the more profitable firms, based

relationships with the community including employees, suppliers

on a number of factors. In general, overall high ESG performing

and customers (Social), and leadership structures including exec-

companies exhibit high financial performance [1, 5]; although we

utive pay and shareholder rights (Governance). Although until

note that the link between high ESG score performance and men-

recently, ESG analyses were almost entirely performed manually

tion of ESG terms is not guaranteed to be straightforward. More

by experts (see e.g. [10]), there has been a large amount of work

specifically, during the period between 2010-2020 analysed here,

Permission to make digital or hard copies of part or all of this work for personal

there was a growing emphasis on corporate social responsibility

or classroom use is granted without fee provided that copies are not made or

(CSR) and sustainability. Investors, consumers, and other stake-

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this holders increasingly prioritised companies that demonstrated a

work must be honored. For all other uses, contact the owner /author(s).

commitment to innovation, diversity, and environmental sustain-

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

ability [11, 2]. Busru and Shanmugasundaram [3] find that firms

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.sikdd.3 closely engaging in fostering innovation, attracting top talent,



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Andrenšek et al.



Year # Reports # Words

’positive’ group included reports with an ROA of at least 0.2,

2012 178 12.5M

reflecting very good yearly performance.

2013 181 14.0M

Subsequently, we employed a statistical n-gram language model

2014 184 15.0M 3 (using NLTK) to analyze the occurrence of each ESG term. For 2015 196 16.3M

each term, we calculated the probability of its occurrence in pos-

2016 198 17.5M + −

itive reports (𝑝 ) and in negative reports (𝑝 ), and the difference

2017 200 18.4M + − −

(𝑝 𝑝 ). Terms with a large difference in these probabilities are

2018 200 19.6M more strongly associated with positive reports than with nega-2019 202 21.2M

tive ones, and vice versa: terms with a large negative difference

total 1539 134.6M are common in negative reports but rare in positive ones. We

Table 1: Number of annual reports available by year conducted this analysis for both unigrams and bigrams.



3 RESULTS AND DISCUSSION

The results for 1- and 2-grams are shown in Figures 1 and 2 below

promoting gender and diversity initiatives, could confer a com- 4

(3- and 4-grams showed no clear interpretable associations). As

petitive advantage over the industry peers. Furthermore, some

hypothesized, many ESG terms show a strong association with

policy and regulatory changes (e.g. the 2018 UK Corporate Gov-

positive performance, with many of these being core terms as-

ernance Code, the 2014 EU Directive on Non-Financial Reporting,

sociated with human resources ( , ), with social innovation talent

Carbon Disclosure Project (CDP)) directly or indirectly encour- gender diversity renewable car-aspects ( , ), environmental aspects ( ,

aged companies to address issues related to diversity, gender

bon footprint, environmental impact) and overall ESG descriptors

equality, and environmental sustainability. ethical

( ). However, many terms are conversely (and contrary to

our general hypothesis) associated with negative performance,

2.2 Data and pre-processing including, again, terms across various ESG categories including

To test this hypothesis, we build on the resources and methods carbon emissions energy efficiency greehouse environmental ( , , ),

of Purver et al. [13], who provide a dataset of annual reports human resources ( , ) and general ESG mental health wellbeing

from FTSE350 companies over the years 2012-2019, based on the governance descriptors ( ).

FTSE350 list as of 25th April 2020 and obtained from the publicly

However, by combining these terms with recent work in clus-

accessible collection at www.annualreports.com. The reports are

tering and describing ESG terms [4], we can shed more light

already converted to plain text, and we use their publicly avail-

on which categories seem to be more positive and which more

able tools to tokenize the collection into words and build ngrams negative. Ferjancic et al. [4], using the same dataset and ESG

of length 1-4 padded with sentence start and end symbols; the

term list [13], perform a further topic analysis using BERTopic

dataset size is reported in Table 1 below (taken from [13]). We [6], in which they derive 30 ESG-related topics and 6 higher-level

use their set of ESG terms, defined via a process of extracting

clusters of ESG concepts; they then examine the correlations

candidate terms from a set of public ESG definitions and tax-

between these ESG topics and company ESG scores as obtained

onomies, asking financial expert annotators to label them as to

from external analysts. We align our ESG terms with Ferjancic

their representativeness as ESG terms and their ESG subcategory,

et al. [4]’s 30 topics by matching against the words most asso-

and keeping the terms with high inter-annotator agreement (see ciated with each topic (if a term appears in the top 10 words

[13] for details).

associated with a topic, we take the term and topic as aligned);

we can then compare our positive/negative associations with Fer-

2.3 Financial performance analysis jancic et al. [4]’s correlations with company ESG scores. Table 2

The reports were then linked to financial indicators for the re- shows this alignment for our most positive and negative bigram

spective year and company. The data on company fundamentals terms here, with the topic labels and an indication of the strength

was obtained from the Refinitiv EIKON Datastream. Each entry and direction of correlation with overall company ESG scores, as 1

contained annual financial indicators, as well as the companies’ given by [4].

industry and sector codes. The main variable of interest was Given this, we see some systematic groupings. , Climate change

normalized, averaged as defined below: as part of the ‘climate risk and policy’ topic, as well as 2 return on assets (ROA) supply

chain and human trafficking as part of the ‘human rights’ topic,

𝑁 𝑒𝑡 𝐼 𝑛𝑐𝑜𝑚𝑒 𝐵𝑜𝑡 𝑡 𝑜𝑚𝐿𝑖𝑛𝑒 −

represent the themes that appear to be, across different industries,

+( (𝐼 𝑛𝑡 𝑒𝑟 𝑒𝑠𝑡 𝐸𝑥 𝑝𝑒𝑛𝑠𝑒𝑂𝑛𝐷𝑒𝑏𝑡 − 𝐼 𝑛𝑡 𝑒𝑟 𝑒𝑠𝑡𝐶𝑎𝑝𝑖𝑡 𝑎𝑙 𝑖𝑧𝑒𝑑)

related to high company ESG scores. A similar observation holds

×(1 − 𝑇 𝑎𝑥 𝑅𝑎𝑡 𝑒))

for , and , which gender balance gender pay environmental impact

𝐴𝑣 𝑒𝑟 𝑎𝑔𝑒𝑂 𝑓 𝐿𝑎𝑠𝑡 𝑌 𝑒𝑎𝑟 ′ ′ 𝑠𝐴𝑛𝑑𝐶𝑢𝑟 𝑟 𝑒𝑛𝑡 𝑌 𝑒𝑎𝑟𝑠𝑇 𝑜𝑡 𝑎𝑙 𝐴𝑠𝑠𝑒𝑡 𝑠 all fall in a group of topics which are strongly and significantly

After extracting financial reports with available ROA data, we correlated with high ESG scores throughout different industries.

Overall high ESG performing companies exhibit high financial

categorized the financial reports into two groups, in order to

examine differences in the associated reports’ use of ESG terms. performance [1, 5], therefore our results for terms such as climate

The distribution of ROA shows a heavy concentration around change supply chain human trafficking , and are not surprising: as

the mean, so in order to derive two distinctive groups we took indicators of topics associated with high ESG, they are good terms

the two extremes and excluded the central group around the for tracking these ESG aspects associated with high financial

mean. The ‘negative’ group comprised reports with a yearly ROA performance.

less than -0.2, indicating very poor performance. Conversely, the

3

https://www.nltk.org/

1 4

https://www.refinitiv.com Note that these figures show differences in absolute probabilities: magnitudes are

2

We use this normalization and averaging to smooth and remove one-off effects. comparable within 1-grams, and within 2-grams, but not between 1- and 2-grams.

What kind of ESG is profitable? Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia





Figure 1: Difference in probability between positive and negative reports 𝑝+ −𝑝− for the most positive and negative unigram

ESG terms.





Figure 2: Difference in probability between positive and negative reports 𝑝+ − 𝑝− for the most positive and negative bigram

ESG terms.



Looking at the terms with low values which are associated . It seems that better performing companies use efficiency carbon

with low RoA, and are instead of , and discuss more on the waste management corporate responsibility footprint carbon emissions

associated with topics, for which in some industries proportion use of than on , renewable energy energy use energy efficiency

of these correlate with ESG scores significantly positively and in and/or . In future work, we plan to analyse the use of fossil fuels other industries this correlation is significantly negative. Based these terms in more depth, including analysis of the lexical and

on overall correlation between ESG scores and topic proportions topical contexts in which they appear, and adding techniques

across different industries, these two topics are among the third such as sentiment and topic analysis to shed more light on these

of the topics for which negative correlation between the topic distinctions.

proportion and ESG score prevails. Due to the aforementioned

correlation between ESG and financial performance it is therefore ACKNOWLEDGEMENTS understandable that these terms are associated with mention in

The authors thank the reviewers for helpful suggestions, and ac-

annual reports of companies with low RoA. Overly extensive

knowledge financial support from the Slovenian Research Agency

discussion on specific topics (such as ‘waste management’ and

for research core funding (No. P2-0103), as well as for funding of

‘corporate responsibility’) can negatively impact ESG score (see Quantitative and qualitative analysis of the

the research project

[4]) which can by analogy of ESG and financial performance [1, unregulated corporate financial reporting (No. J5-2554).

5] hold for companies with low RoA.

There is a surprising number of bigrams in both the high RoA REFERENCES

and low RoA groups which seem to be associated with the same [1] Nisar Ahmad, Asma Mobarek, and Naheed Nawazesh Roni. 2021. Revisiting

topic, namely ‘climate footprint and energy management’. For the impact of ESG on financial performance of FTSE350 UK firms: static and



companies with high RoA, these terms are and Ethics

carbon footprint dynamic panel data analysis. Accounting, Corporate Governance & Business

. doi: 10.1080/23311975.2021.1900500.

renewable energy, and for companies with low RoA, the terms are

fossil fuels, carbon emissions, energy use, air quality and energy



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Andrenšek et al.



2 grams Term/ROA Topic Topic/ESG score

correlation correlation



Supply chain + Human rights ++



Business model + Customer services, People and culture +; -



Gender balance + Diversity and inclusion ++



Environmental impact + General ESG +



Carbon footprint + Climate footprint and energy management =



Gender pay + Diversity and inclusion ++



Climate change + Climate risk and policy ++



Human trafficking + None directly related, in broader context in Hu- ++

man rights



Working environment + People and culture -



Renewable energy + Climate footprint and energy management =



Waste management - Waste management –



Fossil fuels - No explicit match; contextually appears in Cli- =

mate footprint and energy management



Corporate responsibility - Corporate governance –



Carbon emissions - Climate footprint and energy management =



Mental health - Health and safety +



Energy use - Climate footprint and energy management -



Air quality - No explicit match; contextually appears in Cli- -

mate footprint and energy management



Energy efficiency - Climate footprint and energy management -



Product safety - Health and safety =

Table 2: Selected ESG terms with their ROA correlation direction (+/−), topic according to [4], and topic/ESG score correlation

strength (+ + /+/= /−/−−) as calculated by [4].



[2] A. C. Amason and H. J. Sapienza. 2012. The effects of top management team rep. Available from https://iri.hks.harvard.edu/links/transparency- p

size and interaction norms on cognitive and affective conflict. erf ormance- industry- based- sustainability- reporting- key- issues. Journal of

Management, 23, 495–516. Hauser Center for Nonprofit Organizations at Harvard University.

[3] S. A. Busru and G. Shanmugasundaram. 2017. Effects of innovation invest- [11] M. Marzook and B. Al Ahmady. 2022. Linking organisational performance

ment on profitability and moderating role of corporate governance: empiri- and corporate social responsibility. European Jnl. of Business and Manage-

cal study of indian listed firms. , 10, , 7, 335–343, 3. https://doi.org/10.24018/ejbmr.2022.7.3.1466. Indian Journal of Corporate Governance ment Research

97–117, 2. https://doi.org/10.1177/0974686217730938. [12] T. Nugent, N. Stelea, and J. L. Leidner. 2020. Detecting ESG topics using

[4] Ursa Ferjancic et al. forthcoming. Textual analysis of corporate sustainability domain-specific language models and data augmentation approaches. (2020).

reporting and corporate ESG scores. under review. (Forthcoming). http://arxiv.org/abs/2010.08319.

[5] Gunnar Friede, Timo Busch, and Alexander Bassen. 2015. ESG and financial [13] Matthew Purver, Matej Martinc, Riste Ichev, Igor Lončarski, Katarina Sitar

performance: aggregated evidence from more than 2000 empirical studies. Šuštar, Aljoša Valentinčič, and Senja Pollak. 2022. Tracking changes in ESG

Journal of Sustainable Finance & Investment, 5, 4, 210–233. doi: 10.1080/204 representation: initial investigations in UK annual reports. In Proceedings of

30795.2015.1118917. the First Computing Social Responsibility Workshop within the 13th Language

[6] Maarten Grootendorst. 2022. BERTopic: neural topic modeling with a class- . Mingyu Wan and Chu-Ren Huang, Resources and Evaluation Conference

based TF-IDF procedure. (2022). https://arxiv.org/abs/2203.05794 arXiv: editors. Marseille, France, (June 2022), 9–14. https://aclanthology.org/2022.c

2203.05794 . srnlp- 1.2. [cs.CL]

[7] M. Jagannathan, D. Roy, and V. S. K. Delhi. 2022. Application of NLP-based [14] W. Xu and K. Eguchi. 2021. Topic embedding regression model and its appli-

topic modeling to analyse unstructured text data in annual reports of con- cation to financial texts. In Proceedings of the Third Workshop on Financial struction contracting companies. , 10, 2, 97–106. , 15–21. CSI Transactions on ICT Technology and Natural Language Processing

[8] H. Lee, S. H. Lee, K. R. Lee, and J. H. Kim. 2023. Esg discourse analysis

through bertopic: comparing news articles and academic papers. Computers,

Materials & Continua, 75, 3, 6023–6037.

[9] Tristan Lim. 2024. Environmental, social, and governance (esg) and artificial

intelligence in finance: state-of-the-art and research takeaways. Artificial

Intelligence Review, 57, 76. doi: 10.1007/s10462- 024- 10708- 3.

[10] Steve Lydenberg, Jean Rogers, and David Wood. 2010. From Transparency to

Performance: Industry-Based Sustainability Reporting on Key Issues. Tech.



Classification of Patents Into Knowledge Fields: Using a



Proposed Knowledge Mapping Taxonomy (KnowMap)



Elham Motamedi Inna Novalija Luis Rei

elham.motamedi@upr.si inna.koval@ijs.si luis.rei@ijs.si

University of Primorska Jožef Stefan Institute Jožef Stefan Institute

Koper, Slovenia Ljubljana, Slovenia Ljubljana, Slovenia



Abstract Table 1: Example of a sequence of codes across different

levels of the CPC hierarchy

Various platforms, including patent systems and repositories like

GitHub and arXiv, support knowledge dissemination across do-

mains. As knowledge increasingly spans multiple disciplines, CPC Code Title

there is a need to track innovations that intersect various fields.

Section H Electricity

Despite available data, a comprehensive knowledge taxonomy for

Class H03 Electronic circuitry

effectively tracking innovations across domains is lacking. Devel-

Subclass H03C Modulation

oping such a taxonomy and employing automated classification

Group H03C3/00 Angle modulation

methods will enhance the ability to track shared knowledge.

Subgroup H03C3/005 Circuits for asymmetric modulation

In this work, we first developed a knowledge taxonomy based

on the CPC schema. We formulated the classification of textual

data into defined knowledge fields as a multi-label problem. Then,

study, we created a knowledge field taxonomy by merging CPC’s

we evaluated the effectiveness of the classification models by

detailed classes into a more abstract representation. This taxon-

fine-tuning pre-trained transformer language models. The multi-

omy not only serves as a framework for knowledge representa-

label framework enables the tracking of knowledge trends at the

tion but also offers a benchmark for patent classification systems.

intersection of various disciplines.

While some studies address the issue of numerous class labels by



Keywords excluding less-represented classes or truncating hierarchies [24],

a consistent benchmark taxonomy has been lacking. Since our

Knowledge Taxonomy, Knowledge Tracking, Patent Classifica-

proposed knowledge taxonomy aligns with the CPC schema, it

tion, Hierarchical Classification, Multi-label Classification

is able to provide a benchmark for future studies, facilitating the

comparison of different models.

1 Introduction In summary, our paper’s contribution is the proposal of a

According to the World Intellectual Property Organisation (WIPO), knowledge field taxonomy, KnowMap, which aligns with the

a patent is an exclusive right granted for an invention, providing widely used CPC schema. The KnowMap merged several class

legal protection to the inventor while simultaneously benefiting labels within the CPC schema based on the scope of the knowl-

1

society by making the invention publicly accessible . Each year, edge field and the number of patents associated with each class.

2

patent offices receive numerous patent applications that need to The KnowMap taxonomy is available online . In this study, we

be processed [13].To ensure the novelty of patent applications, in- also performed a classification task to categorise patents into the

ventors should also be able to search existing patents. Organising fine-grained classes defined by our proposed taxonomy.

patents with unique codes in a hierarchical structure aids efficient

retrieval and aligns with natural human navigation, starting from 2 Related Work

broad categories and narrowing down to specifics[21]. Among

Patent documents contain various types of information, including

these hierarchical structures, the CPC system is widely recog-

text, diagrams, plots, and references to other patents or scientific

nised [6]. The CPC codes are organised as a taxonomy, meaning

publications [20]. The textual content of a patent is divided into

that each entity in the lower level is the detail group of the parent.

several sections, such as the title, abstract, claim, and description

A patent can be assigned to one or more labels by the experts

[11]. The title and abstract are shorter than the description but

in patent offices [8, 18]. In the first level of the CPC hierarchy,

still provide relevant information for classification. Li et al. [15]

there are nine sections, which are divided into classes, subclasses,

evaluated various lengths of the abstract and title, finding that

groups, and subgroups. Each level of this hierarchy can have

using the first 100 words of title and abstract resulted in the best

several codes ending in approximately 250,000 classification la-

classification performance in their study.

bels [11]. An example of the hierarchical structure of CPC code

Various classification systems exist for organising patents [6].

is provided in Tab. 1.

In this work, we focus on the CPC schema. The hierarchical repre-

The CPC schema’s top level has only nine sections, but the

sentations help organise patents and facilitate efficient searching.

number of groups increases substantially at lower levels. In this

Kamateri et al. [11] discussed several potential challenges that

1 artificial intelligence technologies face in patent classification.

https://www.wipo.int/portal/en/

One such challenge is the extensive number of class labels. As an

Permission to make digital or hard copies of all or part of this work for personal

example, the IPC contains approximately 86,000 classes, while

or classroom use is granted without fee provided that copies are not made or

distributed for profit or commercial advantage and that copies bear this notice and the CPC has around 250,000. the full citation on the first page. Copyrights for third-party components of this

Patent classification is a multi-label classification problem

work must be honored. For all other uses, contact the owner /author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia since every patent can belong to several knowledge fields [18,

© 2024 Copyright held by the owner/author(s).

2

https://doi.org/10.70314/is.2024.sikdd.19 https://github.com/elmotamedi/KnowMap- Taxonomy



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Motamedi et al.



10]. Given the large number of classes at the lowest level of the or higher were considered duplicates. To generate the hash sig-

taxonomy tree, the performance of automatic models in predict- natures in MinHash, we used 128 permutations. For the n-gram

ing such granular categories is limited. Various models have been representation, we used a range of 1 to 3, incorporating 1-grams,

used to classify patents in a multi-label setting, ranging from clas- 2-grams, and 3-grams.

sical machine learning models to deep learning models [15, 5,

8]. Several previous studies have focused on higher levels of the 3.2 Refining Hierarchical Structure Through

hierarchy, limiting classification to broader categories such as Group Merging

sections, classes, or subclasses within the taxonomy [3]. Bekamiri

The hierarchical structure of the CPC groups was refined at each

et al. [3] fine-tuned the SBERT model to predict labels at the sub-

level of the tree. We started with nine sections at the top level (i.e.,

class level (i.e., 663 class labels) using a multi-label formulation. level 1 level 2 ), which were preserved. At subsequent levels (i.e., to

They achieved F1-score of 66%, outperforming previous studies level 4

), groups were merged by manual analysis based on shared

that used the same datasets. Aroyehun et al. [1] similarly trun-

knowledge and the number of documents. Groups with relatively

cated the IPC hierarchy at the subclass level and predicted these level 2

few documents (i.e., groups with fewer than 40,000 for ,

labels by transferring knowledge from two higher levels (section level 3 level 4 20,000 for , and 9,000 for ) were combined with other

and class) to the lower level (subclass), achieving a precision

groups at the same level that shared similar knowledge. As an ex-

score of 0.53. While it remains valuable for patent office experts

ample, at the subclass level of the CPC hierarchy, "A01B" (i.e., Soil

to use an automatic model that can narrow down applications to

working) and "A01C" (i.e., Planting, Sowing, Fertilising) represent

higher levels of the taxonomy tree, this approach has limitations

related steps in agricultural practices, as both are foundational

and challenges. One such challenge is that the choice of target

processes in land preparation and management. We merged them

class labels does not depend on the scope of the knowledge area.

into a single group labelled "Soil working and planting," resulting

More established and expansive areas may benefit from directing

in 162,567 patents in this category. The refinement continued

experts to detailed groups, while less developed areas may be

until the fine-grained classes contained at least 9,000 documents.

adequately served by broader classifications.



3.3 Text Classification

3 Methods and Materials

We formulated the classification problem as a multi-label problem,

In this work, we developed a knowledge taxonomy and classi- in which each document can be assigned to multiple knowledge

fied patents into fine-grained classes by fine-tuning pre-trained fields. In this study, we aimed to classify the patents into the fine-

models. Below, we outline the methods and materials used.

grained classes in the lowest level of the proposed taxonomy (i.e.,

83 classes). To balance performance and computational cost given

3.1 Patent Collection and Preprocessing the large size of the dataset, We used the pre-trained language

models , a distilled version of RoBERTa [16, 19], distilroberta-base

The dataset used in our experiments is the Google Patents Pub-

3 and , a version of MiniLM fine-tuned for seman-all-MiniLM-L6-v2

lic Datasets on BigQuery . Each patent has several pieces of

tic similarity [22, 17]. The pre-trained models were fine-tuned

information, including the publication number, application num-

for the downstream task by adding a classification head. The

ber, CPC code, title, abstract, and detailed description. We have

classification head takes the hidden state of the first token from

expanded the dataset to include the titles associated with each

4 the model and processes it through a fully connected dense linear

CPC code from Espacenet. . In this study, we focused on the tex-

layer, followed by a dropout layer for regularisation and a tanh

tual data. We generated the input text by concatenating the title,

activation function for non-linearity. Since our task is multi-label

followed by the abstract, and then the description. We included

classification, the output logits for each class are converted into

only those documents where the concatenated text is at least 100

probabilities using a sigmoid function.

words long. Previous studies have examined various lengths of

For model training, we used a learning rate of 4e-5 with a

textual data and found that using the first 100 words often results

linear scheduler and a weight decay of 0.1. To prevent overfitting,

in higher performance for classification tasks [15].

the best checkpoint was selected based on evaluation metrics

To create a hierarchical structure where we have enough doc-

on the validation set. We trained the model for up to 5 epochs

uments among leaf-node labels (i.e., avoiding scenarios where

with early stopping criteria based on validation accuracy. The

one group contains only a few hundred documents while others

dataset, consisting of 1,092,991 samples randomly selected after

contain hundreds of thousands as an example), we needed to

deduplication, was split into training, validation, and test sets

count the number of documents which fall into the defined cate-

with ratios of 0.8, 0.1, and 0.1, respectively. To preserve the ratio

gories. As a preprocessing step before counting, we performed

of samples per class in training, validation, and test sets, we used

de-duplication, which involved removing duplicate and near-

5

stratified splitting .

duplicate textual data [4, 12, 14].

Due to the large size of the dataset, we employed MinhHash

Locality Sensitive Hashing (LSH) as a deduplication method to 3.4 Classification Evaluation

efficiently identify similar documents [7, 9, 22]. Specifically, we The F1-score is a common metric for classification tasks. We

used MinHash to approximate the Jaccard similarities between report both Micro-F1, averaged across all instances, and Macro-

sets of n-grams within the documents. MinHash is particularly F1, averaged across all classes.

advantageous for large datasets because it supports parallel com-

putation, enhancing scalability [2]. We set the similarity threshold 4 Results and Analysis

at 0.9, meaning that documents with a Jaccard similarity of 90% In this section, the results are presented in two parts. First, we

present our proposed KnowMap taxonomy. Then, we report the

3 5

https://github.com/google/patents- public- data https://github.com/trent- b/iterative- stratification?tab=readme- ov- file#multilab

4

https://worldwide.espacenet.com/ elstratifiedkfold



Classification of Patents Into Knowledge Fields: Using KnowMap Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



performance of classifiers in categorising patents into the fine- Table 3: Classification Results

grained classes of this taxonomy.

Metric RoBERTa SBERT

4.1 The Proposed Knowledge Mapping

Micro-F1 (Val) 0.76 0.76

Taxonomy (KnowMap) Macro-F1 (Val) 0.86 0.86

The taxonomy, along with the associated CPC sections, classes, Micro-F1 (Test) 0.77 0.76

subclasses, groups, and subgroups are provided in the shared Macro-F1 (Test) 0.90 0.90

online source. An example of detailing the knowledge field of

soil working and planting within the broader knowledge field of

human necessities is illustrated in Fig. 1. 1.0



CPC All groups in 0.8 A A01 A01B, A01C A01B, A01C

162,567 docs 162,567 docs

KnowMap SOIL WORKING AND SOIL WORKING AND

PLANTING PLANTING 0.6 alue

1,543,195 docs HARVESTING AND

PRODUCE PROCESSING

AGRICULTURE

ANIMAL HUSBANDRY malized V

AND CONTROL 0.4 Nor

30,813,838 docs FOODSTUFFS TOBACCO

HUMAN NECESSITIES DAIRY PRODUCTS

PERSONAL OR DOMESTIC

ARTICLES

OPERATIONS AND

TRANSPORTING 0.2

HEALTH AMUSEMENT

d METALLURGY ocs CHEMISTRY AND

2

,02 F1 Macro

49 TEXTILES AND PAPER 0.0 Test Size t

7,7

18 0 FIXED CONSTRUCTIONS 12 20 41 62 82 oo

R Class Index

MECHANICAL

ENGINEERING

PHYSICS Figure 2: Normalised test size along with F1 Macro scores

for each class. The x-axis represents class indices. The y-

ELECTRICITY axis shows normalised values for test size and F1 Macro

NEW TECHNOLOGIES scores (blue dots).



Level 1 Level 2 Level 3 Level4

We demonstrated the experimental results on the two classifi-



Figure 1: An example of a branch extension in KnowMap cation models and in Tab. 3. RoBERTa SBERT

from the root to the lowest level, showing the association As observed from the results, the Macro-F1 score is higher than

of KnowMap classes with corresponding CPC classes at the Micro-F1 score, which may indicate that the model performs

each level. better for minority classes compared to majority classes. To gain

more insights into these results, we generated a plot (see Fig.2),

showing the F1 scores along with the normalised number of

documents for each class in the test set. We used normalised

4.2 Classification Results values to allow both F1 scores and class sizes to be displayed in a

single figure, facilitating better comparison.

The classification task in this study was to classify patents into

The plot shows that the Macro-F1 score is higher for minority

83 fine-grained classes within our proposed KnowMap taxonomy.

classes than for majority classes, also indicating that random

The dataset comprised 1,092,991 documents, which were split

sampling led to an unbalanced dataset. The imbalanced sample

into the train, validation, and test sets with a ratio of 0.8, 0.1,

likely caused the higher Macro-F1 score relative to Micro-F1,

and 0.1 respectively. We preserved the ratio of samples per class

reflecting poorer performance in the majority classes. Future

in all three sets with stratified splitting. The average number

work will focus on using balancing techniques when sampling

of documents in the train set, validation set, and test sets are

to address this issue and enhance model performance.

presented in Tab. 2.

When looking more closely at the lowest F1-Macro scores, we

found that the bottom 10 classes were all leaves under the chem-

Table 2: Overview of sample metrics: total number of sam- istry and metallurgy section. Moreover, the highest F1-Macro

ples, average number of samples per class, and normalised scores (0.996) were achieved by the two classes in the textiles

average number of samples per class across training, vali- and paper section, followed by all 17 leaves from the physics dation, and test sets. section. We suspect this performance difference may be due to

greater variation in the textual data of chemistry and metallurgy

Set Total Avg/ class Normalised Avg class compared to and , leading to more physics textiles and paper

variation between the training and test sets. Analysing this vari-

Train 1,092,991 132,202 0.012

ation in detail remains a task for future work. Additionally, we

Val 874,372 16,476 0.012

believe future work could benefit from adapting the classifier to

Test 218,619 15,543 0.012

a hierarchical structure, prioritising correct predictions at higher



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Motamedi et al.



levels before refining predictions at the leaf level. In our current [4] Gianni Costa, Alfredo Cuzzocrea, Giuseppe Manco, and Riccardo Ortale.

2011. Data De-duplication : A Review Data De-duplication : A Review. Learn-

approach, the classifier does not account for the hierarchy and

ing structure and schemas from documents, January. isbn: 9783642229138.

predicts all leaves directly.

doi: 10.1007/978- 3- 642- 22913- 8.

[5] C. J. Fall, A. Törcsvári, K. Benzineb, and G. Karetka. 2003. Automated cate-

5 Discussion and Conclusions gorization in the international patent classification. ACM SIGIR Forum, 37,

1, 10–25. doi: 10.1145/945546.945547.

In this work, we proposed a knowledge field taxonomy, KnowMap, [6] Juan Carlos Gomez and Marie Francine Moens. 2014. A survey of automated

hierarchical classification of patents. Lecture Notes in Computer Science

which aligns with the widely used CPC schema. The taxonomy

(including subseries Lecture Notes in Artificial Intelligence and Lecture Notes

consists of 83 groups at the lowest level, with fine-grained classes , 8830, 215–249. doi: 10.1007/978- 3- 319- 12511- 4_11. in Bioinformatics)

containing a minimum of 9,000 samples from the original Google [7] Bikash Gyawali, Lucas Anastasiou, and Petr Knoth. 2020. Deduplication of

scholarly documents using locality sensitive hashing and word embeddings.

Patents Public Dataset after preprocessing. KnowMap serves as a

In Proceedings of the 12th Conference on Language Resources and Evaluation

benchmark taxonomy, addressing a gap in the existing literature. . European Language Resources Association, 894–903. (LREC 2020)

[8] Arousha Haghighian Roudsari, Jafar Afshar, Wookey Lee, and Suan Lee.

From the preprocessed original dataset, we randomly selected

2022. PatentNet: multi-label classification of patent documents using deep

1,093,151 samples to fine-tune pre-trained RoBERTa and SBERT Scientometrics learning based language understanding. , 127, 1, 207–231. doi:

models for downstream tasks. However, the random sampling 10.1007/s11192- 021- 04179- 4.

[9] Omid Jafari, Preeti Maurya, Parth Nagarkar, Khandker Mushfiqul Islam,

resulted in an unbalanced dataset, which contributed to higher

and Chidambaram Crushev. 2021. A Survey on Locality Sensitive Hashing

Macro-F1 scores compared to Micro-F1 scores. To enhance clas- Algorithms and their Applications. ACM Computing Surveys. eprint: 2102.0

sification results, we plan to create a balanced dataset from the 8942.

[10] Guik Jung, Junghoon Shin, and Sangjun Lee. 2023. Impact of preprocessing

original data. Additionally, we aim to use larger models than

and word embedding on extreme multi-label patent classification tasks.

those used in this study to further improve the fine-tuning pro- , 53, 4, 4047–4062. doi: 10.1007/s10489- 022- 03655- 5. Applied Intelligence

[11] Eleni Kamateri, Michail Salampasis, and Eduardo Perez-Molina. 2024. Will

cess.

AI solve the patent classification problem? , 78, World Patent Information

June, 102294. doi: 10.1016/j.wpi.2024.102294.

6 Future Work [12] Nikhil Kandpal, Eric Wallace, and Colin Raffel. 2022. Deduplicating Training

Data Mitigates Privacy Risks in Language Models. In International Confer-

Several knowledge platforms, such as news sites and GitHub, ence on Machine Learning, Baltimore number 1. Vol. 162, 10697–10707.

host various types of information shared online. In future work, [13] Jong Wook Lee, Won Kyung Lee, and So Young Sohn. 2021. Patenting trends

in biometric technology of the Big Five patent offices. World Patent Infor-

we aim to incorporate these sources to extend and enhance the

mation, 65, March, 102040. doi: 10.1016/j.wpi.2021.102040.

knowledge taxonomy’s coverage. For example, the All Science [14] Katherine Lee, Daphne Ippolito, Andrew Nystrom, Chiyuan Zhang, Dou-

Journal Classification (ASJC), which organises research publica- glas Eck, Chris Callison-Burch, and Nicholas Carlini. 2022. Deduplicating

Training Data Makes Language Models Better. Proceedings of the Annual

tions by subject area, can be used to identify alignments with Meeting of the Association for Computational Linguistics, 1, 8424–8445. eprint:

the existing taxonomy. This taxonomy alignment can then be 2107.06499. doi: 10.18653/v1/2022.acl- long.577.

[15] Shaobo Li, Jie Hu, Yuxin Cui, and Jianjun Hu. 2018. DeepPatent: patent

further analysed to determine whether to merge or split classes

classification with convolutional neural networks and word embedding.

at various levels. Beyond patents, we plan to evaluate the classi- Scientometrics

, 117, 2, 721–744. isbn: 1119201829. doi: 10.1007/s11192- 018- 2

fier on other data, using domain adaptation methods to transfer 905- 5.

[16] Yinhan Liu et al. 2019. Roberta: a robustly optimized bert pretraining ap-

knowledge from the labelled patent domain to those with limited

proach. , abs/1907.11692. ArXiv

or no labels. Large language models (LLMs) could further aid in [17] Nils Reimers and Iryna Gurevych. 2019. Sentence-bert: sentence embeddings evaluating the classifier’s performance across different domains. using siamese bert-networks. In Conference on Empirical Methods in Natural

Language Processing. Association for Computational Linguistics.

Recent research has shown the potential of LLMs to augment or

[18] Arousha Haghighian Roudsari, Jafar Afshar, Charles Cheolgi Lee, and

even replace human-labeled training data with labels generated Wookey Lee. 2020. Multi-label patent classification using attention-aware

deep learning model. In Proceedings - 2020 IEEE International Conference on

by these models [23].

Big Data and Smart Computing, BigComp 2020, 558–559. isbn: 9781728160344.

Moreover, we plan to enhance the classification task by bal-

eprint: arXiv:1910.01108. doi: 10.1109/BigComp48618.2020.000- 2.

ancing the dataset using balancing techniques for multi-label [19] Victor Sanh, L Debut, J Chaumond, and T Wolf. 2019. Distilbert, a distilled

version of bert: smaller, faster, cheaper and lighter. arxiv 2019. arXiv preprint

problems and leveraging larger pre-trained models. we will also

arXiv:1910.01108.

closely examine the different knowledge fields to better under- [20] Mirac Suzgun, Luke Melas-Kyriazi, Suproteem K. Sarkar, Scott Duke Komin-

stand the variations in classifier performance across them. ers, and Stuart M. Shieber. 2023. The Harvard USPTO Patent Dataset: A

Large-Scale, Well-Structured, and Multi-Purpose Corpus of Patent Applica-



Acknowledgements tions. In 37th Conference on Neural Information Processing Systems (NeurIPS

2023) Track on Datasets and Benchmarks number NeurIPS, 1–39. eprint:

2207.04043.

This work was supported by the Slovenian Research and Inno-

[21] Christoph Trattner, Philipp Singer, Denis Helic, and Markus Strohmaier.

vation Agency under grant agreements CRP V2-2272, V5-2264, 2012. Exploring the differences and similarities between hierarchical decen-

tralized search and human navigation in information networks. In ACM

CRP V2-2146 and the European Union through enrichMyData

International Conference Proceeding Series, 0–7. isbn: 9781450312424. doi:

EU HORIZON-IA project under grant agreement No 101070284.

10.1145/2362456.2362474.

[22] Wenhui Wang, Furu Wei, Li Dong, Hangbo Bao, Nan Yang, and Ming Zhou.



References 2020. Minilm: deep self-attention distillation for task-agnostic compres- Advances in Neural Information Processing sion of pre-trained transformers.

[1] Segun Taofeek Aroyehun, Jason Angel, Navonil Majumder, Alexander Gel- Systems

, 33, 5776–5788.

bukh, and Amir Hussain. 2021. Leveraging label hierarchy using transfer and

[23] Xinru Wang, Hannah Kim, Sajjadur Rahman, Kushan Mitra, and Zhengjie

multi-task learning: A case study on patent classification. , Neurocomputing

Miao. 2024. Human-llm collaborative annotation through effective verifi-

464, 421–431. doi: 10.1016/j.neucom.2021.07.057. Proceedings of the 2024 CHI Conference on Human

cation of llm labels. In

[2] Mehmet Aydar and Serkan Ayvaz. 2019. An improved method of locality- Factors in Computing Systems (CHI ’24) Article 303. Association for Com-

sensitive hashing for scalable instance matching. Knowledge and Information

puting Machinery, Honolulu, HI, USA, 21 pages. isbn: 9798400703300. doi:

Systems, 58, 2, 275–294. isbn: 1011501811995. doi: 10.1007/s10115- 018- 1199 10.1145/3613904.3641960.

- 5.

[24] Junghwan Yun and Youngjung Geum. 2020. Automated classification of

[3] Hamid Bekamiri, Daniel S. Hain, and Roman Jurowetzki. 2024. PatentS- Computers and Industrial Engineering patents: A topic modeling approach. ,

BERTa: A deep NLP based hybrid model for patent distance and classifica-

147, July, 106636. doi: 10.1016/j.cie.2020.106636.

tion using augmented SBERT. , Technological Forecasting and Social Change

206, June, 123536. doi: 10.1016/j.techf ore.2024.123536.



Enhancing causal graphs with domain knowledge: matching



ontology concepts between ontologies and raw text data



Jernej Stegnar Jože M. Rožanec

Jožef Stefan Institute Jožef Stefan International Postgraduate School

Ljubljana, Slovenia Ljubljana, Slovenia

jernej.stegnar@gmail.com joze.rozanec@ijs.si



Gregor Leban Dunja Mladenić

Event Registry d.o.o. Jožef Stefan Institute

Ljubljana, Slovenia Ljubljana, Slovenia

gregor@eventregistry.org dunja.mladenic@ijs.si



ABSTRACT foresight outcomes at such a pace. Nevertheless, this would be

When building a causal graph from textual sources, such as media possible with the use of artificial intelligence.

reports, a key task is to provide an accurate semantic understand- AI enhances strategic foresight by automating the analysis of

ing of the causal variables encoded as nodes and to link them data and detecting patterns that may go unnoticed by human

to existing ontologies with at least two purposes: (i) expand the experts [1]. Machine learning algorithms can continuously mon-

knowledge with the domain knowledge captured in such ontolo- itor emerging trends, geopolitical shifts, and market fluctuations

gies and (ii) provide accurate and different levels of abstraction in near-real time, offering dynamic insights into potential future

of the extracted causal variables. This article describes how we scenarios. Natural language processing (NLP) enables AI to sift

used OntoGPT, a tool for matching raw text to ontology concepts through massive amounts of text, extracting relevant informa-

initially designed for the medical domain, to match concepts from tion from reports, news, and social media, thus accelerating the

media events to relevant ontologies. We build upon our previous forecasting process. By integrating AI into strategic foresight,

work on extracting causal variables and enrich the extraction organizations can adapt more swiftly and make more informed,

pipeline by matching causal variables to concepts from specific data-driven decisions in the face of uncertainty.

domain ontologies. In particular, we describe our work regard- Ontologies provide structured knowledge informing the rela-

ing the GEO ontology. Future work will focus on expanding tionships between concepts within a specific domain. Further-

OntoGPT’s capabilities by utilizing a wider selection of ontolo- more, they describe those concepts through properties and can

gies. Addressing its limitations, such as dealing with multiple link such classes to specific instances observed in the real world.

instances of the same class, will also be crucial for improving its As such, they are of key importance when building a causality

utility. These improvements will allow the tool to better support graph, given they can augment our understanding of the causal

strategic foresight applications by providing more detailed in- relationships between variables with a better understanding of

sights across a multitude of different sectors, further enriching the context and the variable implications [3]. For example, if

causal graphs and facilitating more accurate predictive modeling. the causal relationship reports about the ceasing of an armed

conflict, knowing whether a causal variable relates to a coun-

KEYWORDS try, the location of that country, the neighboring countries, and



strategic foresight, ontology matching, artificial intelligence derstand the magnitude of that event and contextualize other international organizations it is involved in would help to un-

likely outcomes (refugee repatriation, impacts on investments,

1 INTRODUCTION and others).

Strategic foresight is a discipline concerned with anticipating In the scope of the graph massive project, ontology matching

future trends, uncertainties, and disruptions to inform decision- is being used to link the extracted causal relationships from text

making and enable the creation of resilient, long-term strategies. to concepts inside the ontologies, allowing for a more detailed

As such, it is valuable to governments, organizations, and enter- understanding of the concepts that appear in causal relationships

prises, who can use it to remain competitive and adaptable in a and their interconnectivity.

rapidly changing world [4].

The pace of technological advancement, shifting geopoliti- 2 ENRICHING CAUSAL GRAPHS WITH

cal landscapes, environmental crises, and unpredictable market DOMAIN KNOWLEDGE

trends make it essential to react quickly to change. Traditionally, We consider ontologies a framework (an organized and structured

foresight has been based on trend analysis, expert opinion, and system for representing knowledge) used to represent knowl-

qualitative insights. Such approaches lack the agility required to edge within a specific domain by defining the relationships be-

scan real-world events in near-real time and produce strategic tween concepts. They consist of classes (concepts), properties

(attributes), and relationships that connect different concepts.



work must be honored. For all other uses, contact the owner/author(s). Permission to make digital or hard copies of part or all of this work for personal This structure provides a standardized way to organize and in- or classroom use is granted without fee provided that copies are not made or terpret data, ensuring consistent understanding across systems. distributed for profit or commercial advantage and that copies bear this notice and For example, in a medical ontology, concepts like "disease" might the full citation on the first page. Copyrights for third-party components of this be linked to "symptoms," "treatments," and "causes," each with Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia its own defined properties. By formalizing these relationships, © 2024 Copyright held by the owner/author(s). ontologies allow AI systems to better interpret and reason about https://doi.org/https://doi.org/10.70314/is.2024.sikdd.25

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Jernej Stegnar, Jože M. Rožanec, Gregor Leban, and Dunja Mladenić





complex information, leading to more accurate data processing

and decision-making.

Ontologies enhance causality graphs by providing domain-

specific knowledge that improves the accuracy and depth of

relationships represented. When extracting causal relationships

from large datasets, such as media reports, the data can often be

ambiguous or incomplete. Ontologies address this by offering

structured knowledge that defines concepts and their relation-

ships within a specific domain, linking extracted causal rela-

tionships to well-defined entities in the ontology. This enriches

the causality graph, uncovering implicit connections and non-

obvious relationships that may otherwise be missed. In strategic

foresight, for example, ontology-based enrichment helps capture

a broader range of potential future scenarios by incorporating Figure 1: The figure showcases our pipeline for building

knowledge beyond the immediate dataset. This leads to more a causality graph. The sub-figure B showcases how the

reliable predictions, especially when the training data is limited process of ontology linking was executed as a part of our

or domain-specific. Ultimately, ontologies are expected to enable pipeline

the system to generalize better, predict outcomes with higher

accuracy, and improve the overall reliability of causality graphs.

The causality graph pipeline in the Graph Massivizer strategic consistent and accurate representation of complex information

foresight project is designed to automate the extraction, organi- by defining structured relationships between concepts.

zation, and analysis of causal relationships from large datasets, The primary purpose of OntoGPT is to enhance AI systems’

particularly news articles. The Figure 1 showcases the structure understanding, processing, and categorization of data by linking

of our causality graph’s data pipeline. The process begins with extracted information to predefined concepts and relationships

extracting these relationships from news articles, which are then within an ontology. This structured approach ensures greater

organized into a causality graph that maps the interactions be- accuracy and reliability compared to traditional AI systems that

tween various factors and events. The goal is to develop link rely on unstructured data.

prediction models that estimate the likelihood of future events OntoGPT works by connecting data from sources such as text

based on observed patterns. For instance, one use case involves or reports to specific concepts in an ontology, allowing for more

predicting oil price trends by analyzing factors that influence informed and contextually accurate connections. For example, in

pricing. healthcare, OntoGPT can link symptoms from patient records to

Ontology matching is then integrated into the pipeline to link diseases and treatments outlined in medical ontologies, helping

extracted causal relationships with concepts from structured on- to suggest possible diagnoses or treatment plans.

tologies. This enrichment adds layers of context and enables the By combining the language-processing capabilities of LLMs

discovery of connections that may not be evident from raw data with the structured knowledge available in ontologies, OntoGPT

alone. By incorporating ontologies, the pipeline transcends the enables AI systems to go beyond keyword matching and consider

limitations of its training data, identifying causal relationships the relationships between terms. This leads to more intelligent

that may be implied by broader knowledge contained in the on- data interpretation and improved decision-making.

tologies. This not only enhances the accuracy of the graph but OntoGPT is widely used in fields where structured knowl-

also allows it to capture more complex and non-direct relation- edge is critical for high accuracy, such as healthcare, biology,

ships, improving its predictive capabilities. and pharmaceutical research. In medical research, for instance,

As shown in Fig. 1B, the process of ontology linking in our OntoGPT links clinical trial data, medical records, and scientific

pipeline consisted of creating ontology matching templates, then literature to medical ontologies, supporting better analysis and

linking the concepts in text to ontologies, using the information decision-making.

to add additional data to existing causalities, all with the purpose The key advantage of OntoGPT lies in its ability to ground

of finding extra implicit connections based on the information AI outputs in domain-specific, structured knowledge, reducing

provided by the ontologies. the likelihood of errors and improving the relevance of insights.

The main problem that needed solving for that purpose was, This grounding ensures that AI responses are not just based on

how to link ontologies to raw text data. In our case that was patterns but also on well-defined concepts and their relationships.

done using OntoGPT [2], which is a tool used for ontology link- In summary, OntoGPT bridges the gap between the raw data-

ing. Another key challenge is inter-ontology matching, which processing power of LLMs and the structured knowledge in on-

involves linking multiple ontologies through shared concepts. tologies. By leveraging both, it provides a more accurate and

This process expands the knowledge framework, making it even reliable approach to extracting and linking data across various do-

more valuable for our purposes. The challenge of inter-ontology mains, particularly when working with large, complex datasets.

matching hasn’t been addressed yet and remains a matter of

future work. 3.1 OntoGPT’s role

At a lower level, OntoGPT operates using YAML templates that

define how data should be extracted from text and linked to onto-

3 ONTOGPT: A BRIEF OVERVIEW logical concepts. These templates serve as blueprints, specifying

OntoGPT is an advanced tool that integrates large language which types of entities, relationships, and properties to look for

models (LLMs) with ontologies to improve knowledge extraction in the input text. The templates guide the large language model

and organization across various domains. Ontologies provide a by mapping textual data to predefined concepts and relationships

Enhancing causal graphs with domain knowledge Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia





classes inside the ontology, that we are trying to link the text data

to, and their descriptions, which assists OntoGPT in more accu-

rately identifying these classes inside the text. The YAML file also

contains the information of "annotators" which tells OntoGPT,

which ontology to ground the responses to. The generated YAML

templates are saved into a separate file after generation, which

makes them ready for use.

The python code that is used by OntoGPT in the process of

ontology linking, is similarly generated by using the extracted

information to fill in the "general template" and is then saved to

a separate file.



5 LIMITATIONS

Figure 2: A Showcase of the function of OntoGPT

5.1 Multiple Same-Class Concepts

OntoGPT has problems trying to link two or more concepts to a



place in the ontology if the concepts are of the same class. This

happens because both concepts suit the description and similar

criteria that OntoGPT extracts the information based on. This

causes OntoGPT to merge both concepts into a single string and

then try to locate the said string inside the ontology, which fails

because there is no individual inside the ontology class with such

Figure 3: The Process of Templates Generation a name. An example of such a response is shown in Listing 1:



Listing 1: Example of a bad response

from the ontology, ensuring that the extracted information is e x t r a c t e d _ o b j e c t :

both relevant and structured. The figure 2 shows the process c o n t i n e n t : AUTO : Europe%2C%20 A f r i c a

of ontology linking for an example of a simple sentence. Each n a m e d _ e n t i t i e s :



YAML template contains detailed instructions on how to identify − i d : AUTO : Europe%2C%20 A f r i c a key terms, their corresponding ontology classes, and the relation- l a b e l : Europe , A f r i c a ships between them. This allows OntoGPT to recognize when a

piece of text, such as a sentence from a media article, contains If OntoGPT managed to locate the concept inside the text

a concept that aligns with an entity or event in the ontology. in the ontology, it returns its id (an example of this is "sea:

Once identified, the tool links the extracted data to these ontol- GEO:000055471" and "id: GEO:000055471 : White Sea") If the

ogy entries, enabling richer and more meaningful connections concept suits the class criteria, but couldn’t be located inside the

in the data, as it is now grounded in an established knowledge ontology, it returns it as a “AUTO” detection. For the purpose of

framework. ontology linking this is not optimal as it does not give us access

The approach described in this article uses an ontology file to the additional information that is stored inside the ontology’s

as input to create such templates for data extraction and link- individual information. The ontology’s individual information is

ing. This enables for a broader range of ontology linking, as the a set of predefined relationships and properties, that an individ-

templates can be created on demand. ual concept has. For example, if the individual "Africa" is defined

inside the ontology, the individual’s data would include its size,

4 TEMPLATES AND PYTHON CODE countries on the continent, population, and climates, among oth-

GENERATION ers. This information gives us reliable information about a certain



The approach works by using the information defined inside To solve this problem, the approach of creating "buffer" classes concept, allowing for more contextual understanding. the ontology, to generate the YAML templates. The Figure 3 was taken, where a certain class from ontology would be used to showcases the process of how this is done. generate three classes describing the different occurrences of the First the class information, for each class inside the ontology, ontology class and a description that would provide sufficient



is extracted. This is done by using the "owlready2" python library context to OntoGPT to separate the same class concepts into to parse the ontology into an object, and then extract the relevant different entities. The corrected response is showcased in Listing information from the new object. 2: Every class inside the ontology is used to create a correspond-

ing template class, which is optimal, as it covers all parts of the Listing 2: Example of a corrected response



ontology that could potentially be linked. A small portion of e x t r a c t e d _ o b j e c t : the data extraction process is ontology-specific and was custom-c o n t i n e n t : GEO: 0 0 0 0 0 0 3 4 0 tailored to the individual ontology, as some information (like c o n t i n e n t _ 2 : GEO: 0 0 0 0 0 0 3 4 2 class descriptions) is saved in different parts. n a m e d _ e n t i t i e s : Secondly the data extracted from the ontology is processed − i d : GEO: 0 0 0 0 0 0 3 4 0 and used to create custom YAML templates. This is done by sim-l a b e l : A f r i c a ply using the extracted information to fill in a "general template"

we used for generation. Specifically the class names and descrip- − i d : GEO: 0 0 0 0 0 0 3 4 2

tions are used, to do so. This gives OntoGPT the names of the l a b e l : Europe



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Jernej Stegnar, Jože M. Rožanec, Gregor Leban, and Dunja Mladenić



While this approach deals with a high percentage of this type ACKNOWLEDGMENTS

problem, it does not cover the cases where more than three same- The Slovenian Research Agency supported this work. This re-

class concepts are inside the piece of text being analyzed. search was developed as part of the Graph-Massivizer project

funded under the Horizon Europe research and innovation pro-

6 CONCLUSIONS gram of the European Union under grant agreement 101093202.

Using OntoGPT in the Graph Massivizer strategic foresight project

will prove valuable for enriching causal graphs with linked on- REFERENCES



future events. Despite OntoGPT’s initial focus on medical data, foresight–Current practices and future application potentials: current practices and future application potentials. In Proceedings of the 2021 12th International some custom adaptations were successfully implemented to suit tology data, aiming to improve predictive accuracy in predicting [1] Patrick Brandtner and Marius Mates. 2021. Artificial intelligence in strategic

Conference on E-business, Management and Economics. 75–81.

a portion of different domains. However, limitations persist in [2] J Harry Caufield, Harshad Hegde, Vincent Emonet, Nomi L Harris, Marcin P

distinguishing between multiple instances of the same concept Joachimiak, Nicolas Matentzoglu, HyeongSik Kim, Sierra Moxon, Justin T Reese,

class. These challenges highlight the need for further develop- Melissa A Haendel, et al. 2024. Structured prompt interrogation and recursive

extraction of semantics (SPIRES): A method for populating knowledge bases

ment to enhance the tool’s versatility across a broader array of using zero-shot learning. Bioinformatics 40, 3 (2024), btae104.

applications and ontologies. [3] Fatma Özcan, Chuan Lei, Abdul Quamar, and Vasilis Efthymiou. 2021. Semantic

enrichment of data for AI applications. In Proceedings of the Fifth Workshop on

Data Management for End-To-End Machine Learning. 1–7.

[4] David Sarpong and Nicholas O’Regan. 2014. The Organizing Dimensions of

Strategic Foresight in High-Velocity Environments. Strategic Change 23, 3-4

(2014), 125–132.



Measuring and Modeling CO2 Emissions in Machine Learning



Processes



Ivo Hrib Oleksandra Topal

Jožef Stefan Institute Jožef Stefan Institute

Ljubljana, Slovenia Ljubljana, Slovenia

ivo.hrib@gmail.com Oleksandra.Topal@ijs.si



Jan Šturm Maja Škrjanc

Jožef Stefan Institute Jožef Stefan Institute

Ljubljana, Slovenia Ljubljana, Slovenia

jan.sturm@ijs.si maja.skrjanc@ijs.si



Abstract offer insights into a model’s emissions before its construction or

use. The service we aim to provide addresses this gap by offering

With the rapid expansion of the computing industry, efficient

an estimation of emissions and power consumption for differ-

energy utilization and reduction of CO2 emissions are critically

ent models before they are selected for specific use cases. This

important. This research develops analytical tools to predict CO2

forward-looking approach allows for more informed decisions

emissions from various machine learning processes. We present a

when choosing models, potentially reducing their environmental

novel methodology for data acquisition and analysis of CO2 emis-

footprint.

sions during model training and testing. Our results demonstrate

the environmental impact of different algorithms and provide

insights into optimizing energy consumption in artificial intelli- 2 Related Work

gence applications. The environmental impact of machine learning models has been

a growing concern in recent years. Several studies have focused

Keywords on quantifying and reducing the carbon footprint of artificial

CO2 Emissions, Machine Learning, Energy Consumption, Envi- intelligence (AI) processes. For instance, [12] highlighted the en-

ronmental Impact, AI Model Optimization, Green AI, Sustainable ergy consumption of training large neural models and suggested

Computing, Carbon Footprint methods for minimizing emissions. Similarly, tools like CodeCar-

bon [2] and eco2AI [3] have emerged to measure real-time CO2

1 Introduction emissions from computational tasks. However, these tools often

lack predictive capabilities for assessing emissions before model

The global computing industry significantly contributes to CO 2

selection, as pointed out by. Our work builds on these existing

emissions, with data centers accounting for 2.5 to 3.7 percent

methodologies, concretely on the work of eco2AI[3], by providing

of global greenhouse gas emissions [1]. These emissions exceed

a forward-looking approach that estimates emissions during the

those of the aviation industry due to continuous operations and

model selection phase, thus complementing real-time monitoring

heavy reliance on fossil fuels [11]. Given the growing demand for

tools. This is achieved through heavy dependency on eco2AI[3]

artificial intelligence (AI) applications, there is an urgent need

measuring systems for data collection, later used for modeling

for CO-conscious solutions.

2

based on the collected data and registered hyperparameters.

This research aims to develop tools for predicting CO 2 emis-

sions associated with machine learning processes, thus enabling

the reduction of the environmental impact of AI models. In col- 2.1 Research Gap and Contribution

laboration with Eviden (Spain) and under the FAME EU project, Despite the growing availability of tools like CodeCarbon [2]

we have developed a CO2 emissions analysis system using tools and eco2AI [3], a significant gap remains in the preemptive eval-

like CodeCarbon [2] and eco2AI [3]. uation of environmental impact during the machine learning

(ML) model selection phase. The mentioned tools are valuable for

1.1 Research Goals post hoc analyses but do not assist ML practitioners in making

The primary goal of this research is to develop a service that pre- —before model development—on informed decisions upfront

dicts CO2 the environmental footprint of different model architectures or

emissions and power consumption of different machine

learning models during both training and evaluation phases with hyperparameters.

emphasis on hyperparameter dependency. The CO 2 This gap is crucial, as the model selection phase often involves

emissions

𝑘𝑔 trial-and-error across multiple models and configurations, po-

are measured in kilograms per second ( ) , while the power

𝑠

tentially leading to unnecessary resource consumption. Without

consumption is measured in kilowatt-hours (kWh).

predictive capabilities, practitioners have limited insight into

While existing services, such as CodeCarbon [2] or eco2AI

which models will have the lowest environmental impact before

[3], provide real-time measurement of emissions, they do not

engaging in resource-intensive training.

Permission to make digital or hard copies of all or part of this work for personal predictive

Our research aims to fill this gap by introducing a

or classroom use is granted without fee provided that copies are not made or

distributed for profit or commercial advantage and that copies bear this notice and service

that estimates the environmental footprint of different

the full citation on the first page. Copyrights for third-party components of this ML models before they are trained or used. This service leverages

work must be honored. For all other uses, contact the owner /author(s).

the data collected from existing tools like eco2AI [3], incorporat-

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

ing key features such as hyperparameters, and model architec-

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.sikdd.23 turre into predictive models. By doing so, we enable developers



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Hrib et al.



to make at the model selection stage, more sustainable choices 3.3 CO2 Emission Measurement reducing carbon emissions from the start of the ML lifecycle.

We measure CO2 emissions produced during both the training

The table 1 below presents a feature matrix comparing our

and testing phases of the machine learning models. This involves

proposed service with current tools, showing how our approach

using tools like eco2AI [3] to track energy consumption and

addresses unmet needs:

convert it into equivalent CO2 emissions. The measurements

are taken for various models, such as Decision Trees, Random

3 Methodology Forests, Logistic Regression, and Neural Networks, to assess their

environmental impact under different computational loads.

Due to the lack of suitable data on CO emissions of machine

2

learning models, we began by developing an infrastructure for

data collection. This infrastructure is composed of the following 3.4 Feature Extraction

steps: To gain deeper insights, we extract various features that could im-

pact CO 2 emissions and energy consumption. These features in-

• Dataset Generation: Creating synthetic datasets using clude project identifiers, detailed descriptions of each experiment,

random data generation methods.

the duration of each training epoch, power consumption metrics,

• Data Preprocessing: Cleaning and preparing the data for hardware configurations (such as the type of CP U/GP U used), and

analysis.

hyperparameters. The project identifiers refer to unique alphanu-

• CO2 Emission Measurement: Recording CO2 emissions meric codes assigned to each machine learning experiment upon

during both training and testing phases using different

execution. These identifiers help differentiate between various

machine learning algorithms.

model configurations and experimental setups. They are gener-

• Feature Extraction: Extracting relevant features such ated and stored automatically by our system during the dataset

as project ID, experiment details, epoch duration, power

generation process to ensure traceability and reproducibility of

consumption, and hardware configurations.

the experiments.

• Adding Hyperparameters to Final Dataset: Document-

ing hyperparameters used in each experiment to assess 3.5 Adding Hyperparameters to Final Dataset their impact on emissions.

• We document the hyperparameters used in each machine learn- Containerization:

Utilizing Docker for containerization

ing experiment, such as learning rates, batch sizes, and the num-

to ensure reproducibility and scalability of the experi-

ber of layers in neural networks. This allows us to evaluate how

ments.

• these hyperparameters influence CO 2 emissions and energy con- Data Storage:

Storing all datasets, features, and emission

sumption.

records systematically in a database for further analysis.

• Modeling: Developing and training machine learning

models to predict CO 2 emissions and power consumption. 3.6 Containerization

To ensure reproducibility and scalability of our experiments, we

The software implementation uses Python, with dependencies

employ Docker for containerization. This approach encapsulates

including pandas [7], scikit-learn [10], matplotlib [5], eco2AI [3],

the code, dependencies, and environment settings, allowing the

TensorFlow [ ], Keras [ ], abadi2016tensorflow chollet2015keras

experiments to be easily replicated and deployed across different

and Docker for containerization [ ]. merkel2014docker

platforms.



3.1 Dataset Generation 3.7 Data Storage

In this step, we created a synthetic dataset by generating random All datasets, extracted features, hyperparameter configurations,

data points using tools like and CO2 sklearn.datasets.make_regression emission records are systematically stored in a database.

or . The primary objective here is not This central repository facilitates efficient querying, retrieval, make_classification

to reflect real-world data scenarios but to produce a controlled and analysis of data to support ongoing and future research.

environment where the focus is on measuring CO2 emissions

and power consumption during model training and evaluation. 3.8 Modeling Datasets generated vary in size from ranges of 250 to 15000 sam-

In this step, we develop and train machine learning models to

ples and 5 to 2000 features. In classification cases additionally the

predict CO2 emissions and power consumption based on various

number of classes ranges from 2 to 50. These parameter ranges

features, such as the type of algorithm used, hardware configura-

were selected to mitigate the risk of computational overload, en-

tion, and model parameters. This modeling allows us to estimate

suring that the experiments remain feasible within the available

emissions for different machine learning workflows before their

computational resources while maintaining the integrity of the

actual deployment. The models help identify the most efficient

analysis.

algorithms and configurations, thus guiding the selection of en-

vironmentally friendly AI solutions.

3.2 Data Preprocessing The general pipeline for the previously mentioned steps can

be seen below (see Figure 1).

Before analysis, the dataset must be cleaned and prepared. This

A more thorough view of the workings of this can be seen as

includes handling missing values, normalizing or standardizing

shown below for running a single measurement (see Figure 2).

data, encoding categorical variables, and splitting the data into

training and testing sets. Proper preprocessing ensures that the

data is in the optimal format for the models to learn from and min- 4 Model Architecture

imizes biases that may affect model performance and emission In this section, we explain the architecture of the model used

measurements. for predicting CO2 emissions and power consumption based on

Measuring and Modeling CO2 Emissions in Machine Learning Processes Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



Tool/Technology Platform Model Cov-Metric Carbon Energy Additional Real-time Forward-

Compati- erage Granular- Metrics Metrics Features measure- looking

bility ity ment Prediction

CodeCarbon Cloud, On-All ML mod-Per training CO emis- Energy con- Dashboard Yes No 2

Premise els session sions (kg) sumption Visualiza-

(kWh) tion

eco2AI Cloud, On-All ML mod-Per training CO emis- Energy con- Not RAPL Yes No 2

Premise els session sions (kg) sumption based

(kWh)

Proposed On-Premise Specific Per model, CO2 emis-Energy con- Predictive No Yes

Service kg

models per selec- sions sumption modeling

s

(mentioned tion phase (kWh)

bellow)

Table 1: Feature comparison of existing tools and the proposed service





• Output Layer: A single neuron that outputs the predicted

value for either CO2 emissions or power consumption.



4.1 Model Training

The model is compiled using the Adam optimizer [6] and the

Mean Squared Error (MSE) loss function. Seeing as we were un-

able to gather adequate real-time environmental data of factors

that may influence our predictions (e.g. Distribution of energy

sources, real time CO 2 per kWh), our model relies on static yearly

averages of these values[8] [9] . Our model uses the aforemen-

tioned features for the purpose of regression with the goal of

Figure 1: General Measurement Pipeline predicting power consumption and CO2 emissions gathered by

previously mentioned random tests. Each model is trained for 25



epochs using the preprocessed data. After training, the models,

along with their respective scalers and encoders, are saved to

disk for later use.



4.2 Prediction

Once trained, the model can predict CO2 emissions and power

consumption for new data points by loading the appropriate

model, scaler, and one-hot encoder. The input data is prepro-

cessed in the same manner as during training, and the predictions

are obtained by applying the trained models.

This modular approach allows for easy extension to additional

models or data sources and provides a scalable solution for ana-

lyzing the environmental impact of machine learning processes.



5 Web Application Interface for CO 2

Figure 2: Single Model Measurement Pipeline Emissions and Power Consumption

Prediction

In addition to the backend model developed for predicting CO

various features such as CP U type, GP U type, region, and other 2

experiment-specific details. The model implementation is en- emissions and power consumption of various AI models, a web

capsulated within a Python class named , which is

MultiModel application was created to provide a user-friendly interface for

responsible for managing the entire process from data prepro- real-time predictions. The web app, as shown in Figure 3, allows

users to select different machine learning models and configure

cessing to training and prediction.

The model employs two separate neural networks for predict- parameters to estimate the associated environmental impacts.

ing CO2 emissions and power consumption. The architecture for

each neural network is as follows: 5.1 Key Features of the Web Application



• Input Layer: The web application interface is designed with simplicity and Receives the scaled and encoded features.

• Hidden Layers: functionality in mind. It includes several key components: Consist of multiple Dense layers with

ReLU activation functions. The CO2 emissions model in- Users can choose the type of machine • Model Selection:

cludes three hidden layers with 128, 64, and 128 neurons, learning model they are interested in evaluating (e.g., Lo-

respectively, while the power consumption model has gistic Regression ( abbr. LogR ), Decision Tree Classifier (

three hidden layers with 64, 64, and 128 neurons. abbr. DTC ), Decision Tree Regression ( abbr. DTR ), Neural

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Hrib et al.





6 Results



6.1 Model Error

To evaluate the performance and accuracy of the models, we

conducted a 10-fold cross-validation to estimate the errors in

predicting CO 2 emissions and power consumption. The results

are presented in Table 2. The errors for both CO 2 emissions

and power consumption were computed for both training and

evaluation phases of each model type.

Note: In this context, "Train." refers not to the error on the

training set, but rather to the error made by our model in predict-

ing the CO2 emissions / Power Consumption during the training

Figure 3: Web App Interface phase of the listed model. Similarly, "Eval." refers not to the error

on the evaluation set, but rather to the error made by our model

in predicting the CO2 emissions / Power Consumption when

Network Classifier ( abbr. NNC ), Neural Network Regres- the listed model makes predictions. This distinction is crucial to

sion (abbr. NNR ), Linear Regression ( abbr. LinR ), Random understanding the results accurately.

Forest Classifier ( abbr. RFC ) and Random Forest Regres-

sion ( abbr. RFR ) ). The dropdown menu in the upper-left

corner of the interface provides a list of available models. Table 2: Model Scaled Error Estimates from 10-Fold Cross-

• Model Parameters Configuration: A section labeled Validation

"Model Parameters" allows users to specify various inputs:

– Train or Evaluate: Users can choose whether to esti-

Model Phase CO2 Error Power Er-

mate emissions for the training or evaluation phase of

ror

the model.

– Dataset Samples and Features: Input fields are pro- DTC Eval. 0.0036 0.0043

vided for users to define the size of the dataset in terms DTC Train. 0.0631 0.0649

of the number of samples and features. DTR Eval. 0.0032 0.0034

– CPU and GPU Specifications: The app allows the DTR Train. 0.0133 0.0517

selection of the CP U and GP U type, reflecting differ- RFC Eval. 0.0094 0.0098

ent hardware configurations, such as "Intel(R) Xeon(R) RFC Train. 0.3242 0.3582

Gold 6246R CP U @ 3.40GHz/1 device(s), TDP:205.0" or RFR Eval. 0.0087 0.0081

"AMD Ryzen 7 4800H with Radeon Graphics/1 device(s), RFR Train. 0.2565 0.2779

TDP:45.0". LogR Eval. 0.0063 0.0057

– Region/Country Selection: A dropdown to select the LogR Train. 0.0055 0.0043

geographic location where the model is being executed, LinR Eval. 0.0099 0.0105

which influences the CO LinR Train. 0.0104 0.0095 2 emissions based on local en-

ergy sources. NNC Eval. 0.0018 0.0030

• Real-Time Predictions: Once all parameters are config- NNC Train. 0.1083 0.1216

ured, the application dynamically calculates and displays: NNR Eval. 0.0045 0.0112

– CO NNR Train. 0.1051 0.1008 2 Emissions: The predicted emissions are shown in

kilograms per second (kg/s).

– Power Consumption: The power consumption is pro-

vided in kilowatt-hours (kWh).

Based on the results obtained through the 10-fold cross-validation,

• Electricity Source Distribution: A graphical representa- it is evident that the model performance varies significantly

tion is provided for the distribution of electricity sources,

across different algorithms and phases. One notable observa-

such as coal, gas, and oil, in the selected region. This in-

tion is that the errors in predicting CO2 emissions and power

formation is crucial for understanding the environmental

consumption are relatively higher during the training phases,

impact of power consumption based on the local energy

particularly for more complex models like Neural Networks and

mix.

Random Forests [4].

This discrepancy in model performance can be attributed to

5.2 User Experience and Accessibility the sparsity of the data collected during the measurement phase.

The web application is developed with accessibility in mind, en- The limited data points lead to substantial gaps in the attribute

suring that users, regardless of technical background, can interact space covered by the models, resulting in erratic behavior when

with the model’s predictive capabilities. By offering a clear and predicting outside these ranges. Consequently, the models show

intuitive interface, it aims to make the process of estimating CO diminished accuracy and reliability when confronted with input

2

emissions and power consumption transparent and straightfor- configurations that fall beyond the scope of the original data.

ward. Future research should focus on enhancing the robustness of

Figure 3 illustrates the application’s main screen, where the these models by expanding the dataset to include a broader range

model type, parameters, and results are all visible at a glance. This of scenarios and conditions. This would help mitigate the effects

real-time feedback loop allows users to make informed decisions of sparsity and improve the model’s generalizability, ensuring

based on the predicted environmental impact. more reliable predictions across diverse settings.

Measuring and Modeling CO2 Emissions in Machine Learning Processes Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia





8 Limitations

This study presents several limitations, particularly regarding

the data, model evaluation, and hardware configurations, which

must be considered when interpreting the results.



8.1 Training Duration and Model Learning

The models were trained for a fixed number of epochs (e.g., 10 or

20), prioritizing computational cost over learning performance.

The focus was on estimating CO2 emissions rather than model

accuracy or convergence, meaning the models may not have fully

captured patterns in the data. As such, the reported emissions

reflect standardized training durations (with an upper limit for

computational efficiency), not optimized learning outcomes.



Figure 4: Logarithmically scaled mean emissions across

different models 8.2 Lack of Meaningful Learning Objective

The use of randomly generated data limits the evaluation of model

learning. Since the data lacked inherent structure, the models’

ability to learn was not assessed. Instead, the models were pri-

marily evaluated on their resource consumption during training,

reducing the focus on generalization or predictive power.



6.2 CO2 Emission Analysis Across Different

Models 8.3 Hardware and Software Considerations

Figure 4 provides a comparative analysis of the mean CO emis- The experiments were conducted on specific hardware (e.g., GP U/CP U

2

sions generated by different machine learning models during configurations), and variations in hardware were not examined.

their operation, represented on a logarithmic scale to accommo- Different hardware setups, especially energy-efficient systems,

date the wide range of emission values. could significantly impact CO2 emissions and energy consump-

The chart highlights significant variations in CO tion. Therefore, the findings may not generalize across all hard- 2 emissions

among models, with the Neural Network Classifier and Neu- ware environments. However, we would like to point out that this

ral Network Regressor exhibiting the highest emissions by a was due to lack of infrastructure for broader experimentation.

considerable margin. This is expected due to the intensive com-

putational requirements and numerous parameters these models

necessitate, resulting in elevated power consumption and conse- 9 Future Work

quently higher CO 2 output. Future research should incorporate real-world datasets, optimize

In contrast, simpler models like Logistic Regression, Linear hyperparameters, and evaluate diverse hardware configurations

Regression, and Decision Tree models show substantially lower to extend these findings to broader machine learning scenarios.

CO2 emissions, reflecting their reduced computational complex- The exploration of more complex architectures and learning

ity and lower resource demand. objectives will provide a deeper understanding of the trade-offs

Interestingly, the Random Forest models, particularly the Re- between performance and environmental impact.

gressor, present moderate emissions, illustrating that even ensem-

ble methods, which typically involve training multiple decision

trees, can maintain reasonable emission levels depending on their 10 Conclusion configuration.

Our study presents a methodology for monitoring and analyzing

This analysis underscores the importance of model selection

CO 2 emissions during machine learning processes. The find-

not only for performance but also for minimizing environmental

ings demonstrate that different machine learning models exhibit

impact, particularly when scaling up operations or deploying in

significant variability in their energy consumption and CO emis-

2

resource-constrained settings.

sions, with complex models like neural networks having a higher

environmental impact. By providing predictive insights into these

emissions, our approach enables more informed decision-making

7 Discussion during model selection, thus contributing to the broader goal of

The results highlight the significant environmental impact of reducing the carbon footprint of AI applications.

training complex AI models, particularly neural networks. The Future work will focus on expanding the dataset to include

variability in emissions suggests that optimizing model hyperpa- more diverse models and configurations. Additionally, we plan

rameters and selecting appropriate hardware configurations can to integrate real-time monitoring tools to compare predictions

reduce CO output. with actual emissions, further refining our predictive capabilities.

2

Future research should focus on model improvement for better Moreover, optimizing model hyperparameters and exploring al-

and more accurate prediction, expanding the range of algorithms ternative, more sustainable hardware configurations will be key

studied, as well as intensive data collection to accommodate gaps areas of investigation for minimizing the environmental impact

in training data. of machine learning workflows.



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Hrib et al.



Acknowledgements [4] Ian Goodfellow, Yoshua Bengio, and Aaron Courville. 2016. Deep Learning. A

comprehensive resource on machine learning algorithms, including neural

This work was supported by the FAME project, funded by the

networks. MIT Press. http://www.deeplearningbook.org.

European Union’s Horizon 2023 Research and Innovation Pro- [5] John D Hunter. 2007. Matplotlib: a 2d graphics environment. Computing in gramme under grant agreement nº 101092639. science & engineering, 9, 3, 90–95.

[6] Diederik P Kingma and Jimmy Ba. 2014. Adam: a method for stochastic

optimization. . arXiv preprint arXiv:1412.6980

References [7] Wes McKinney. 2010. Data structures for statistical computing in python.

In . Vol. 445, 51–56. Proceedings of the 9th Python in Science Conference

[1] Climatiq. 2023. Climatiq: emissions intelligence platform. Provides data

[8] OurWorldInData. [n. d.] (). https://ourworldindata.org/grapher/carbon- inte

on the carbon emissions of various activities, including computing. (2023).

nsity- electricity?tab=table.

https://www.climatiq.io/.

[9] OurWorldInData. [n. d.] (). https://ourworldindata.org/electricity- mix.

[2] CodeCarbon Development Team. 2023. Codecarbon: an open source tool

[10] Fabian Pedregosa et al. 2011. Scikit-learn: machine learning in python. Jour-

for tracking the carbon emissions of machine learning experiments. An

nal of machine learning research, 12, 2825–2830.

open-source tool for tracking and reducing carbon emissions in machine

[11] Roy Schwartz, Jesse Dodge, Noah A. Smith, and Oren Etzioni. 2020. Green

learning models. (2023). https://github.com/mlco2/codecarbon.

ai. . 12 pages. doi: 10.48550/arXiv.1907.10597. arXiv preprint arXiv:1907.10597

[3] Eco2AI Development Team. 2023. Eco2ai: real-time co2 emission tracking

[12] Emma Strubell, Ananya Ganesh, and Andrew McCallum. 2019. Energy

for machine learning. A tool for real-time tracking of CO2 emissions during

and policy considerations for deep learning in nlp. In Proceedings of the

machine learning processes. (2023). https://github.com/sb- ai- lab/Eco2AI.

57th Annual Meeting of the Association for Computational Linguistics (ACL).

Association for Computational Linguistics, 3645–3650.



Enhancing Ontology Engineering with LLMs: From Search



to Active Learning Extensions



Ganna Kholmska Klemen Kenda Joze Rozanec

Jožef Stefan Institute Jožef Stefan Institute Jožef Stefan Institute

Ljubljana, Slovenia Ljubljana, Slovenia Ljubljana, Slovenia

anna.kholmska@gmail.com klemen.kenda@ijs.si joze.rozanec@ijs.si



Abstract Recent studies show that leveraging Large Language Models

(LLMs) can streamline ontology construction by reducing

This paper explores the use of LLMs in ontology engineering manual effort and improving consistency and quality. For

within the HumAIne project, focusing on the discovery, analysis, instance, [1] demonstrates semi-automatic knowledge graph

and extension of ontologies in Data Mining, Machine Learning, construction using open-source LLMs, while [2] proposes

and manufacturing. The methodology leverages fine-tuned methods for automatic concept hierarchy generation through

prompts and combines LLMs with traditional tools like Protege LLM queries. Building on this research, this paper contributes a

for validation. A multi-LLM approach improved domain- methodology that integrates LLMs with traditional tools like

specific concept coverage and reduced errors, though challenges Protege to streamline the discovery, analysis, and extension of

remain in addressing deep domain-specific gaps and ensuring ontologies. By employing a multi-LLM approach, we address

logical consistency. challenges in domain-specific concept identification and ensure



Keywords fields like Data Mining, Machine Learning, and manufacturing. more consistent, accurate results in ontology development for

LLMs, Ontology Engineering, Active Learning, Data Mining,

Machine Learning, Ontology Selection, Ontology Extension

2 LLM-Assisted Search and Analysis of Domain

Ontologies

1 Introduction

Our experimentation with methodologies and tools for

The HumAIne project, funded by the European Commission efficient web search and ontology analysis in Data Mining (DM),

under the Horizon Europe program, aims to develop a platform Machine Learning (ML), and manufacturing domains led to the

integrating advanced AI paradigms such as Active Learning development of the LLM-leveraging algorithm shown in Fig. 1.

(AL), Neuro-Symbolic AI, Swarm Learning, and Explainable AI. This algorithm uses carefully crafted prompts to guide LLMs in

This platform is designed to enhance human-AI collaboration in generating accurate, targeted queries. Before each step, the initial

dynamic, unstructured environments, with applications spanning prompt is optimized through several iterations in a dialogue with

healthcare, manufacturing, finance, energy grids, and smart cities. the LLM to improve accuracy and relevance. Further details on

Its primary goal is to support decision-making by combining the iterative query refinement process are provided in the

human expertise with AI capabilities. Discussion section.

One of the project's key challenges is developing multiple Step 1: Define the Search Objective. At this stage, LLMs like

ontologies that provide a structured framework for integrating Bing Chat, Google’s Bard, or ChatGPT with Web Browsing are

domain-specific knowledge. This framework is essential for employed to iteratively refine the search objectives initially

enhancing the clarity and reliability of AI-driven decisions, while formulated by the researcher, along with relevant keywords,

ensuring adaptability across diverse applications. To construct phrases, and terms describing the ontologies or concepts of

these ontologies, we first explored publicly available ontologies interest. For instance, our initial search objective for DM and ML

relevant to the project's scope, then extended selected ones with ontologies was to "Find ontologies that offer up-to-date, detailed

concepts from HumAIne’s AI paradigms, starting with Active descriptions of the DM and ML domains, following best

Learning practices in ontology engineering." Keywords included "Active

However, manual ontology construction is a complex, Learning" and "CRISP-DM standard."

resource-intensive process that requires expertise across multiple Step 2: Formulate Search Queries Using LLMs. Based on the

domains, collaboration among stakeholders. Ensuring refined search objectives and keywords, and using a carefully

modularity, reusability, and scalability adds to this complexity. crafted prompt, LLMs generate targeted search queries. These

queries are fine-tuned through feedback or early search results to

maximize relevance and accuracy. For example, for a DM

Permission to make digital or hard copies of part or all of this work for personal or ontology, the LLM generated queries such as "Data Mining

classroom use is granted without fee provided that copies are not made or distributed ontology for semi-supervised machine learning," which were

for profit or commercial advantage and that copies bear this notice and the full further refined before finalizing the query. citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s). Step 3: Conduct Web Search. This step involves real-time

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia browsing tools like Copilot in Microsoft Edge (GPT-4) and

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.sikdd.28 Perplexity AI to execute searches and identify relevant sources.

Our study prioritized high-quality sources like ontology

repositories (e.g., BioPortal, OBO Foundry) and academic OWL and RDF code, were used alongside ontology tools like

databases (Google Scholar, IEEE Xplore, ACM Digital Library). Protege. This combination ensured that the ontologies addressed

It is important to acknowledge that LLM-driven web relevant concepts and aligned with frameworks like CRISP-DM.

searches are generally confined to public repositories and a GPT-4 helped significantly in bridging the gap between textual

limited range of academic databases. As a result, proprietary or descriptions and formal ontology representations.

lesser-indexed ontologies may require manual exploration to Step 6: Cross-Reference and Compare Findings. LLMs with

ensure a more thorough search. contextual understanding were employed to integrate and refine

information from multiple sources. For this task, we used.



Additionally, ChatGPT (GPT-4) categorized 65 manufacturing

ontologies categorized 65 manufacturing ontologies, assessing

them for relevance to process planning, standardization, industry

adoption, interoperability, and support for advanced

manufacturing concepts. Further exploration of the top 8 LLM-

scored ontologies showed strong alignment with expert

evaluations, but domain-specific tasks required carefully crafted

prompts and human oversight for effectiveness.

Step 7: Provide Recommendations for Further Exploration.

LLMs generated recommendations for the most suitable

ontologies or areas for additional research based on the previous

step's results. This includes identifying underexplored concepts

and areas needing further investigation.

Step 8: Validate and Document Findings. The findings were

manually validated for accuracy and relevance, then

systematically documented. ChatGPT (GPT-4) was used to

summarize and structure the documentation.

Step 9: Iterate and Refine Search (if needed). When results

were too broad or irrelevant (e.g., Active Learning

misinterpreted as an educational method), we refined the search

prompt by adding more context.

By using this LLM-based algorithm, we conducted

comprehensive web searches and extracted relevant information

to identify the most suitable ontologies for the HumAIne project.

In the DM and ML domains, we selected the OntoDM suite

(OntoDM-Core, Onto-KDD, and OntoDT). For the

manufacturing domain, we identified the Industrial Ontologies

Foundry Core (IOF Core) as the best fit.



Figure 1: Key steps of LLM-leveraging algorithm 3 LLM-Assisted Ontology Extension with

Active Learning Concepts

Step 4: Retrieve and Summarize Information. LLMs (Google

Bard, Copilot (GPT-4), Perplexity AI) were employed to extract Integrating Active Learning (AL) into an ontology requires

and summarize key information from ontology descriptions extending it with new classes, properties, and relationships



documentation identified during the search. Using a specifically building and extending ontologies are well-documented, we leveraged GPT-4 for this task using iteratively refined prompts tuned prompt, LLMs extracted 11 characteristics for each of the found in publications, technical papers, and repository representing key AL concepts. While traditional methods of



included purpose, availability, ontology metrics, reused particularly GPT-4, were used to extend the IOF Core ontology with AL concepts. ontologies, software editors, representation language, and 34 identified DM and ML ontologies. These characteristics (see Discussion section). This section outlines how LLMs,

evaluation methodologies. This structured data, organized in Step 1: Define the Problem and Objectives. Through

table format, provided valuable insights into each ontology’s iteratively refined prompts, LLMs formulated clear objectives,

scope, quality, and reusability. From these results, we selected 6 specifying the domain (e.g., manufacturing) and key concepts

ontologies for further exploration, prioritizing comprehensive (e.g., Active Learning). These outputs were used to guide further

coverage of DM and ML concepts, adherence to ontology steps, with LLMs leveraging contextual understanding,



standards in these domains. relevant AL applications such as adaptive scheduling. Queries like "How can Active Learning improve adaptive scheduling in Step 5: Analyze and evaluate ontologies. engineering best practices, and alignment with established knowledge synthesis, and language generation to suggest



utilized to assess the relevance, content, and structure of the manufacturing?" generated valuable insights into potential use cases. where AL would be most beneficial. LLMs were further



LLMs such as GPT-4, which can process, explain, and generate Step 2: Analyze the Ontology to be Extended. By combining selected ontologies. In our study of DM and ML ontologies,

Protege’s visualization and navigation tools with GPT-4’s ability

to process textual and machine-readable data (e.g., OWL/RDF), efficiency and accuracy of reviewing, debugging, and validating

we thoroughly examined the IOF Core ontology structure and OWL code.

identified areas for introducing AL concepts. For example, GPT-



4 helped uncover key classes like “Process,” “Resource,” and

“PerformanceMetric” within IOF Core, highlighting relevant

properties for AL integration. Queries such as "What aspects of

IOF Core can benefit from AL integration?" and "What key

concepts are missing from the IOF Core ontology for integrating

Active Learning in manufacturing?" guided us in identifying

areas for improvement, including handling uncertainty and

adjusting dynamic processes.

Step 3: Identify Active Learning Concepts. The main tasks of

this step and the role of LLMs in supporting each task are

summarized in the Table 1:



Table 1: LLMs applications for Identifying AL Concepts



Task LLM Application Example Output Figure 2: Screenshot of LLM-generated code defining the

1. Identify “LearningAlgorithm” class with properties “trainingData” Use LLMs to Concepts like“Uncertainty

fundamental and “validationData” generate a list of sampling,” “Query-by-

AL concepts core AL strategies committee”

and techniques Step 5: Ensure Semantic Consistency. LLMs, such as GPT-

2. Extract 4, assisted in ensuring semantic consistency by reviewing new Query LLMs about Concept like "Query

domain- and existing ontology elements and suggesting how new AL in specific Efficiency" in decision-

specific AL concepts could align with the existing framework. For example, industrial contexts making for manufacturing

concepts an LLM suggested how an AL “QueryStrategy” class fits within

the IOF Core ontology.

3. Mine AL Example Prompt: " Review the new QueryStrategy class and Process academic Concepts like “Stream-based

concepts from suggest how it can align with the existing classes in IOF Core." papers, reports to selective sampling” from

literature LLM Output: The QueryStrategy class aligns with decision- extract relevant AL papers on AL in

terms making aspects of the Process concept. Strategies such as manufacturing

“UncertaintySampling,” “QueryByCommittee,”

4. Assign “ExpectedModelChange,” and “ExpectedErrorReduction” can Generate properties QueryStrategy class

properties to be viewed as specialized decision-making processes within the for AL ontology properties:

new classes broader process framework of IOF Core. classes “hasuncertaintySampling”

“queryByCommittee” However, LLMs cannot guarantee logical consistency and face

5. Refine and limitations in handling complex relationships, making it necessary Ensure definitions, Refined and validated terms

validate to use ontology reasoners, such as Protege or HermiT, to perform resolve overlaps based on domain-specific

terminology consistency checks. standards

Step 6: Map to Existing Ontologies. LLMs, such as GPT-4,

By prompting, LLMs generated nearly 200 fundamental AL assist in generating initial mapping suggestions by analyzing

concepts, structuring them into a hierarchy by leveraging their similarities in definitions, relationships, and properties between

vast training data. Additionally, LLMs helped generate new and existing concepts. This involves creating explicit

definitions, assisting in verifying and refining concepts. relationships like “owl:sameAs,” “owl:equivalentClass”, and

However, after a point, LLMs began repeating concepts or “owl:equivalentProperty”.

producing less relevant terms. LLMs were also effective in Example LLM Output:

generating domain-specific concepts through targeted queries. :FeedbackMechanism a owl:Class ;

For instance, querying AL in manufacturing led to concepts like owl:equivalentClass :ControlSystem ;

"uncertainty management" and "query efficiency." More rdfs:label "Feedback Mechanism" ;

specialized concepts required extraction from academic papers, rdfs:comment "Mechanisms that provide feedback in

which were cross-referenced with existing standards in DM, ML, Active Learning to control systems."

and manufacturing (e.g., CRISP-DM, IEEE 7000 Series, ISA-95, While LLMs are effective in identifying high-level

ISO 15531). Ontology learning tools like Text2Onto and similarities, they may face challenges with complex or domain-

OntoLearn were combined with LLMs for cross-verification. specific relationships, requiring further refinement. Although we

Step 4: Develop Ontology Extensions. didn’t encounter these issues during our initial work extending LLMs helped create

AL-related classes, properties, and relationships based on the IOF Core with AL concepts, we used Protege’s alignment plug-

identified concepts, using OWL-compliant syntax (see Fig. 2). ins to refine LLM-generated mappings. For more complex

By combining GPT-4’s knowledge synthesis with Protege’s mappings, tools like AgreementMaker or COMA can further

structural reasoning and consistency checking, we improved the refine the suggestions.



Step 7: Prototype and Test. LLMs, such as GPT-4, were prompt (clear, concise, and easily understood by you), b)

prompted to generate validation scenarios, competency questions, Suggestions (on what details to include in the prompt to improve

and SPARQL queries based on the integrated AL concepts. For it), and c) Questions (ask any relevant questions to improve the

instance, a prompt like "Suggest validation scenarios for adaptive prompt). We will continue this iterative process with me

scheduling with Active Learning" helped us produce realistic test providing additional information to you and you updating the

cases, including prototype code, descriptions of initial setup, prompt until it's complete.”

process flows, validation steps, and queries based on newly After 4-5 cycles, the prompts were highly optimized,

integrated concepts. ensuring relevant outputs. This refinement process reduced

SPARQL queries generated by LLMs were executed using inconsistencies and improved LLM-generated content across

Protege with SPARQL plugins to assess the ontology’s ability to both search and extension phases.

retrieve relevant information and answer competency questions. We integrated multiple LLMs, including Bing Chat (GPT-4),

However, some LLM-generated scenarios revealed Google’s Bard, and Perplexity AI, to cross-validate outputs,

limitations in domain-specific knowledge, resulting in generic reducing errors and refining results. This ensured consistency in

outputs that required refinement. Additionally, LLMs struggled LLM-generated ontologies and mappings.

with modeling intricate relationships or complex data retrieval To evaluate this multi-LLM approach, we propose the

conditions, making human oversight essential for ensuring following metrics: Inter-Model Consistency (measures

accuracy and thorough testing. alignment between LLM outputs). Error Rate Reduction (Tracks

Step 8: Iterative Refinement. Following initial prototyping how often one LLM corrects another’s errors),.Coverage of

and testing, we gathered feedback from domain experts and users Relevant Concepts (assesses LLMs' ability to capture domain-

to further refine the ontology. Validation reports were uploaded specific concepts). Although these metrics provide a framework,

to AskPDF Research Assistant (GPT-4), where LLMs reviewed formal measurements are yet to be implemented.

the reports, extracted key improvement suggestions, and refined Future stages will involve applying these metrics to validate

task lists. The LLM provided insights into areas where ontology outputs and testing extended ontologies in real-world

relationships or properties required adjustments and identified applications. This hybrid method combines LLMs and traditional

additional concepts that might have been overlooked. tools, ensuring both efficiency and accuracy in scalable ontology

Step 9: Document and Disseminate. LLMs like ChatGPT or development.

Bard were instrumental in generating comprehensive

documentation, including details on the ontology extensions.

Additionally, LLMs contributed to drafting technical reports and 5 Conclusions

research papers. This study demonstrates how LLMs can streamline ontology

Using this methodology, we successfully extended the IOF engineering by automating the search, analysis, and extension of

Core ontology with Active Learning (AL) concepts. Future domain-specific ontologies. Leveraging multiple LLMs, we

stages of the HumAIne project will focus on further validation successfully identified and extended key ontologies, including

and refinement, particularly during pilot case implementations. OntoDM and IOF Core, for the HumAIne project, improving

efficiency in generating classes, properties, and relationships.

While LLMs significantly enhance the process, they face

4 Discussion

challenges in domain-specific precision and require human

This study highlights LLMs' potential in ontology oversight, particularly for complex relationships. Traditional

engineering by reducing manual effort and increasing efficiency. tools like Protege and ontology reasoners remain critical for

LLMs rapidly identified key ontologies like OntoDM and IOF ensuring logical consistency and validation.

Core and generated structured classes, properties, and Future work will focus on refining these extended ontologies

relationships, reducing the need for manual OWL/RDF code through real-world pilot tests and applying evaluation metrics to

generation and concept mapping. However, LLMs face LLM-generated outputs. This hybrid approach, combining LLM

challenges in domain-specific precision, requiring human automation with traditional validation tools, offers a scalable

oversight to refine outputs and address nuances in specialized solution that balances efficiency with the need for human

fields. While tools like Protege excel at ensuring logical expertise. consistency, LLMs offer dynamic capabilities for generating new



tools like AgreementMaker and COMA are still necessary to This work was supported by the European Commission under the refine and validate LLM-generated mappings. concepts and relationships. Despite these advantages, traditional Acknowledgments

Horizon Europe project HumAIne, Grant Agreement No.

One strategy to mitigate LLM limitations was iterative

101120218.

prompt engineering. We refined prompts for ontology search and



extension tasks through multiple cycles of improvement. These References cycles, with LLMs like GPT-4, involved clarifying questions,

refining queries, and generating more focused outputs. Initial [1] Kommineni, Vamsi Krishna, Birgitta König-Ries and Sheeba Samuel.

“From human experts to machines: An LLM supported approach to

prompt for starting the cycle can be the following: ontology and knowledge graph construction.” ArXiv abs/2403.08345

“Your role is my Prompt Creator. Your goal is to craft the (2024): n. pag. DOI: https://doi.org/10.48550/arXiv.2403.08345

[2] Funk, Maurice, Simon Hosemann, Jean Christoph Jung and Carsten Lutz.

best possible prompt for my needs. The prompt will be used by “Towards Ontology Construction with Language Models.” ArXiv

you, [LLM's name]. I want to write about: [keyword/topic]. abs/2309.09898 (2023): DOI: https://doi.org/10.48550/arXiv.2309.09898

Based on my input, you will now generate 3 sections. a) Revised





On the Brazilian Observatory for Artificial Intelligence




Rafael Meira Silva, Joao Paulo Candia Vieira Joao Pita Costa* Luiz Costa, Alexandre Barbosa



CIAAM, C4AI, Univ. of São Paulo Cristina Godoy Oliveira CETIC, OBIA CIAAM, C4AI, Univ. of São Paulo IRCAI, Quintelligence São Paulo, Brazil São Paulo, Brazil Ljubljana, Slovenia tuca@nic.br candia@usp.br Joao.pitacosta@quintelligence.com São Paulo, Brazil alexandre@nic.br rafael@meirasilva.com.br

cristinagodoy@usp.br





ABSTRACT

Artificial Artificial Intelligence (AI) is rapidly transforming

industries and economies worldwide, with Brazil and South

America emerging as significant players in this global shift. The

fundamental need to monitor the impact of artificial

intelligence (AI) in the verticals for sustainable development,

government engagement, investment and society at large

motivated the Brazilian Artificial Intelligence Observatory

(OBIA). It is also an integral part of the Brazilian Artificial

Intelligence Plan (PBIA), and a former objective of the Brazilian



monitoring the uses of AI in the country. OBIA is part of Axis 5 Figure 1: Screenshot of OBIA showing some results on the Strategy of AI aims to become the leading platform for

preparedness of Brazilian industry to adopt AI workflows.

of the PBIA focused on supporting the regulatory and

governance process of AI. This research paper explores the The objectives of OBIA include compiling, recording, and

current state, challenges, and potential of AI development in providing information related to Artificial Intelligence in Brazil,

the region, examining how technological advancements are enabling analyses of its adoption and its main impacts on

influencing economic growth, societal change, and policy- society. It also has the mission of consolidating and

making across South America, with a particular focus on Brazil disseminating knowledge about the repercussions of this

as a leading hub of innovation. It is also investigating common technology, providing support to guide policies, strategies, and

aspects of the research agendas as with IRCAI’s SDG actions in promoting development and responsible use of AI.

Observatory, particularly in what regards machine learning The observatory gathers Brazilian data on the use and adoption

workflows and approaches complementing traditional and of Artificial Intelligence by different sectors, such as education,

crowdsourced heterogeneous data collection and analysis. business, government, health, and others (see Figure 2).



KEYWORDS data sources for analysis, such as surveys and data sets made The currently available indicators rely mostly on traditional

Artificial Intelligence, Observatory, Survey Data Analysis, Complex Data available for the team. The first product of OBIA is the book

Visualization, Multidisciplinary Collaboration. “Artificial Intelligence in Healthcare - Potentialities, Risks and



1 Perspectives”, published in July 2024. In a second line of action, Introduction it functions as a repository of guiding documents in the area, AI is increasingly shaping the economic landscape and societal originating from all parts of the world. In a third line, it acts as dynamics across Brazil and South America, positioning the an "information exchange point" between AI centers operating region as a growing hub for technological innovation. Despite in Brazil: the IAX. All indicators collected will be public and can challenges such as uneven infrastructure and regulatory be accessed on the OBIA portal [4]. hurdles, Brazil is making significant strides in AI research and The Center for Artificial Intelligence (C4AI) at the University of development, contributing to the regulation and better São Paulo, funded by FAPESP (the public agency for research understanding of the impact of AI in South America. OBIA [5] is funding in the State of São Paulo) and IBM, participates in the answering this need, serving as a platform to support the OBIA through its Humanities area. C4AI will contribute with strategy and other government actions with data on the uses qualitative research in the horizontal axes of "Legislation, and impacts of AI (see Figure 1). Regulation, and Ethical Use" and "AI Governance," while also

conducting studies across various vertical axes to be

Permission to make digital or hard copies of part or all of this work for personal monitored. The research group dedicated to this effort

or classroom use is granted without fee provided that copies are not made or comprises scholars from the fields of law, computer science,

distributed for profit or commercial advantage and that copies bear this notice

and the full citation on the first page. Copyrights for third-party components of electrical engineering, sociology, and political science, allowing

this work must be honored. For all other uses, contact the owner/author(s). for an interdisciplinary analysis of the key topics monitored by

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.sikdd.18



259

OBIA. This interdisciplinary approach will provide a Python and R programming languages, based on the TJSP API,

comprehensive view of the current state of AI development and by Jesus Filho (github.com/jjesusfilho/tjsp). For the Executive

implementation in Brazil. Various reports, articles, and data Power, a script was developed to scrape data from the Data

will be provided to support OBIA in fulfilling its mission. Download section of the Brazilian Transparency Portal

In addition to the participation of professionals from various (portaldatransparencia.gov.br/download-de-dados).

NIC.br departments, the Observatory has a network of external Currently, we are developing an automation tool, based on NLP

partners, including the Center for Management and Strategic techniques, to enhance the qualitative analysis of these court

Studies (CGEE), the São Paulo State System Data Analysis rulings, allowing for more efficient identification and

Foundation (SEADE), C4AI, CIAAM (Center of Artificial categorization of data relevant to AI research. The first

Intelligence and Machine Learning) and others. The following approach for this automation tool is using a NER (Named Entity

will explore how C4AI contributes to OBIA through a Recognition) model, to automate the identification of relevant

complementary approach, focusing on the qualitative analysis entities, including litigants and court judgments. The next step

of decisions by the São Paulo Court of Justice related to AI. would be to apply a classification model, yet to be chosen, to

filter out noise data. The process of constructing the terms for

2. web scraping is a critical step to ensure the relevance and Data and Methodology

accuracy of the data collected for AI research. This process

2.1. Legislation, Regulation, and Ethical Use: A Qualitative begins with the development of a comprehensive list of AI-

Analysis related terms, which is built using multiple authoritative

The research presented in this paper is the base of an action sources. One primary source is the OECD's report "Identifying

contributing to implement the PBIA strategy [7], responsible and Measuring Developments in Artificial Intelligence," which

for monitoring AI regulation and legislation. It has divided its offers a foundation of 226 AI-related terms identified through

research into three main areas: the Executive, the Judiciary, and extensive analysis of scientific articles, open-source systems,

the Legislative branch, combining traditional and modern data and patents. Another source is the ISO/IEC 22989:2022

collection methods. Regarding the Executive branch, standard [3], which provides a framework for AI concepts and

monitoring is being conducted through data scraping of terminologies. These terms are carefully selected, refined, and

government transparency websites based on a curated and translated into Portuguese by experts working within the

continuously updated list of AI-related terms developed by the Brazilian Technical Standards Association (ABNT) to ensure

group. This monitoring aims to understand what AI systems are that only those terms that are highly relevant and specific to AI

being purchased or contracted by public authorities. For the are included. Terms that are too general or contextually

Judiciary, we have been analyzing court decisions from the São irrelevant—such as "transparency," which could result in

Paulo Court of Appeal (TJSP) related to AI, to understand unrelated hits concerning Brazil's Access to Information Law—

judicial interpretations and rulings in the absence of specific AI are excluded to avoid false positives in the scraping process.

legislation [2]. As of the latest data scraping in August 2024, The final list of terms, consisting of 103 terms in both English

more than 13.000 relevant decisions have been identified. and Portuguese, is used to guide the web scraping data

Lastly, in relation to the Legislative Branch, the group is closely collection processes, allowing a focused and efficient retrieval

following the progress of discussions on Bill 2338/2023, which of information that aligns with specific research objectives.

focuses on AI regulation, by participating in public hearings and



issuing technical notes to guide legislators. The goal is to

expand this research to monitor AI-related legislation at the

state and municipal levels, as many municipalities are

legislating on the matter to prepare their cities to assume roles

of “smart cities”.



2.2. Monitoring and exploring the local data Figure 2: Current Dimensions of OBIA’s monitoring topics

To effectively monitor developments in AI, it is essential to

establish a comprehensive list of AI-related terms that can 2.3. How to implement and classify repositories with

guide data collection efforts. This list is derived from multiple reference documents and statistics?

sources, including scientific articles, standards like [3], and As part of the data collection and structuring process for

reports such as OECD's [1]. The monitoring process involves qualitative analysis, we are implementing and classifying

monthly web scraping of court rulings, based on the AI-related repositories containing reference documents and statistics.

terms list, from TJSP (Judiciary Power) and data from the These repositories will focus on key thematic areas, such as

Brazilian Transparency Portal (Executive Power), which occurs "Legislation, Regulation, and Ethical Use" and "AI Governance,"

on the 15th of each month. For the Judiciary Power, the scrapes and will be populated with data from sources like TJSP, the

and data treatment are performed with scripts developed in



260

Transparency Portal, and other relevant databases. By the latter focusing on case-specific factors to draw broader

combining different methods, data retrieval becomes more generalizations. From TJSP’s website, 597 rulings were

efficient and targeted, ensuring the collection of relevant reviewed: "Facial Recognition" (1), "Facial Expression

information. Web scraping supplements this process by Recognition" (1), "Machine Learning" (7), "Artificial

capturing data unavailable through APIs, ensuring Intelligence" (163), "Artificial Intelligence" in English (4),

comprehensive coverage. The data is regularly updated, with "Machine Learning" in Portuguese (3), "Learning Agent" (1),

documents classified by relevance to AI terms, creating a and "Facial Biometrics" (417).

dynamic and organized repository (see Fig 3) described in [6].





Figure 4: Nr. of Decisions per Month from Jan 2018 to Jun 2024.





Figure 3: OBIA’s guiding principles and expected results [6]





2.4. How to establish and maintain cooperation networks?



Establishing and maintaining cooperation networks requires

fostering collaboration among interdisciplinary researchers

from fields such as law, computer science, engineering,

sociology, and political science. These networks are essential

for sharing insights and methodologies related to AI

monitoring. Using APIs and web scraping tools enables access

to current data, supporting continuous knowledge exchange.

Regular workshops, webinars, and joint research projects help

keep participants engaged. Publishing reports, articles, and

datasets strengthens the network and supports OBIAs mission

to monitor AI developments comprehensively.



3 Figure 5: Number of results per AI term. Discussion of initial results



As of June 28, 2024, a total of 13,064 decisions were scraped

from the São Paulo State Court of Justice based on AI-related

terms. Out of 103 terms searched, 45 returned at least one

result. Graph 1 shows the monthly distribution of all results,

while Figure 5 (logarithmic scale) displays the distribution of

results by AI term. Both Portuguese and English terms were

used for scraping. The top 15 terms with the most occurrences

were analyzed over time, and Figure 6 presents the temporal

evolution of these results by publication date. A qualitative

review of 597 decisions from the São Paulo Court of Justice



terms like "Facial Recognition" and "Facial Biometrics," Figure 6: Evolution of results by year for top 15 terms. (TJSP) using a detailed list of AI-related terms, focused on

showing they are often used in various legal contexts, The rulings followed a structured format, and the analysis

sometimes diverging from their technological meanings. included 14 categories, such as case number, appeal type,

Terms like "Facial Expression Recognition" and "Learning judge, district, and the context of term usage. Key findings

Agent" were often interpreted in psychological or social highlighted the use of "Artificial Intelligence" and "Machine

contexts rather than purely technological ones. The analysis Learning" in commercial disputes and credit issues rather than

used analytical, comparative, and monographic methods, with solely technological matters. The rulings analyzed represent



261

decisions, rendered by collegiate bodies composed of multiple study when, e.g., capturing the attention of media on the terms

magistrates. Each ruling follows a structured format: “criminal law” and “AI” in “Brazil” in the past 12 months, where

Description and Qualification, covering aspects such as appeal, 1.4% exhibits discussions on Human Rights, and terms like

case number, judicial district, presiding judge, and parties “democracy” and “discrimination” are within the top 30. When

involved; Summary of the ruling; Report, offering a brief performing sentiment analysis over these results we can see

description of the facts; Majority Opinion; and Dissenting large variations after the summer of 2022 with a

Opinion (if applicable). The analysis was conducted with each predominantly negative sentiment regarding this search topic.

of the 14 subcategories corresponding to columns in a single



row: case number; type of appeal; reporting judge; district;

judicial body; subject matter; judgment date; publication date;

summary; parties; reasoning; final decision; context of term

usage in the full text; and relevant jurisprudence. While the first

nine categories were predefined based on the complete

jurisprudence search, the remaining five were more subjective,

created to enhance the understanding of the rulings' content

and improve data visualization. Significant findings were noted

in cases involving "Artificial Intelligence" and "Machine

Learning," where the terms were often associated with

commercial disputes, service contracts, or credit-related issues



rather than purely technological applications. A recurrent

theme in cases involving "Facial Biometrics" was the legality

and validity of loan contracts signed through biometric

recognition. The majority of decisions upheld the legality of

such contracts, highlighting issues of consent and the technical

reliability of biometric systems [1]. However, inconsistencies in

judicial reasoning were identified, where similar cases had

varying outcomes depending on the presiding judge. Overall, Figure 7: Significance of criminal law and AI in the news.

the analysis highlighted several gaps and challenges in the legal

treatment of AI-related technologies, particularly concerning ACKNOWLEDGMENTS



transparency, fairness, and consumer protection. The study We would like to express our sincere gratitude to the Center for

underlined the need for more consistent legal standards and Artificial Intelligence (C4AI) at the University of São Paulo

better understanding among judges of the technical nuances (USP), supported by FAPESP and IBM, for their invaluable

involved in AI applications to ensure fair and equitable rulings. support to the AI Observatory team. We thank to the CIAAM for

their continued collaboration and contributions to this

4 Conclusions and further work research. We thank the support of the European Commission

The qualitative research findings from the analysis of court project ELIAS - Lighthouse of AI for Sustainability (10080425).



decisions related to AI reveal several key conclusions. AI- REFERENCES related terms such as "Facial Recognition," "Voice Recognition,"

and "Autonomous Systems" are frequently used in judicial [1] Baruffaldi, Stefano, et al. (2020) Identifying and measuring developments

in artificial intelligence: Making the impossible possible. OECD.

contexts that extend beyond their traditional technological [2] Cristina Godoy B. de Oliveira, Otávio de Paula Albuquerque, Emily Liene

meanings, intersecting with areas like consumer protection, Belotti, Isabella Ferreira Lopes, Rodrigo Brandão de A. Silva, Glauco Arbix.

contract law, and fraud. The inconsistency in judicial reasoning Horizonte, Brazil, September 25 Intelligent Systems: 12th Brazilian Conference, BRACIS 2023, Belo –29, 2023, Proceedings, Part I, pp 18 – 32.



and varying outcomes in similar cases highlight the need for [3] ISO (2022) Information technology — Artificial intelligence — Artificial intelligence concepts and terminology. ISO/IEC 22989:2022. [Online]. clearer legal frameworks and a deeper understanding of AI's Available: https://www.iso.org/standard/74296.html/ [27 8 2024]

technological implications among judges. Moving forward, the [4] Luiz Costa et al. (2024) The Brazilian Artificial Intelligence Observatory

incorporation of NLP techniques into the analysis will help (OBIA). [Online]. Available: https://www.obia.nic.br/ [27 8 2024]

[5] MCTI (2021). Brazilian Strategy of Artificial Intelligence. [Online]. Available:

extract key arguments from judicial decisions, providing ebia-documento_referencia_4-979_2021.pdf (www.gov.br) [07 9 2024]

deeper insights into the legal discourse on AI. This will enhance [6] MCTI (2023). OBIA: Observatório Brasileiro de Inteligência Artificial,.

the robustness of future research on AI regulation and its Available: https://www.gov.br/mcti/pt-br/acompanhe-o-

mcti/transformacaodigital/arquivosinteligenciaartificial/1_ebia-reuniao-

implications for public policy. ro_7_24_05_2023_anexo_2_eixo2-pdf.pdf [27 8 2024]

Furthermore, a preliminary analysis of news using the NLP [7] PBIA (2024). Brazilian Artificial Intelligence Plan . [Online]. Available:

https://www.gov.br/mcti/pt-br/acompanhe-o-

capabilities of the Eventregistry.org system (see Figure 7) show mcti/noticias/2024/07/plano-brasileiro-de-ia-tera-supercomputador-e-

how this source can provide complementary results to the investimento-de-r-23-bilhoes-em-quatro-

anos/ia_para_o_bem_de_todos.pdf/view [07 09 2024]



262





Pojavljanje incidentov ob uporabi Umetne Inteligence




Marko Grobelnik Besher M. Massri Alenka Guček Dunja Mladenić

Department for Artificial Department for Artificial Department for Artificial Department for Artificial

Intelligence, Intelligence, Intelligence, Intelligence,

Jozef Stefan Institute Jozef Stefan Institute Jozef Stefan Institute Jozef Stefan Institute

Ljubljana Slovenia Ljubljana Slovenia Ljubljana Slovenia Ljubljana Slovenia

marko.grobelnik@ijs.si m.besher.massri@gmail.com alenka.gucek@ijs.si dunja.mladenic@ijs.si



Povzetek takšne incidente preprečujejo ali vsaj zmanjšujejo. Predstavljeni

sistem deluje kot orodje, ki pomaga uporabniku, ki si prizadeva v

Prispevek predstavi prve rezultate ob uporabi sistema, ki je bil em času slediti dejanskim incidentom, povezanim z umetno

realn

zasnovan in razvit v sodelovanju z OECD za spremljanje

incidentov, povezanih z umetno inteligenco. Glavna motivacija teh inteligenco, ter zagotavljati dokazno osnovo za oblikovanje okvira

prizadevanj je podpora zakonodaji, povezani z umetno inteligenco, poročanja o incidentih in povezanih političnih razpravah o UI. Z

in učinkovitemu oblikovanju politik, saj sistem zagotavlja vpoglede zbiranjem podrobnih vpogledov v vsak incident omogoča učenje iz

na podlagi zbranih podatkov. OECD AI Incidents Monitor za preteklih napak ter spodbuja varnejši in bolj odgovoren razvoj ter

spremljanje incidentov, povezanih z umetno inteligenco, uporabo umetne inteligence. Koristi skupnosti, ki se ukvarja z

dokumentira incidente in nevarnosti v zvezi z umetno inteligenco, umetno inteligenco, saj izpostavlja trende in področja, ki

da bi oblikovalcem politik, strokovnjakom za umetno inteligenco potrebujejo pozornost ali regulativni poseg. in vsem zainteresiranim stranem po vsem svetu pomagal pridobiti

dragocen vpogled v tveganja in škodo, ki jo povzročajo sistemi Prednost sistema je, da je zbiranje podatkov avtomatizirano,

umetne inteligence. Ideja je, da bo sistem sčasoma pomagal kar je prednost v primerjavi s podobnimi repozitoriji, ki so

povečati ozaveščen urednikovani ročno, kot je na primer AIAAIC Repository [2]. ost javnosti in vzpostaviti skupno razumevanje

incidentov in nevarnosti umetne inteligence, in tako prispeval k Repozitorij je prosto dostopen in namenjen tako oblikovalcem

zaupanja vredni umetni inteligenci. politik, kot razvijalcem UI, raziskovalcem, pravnikom in javnim

organizacijam.

Ključne besede V nadaljevanju predstavimo metodologijo za spremljanje

umetna inteligenca, analiza podatkov, oblikovanje politik,incidenti incidentov, prikažemo delovanje sistema na nekaj realnih primerih,

Abstract predstavimo deležnike in nekaj zaključkov.



This paper presents a system designed and developed in



collaboration with OECD for monitoring of AI-related incidents. 2 Metodologija



The main motivation behind the efforts is in supporting AI-related Metodologija OECD za spremljanje AI incidentov se osredotoča na

legislation and effective policymaking, as the system provides identifikacijo in klasifikacijo incidentov, s čimer zagotavlja

evidence based on the collected data. The OECD AI Incidents vpogled v realno dogajanje in podpira razvoj okvira za poročanje o

Monitor documents AI incidents and hazards to help policymakers, incidentih. Začetna točka je identifikacija in klasifikacija

insights into the risks and harms of AI systems. The idea is that over i AI practitioners, and all stakeholders worldwide gain valuable ncidentov, ki so poročani v uglednih mednarodnih medijih, s

time the system will help to raise awareness and establish a pomočjo modelov strojnega učenja, kar omogoča gradnjo

collective understanding of AI incidents and hazards contributing zanesljive baze podatkov (incidenti so zajeti od 2014 naprej).

to trustworthy AI. Kljub prizadevanjem, ti incidenti predstavljajo le

podmnožico vseh globalnih AI incidentov. Incidenti so razvrščeni

Keywords glede na resnost, industrijo, povezane AI principe (OECD AI

Artificial Intelligence, data analysis, policy making, AI incidents Principles [3]), vrste škode in prizadete deležnike. Analiza temelji

na naslovih, povzetkih in prvih odstavkih novinarskih člankov, pri

čemer se pridobljeni podatki uporabljajo za izgradnjo zanesljive,

1 Uvod objektivne in kakovostne baze podatkov o incidentih, povezanih z

AI. Kot vir novic služi sistem Event Registry [4].

Ob vse širši uporabi umetne inteligence (UI) prihaja tudi do

incidentov ob njeni uporabi. Spremljanje teh incidentov je nujno za Razvoj sistema, h kateremu smo prispevali, nadgrajuje delo

zagotavljanje preglednosti, nadzora in razvoj politik, ki lahko mednarodne skupine strokovnjakov (OECD Expert group), ki

razvija teoretično ogrodje za poročanje o incidentih, definira pojem

∗Article Title Footnote needs to be captured as Title Note AI incidenta in oblikuje povezano terminologijo, kot je AI

†Author Footnote to be captured as Author Note

nevarnosti in njene potencialne posledice. Podrobna metodologija

Permission to make digital or hard copies of part or all of this work for personal or in definicije so razložene na spletni strani OECD: classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full https://oecd.ai/en/incidents-methodology.

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.sikdd.26

263





Slika 1 https://oecd.ai/en/incidents) prikazuje vmesnik za iskanje po Prikaz začetne strani OECD monitorja AI incidentov (

konceptih, vizualizacijo incidentov v času (spodaj levo; y os: število incidentov; x os: čas (2014-danes) in statistični povzetek (spodaj

incidentov glede na izbrano področje (12 883 incidenti in

3 AI Incidents Monitor nevarnosti o katerih je poročalo 70612 novičarskih člankov),

statistiko za zadnji mesec in mesece z največjimi vrednostmi

AI Incidents Monitor je do konca avgusta 2024 zaznal preko 12 000 (februar 2024). Iz statistik o spremembi glede na mesec, na

incidentov in nevarnosti v zvezi z UI, Kot je razvidno iz Slike 1. in na leto, vidimo padec števila incidentov in nevarnosti o četrtletje

Sistem je popolnoma avtomatski in zaznava incidente s katerih so mediji poročali v zadnjem mesecu glede na prejšnji

skeniranjem številnih podatkov objavljenih v novicah, ter nato s mesec oz. prejšnje četrtletje. pomočjo UI določa kaj se označi kot incident ali nevarnost. Na



incidentov v času (levo) in pripadajoča statistika (desno). Sistem omogoča napredno filtriranje po incidentih UI za sledeče Uporabnik lahko izbira med absolutnim prikazom incidentov (kot kategorije: čas, država, industrija, princip UI, resnost, tip škode, naslovni strani (Slika 1) je prikazan črtni diagram naraščanja 3.1 Primer analize pojavitve incidentov UI na Sliki 1) ali v ustreznem meniju izbere pod-področja. Če se

poglobimo v prikaz na Sliki 1, vidimo z

različnimi barvami oškodovanci, tip iskanja po vsebini (glej Sliko 1). Tako so na

označeni kumulativni incidenti (vijolična) oz. njihovo trimesečno primer možne vrednosti za resnost: smrt, poškodba, nevarnost, ne-

povprečje (modra). St

atistika na desni prikazuje absolutno število fizična nevarnost, možni tipi škode pa so: fizična, psihološka,

ekonomska, ugled, javni interes, človekove pravice, neznana.



264



Sistem omogoča napredno iskanje po konceptih, recimo za njihovih pravnih posledic lahko usmerja razvoj robustnih okvirov

primer generativne UI, s e, ki kažejo 2302 upravljanja AI. istem poroča statistik

incidenta in nevarnosti, en od primerov incidentov, ki jih je sistem Nazadnje lahko javne organizacije in zagovorniške skupine

zaznal pa se nanaša na Apple in razvoj »AI personality«, ki naj bi uporabljajo AIM za spremljanje družbenih vplivov umetne

nadomestil obstoječi Applov Siri. inteligence, s čimer zagotavljajo, da so interesi javnosti zaščiteni.

Poleg konceptov uporabnik lahko nadalje izbere tudi To lahko vključuje analizo vzorcev incidentov z umetno

napredno iskanje za natančnejšo identifikacijo želene podskupine inteligenco za zagovarjanje boljše zaščite potrošnikov in etičnih

incidentov, ki ga zanimajo. Tako lahko recimo izbere državo, ki je standardov pri uvajanju AI.

povezana s poročanjem o incidentih in nevarnostih UI. Na Sliki 2

je tako prikazan primer iskanja po kategoriji države, za Slovenijo.

Sistem najde dva incidenta, ki sta bila povezana s Slovenijo. Prvi 5 Diskusija

incident se nanaša na Microsoftov povečan prispevek k emisiji V prispevku smo predstavili OECD-jev monitor incidentov umetne

CO2. inteligence, pri razvoju katerega smo sodelovali. Sistem služi kot Na prvi pogled ni očitna povezava s Slovenijo, ko pa

pogledamo povezane novice naletimo na omembo Slovenije: dober vir za širok nabor uporabnikov, ki želijo razumeti in

»…But the tech giant’s electricity consumption last year rivaled upravljati tveganja, povezana s tehnologijami UI. Sistem se

that of a small European country—beating Slovenia easily.« [6]. nadgrajuje z dodatnimi podatkovnimi viri.

Vsak primer je tudi semantično označen. Tako je na Sliki 2 za prvi V prihodnosti je predvideno, da bo omogočen odprt postopek

primer označena povezanost s principi UI učinkovitost, trajnostni oddaje podatkov, ki bo dopolnil informacije o incidentih,

razvoj. kov: splošno pridobljene iz trenutnih virov. Nadaljnje delo zajema tudi Microsoft s tem lahko prizadene več deležni

javnost, podjetja, delavce, vlade (Affected Stakeholders, Slika 2). avtomatsko analizo podatkov o incidentih za namen bolj celovitega

Poleg tega predstavlja nevarnost za okolje, javne interese in vpogleda. To vključuje avtomatsko odkrivanja vzorcev, kot so

človekove pravice (Harm type, Slika 2). Klasificirano je kot ne- verižne reakcije ali učinki na več industrij hkrati. Za potrebe

fizična nevarnost (Severity, Slika 2). preverjanja resničnosti poročanih incidentov, bi lahko vključili

Iz podrob nih analiz, ki so zbrane v nedavnem poročilu kombiniranje informacij iz več neodvisnih virov in uporabljal

»Observatory of the social and ethical impact of artificial algoritme za odkrivanje lažnih novic, kot tudi ročno preverjanje.

intelligence« [5] , je razvidno, da večina incidentov (96%) spada

pod kategorijo ne-fizično nevarnih, a imajo lahko zelo resne

psihološke in finančne posledice, vključujoč nadlegovanja, Zahvala

odvisnosti in škodo ugledu tako posameznikom kot tudi Delo, opisano v tem prispevku, so podprli OECD in številni

inštitucijam. mednarodni eksperti, Ministrstvo za digitalno preobrazbo in Javna

agencija za raziskovalno dejavnost Republike Slovenije v okviru

CRP V2-2272 in V5-2264.

4 Deležniki

Acknolwedgements

OECD-jev monitor incidentov AI (AIM) je dragoceno orodje,

zasnovano za različne deležnike, ki sodelujejo pri razvoju, The described work was supported by OECD and many os its

regulaciji in uporabi umetne inteligence. Potencialni uporabniki international experts, Slovenian Ministry of Digital Transformation

tega orodja vključujejo oblikovalce politik, razvijalce AI, and Slovenian Research and Innovation Agency under CRP V2-

raziskovalce, pravne strokovnjake in javne organizacije.

2272 and V5-2264.

Oblikovalci politik lahko AIM uporabljajo za sledenje in

analizo podatkov v realnem času o incidentih, povezanih z AI, po

vsem svetu, kar jim pomaga pri oblikovanju informiranih in na Literatura

dokazih temelječih predpisov. Zmožnost orodja za kategorizacijo [1] OECD AI Incidents Monitor (AIM), https://oecd.ai/en/incidents. August

incidentov glede na resnost, industrijo in vrste škode je ključna za 2024

razumevanje širših posledic tehnologij umetne inteligence in [2] AIAAIC Repositoryhttps://www.aiaaic.org/aiaaic-repository. August 2024



oblikovanje politik, ki zmanjšujejo tveganja. August 2024 [3] OECD AI Principles for trustworthy AI https://oecd.ai/en/ai-principles

Razvijalci AI in raziskovalci lahko koristijo AIM, da [4] Gregor Leban, Blaz Fortuna, Janez Brank, and Marko Grobelnik.

prepoznajo pogoste težave, povezane s sistemi umetne inteligence. 2014. Event registry: learning about world events from news. In



S preučevanjem incidentov, zabeleženih v AIM, lahko izbo Proceedings of the 23rd International Conference on World Wide Web, ljšajo 107–110. svoje modele, da bi se izognili podobnim težavam in povečali [5] Richard Benjamins, Another Inconvenient Truth: The Societal

varnost ter zanesljivost aplikacij umetne inteligence. Emergency of AI Incidents - We Should Do Something About It

Pravni strokovnjaki lahko uporabljajo AIM za pridobitev https://www.odiseia.org/post/another-inconvenient-truth-the-societal-emergency-of-

vpogledov v spreminjajočo se pokrajino tveganj, povezanih z ai-incidents-we-should-do-something-about-it

[6] Microsoft’s AI Push Imperils Climate Goal as Carbon Emissions Jump

umetno inteligenco, kar bi lahko bilo koristno v pravnih primerih 30% https://tanaka-preciousmetals.com/en/elements/news-cred-20240821/

ali ocenah skladnosti. Razumevanje preteklih incidentov in



265





Slika 2 Prikaz naprednega iskanja na OECD monitorju AI incidentov (https://oecd.ai/en/incidents) filtrirano po državi za Slovenijo.

Podane so statistike dveh incidentov o katerih je poročalo 25 novičarskih člankov, in spodaj sta prikazana oba incidenta.



266





Perception of AI in Slovenia




Abdul Sittar Alenka Guček Dunja Mladenić

abdul.sittar@ijs.si alenka.gucek@ijs.si dunja.mladenic@ijs.si

Jožef Stefan Institute Jožef Stefan Institute Jožef Stefan Institute and Jožef

Jamova cesta 39 Jamova cesta 39 Stefan Postgraduate School

Ljubljana, Slovenia Ljubljana, Slovenia Jamova cesta 39

Ljubljana, Slovenia



Abstract Following are the main scientific contributions of this paper:

This paper introduces the AI News Monitor system developed (1) We present a methodology to understand public perception

for real-time monitoring and analysis of artificial intelligence about AI in news.

(AI) perception in global and local news media. Leveraging data (2) We analyse some trends in AI’s Perception.

from the Event Registry platform, the AI News Monitor tracks The remainder of the paper is structured as follows. Section 2

AI-related news articles across multiple dimensions, providing describes the methodology to collect historical data, AI news

insights through three key views: a global historical overview, categories and gaining insights in public perception about AI in

current global trends, and local trends specific to Slovenian media. news. Section 3 presents the analysis of trends in AI’s Perception.

The system facilitates both passive observation of AI discourse We present different user scenarios and possible applications

and active exploration of specific AI-related events. Our illustra- of AI News Monitoring in Section 4 and discussion in Section

tive analysis reveals significant global trends, including height- 5. Section 6 concludes the paper and outlines possible areas of

ened media focus on deep learning, generative AI, and robotics, future work.

and examines the implications of these trends on public trust in Plotly Graphs

AI. Additionally, the paper discusses the practical applications BERT Topic Modeling Global Overview of the AI News Monitor for stakeholders such as policymakers, Global Trends Front End requests Back End Database

journalists, business leaders, and researchers. We conclude with a Local Trends

discussion on the impact of media coverage on public perception requests

of AI and propose possible future enhancements of the system,

including broader language and source coverage. Users





Keywords Figure 1: Architecture for Real-Time AI News Monitor-



datasets, artificial intelligence, media monitoring, perception ing and Visualization based on Event Registry and imple-

mented using Flask and Plotly.

1 Introduction

Artificial Intelligence (AI) is increasingly becoming an integral

part of society, influencing various aspects of daily life and in- 2 Methodology

dustries [4]. As AI continues to evolve, so does its portrayal in



tion and trust. Understanding how AI is perceived globally and lic perception involves two key steps: 1) identifying AI-related categories and gathering news within these categories, and 2) the media, which plays a critical role in shaping public percep- The proposed approach to creating a web service to analyze pub-



ways that are socially acceptable and trustworthy [3, 4]. egories, news publishers, and highlights current trends among both global and local (Slovenian) news sources (see Figure 1). In response to this need, we have developed the AI News Mon- Firstly, we selected AI-related categories based on the Slovenian itor system designed for real-time monitoring and exploratory to ensure that AI technologies are developed and deployed in developing a web service that displays trends across these cat- locally is essential for policymakers, businesses, and researchers

analysis of AI-related news coverage. The AI News Monitor of- AI observatory1 2 and Wikipedia. The key categories associated



The AI News Monitor system is structured around three main fers a comprehensive view of how AI is discussed in the media, with Artificial Intelligence include ‘Generative AI’, ‘Artificial In- telligence’, ‘NLP’, ‘Chat-GPT’, ‘Deep Learning’, ‘Robotics’, ‘Com- capturing data from the Event Registry platform on a monthly puter Vision’, ‘Neural Networks’, ‘Graph Neural Networks’, ‘Self- basis [7]. supervised Learning’, and ‘Zero-shot Learning’. Next, we collected news articles from the last year related to views: a global overview that presents historical data from the these categories. These articles were classified into the appropri- past year, global trends that highlight recent AI-related events, ate categories based on Wikipedia concepts, and we also obtained and local trends focusing on mentions of AI by Slovenian news sentiment data from Event Registry. The portrayal of AI-related sources. These views allow the users to either passively monitor news significantly impacts public perception, with the emphasis ongoing developments in AI or actively explore specific events on risks, benefits, or ethical concerns shaping public opinion and and trends that may influence public opinion. driving narratives that can either build trust or instill fear[8],[12],

Permission to make digital or hard copies of all or part of this work for personal [1].

or classroom use is granted without fee provided that copies are not made or To understand global trends, we retrieved news events published

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this globally in the last month. For local trends, we focused on news

work must be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia 1 http://siai.ijs.si/dashboards/Main/SlovenianObservatoryIntro?globalCountry=SV

© 2024 Copyright held by the owner/author(s). N

https://doi.org/10.70314/is.2024.sikdd.14 2http://country-dashboards.ijs.si/dashboards/Main/Index?

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Trovato et al.





Figure 2: Time series of the number of news articles by specific areas (in colors, at the top). Detailed view upon precise

exploration (middle) and corresponding sentiment of news from specific areas (at the bottom).





articles published by the top 50 Slovenian news publishers.

Finally, we employed topic models to analyze the corpus of news

articles and extract underlying themes [9], [2].



3 Analysis of trends of AI’s Perception



3.1 Global Overview

The global overview provides a historical review of global AI-

related news (see Figure 2). Users can explore the number of news

articles across 13 AI fields (Generative AI, Chat-GPT, Deep Learn-

ing, Robotics, Computer Vision, Neural Networks, Graph Neural

Networks, Artificial Intelligence, Federated Learning, Few-shot

Learning, Meta Learning, Self-supervised Learning, and Zero-

shot Learning) or by news providers and have an overview of

the sentiment of the news.

Global trends allow for the review and exploration of global AI- Figure 3: A detailed view of Global Trends, showing the

related trends based on captured events from the last month. option to select news events based on chosen AI fields.

Figure 3 shows a detailed view of the Global Trends, where a

written report of the number of news articles and events, a his-

togram of the number of AI-related news articles over time, and 3.3.1 Global Overview. In the historical overview of AI trends in

the ability to explore the last 10 events in a selected field. March 2024 (Figure 2), there was a significant increase in the num-

ber of news articles and interest in deep learning, generative AI,

and robotics. Specifically, on March 18th, there were 1,800 news

articles about generative AI, 970 about robotics, and 274 about

3.2 Local Trends deep learning. This spike in news highlights several key events:

Local trends allow for the review of news from Slovenian news one of the standout stories was the launch of Gen-2 by Runway,

providers for the last month. The local trends show the detailed a generative video model capable of creating high-quality short

view, where a written report of the number of news articles and clips. An important topic was the use of AI in political campaigns,

events, a histogram of the number of AI-related news articles particularly the creation of deepfakes and misinformation. This

over time, and the ability to explore (see Figure 4). raised concerns about AI’s impact on elections and voter trust. In

the field of robotics, researchers were inspired by advancements

in generative AI to develop more versatile robots. These new

3.3 EXAMPLES OF TRENDS robots can perform various tasks using a single, comprehensive

Perception of AI in Slovenia Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



model, demonstrating significant progress in robotic capabili- News Monitor can support their specific goals. Other potential

ties. Overall, the sentiment in March 2024 was positive (as seen stakeholders are business executives, NGOs, researchers and ed-

from the sentiment analysis), reflecting enthusiasm and optimism ucators.

regarding this technological progress. The increased media atten- Policy Makers: Scenario: A policymaker uses AI News Monitor

tion highlights the rapid development and growing importance to track trends in robotics.

of AI in various fields. Background: Jure, a decision-maker at a government agency for

technology and innovation, is tasked with drafting new guide-

lines for the development and implementation of robotics in



Slovenia. To understand the broader context and local trends, he

needs to explore the global perception of robotics and compare

it with local perspectives.

Steps: Step 1: Searching for a Global Overview: Jure logs into

AI News Monitor and searches for "robotics" under the global

overview section. The system displays a line chart showing how

robotics has been mentioned over time, along with a sentiment

graph for the past year. He finds that robotics is globally discussed

with mostly positive sentiment, particularly in Asia and North

America. Step 2: Global Trends: Jure selects "robotics" among the

topics and reviews recent events on this subject. He chooses an

event focusing on robotics in the EU and examines the sentiment

of the publications and the main themes. In his browser, he looks

at the specific articles and discovers that discussions predomi-

Figure 4: A detailed view of Local Trends, showing the nantly revolve around automation and industrial robotics. Step

option to select news events based on chosen AI fields. 3: Local Trends in Slovenia: Next, Jure is interested in a review

for Slovenia to understand how robotics is perceived at the local



selected the news story "AI and heat waves pose dual threats to of recent articles from Slovenian media. By using the browser, he discovers that discussions mainly focus on the impact of robotics 3.3.2 level. The dashboard for the selected topic displays an analysis Global Trends. In our examination of global trends, we



these articles, as shown in the middle graph (Figure 4), fluctuates Jure finds that local concerns are more focused on social and economic impacts. He includes these insights in his preparatory between positive and neutral. Upon delving into the content of documents for the new guidelines. Step 4: Compiling the Report these publications, we found that Forbes focused on the issue of more articles on this topic compared to others. The sentiment of on employment and the potential use of robots in healthcare. the power grid" and found that two specific newspapers published



applications in various fields. sentiment graphs and media summaries, from AI News Monitor. He compiles a report that summarizes global trends and local fake news generated by AI, while Lexology explored future AI and Recommendations: Finally, Jure exports key data, including

3.3.3 Local Trends. In the last month (at the time of writing the concerns and proposes balanced guidelines that promote innova-

report, this was June 2024), there was an increase in AI-related tion in robotics while addressing social impacts.

news from Slovenian news providers, particularly from Delo.si Journalists: Scenario: A policymaker uses AI News Monitor to

and Sta.si (Figure 5). When analyzing the sentiment of these arti- track trends in robotics.

cles, most were neutral, with a few expressing positive opinions Background: Ana, a journalist at a technology magazine, is tasked

about AI. Delo.si focused on the growing adoption of AI by com- with writing an article on the growing trend of using genera-

panies in Slovenia, highlighting discussions on the potential of tive AI to create videos. She needs to explore both global trends

quantum computing and recent advancements in AI technology. and local perspectives in Slovenia to provide a comprehensive

This coverage indicates a balanced view of AI’s impact and po- overview.

tential. Sta.si reported on the construction of a state-of-the-art Steps: Step 1: Searching for a Global Overview: Ana searches for

data center in Maribor, which will also house a supercomputer. "generative AI" under the global overview section. The system

This event represents a major development in Slovenia’s techno- displays a line chart showing that this topic is on the rise, identi-

logical infrastructure. Additionally, Sta.si wrote about AI trends fies the media outlets reporting on generative AI, and provides

that benefit semiconductor manufacturers, reflecting a positive a sentiment graph for the past year. Step 2: Global Trends: Ana

outlook on the economic impact. selects "generative AI" and reviews recent events on this topic.

She focuses on deepfake video generation, checking who has

written about it and what the main themes are. She then looks

4 User Scenarios and Applications up these articles in her browser. Step 3: Local Trends in Slovenia:

The AI News Monitor can cater to a range of stakeholders with Ana shifts her focus to Slovenia to understand local views. The

varying use case objectives [10], [6], [5]. Policy makers can utilize dashboard reveals that Slovenian media coverage is largely posi-

the developed system to track global and local trends in AI-related tive, particularly for certain providers. However, Ana realizes the

topics, enabling them to craft data-driven policies that balance need to include concerns about authenticity and misinformation

innovation with societal concerns. Journalists can leverage the to provide a balanced perspective. Step 4: Compiling and Writing:

system to gather comprehensive insights into public sentiment Ana exports key data, including sentiment graphs and media

and media coverage, enriching their reporting with accurate and summaries, from AI News Monitor. She drafts her article, start-

timely information [11]. Detailed scenarios for both policy mak- ing with global trends and then delving into specific concerns in

ers and journalists are explained below, illustrating how the AI Slovenia, enriched with visual data.

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Trovato et al.





Figure 5: Time series of the number of news articles by news provider in Slovenia (at the top). Sentiment analysis (in the

middle) and frequency of topics for this period (at the bottom).



5 Discussion of Digital Transformation and Slovenian Research and Innovation

Services like AI News Monitor can play a role in fostering greater Agency under CRP V2-2272.

transparency around AI by offering detailed insights into how AI



public sentiment and highlighting both positive and negative [1] Iyad AlAgha. 2021. Topic modeling and sentiment analysis of twitter dis- cussions on covid-19 from spatial and temporal perspectives. Journal of trends, it helps to ensure that the development and deployment Information Science Theory and Practice , 9, 1, 35–53. is being discussed across various media platforms. By tracking References



While AI News Monitor offers valuable insights, it has limitations, of AI technologies are aligned with public concerns and expecta- [2] David Alvarez-Melis and Martin Saveski. 2016. Topic modeling in twitter: Tenth international AAAI conference aggregating tweets by conversations. In tions. on web and social media . [3] Stephen Cave, Kate Coughlan, and Kanta Dihal. 2019. " scary robots" exam-



such as its reliance on media reporting, which may not capture ining public responses to ai. In Proceedings of the 2019 AAAI/ACM Conference on AI, Ethics, and Society , 331–337. the full spectrum of public opinion. Additionally, potential biases [4] Ethan Fast and Eric Horvitz. 2017. Long-term trends in the public perception in media sources or the algorithms used for sentiment analysis of artificial intelligence. In Proceedings of the AAAI conference on artificial number 1. Vol. 31. could skew the results, presenting challenges in ensuring a fully intelligence [5] Fabian Gilson, Matthias Galster, and François Georis. 2020. Generating use accurate and balanced representation of public perception. case scenarios from user stories. In Proceedings of the international conference

on software and system processes, 31–40.

[6] Debasish Kundu and Debasis Samanta. 2007. A novel approach of prioritizing



6 Conclusions (APSEC’07) use case scenarios. In 14th Asia-Pacific Software Engineering Conference

. IEEE, 542–549.



AI News Monitor was developed to understand and track pub- [7] Gregor Leban, Blaz Fortuna, Janez Brank, and Marko Grobelnik. 2014. Event registry: learning about world events from news. In Proceedings of the 23rd lic sentiment around AI, offering policymakers, journalists, and International Conference on World Wide Web , 107–110.

other stakeholders the insights needed to make informed de- [8] Kalle Lyytinen, Heikki Topi, and Jing Tang. 2021. Information systems cur-

cisions. AI perceptions can be monitored globally and locally, riculum analysis for the macude project. Communications of the Association

for Information Systems, 49, 1, 38.

for the context of Slovenia. However, there are opportunities [9] Rishabh Mehrotra, Scott Sanner, Wray Buntine, and Lexing Xie. 2013. Im-



for future work to enhance its capabilities. Expanding the its proving lda topic models for microblogs via tweet pooling and automatic labeling. In Proceedings of the 36th international ACM SIGIR conference on coverage to include more languages and diverse sources would Research and development in information retrieval , 889–892. provide a more global perspective, while refining sentiment anal- [10] Frank Moisiadis. 2000. Prioritising use cases and scenarios. In Proceedings

ysis techniques could improve accuracy and reduce potential 37th International Conference on Technology of Object-Oriented Languages

biases. and Systems. TOOLS-Pacific 2000 . IEEE, 108–119.

[11] Abdul Sittar, Daniela Major, Caio Mello, Dunja Mladenić, and Marko Grobel-

nik. 2022. Political and economic patterns in covid-19 news: from lockdown

to vaccination. IEEE Access, 10, 40036–40050.

[12] Abdul Sittar, Dunja Mladenić, and Marko Grobelnik. 2022. Analysis of

7 Acknowledgments information cascading and propagation barriers across distinctive news

This work was supported by the European Union through AI4Gov events. Journal of Intelligent Information Systems, 58, 1, 119–152.

(101094905) and TWON (101095095) EU HE projects and Ministry



What will happen tomorrow? Predicting future event types for



businesses



Tesia Šker Jože M. Rožanec

Jožef Stefan Institute Jožef Stefan International Postgraduate School

Ljubljana, Slovenia Ljubljana, Slovenia

tesia.sker@gmail.com joze.rozanec@ijs.si



Gregor Leban Dunja Mladenić

Event Registry d.o.o. Jožef Stefan Institute

Ljubljana, Slovenia Ljubljana, Slovenia

gregor@eventregistry.org dunja.mladenic@ijs.si



ABSTRACT location prediction, semantics prediction, and a combination of

Strategic foresight helps organizations anticipate future chal- these. Each goal is divided into subgoals for which various tech-

lenges and opportunities, allowing them to handle uncertainty niques can be applied. According to the classification provided

better. While strategic foresight is becoming more widely adopted by Zhao, our technique can be classified as a semantic prediction.

across organizations, the process still heavily relies on expert In this research, we explore how graphs can be used to model

knowledge, and little of it has been automated through artificial media news events and to forecast event types in the near future.

intelligence. In this research, we explore how media news events By doing so, we provide a valuable tool for decision-makers,

can be analyzed to forecast event types that will take place in offering them a clearer view of potential outcomes. Specifically,

the near future. In particular, we consider it a supervised ma- our research focuses on using a JSON dataset containing a variety

chine learning problem with a well-defined set of event types and of articles about a particular business company. We create a

leverage graph representation of the media news events to create graph representation of the articles and use Graph2Vec to create

graph embeddings, train a classifier, and predict event types that embeddings that can be used downstream to fit other machine-

will likely occur one day ahead. We validated our approach on a learning models. Using this information, we apply a Random

real-world dataset of an American multinational conglomerate Forest Classifier to predict the categories of articles about the

operating in industry, worker safety, healthcare, and consumer company for the following day.

goods. In particular, we expect this to be useful to give organizations

a competitive advantage in fast-changing markets [5]. While

KEYWORDS human expertise is valuable, it varies from person to person,



strategic foresight, event prediction, machine learning, graphs leading to inconsistent predictions. Manually analyzing large datasets is also time-consuming and prone to errors. AI, however,



1 can process vast amounts of data, spot patterns, and predict future INTRODUCTION event types more accurately. Strategic foresight helps organizations anticipate future chal- This work is structured as follows. Section 2 presents related lenges and opportunities, allowing them to handle uncertainty work that is relevant for this paper. Section 3 describes the data in better [9]. Therefore, predicting future event types as a part of the dataset, and the data extraction process. Section 4 introduces strategic foresight became necessary for businesses to manage a new approach to predict future event types. Section 5 presents their operations without significant losses. Various events on the results of this research. Section 6 concludes this work and a major scale, such as floods, earthquakes, internet failures, or proposes future improvements. pandemics, as we are witnessing recently, or on a minor scale,

such as road closures due to sports events or promotions at fairs,

can have a major impact on business operations. By predicting

the next event type, businesses can adjust prices, reschedule 2 RELATED WORK

staff, manage stocks, reschedule transportation routes to avoid In recent decades there has been an increasing interest in strategic

delays, and more, and thus reduce losses or increase their sales foresight in the academic field. According to Fergnani (2020)

and profits. [2] this is because by "using corporate foresight, organisations

There is currently a massive number of articles written on Fu- can reconfigure their strategy based on the analysis of business

ture Event Predictions. Based on Zhao [11], the event prediction opportunities suggested by future possibilities". Even in academia

methods can be classified in terms of goals into time prediction, "one of the domains heavily impacted by Artificial Intelligence is

innovation management and in this context especially the area



Jože M. Rožanec and Tesia Šker are co-first authors with equal contribution and of Strategic Foresight (SF)" as per Brandtner et. al (2021) [1]. importance. However it seems that strategic foresight methods related to Corresponding author: Jože M. Rožanec: joze.rozanec@ijs.si. AI only end up being used by bigger companies with a larger



work must be honored. For all other uses, contact the owner/author(s). Permission to make digital or hard copies of part or all of this work for personal number of resources. As noted by Kim and Seo (2023) [6], "except or classroom use is granted without fee provided that copies are not made or for AI start-ups and players in the consumer electronics and infor- distributed for profit or commercial advantage and that copies bear this notice and mation and communication industry, small- and medium-sized the full citation on the first page. Copyrights for third-party components of this enterprises (hereafter SMEs) in other industries do not demon- Information Society 2024, 7–11 October 2023, Ljubljana, Slovenia strate competence in AI." Therefore, effective implementation of © 2023 Copyright held by the owner/author(s). AI solutions for strategic foresight in smaller and medium sized https://doi.org/https://doi.org/10.70314/is.2024.sikdd.24

Information Society 2024, 7–11 October 2023, Ljubljana, Slovenia Tesia Šker, Jože M. Rožanec, Gregor Leban, and Dunja Mladenić





companies would be one of the topics to be explored in future

research.

In this research however, we focus more on the general imple-

mentation of strategic foresight by means of next event predic-



tion. Exploring similar fields, we found that there was already Figure 1: Sample of relevant data considered when parsing some research exploring the field of event predictions, which an event type to build the dataset. rather than focusing on businesses focused on other domains. In

the field of sequential event prediction, several researchers are



exploring diverse methods. Although the methods share some

conceptual similarities with our research, they differ significantly

in methodology and focus. Letham, Rudin, and Madigan (2013)

[7] developed a model that predicts the next event using an

ERM-based approach with logistic regression, focusing on the

presence of events rather than their order. On the other hand, our

work uses labeled article databases and considers the sequence

of past events, using techniques like graph construction, random

walks, and random forests. Yeon, Kim, and Jang (2015) [10] focus

on predicting event flow through visual analytics, using LDA

for topic extraction and emphasizing specific keywords, while

our approach is entirely text-based and relies on graphs. On the

other hand, Hu et al. (2017) [4] use LSTM networks for predict-

ing future subevents, which offers an alternative method to our

non-LSTM-based text analysis.

Although these studies provide useful insights and have of-

fered significant improvement in sequential event prediction,

they may face certain challenges. For instance, Letham, Rudin,

and Madigan (2013) [7] emphasize event presence over sequence,

potentially missing key temporal relationships, while Yeon, Kim,



and Jang (2015) [10] depend heavily on keywords, overlooking Figure 2: Event Type Taxonomy broader context. Additionally, LSTM-based models like those

used by Hu et al. (2017) [4] are powerful however they require



significant computational power. In contrast, our work addresses 3.2 Data Description these limitations by employing a graph-based approach that pri-For our research, we used a dataset of events provided by Event oritizes event sequences and leverages standardized data from Registry, with media events encoded in JSON format. Specifically, sources like DMOZ and Wikipedia. This enables us to make more we analyzed 4,216 events related to the company 3M, recorded accurate and efficient predictions, offering a practical and scalable between June 23, 2021, and July 23, 2024. We used a URI to clas-solution that enhances predictive accuracy. sify each event, drawing from DMOZ and Wikipedia categories

(Fig. 1). These were selected because they provide standardized

descriptions of the events being reported, which makes the data

consistent and reliable. The events are categorized into 94 distinct

types, which are further grouped into three primary domains:

3 DATASET business, environment, and society. The business domain makes

3.1 Data Extraction Pipeline up the largest proportion of events, accounting for 65 types (69%

The event detection pipeline processes about 300.000 English of the total), while the environment and society domains contain

news articles per day. Each news article is first annotated using 13 types (14%) and 16 types (16%), respectively. Within these

tools like entity linking, topic classification and sentiment detec- domains, the event types are further divided into smaller subdo-

tion. Each article is then split into sentences where each sentence mains, which can be aggregated into larger subdomain units as

retains it’s annotations and other meta-data. For each pair of the demonstrated in the event type taxonomy (Fig. 2).

entities in the sentence, an event classifier then determines if

there is a particular relation of interest expressed in the sentence 4 METHODOLOGY

between the two entities. The predefined taxonomy currently This study uses graph-based techniques to predict future event

includes 133 event types of interest, ranging from security, en- types from news articles about a specific company. The process

vironment, natural disasters, accidents, politics, and other areas. starts by building a graph that maps relationships between event

To classify the events, a neural network transformer architecture types and concepts from Wikipedia and DMOZ. Random walks

with a pretrained encoder is used. The entire network, including are then performed on this graph to extract key information

the encoder, is trained on our supervised dataset using best prac- such as URIs, dates, and event types, which are then transformed

tices like online hard example mining, class balancing, dropout, into embeddings using Graph2Vec [8]. Next, the event types are

and consistency regularization. The sentences for which the clas- encoded and adjusted through a process called target shifting.

sifier finds that it mentions a relation of interest are then stored This step aligns the features to better forecast future outcomes

in a database, together with the pair of associated entities and based on previous data. The predictions are made using a Random

other available meta-data. Forest classifier, which is then validated through stratified k-fold

What will happen tomorrow? Predicting future event types for businesses Information Society 2024, 7–11 October 2023, Ljubljana, Slovenia





4.4 One Hot Encoding & Target Shifting

To transform the categorical event types into binary vectors, One

hot encoding is applied. This allows the model to treat each event

type as a separate class. After extracting relevant column names,

the encoded target data is concatenated with the feature embed-



a) b) c) dataset is then aggregated by averaging out the embeddings and dings, creating a dataset for model training and evaluation. The

Figure 3: Event Type Graphs calculating the maximum value of the encoded target columns

for a given day. Finally the ’target’ data is shifted by one day,

which allows the embeddings to forecast the event types for the

cross-validation for higher accuracy. The following sections will following day.

present each step of this process in more detail (see Fig. 4).

4.5 Random Forest Classification & Stratified

4.1 Graph Construction K-Fold Cross Validation



wiki and dmoz addresses that are directly related to the event For each article in the JSON dataset, a detailed graph G is gen- To ensure an effective classification and prediction of the data, A erated using the NetworkX library [3]. The graph construction Random Forest classifier is created. When employing this method, process starts by extracting key information such as the article’s embeddings are used as features and the one-hot encoded event URI(unique identifier), as well as the date associated with the types are used as labels. The data itself is split into testing and article and the event types, which are represented by specific training sets, followed by the incorporation of the Stratified K- URIs. In addition to these elements, each article also includes two Fold cross validation. This technique splits the data into 10 folds, important lists: ’slots’ and ’categories’. The ’slots’ list contains while ensuring that the event type proportion in each fold re- mains equal. The model is then trained on 9 folds, with the re- described in the article, while the ’categories’ list includes vari- maining fold being used for validation. This ensures balanced ous classifications of the event. To complete the graph, labels are representation of each class across the folds resulting in a more created by extracting URIs from the ’slots’ list and filtering the effective performance. ’categories’ to focus on those with the "dmoz" prefix.



4.2 5 RESULTS Random Walks for Feature Extraction



are performed, starting at a given node (event type) and moving and then evaluated on a test set. The training set included ap- proximately 508 samples for each fold, and the test set included Once the graphs for each article are constructed, random walks As mentioned above, the model was trained on a training set,

walks are generated for each node, forming the foundation for about 10% of the whole set, which amounted to 56 samples per to adjacent nodes based on specific probabilities. Several random

feature extraction processes. A single random walk begins by fold. Using this, the model then predicted the probabilities for



specified path length. At each step, a random number is compared we noticed certain classes did not have enough occurrences to have at least one entry of such a class per dataset fold and were initializing the path with the starting node and iterating over a event types for each set. When training the model for each class,

with a probability p. If the number is less than p, the walker stays at the current node, otherwise it moves to a random neighbor. If skipped. We, therefore, trained the model and predicted for a



no neighbors are available, the walk ends. total of 45 classes. Generating multiple random walks for every node follows To evaluate the discriminative performance of the model, the ROC AUC score was used. The results produced showed us how

current node (set at 0.05). The process involves creating an empty well the model distinguishes between different classes, as well a similar approach, using p as the probability of staying at the

list to store all random walks and iterating through each node in as the model’s ability to predict future event types. The ROC



is generated, and each walk is appended to the list. was around 0.5674, and the median was close to it, with an AUC ROC score of 0.5559, with the highest score reaching 0.8194 and the graph. For each node, the specified number of random walks AUC score showed us that the average performance of the model



The random walks from the graphs are processed similarly to demonstrate we can effectively forecast event types ahead of time, further work is required to enhance results, which in most word sequences in a document. The ’embedding_data’ function 4.3 the lowest reaching a value of 0.3338. While the best scores Embedding Generation Using Graph2Vec

generates vector embeddings for graph data using the Doc2Vec cases remain close to 0.5.

model. It begins by converting each random walk into a Tagged-

Document, storing these in ’documents_gensim’. The Doc2Vec 6 CONCLUSIONS

model, with a vector size of 5 is trained on these documents, This study was used to develop a graph-based approach to pre-

creating a vector space where similar sequences are positioned dicting event types in articles. In the process, we utilized ran-

close together. dom walks for feature extraction and Doc2Vec for embedding

The function then processes each graph in the graphs dictio- generation. Then, we trained the resulting model on a Random

nary, extracting uri, date, and event type, and generating addi- Forest classifier and evaluated it with a Stratified K-Fold Cross

tional random walks. These walks are converted into embeddings Validation. The model demonstrated solid performance with an

using the ’infer_vector’ method, and the resulting vectors are average ROC AUC score of around 0.5674, reaching a peak at

averaged into one final embedding. This embedding is stored in approximately 0.8194. This indicates the model’s effectiveness

a dictionary across ’embedding1’ to ’embedding5’, alongside the in capturing relationships within the data and predicting future

graph’s metadata. event types.

Information Society 2024, 7–11 October 2023, Ljubljana, Slovenia Tesia Šker, Jože M. Rožanec, Gregor Leban, and Dunja Mladenić





Figure 4: Data Extraction Pipeline



However, while the model performed well overall, occasional [11] Liang Zhao. 2021. Event Prediction in the Big Data Era. Comput. Surveys 54, 5

fluctuations in accuracy suggest space for further improvement. (2021), 1–37.

We are currently striving to find ways to make graphs more in-

formative. In future work we could refine the feature extraction

process by incorporating larger datasets, with a wider variety

samples and a larger number of companies.



ACKNOWLEDGMENTS

The Slovenian Research Agency supported this work. This re-

search was developed as part of the Graph-Massivizer project

funded under the Horizon Europe research and innovation pro-

gram of the European Union under grant agreement 101093202.



REFERENCES

[1] Patrick Brandtner and Marius Mates. 2021. Artificial intelligence in strate-

gic foresight–Current practices and future application potentials: current

practices and future application potentials. In Proceedings of the 2021 12th

International Conference on E-business, Management and Economics. 75–81.

[2] Alex Fergnani, Andy Hines, Alessandro Lanteri, and Mark Esposito. 2020.

Corporate foresight in an ever-turbulent era. European business review 25

(2020), 26–33.

[3] Aric Hagberg, Pieter J Swart, and Daniel A Schult. 2008. Exploring network

structure, dynamics, and function using NetworkX . Technical Report. Los

Alamos National Laboratory (LANL), Los Alamos, NM (United States).

[4] Linmei Hu, Juanzi Li, Liqiang Nie, Xiao-Li Li, and Chao Shao. 2017. What

happens next? Future subevent prediction using contextual hierarchical LSTM.

In Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 31.

[5] Jon Iden, Leif B. Methlie, and Gunnar E. Christensen. 2017. The nature of

strategic foresight research: A systematic literature review. Technological

Forecasting and Social Change 116 (2017), 87–97. https://www.sciencedirect.

com/science/article/pii/S0040162516306035

[6] Jong-Seok Kim and Dongsu Seo. 2023. Foresight and strategic decision-making

framework from artificial intelligence technology development to utilization

activities in small-and-medium-sized enterprises. foresight 25, 6 (2023), 769–

787.

[7] Benjamin Letham, Cynthia Rudin, and David Madigan. 2013. Sequential event

prediction. Machine learning 93 (2013), 357–380.

[8] Annamalai Narayanan, Mahinthan Chandramohan, Rajasekar Venkatesan,

Lihui Chen, Yang Liu, and Shantanu Jaiswal. 2017. graph2vec: Learning

distributed representations of graphs. arXiv preprint arXiv:1707.05005 (2017).

[9] Freija van Duijne and Peter Bishop. 2018. Introduction to strategic foresight.

Future 1 (2018), 67.

[10] Hanbyul Yeon, Seokyeon Kim, and Yun Jang. 2015. Visual Analytics using

Topic Composition for Predicting Event Flow. KIISE Transactions on Computing

Practices 21, 12 (2015), 768–773.





Generating Non-English Synthetic Medical Data Sets




Lenart Dolinar Erik Calcina Erik Novak

University College London Jožef Stefan International Jožef Stefan International

London, United Kingdom Postgraduate School Postgraduate School

Jožef Stefan Institute Jožef Stefan Institute

Ljubljana, Slovenia Ljubljana, Slovenia



Abstract The experiment setting is presented in Section 4, followed by

the experiment results in Section 5. We discuss the results in

Using synthetic datasets to train medicine-focused machine learn-

Section 6 and conclude the paper in Section 7.

ing models has been shown to enhance their performance, how-

ever, most research focuses on English texts. In this paper, we ex-

plore generating non-English synthetic medical texts. We propose 2 Related Work a methodology for generating medical synthetic data, showcasing

This section describes the related work, focusing on large lan-

it by generating Greeklish medical texts relating to hypertension.

guage models and methods for generating synthetic data.

We evaluate our approach with seven different language models

and assess the quality of the datasets by training a classifier to

distinguish between original and synthetic examples. We find 2.1 Large language models

that the performs best for our task. Large Language Models (LLMs) are models that were trained Llama-3

to generate human-like texts based on an extensive process of

Keywords training on vast amounts of data. Models, such as Llama 3 [2],

GPT-4 [9], Aya 23 [3] and Mistral [7], are often easy to work

Synthetic data, healthcare data, multilingual data, large language

with by providing an input textual prompt, based on which the

models, classification

models respond. The LLMs are helpful in specialized fields, such



1 Introduction as medicine, since they can be fine-tuned on extensive data sets

containing medical terms and concepts. This enables them to per-

The healthcare domain produces a lot of medical data that can be

form well in tasks such as medical synthetic data generation [12].

used to train machine-learning models to help medical person-

Despite that, they are sometimes unable to follow the instruc-

nel. For example, a machine-learning model designed to perform

tions in the prompt accurately, leading them to hallucinate, i.e.

Named Entity Recognition (NER) on electronic health records

confidently produce wrong responses [5].

(EHRs) needs extensive labeled datasets to accurately identify

In our experiments, we investigate the LLMs’ performance

medical terms like diseases, treatments, and patient details. How-

in generating synthetic medical data given specific constraints

ever, the data contains a lot of personal information, and hospitals

and detailed prompts to simulate the original data set as best as

cannot share it freely due to data protection. In addition, there

possible.

are not enough examples to train the models for some problems,

such as those relating to rare diseases. Because of this, synthetic

data is being used as a substitute to train the models. 2.2 Synthetic medical data generation

Most synthetic data generation approaches focus on generat- Recently, synthetic medical data, generated using LLMs, has been

ing English texts. These usually utilize large language models used to enhance the performance of models for solving different

trained on predominantly English documents retrieved from the natural language processing tasks in medicine.

web. However, there are few examples of using them to gener- One work focuses on generating a synthetic dataset of elec-

ate non-English texts. Furthermore, the language models have tronic health records of Alzheimer’s Disease (AD) patients based

difficulties generating texts that do not reflect the distributions on a label that is provided [8]. They find that the performance of

found in the training sample. This includes medical texts, which their system for detecting AD-related signs and symptoms from

are usually not accessible to the general public. EHRs improves vastly when trained on synthetic and original

This paper proposes a methodology for generating medical data sets as opposed to training the system only on the origi-

synthetic data using open-source large language models. We nal one. Another work investigated using LLMs for extracting

apply the methodology to a medical data set written in Greeklish, structured information from unstructured healthcare text [13].

a combination of Greek and English scripts. We test it with seven By generating synthetic data using LLMs and fine-tuning the

large language models and assess performance by training a model, they significantly improved the models’ performance for

classifier to distinguish original examples from synthetic ones. medical-named entity extraction and relation extraction tasks.

Using the same prompt, we find that the open-source Llama-3 Most related works focus on English synthetic data due to

model best generates synthetic data that reflects the original data scarce non-English training data and the dominance of English

set. in medical terminology [6]. This paper focuses on generating

The remainder of the paper is as follows: Section 2 presents the non-English texts, specifically medical texts written in Greeklish

related work on generating synthetic data using large language about hypertension.

models. Next, the proposed methodology is described in Section 3.



Permission to make digital or hard copies of part or all of this work for personal 3 Methodology or classroom use is granted without fee provided that copies are not made or

This section outlines our research methodology. We first present

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this the pre-processing of the data set, followed by describing the syn-

work must be honored. For all other uses, contact the owner /author(s).

thetic data generation process. Finally, we present the description

Information Society 2024, 10–14 October 2024, Ljubljana, Slovenia

of synthetic dataset evaluation using a classifier. Figure 1 shows

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.sikdd.4 the diagram overviewing the proposed methodology.

Information Society 2024, 10–14 October 2024, Ljubljana, Slovenia Lenart Dolinar, Erik Calcina, and Erik Novak



evaluation 3.3 Technical details

with classifier

In this section, we describe the models and the parameters used

in the experiment. All models used are available via the Hugging-

Face’s transformer library [15].



We tested five open-source models to generate the synthetic



LLM and prompt data sets, all of which can be run on a 32GB GP U: Llama-3 [2]



choice

only has support for the English language but has been fine-tuned

translate to Latin script DataDreamer synthetic patient records to understand user prompts, which is a feature we expected would

2 Aya-23

help a lot with the synthetic data generation. [3] is a

multilingual language model and offers support for 23 languages,



3 Mistral

including Greek. [7] supports a variety of languages



extraction LLM train entity 4 but omits Greek. The models Gemma-2 [4] and Phi-3 [1] were

5 6

patient records also tested and compared in the experiments. In addition, we

experimented with [9] and , which are GPT-4o GPT-3.5-Turbo

Figure 1: An overview of the methodology. The image was accessible via the OpenAI API.

designed using resources from flaticon. All models were given the same prompt containing instruc-

tions that included (1) generating Greek texts written in Latin

script and (2) containing a label randomly selected from the orig-

inal data set, (3) examples are supposed to be at most 6 words

3.1 Data pre-processing long, (4) should provide concise responses, (5) structured format

The data set used consisted of 1,299 examples of medical history (all text must be in a single line, must use // and commas as sepa-

in Greeklish, where the Latin and Greek scripts were used inter- rators, and must be similar in format as the provided few-shot

changeably. It also contained 1,495 labels, most of which were examples). To stress some more important instructions, some

in English. The labels consisted of drugs, medical events, and instructions were given in capital letters and were also repeated.

measurements.

To translate the labels into Greek, we used the [14] NLLB-200 4 Experiment Setting

1

translation model . Since LLMs were predominantly trained on

This section describes the experiment setting, which consists

texts written in Latin script, we decided to transliterate both the

of the evaluation process and the metrics used to measure the

labels and examples from Greek to Latin script. This allowed the

approach’s performance.

LLMs to generate longer tokens with richer information.



We split the original data set into two subsets to ensure no 4.1 Evaluation approach

data leakage. The first one, consisting of 930 examples, was used

The quality of the generated synthetic data was measured in two

for synthetic data generation. The second one, containing the

parts. The first consisted of statistical measurements, such as

remaining 369 examples, was used for evaluation.

calculating the average length of the generated examples and

finding the proportion of examples that included the required

3.2 Synthetic data generation labels. These statistics were then compared to the original data

We utilized the datadreamer library [10] to generate the synthetic set.

data set. The library enables open-source models to create syn- The second part consisted of training a classifier to discern if

thetic data sets and was developed to work in research settings, the input text was from the original or from the synthetic data set.

supporting prompt templates and few-shot learning. The data set used to train and evaluate the classifier involved 369

We developed a prompt containing the instructions and re- randomly selected synthetic examples and 369 examples from

strictions on generating the examples. To better showcase the the original data set, transliterated into Latin script. We chose

structure of the generated text, we also provided five random 5-fold validation as our classification procedure and calculated

examples from the original data set as few-shot examples. Next, the mean performance across all trials.

using datadreamer, we sent the prompt to the chosen LLM. We The classifier was trained using the BERT [11] language model,

experimented with multiple LLMs, and about 800 examples were specifically the variant . The

bert-base-multilingual-cased 7

generated for each used LLM. When experimenting with LLMs classifier was trained using the following parameters: batch size

that required calling an external provider (e.g., OpenAI), we pro- = 16, epochs = 3, and learning rate = 2e-5. The same parameters

vided five static few-shot examples that did not include any pa- were used for all synthetic data sets.

tient personal data due to data privacy concerns.

To ensure the quality of generated data, we implemented a 4.2 Metrics

post-processing step. This included formatting the generated

To assess the quality of the generated synthetic data sets, we used

text into one line and excluding examples where the length was

the F1 score as our main metric for evaluating the classifier’s

too long or where the model started repeating words meaning-

performance. The target value was 0.5; if the performance is

lessly. This ensured that all generated examples followed the

greater than 0.5, the classifier can discern the original from the

same format and could be used for evaluation.

synthetic examples. Hence, the synthetic data does not reflect

Table 2 presents generated examples for the label "OSTEO-

the original data set. If the performance is less than 0.5, the

POROSH". Similarities in the examples highlight the need for

rigorous methods to evaluate how closely they resemble the 2

https://huggingface.co/meta- llama/Meta- Llama- 3- 8B- Instruct

original data set. The methods are explained in Section 4.1. 3 https://huggingface.co/CohereForAI/aya- 23- 8B

4

https://huggingface.co/mistralai/Mistral- 7B- Instruct- v0.3

5

https://huggingface.co/google/gemma- 2- 9b- it

6

https://huggingface.co/microsoft/Phi- 3- medium- 4k- instruct

1 7

https://huggingface.co/facebook/nllb- 200- distilled- 600M https://huggingface.co/google- bert/bert- base- multilingual- cased Generating Non-English Synthetic Medical Data Sets Information Society 2024, 10–14 October 2024, Ljubljana, Slovenia



classifier has difficulties separating the synthetic from the original 6.1 LLM performance data, which can be because the synthetic data contains copies of

Results in Table 1 show significant quality differences among

the original examples. In addition to the F1 score, we measured

synthetic datasets from different LLMs, with label occurrence

the classifier’s accuracy, precision, and recall, which are also Mistral GPT-4o ranging from 0.740 for to 0.996 for , and average

reported. GPT-4o Aya-23

example length from 3.691 for to 8.040 for .

However, Table 3 indicates no significant performance differ-

5 Results ences within a single synthetic dataset, with a maximal standard

In this section, we present the results of our experiment. We deviation of the metrics being 0.021 for the dataset. Llama-3 first present the statistical results, followed by the classifier’s We can also notice that the F1 and accuracy scores are very

evaluation. close for all synthetic data sets. This means the classifier was

likely performing relatively similarly on both classes (synthetic

5.1 Statistical analysis and original datasets) without significant bias to either class.

Table 1 compares the synthetic data sets and the original one We can observe much better performance on the Llama-3

regarding label occurrence and average example length. The label data set, which is primarily trained on English data, than on the

occurrence is 1.000 in the original data set, as all examples from data set, which is also trained on Greek data. This shows Aya-23

the original data set are assumed to include relevant labels and that a model does not need to be extensively trained in Greek

information. texts to generate this type of synthetic medical data well.

The most aligned synthetic data set regarding label occurrence



was generated using , followed by . However, in GPT-4o Llama-3 6.2 Limitations terms of average example length, the data set generated using

Due to limited computing power, only one GP U with 32GB of

Gemma-2 performed the best, followed by Llama-3.

space was available, restricting the testing of larger LLMs. To ad-

The worst-performing models, in terms of label occurrence,

dress these challenges, using cloud-based resources or distributed

were and , which in about 25% did not include the Mistral Phi-3

computing could help run larger models and improve the variety

selected label. The data set generated using the had the Aya-23

of synthetic data generated.

largest difference in terms of average example length, on average

Due to privacy concerns, when using and GPT-4o GPT-3.5-Turbo

generating examples with three extra words.

models, which are not locally-run models, we had to use five fixed

Table 1: Statistical comparison between the original and examples when generating synthetic data instead of a larger vari-

synthetic data sets. The bold and underlined values repre- ety. This potentially led to larger similarities of the synthetic GPT

sent the best and second-best statistics, respectively. datasets to the examples instead of the original dataset and, con-

sequently, worse performance.



LLM Label occurrence Avg example length

6.3 Potential improvements

original dataset 1.000 4.682

The prompt was the same for all seven LLMs and was primarily

Llama-3 0.990 5.330 (+0.648) tested on Llama-3. Hence, the performance might be biased to-Aya-23 0.949 8.040 (+3.358) wards the model. The method could be improved by tailoring the

Mistral 0.740 6.376 (+1.694) prompts to each model individually.

Gemma-2 0.988 4.207 (-0.475) The evaluation of synthetic datasets could be further extended

Phi-3 0.782 6.071 (+1.389) by checking for repeating examples in the synthetic dataset or

GPT-4o 0.996 by checking how different the generated example is from the five 3.691 (-0.991)

GPT-3.5-Turbo provided examples. The evaluation could also be improved by

0.867 6.764 (+2.082)

checking for overfitting to the original data set.

Looking at both statistics, we can conclude that had Llama-3

the best alignment to the original data set in terms of label oc- 7 Conclusion and Future Work currence and example length, closely followed by . GPT-4o

This paper presents a method for generating Greek synthetic

To better imagine the differences between the generated ex-

medical data sets. To synthetically create datasets similar to the

amples, we handpicked an example from each synthetic data set

original, we carefully craft a prompt and perform pre-processing

related to the label “OSTEOPOROSH”, shown in Table 2.

and post-processing of the data to increase performance and

eliminate the effect of hallucinations.

5.2 The classifier evaluation

Using a classifier and considering the inclusion of labels and

Table 3 shows the F1, Precision, Recall, and Accuracy perfor- generated text length, we conclude that Llama-3 is best for gen-mances of the trained classifier on different synthetic data sets.

erating examples that most closely resemble the original dataset.

The best performance was achieved by with approxi-Mistral

In the future, we plan to explore the underlying architectures

mately 0.85 scores in all four metrics, followed by , with Llama-3

of the models to understand their performance differences in

approximately 0.88 scores in all metrics. The worst performances

multilingual contexts. This will allow us to further refine our

were on data sets generated by the and Aya-23 GPT-3.5-Turbo

methods and create more accurate data sets.

models. Surprisingly, the is a language model supporting Aya-23

Furthermore, we intend to use the synthetic dataset to train a

Greek; thus, it was expected to generate better examples. named entity recognition (NER) system to recognize medical la-

bels from medical history examples. Measuring the performance

6 Discussion of the NER trained on synthetic datasets will give us another

This section discusses the synthetic data generation performance, way of evaluating their quality. We also intend to create a more

outlines our methodology’s limitations and drawbacks, and pro- general pipeline enabling the code to generate synthetic medical

poses potential improvements to the approach. data in a wider variety of languages and formats.

Information Society 2024, 10–14 October 2024, Ljubljana, Slovenia Lenart Dolinar, Erik Calcina, and Erik Novak



Table 2: Generated examples for label "OSTEOPOROSH".



LLM Examples



original dataset APO 2O ETON YPERTASH ME AGOGI// OSTEOPOROSH // YPOTHYROIDISMOS

Llama-3 YPOTHYROEIDISMOS, OSTEOPOROSH, APO//

Aya-23 CA ORTHOU, ANEYRISMA KOILAKHS AORTHOU, OSTEOPOROSH.

Mistral OSTEOPOROSH, APO 60 ETOS, APO 2 MHNES KAI APO 10 GRAMM

Gemma-2 OSTEOPOROSH, ARTHROSITIS, ETOVIR

Phi-3 OSTEOPOROSH, XAROSTHROMA, ALPHA-BISFIOVITINI, 2018, DIATHRHSH, DIA

gpt-4o OSTEOPOROSH, ANEMIA

gpt-3.5-Turbo OSTEOPOROSH, GASTREKTOMH, EMFISIMA, YDRONERFOSI, PSIXROS.



Table 3: Mean performance metrics of the classifier for synthetic data sets, with standard deviation. Performances that

are closer to 0.5 are considered better. The bold and underlined values represent the best and second-best performances,

respectively.



LLM F1 Precision Recall Accuracy



Llama-3 0.875 ± 0.021 0.881 ± 0.020 0.875 ± 0.020 0.875 ± 0.020

Aya-23 0.945 ± 0.005 0.947 ± 0.004 0.945 ± 0.005 0.945 ± 0.005

Mistral 0.848 ± 0.012 0.856 ± 0.001 0.849 ± 0.011 0.849 ± 0.011

Gemma-2 0.928 ± 0.005 0.930 ± 0.005 0.928 ± 0.005 0.928 ± 0.005

Phi-3 0.927 ± 0.009 0.932 ± 0.008 0.927 ± 0.009 0.927 ± 0.009

GPT-4o 0.906 ± 0.014 0.912 ± 0.012 0.907 ± 0.014 0.907 ± 0.014

GPT-3.5-Turbo 0.940 ± 0.013 0.944 ± 0.011 0.940 ± 0.013 0.940 ± 0.013



Acknowledgments [7] Albert Q. Jiang et al. Mistral 7B. 2023. arXiv: 2310.06825

This work was supported by the Slovenian Research Agency. [cs.CL]

. url: https://arxiv.org/abs/2310.06825.

[8] Rumeng Li, Xun Wang, and Hong Yu. “Two Directions for

Funded by the European Union. UK participants in Horizon Eu-

Clinical Data Generation with Large Language Models:

rope Project PREPARE are supported by UKRI grant number

Data-to-Label and Label-to-Data”. In: Findings of the Asso-

10086219 (Trilateral Research). Views and opinions expressed are

however those of the author(s) only and do not necessarily reflect ciation for Computational Linguistics: EMNLP 2023. 2023,

pp. 7129–7143. doi: 10.18653/v1/2023.findings- emnlp.474.

those of the European Union or European Health and Digital Ex-

[9] OpenAI et al. . 2024. arXiv: 2303 . GPT-4 Technical Report

ecutive Agency (HADEA) or UKRI. Neither the European Union

08774 . url: https://arxiv.org/abs/2303.08774. [cs.CL]

nor the granting authority nor UKRI can be held responsible for

[10] Ajay Patel, Colin Raffel, and Chris Callison-Burch. DataDreamer:

them. Grant Agreement 101080288 PREPARE HORIZON-HLTH-

2022-TOOL-12-01. A Tool for Synthetic Data Generation and Reproducible LLM

Workflows. 2024. arXiv: 2402.10379 [cs.CL]. url: https:

References //arxiv.org/abs/2402.10379.

[11] Telmo Pires, Eva Schlinger, and Dan Garrette. “How Mul-

[1] Marah Abdin et al. Phi-3 Technical Report: A Highly Ca-

tilingual is Multilingual BERT?” In: Proceedings of the 57th

pable Language Model Locally on Your Phone. 2024. arXiv:

2404 . 14219 . url: https : / / arxiv . org / abs / 2404 .

[cs.CL] Annual Meeting of the Association for Computational Lin-

guistics. Association for Computational Linguistics, 2019,

14219.

pp. 4996–5001. doi: 10.18653/v1/P19- 1493.

[2] AI@Meta. “Llama 3 Model Card”. In: (2024). url: https :

[12] Karan Singhal et al. “Large language models encode clin-

/ / github. com / meta - llama / llama3 / blob / main / MODEL _

ical knowledge”. In: 620 (2023), pp. 172–180. doi: Nature

CARD.md.

10.1038/s41586- 023- 06291- 2.

[3] Viraat Aryabumi et al. Aya 23: Open Weight Releases to

[13] Ruixiang Tang et al. Does Synthetic Data Generation of

Further Multilingual Progress. 2024. arXiv: 2405 . 15032

[cs.CL] LLMs Help Clinical Text Mining? 2023. arXiv: 2303.04360

.

[cs.CL]. url: https://arxiv.org/abs/2303.04360.

[4] Google DeepMind Gemma Team. Gemma 2: Improving

[14] NLLB Team et al. No Language Left Behind: Scaling Human-

Open Language Models at a Practical Size. 2024. url: https:

Centered Machine Translation. 2022. arXiv: 2207 . 04672

/ / storage . googleapis . com / deepmind - media / gemma /

[cs.CL]. url: https://arxiv.org/abs/2207.04672.

gemma- 2- report.pdf .

[15] Thomas Wolf et al. “Transformers: State-of-the-art natural

[5] Xu Guo and Yiqiang Chen. Generative AI for Synthetic

language processing”. In: Proceedings of the 2020 Confer-

Data Generation: Methods, Challenges and the Future. 2024.

arXiv: 2403 . 04190 . url: https : / / arxiv. org / abs /

[cs.LG] ence on Empirical Methods in Natural Language Processing:

System Demonstrations. Association for Computational

2403.04190.

Linguistics, 2020, pp. 38–45. doi: 10.18653/v1/2020.emnlp-

[6] Rainer Hamel. “The dominance of English in the inter-

demos.6.

national scientific periodical literature and the future of

language use in science”. In: 20 (Dec. 2007), AILA Review

pp. 53–71. doi: 10.1075/aila.20.06ham.

LLNewsBias: A Multilingual News Dataset for Lifelong



Learning



Swati Swati Dunja Mladenić

swati.swati@unibw.de dunja.mladenic@ijs.si

Jožef Stefan International Postgraduate School Jožef Stefan Institute and

Ljubljana, Slovenia Jožef Stefan International Postgraduate School

Ljubljana, Slovenia



Abstract In this study, we address these challenges by introducing a

The rise of digital media enhances information accessibility but novel dataset LLNewsBias specifically designed for the detection

also introduces challenges related to the quality and impartiality and analysis of political bias in multilingual news headlines. Our

of news reporting, particularly regarding biases that influence dataset spans four major global events from 2019 to 2022: Brexit,

public perception during key global events. In response, this COVID-19, the 2020 U.S. election, and the Ukraine-Russia war,

study introduces capturing a wide range of political discourse across 17 languages. LLNewsBias , a dataset designed to detect and

analyze political bias in multilingual news headlines, covering To collect this dataset, we use Media Bias/Fact Check for the

four major events from 2019 to 2022 — Brexit, COVID-19, the assignment of bias labels, and Event Registry [2] for the extrac-

2020 U.S. election, and the Ukraine-Russia war. With over 350,000 tion of relevant headlines and metadata. The resulting dataset is

headlines in 17 languages, annotated with bias labels, this dataset not only comprehensive in its linguistic diversity but also struc-

is compiled using Media Bias/Fact Check and Event Registry. Our tured to support both event-wise and year-wise analyses, with

contributions include a structured framework for data collection an emphasis on lifelong learning.

and organization, enabling event-wise and year-wise analysis

while supporting lifelong learning. We also highlight potential 1.1 Contributions

use cases that demonstrate the dataset’s utility in advancing bias Our study makes the following contributions: prediction models, multilingual adaptation, and model robustness.

• Multilingual bias-annotated dataset: We introduce a

Additionally, we discuss the dataset’s limitations, addressing po-

multilingual bias-annotated dataset containing over 350,000

tential biases, sample size constraints, and contextual factors. This

news headlines in 17 languages, each annotated with po-

work provides a valuable resource for improving bias detection

litical bias labels.

in dynamic, multilingual news environments, contributing to the

• Data collection and organization framework: We pre-

development of more accurate and adaptable models in natural

sent a structured framework for data collection and or-

language processing and media studies. For code and additional

ganization, enabling event-wise and year-wise analysis

insights, visit: https://github.com/Swati17293/LLNewsBias

while ensuring adaptability for lifelong learning.



Keywords • Potential use-cases: We outline several potential applica-

tions of our dataset, highlight its potential for advancing

Dataset, News, Bias, Multilingual, Headline, Low-resource, Media lifelong learning models, particularly in bias prediction,

Bias, News Bias, Continual Learning, Lifelong Learning multilingual adaptation, and model robustness.

• Discussion of limitations: We identify and discuss the

1 Introduction dataset’s limitations, such as biases in data collection, sam-

The rapid growth of digital media has greatly enhanced the ac- ple size constraints, and contextual influences, offering a

cessibility of information, but it has also introduced significant transparent assessment of its applicability.

challenges concerning the quality and impartiality of news re-

In summary, our paper introduces a comprehensive dataset

porting. Political bias in news content is particularly concerning,

and a framework for the study of political bias in multilingual

as it has the potential to influence public perception and shape

news headlines. By focusing on key global events and providing

societal narratives, especially around key global events. Under-

support for lifelong learning, our study contributes to the ongoing

standing and predicting such biases, particularly in multilingual

effort to develop more accurate and adaptable models for bias

contexts where biases can manifest differently across cultural and

detection in diverse linguistic and cultural contexts.

linguistic boundaries, is essential for promoting fair and balanced

journalism. Traditional approaches to bias detection often rely

on monolingual datasets and static models that may not effec- 2 Related Work



tively capture the evolving nature of news content [6]. These Several datasets focus on news articles and political bias [5],

limitations underscore the need for more robust datasets and but there is a notable scarcity of multilingual, bias-annotated

methodologies that can adapt to the dynamic and multilingual datasets designed for lifelong learning [4]. While resources like

landscape of modern news reporting. the media bias chart by Ad Fontes Media and PolitiFact provide

insights into bias, they are often limited to English-language

Permission to make digital or hard copies of all or part of this work for personal sources or specific fact-checked claims, lacking the continuous,

or classroom use is granted without fee provided that copies are not made or event-centric data necessary for broader analysis. GDELT [3],

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this a large-scale event-oriented news dataset, covers multiple lan-

work must be honored. For all other uses, contact the owner/author(s). guages but focuses on location, network, and temporal attributes

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia rather than political bias or the event-outlet relationship. Exist-

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.sikdd.8 ing multilingual datasets are often domain-specific [1], limiting



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Swati et al.



their utility for general bias analysis. In contrast, LLNewsBias exclude outlets labeled as questionable and assign each remain-

dataset fills these gaps by offering a generalized, multilingual, ing outlet 𝑜𝑖 ∈ 𝑂 a bias label 𝑏𝑖 ∈ 𝐵, where 𝐵 = {𝑏 1, 𝑏2, ..., 𝑏𝑞}

and bias-annotated data designed for event-wise and year-wise represents the set of bias labels, with 𝑞 representing the number

analyses, particularly suited for lifelong learning models. of distinct bias labels.

Next, we define a temporal query 𝑄𝑡 to extract article headlines

3 (𝐻 = {ℎ 1, ℎ2, ..., ℎ𝑟 }), where 𝑟 represents the total number of Dataset Description

headlines retrieved from the Event Registry (ER). The query 𝑄𝑡

In this section, we introduce our dataset LLNewsBias and describe

is formulated as:

the framework used for its collection and organization. We begin

by detailing the primary data sources that form the foundation of 𝑄 , 𝑄 𝑡 = { 𝑄 , 𝑄 , 𝑄 𝑒 𝑜 𝑐𝑎𝑡𝑑 𝑡 } (1)

this dataset. Following this, we present a comprehensive overview where 𝑄𝑒, 𝑄𝑜 , 𝑄𝑐𝑎𝑡 specify the event, media outlet, and news

of the data collection process, with a focus on the methodologies categories (limited to those classified as ’news’ by ER 𝑄𝑐𝑎𝑡 =

employed to ensure robustness and reliability. Finally, we provide {‘𝑝𝑜𝑙 𝑖𝑡 𝑖𝑐𝑠’, ‘𝑏𝑢𝑠𝑖𝑛𝑒𝑠𝑠’, ‘𝑠 𝑝𝑜𝑟 𝑡 𝑠’, ‘𝑎𝑟 𝑡 𝑠 𝑎𝑛𝑑 𝑒𝑛𝑡 𝑒𝑟 𝑡 𝑎𝑖𝑛𝑚𝑒𝑛𝑡’,

an in-depth overview of the dataset’s structure, including its ‘𝑠𝑐𝑖𝑒𝑛𝑐𝑒’, ‘𝑡 𝑒𝑐ℎ𝑛𝑜𝑙𝑜𝑔𝑦’, ‘ℎ𝑒𝑎𝑙 𝑡 ℎ’, ‘𝑒𝑛𝑣𝑖𝑟 𝑜𝑛𝑚𝑒𝑛𝑡 ’}), respectively.

directory organization, file contents, and the various ordering The time constraint is represented as 𝑄𝑑 𝑡 = [𝑄 , 𝑄 𝑠𝑑𝑒𝑑] , where

methods applied to facilitate detailed analysis. Our dataset is 𝑄 𝑄 and 𝑠𝑑𝑒𝑑 denote the start and end dates. To scrape all the

documented in accordance with the FAIR Data Principles. article headlines (𝐻 ), we utilize 𝑄𝑡 to query ER.

We then associate the extracted headlines 𝐻 with the corre-

3.1 Primary Data Sources sponding bias labels in 𝐵 and structure the dataset according to

In this section, we outline the two primary data sources used in two classification types: event-wise and year-wise. To organize

our study: Media Bias/Fact Check (MBFC) and Event Registry the data, we define an event-based order 𝑂𝑒 𝑣𝑒𝑛𝑡 and a year-based

(ER). MBFC serves as the bias rating portal, providing bias la- order 𝑂𝑦𝑒𝑎𝑟 as follows:

bels for selected media outlets, while ER is used to extract the 𝑂 𝑒 𝑣𝑒𝑛𝑡 = {𝑒 1 → 𝑒2 → ... → 𝑒𝑛 } (2) headlines and corresponding metadata from articles published

by these outlets. 𝑂𝑦𝑒𝑎𝑟 = {𝑦1 → 𝑦2 → ... → 𝑦𝑚 } (3)



3.1.1 . For bias labeling in this study,

Media Bias/Fact Check For lifelong learning, we design the dataset to be extendable,

allowing for the integration of new events and years as they

we utilized Media Bias/Fact Check (MBFC), a well-established ′ ′ ′ ′

emerge, denoted by 𝐸 ⊆ 𝐸 and 𝑌 ⊆ 𝑌 , where 𝐸 and 𝑌 repre-

platform known for its comprehensive coverage and frequent

sent the sets of newly added events and years.

updates. Although other platforms like allsides.com and adfontes-

We designed the dataset with a flexible framework that allows

media.com also provide bias ratings, MBFC was selected for its

reliability and particular focus on low-resource languages. MBFC ′ for the seamless integration of new events and years as they

emerge, represented as ′ ′ ′ ⊆ and ⊆ , where and de-

𝐸 𝐸 𝑌 𝑌 𝐸 𝑌

assigns bias labels based on political orientation and evaluates

note the newly added events and years. This structured approach

outlets for credibility and factual accuracy. These labels are de-

ensures scalability for continuous learning without requiring

termined by a team of contractors and volunteers who follow

major restructuring and supports the training of adaptive mod-

a standardized methodology, ensuring that the ratings are both

els capable of integrating new information effectively. Unlike

consistent and dependable for our analysis.

standard multi-year datasets, our dataset includes annotations

3.1.2 that facilitate contextual understanding, enabling models to learn Event Registry . In this study, we use Event Registry [2]

platform as the primary source for collecting multilingual news from historical data while adapting to evolving trends and pat-

headlines. It aggregates content from over 150,000 news sources terns in news reporting. This ensures that the models remain

across more than 60 languages, making it an ideal resource for relevant as new information becomes available.

analyzing bias in diverse and low-resource languages. Apart from Finally, we split the dataset into training and test sets using a

the headlines, it allows access to numerous metadata such as stratified sampling approach to ensure the preservation of bias

publication date, news category, and political bias. By leveraging label distributions across both events and years. We perform this

its Python API, we efficiently filtered and extracted headlines step as it is critical for maintaining the integrity of the model

relevant to our study. This ensured a comprehensive dataset training process in a lifelong learning context.

that supports the analysis of bias in a lifelong learning setup,

exploring how emerging events and domain shifts influence the 3.3 Data Synopsis and Structure

performance of bias prediction models over time. In this section, we present an overview of the data and explain

how it is systematically organized, making it easier to understand

3.2 both the content and format of our dataset. Data Collection Framework

Our data collection framework as depicted in Figure 1, is designed 3.3.1 Data Synopsis. The dataset features 356,060 headlines

to support both event-wise and year-wise analyses, with the on four major events from 2019 to 2022: Brexit, COVID-19, the

additional capability of facilitating lifelong learning. election, and the Ukraine-Russia war. These headlines, sourced

For data collection, we begin by defining two sets: a set of from 45 unique news outlets in 17 different languages, are anno-

significant global events (𝐸 = {𝑒1, 𝑒2, ..., 𝑒𝑛 }), and a set of years tated with 3 political bias labels: Left Centre, Least Biased, and

(𝑌 = {𝑦1, 𝑦2, ..., 𝑦𝑚 }), where 𝑛 and 𝑚 represent the total num- Right Centre covering diverse topics such as politics, business, arts

ber of events and years, respectively. We then use the Media and entertainment, sports, science, technology, health, and environ-

Bias/Fact Check (MBFC) platform to select media outlets (𝑂 = ment. The dataset is structured into 7 distinct columns within .csv

{𝑜 1, 𝑜2, ..., 𝑜𝑝}) and determine their respective political bias, with files. Table 1 presents a comprehensive summary of the dataset

𝑝 as the total number of outlets. To maintain data reliability, we statistics.

LLNewsBias: A Multilingual News Dataset for Lifelong Learning Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia





Figure 1: Data Collection Framework. The framework uses MBFC for bias labeling and ER for headline retrieval.



Table 1: Summary of Dataset Statistics. • article_ID: A unique identifier for the raw news article

in the Event Registry platform from which the headlines

Language-wise Distribution are extracted.

• language: The source language of the published news

Catalan 882 Romanian 17,038

article.

Croatian 13,929 Russian 10,511 • date: The date on which the news was published.

Czech 1,876 Slovak 5,642 • headline_text: The text of the news headline.



Dutch 10,905 Swedish 6,441 • political_bias: The political bias of the news outlet as Finnish 1,512 Ukrainian 10,616 provided by the bias rating portal Media Bias/Fact Check. Danish 4,330 Spanish 83,940 • news_category: The category assigned by Event Registry. French 85,007 Italian 48,450

Hungarian The dataset is annotated with bias labels: Left Centre (LC), 105

Least Biased (LB), and Right Centre (RC). To ensure model ro-

Event-wise Distribution

bustness across varying data distributions, we concatenate and

Brexit 32,286 COVID 309,329 shuffle files for each event and year in four distinct random orders.

Election 3,829 Ukraine 10,616 This prevents overfitting to specific sequences and helps evaluate



Year-wise Distribution generalization across diverse configurations. While chronolog-

ical order is ideal for practical use, this randomized approach

2019 20,664 2021 4,638 tests broader performance, with the original event and year splits

2020 258,871 2022 71,887 provided for user flexibility.

Event-wise Ordering:



3.3.2 Directory Structure. The dataset is organized in a main → (1) brexit → covid → election → ukr-rus-war

‘data’ directory with subdirectories categorized by events (‘brexit’, → (2) election covid → ukr-rus-war → brexit

(3) brexit ukr-rus-war → election → covid

‘covid’, ‘election’, ‘ukr-rus-war’) and years (2019-2022). Addi-

(4) covid → brexit → ukr-rus-war → election

tional subdirectories consolidate data across all events (ordered_events)

and all years (ordered_years). Each subdirectory contains .csv Year-wise Ordering:

files for training and testing, structured across the following (1) 2019 → 2020 → 2021 → 2022

columns. (2) 2021 → 2020 → 2022 → 2019

• news outlet: The name of the news outlet. (3) 2019 → 2022 → 2021 → 2020



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Swati et al.



(4) 2020 → 2019 → 2022 → 2021 Event Registry (ER). MBFC is publicly accessible, while ER

The dataset captures the distribution of headlines related to provided comprehensive but limited coverage, potentially

various events over the years, reflecting the temporal dynamics missing relevant articles. The use of ER’s paid version also

of news coverage and the evolving reporting on these events. restricted the extent of data collection.

The differences in coverage levels reveal important patterns in • Sample Size: The dataset is constrained by its focus on

media attention, which are essential for developing datasets that four major events over a span of four years. This limited

support lifelong learning models. number of events and time frame may not fully capture

the broader spectrum of news and media biases, affecting



4 the diversity of the samples. Potential Use-Cases • Biases: Selection bias is a significant factor, as only news

Our dataset introduced in this study has a wide range of potential outlets labelled by MediaBiasFactCheck were included.

use-cases, particularly in the fields of natural language processing This restriction may limit the number of languages and

and media studies. It is particularly valuable for research and perspectives represented in the dataset, thereby influenc-

applications that require understanding and predicting news bias ing the overall analysis.

in a continual, multilingual environment. Below we list some • Contextual Factors: The dataset is limited by its tem-

potential use cases: poral scope, covering only four specific events over four

• Lifelong learning for news bias prediction: Our dataset years. While it reflects the dynamic nature of news media,

is ideal for developing and testing lifelong learning mod- it does not account for all future events and years to come.

els. It allows models to adapt to new events and evolving

entities. With its year-wise structure from 2019 to 2022, 6 Conclusions

the dataset addresses the challenges of emerging events In this study, we present LLNewsBias, a comprehensive dataset

and domain shifts (e.g., Brexit, COVID-19, Ukraine-Russia designed to tackle the challenges of detecting and analyzing polit-

War), providing the data needed to develop and evaluate ical bias in multilingual news headlines. By spanning four major

robust models. global events from 2019 to 2022 across 17 languages, this dataset

provides a valuable resource for research in natural language

• Domain Adaptation in Multilingual Contexts: Our processing and media studies. Our framework supports both

dataset enables researchers to investigate domain adapta- event-wise and year-wise analysis, emphasizing lifelong learning

tion techniques in a multilingual context, featuring head- and enabling models to adapt continuously to new data. The

lines in 17 languages. This facilitates the development of dataset’s potential use cases include enhancing bias prediction

models that generalize across languages and adapt to vari- models, facilitating domain adaptation in multilingual contexts,

ous cultural and political contexts, ensuring accurate bias and improving model robustness. While LLNewsBias offers sig-

prediction. It addresses the challenges faced by generic nificant contributions, we also acknowledge limitations such as

models in the news domain, which often struggle with potential biases in data collection, sample size constraints, and

topic and language diversity. contextual factors. Addressing these challenges in future work

will be crucial for maximizing the dataset’s impact, ultimately

• Sparse Experience Replay for Continual Learning: contributing to fairer and more balanced journalism.

Our dataset is particularly well-suited for the news do-

main, supporting efficient experience replay by allowing 7 Acknowledgments

the selection of specific topics and categories. With its This work was supported by the Slovenian Research Agency and

event-wise and year-wise classifications, our dataset en- National grants (CRP V2-2272; V5-2264; CRP V2-2146) and by the

hances memory utilization, improves generalization, re- European Union through enrichMyData EU HORIZON-IA project

duces catastrophic forgetting, and ensures that models under grant agreement No 101070284 and ELIAS HORIZON-RIA

remain accurate and up-to-date in real-time applications. project under grant agreement No 101120237.

In a nutshell, our dataset serves as a valuable resource for

advancing news bias prediction, particularly in the context of References

lifelong learning, by providing a flexible framework for integrat- [1] Jason Armitage, Endri Kacupaj, Golsa Tahmasebzadeh, Swati, Maria Maleshk-

ing new events and years. Unlike many news-based datasets with ova, Ralph Ewerth, and Jens Lehmann. 2020. Mlm: a benchmark dataset

for multitask learning with multiple languages and modalities. In Proceed-

timestamps, it offers structured annotations and contextual in- ings of the 29th ACM International Conference on Information & Knowledge

formation that enhance the understanding of evolving trends Management, 2967–2974.

in news coverage, making it particularly suitable for lifelong [2] Gregor Leban, Blaz Fortuna, Janez Brank, and Marko Grobelnik. 2014. Event

registry: learning about world events from news. In Proceedings of the 23rd

learning applications. It supports a range of research activities, International Conference on World Wide Web, 107–110.

from model development and evaluation to the exploration of [3] Kalev Leetaru and Philip A Schrodt. 2013. Gdelt: global data on events, loca-

new techniques for handling dynamic and multilingual news tion, and tone, 1979–2012. In ISA annual convention. Vol. 2, 1–49.

[4] Swati Swati, Adrian Mladenić Grobelnik, Dunja Mladenić, and Marko Grobel-

environments. nik. 2023. A commonsense-infused language-agnostic learning framework

for enhancing prediction of political bias in multilingual news headlines.



5 Knowledge-Based Systems, 277, 110838. Limitations

[5] Swati Swati, Dunja Mladenić, and Tomaž Erjavec. 2021. Eveout: an event-

Several limitations are associated with the dataset presented in centric news dataset to analyze an outlet’s event selection patterns. Informat-

ica, 45, 7.

this article and should be carefully considered in any further [6] Swati Swati, Dunja Mladenić, and Marko Grobelnik. 2023. An inferential

research or analysis: commonsense-driven framework for predicting political bias in news head-

lines. IEEE Access.

• Data Collection Issues: The dataset was gathered using

Media Bias Fact/Check (MBFC) and the paid version of





Creating Local World Models using LLMs




Mark David Longar Erik Novak Marko Grobelnik

Jožef Stefan Institute Jožef Stefan Institute Jožef Stefan Institute

Ljubljana, Slovenia Ljubljana, Slovenia Ljubljana, Slovenia



Abstract e.g. by providing LLMs a framework for responding with logi-

cally consistent and pedagogically sound explanations. Moreover,

A key limitation of state-of-the-art large language models is their

by modifying some of the components, the approach can also be

lack of a consistent world model, which hinders their ability to

applied to other domains, such as industry, finance, and law.

perform unseen multi-hop reasoning tasks. This paper addresses

The remainder of the paper is as follows: Section 2 presents

this by extracting local world models from text into a system-

the related work on LLMs and creating world models. Next, the

atic first-order logic framework, enabling structured reasoning.

proposed approach is described in Section 3. The experiment set-

Focusing on the educational domain, we present a multi-step

ting is presented in Section 4, followed by the experiment results

approach using Prolog to represent and reason with these mod-

in Section 5. We discuss the results in Section 6 and conclude the

els. Our method involves segmenting educational texts, generat-

paper in Section 7.

ing Prolog definitions, and merging them into a comprehensive

knowledge graph. We successfully extracted several small models

and manually verified their accuracy, demonstrating the poten- 2 Related Work

tial of this approach. While promising, our results are currently The recent surge in large language models, such as GPT-3 [3] and

limited to small-scale models. GPT-4 [1], has significantly advanced natural language process-

ing, showing emergent reasoning abilities across various tasks.

Keywords However, despite their impressive performance, LLMs are often

criticized for lacking factual consistency, interpretability, and

Large language models, local world models, knowledge represen-

logical coherence, especially in complex, multi-hop reasoning

tation, educational technology, structured reasoning, knowledge

tasks [8]. To address these shortcomings, efforts have been made

graphs

to integrate LLMs with structured knowledge frameworks, like

knowledge graphs (KGs) and ontologies, to enhance reasoning

1 Introduction and knowledge flow between structured data and language mod-

In recent years, Large Language Models (LLMs) have revolu- els [9].

tionized the field of Natural Language Processing (NLP), offer- In the field of ontology and KG development, early initiatives

ing unprecedented capabilities in understanding, reasoning over, like Cyc [6] laid the groundwork for large-scale structured knowl-

and generating human-like text. Despite their impressive per- edge representation. More recent efforts [8, 5] have explored

formance across various language tasks, a significant limitation using LLMs to assist in ontology generation and KG construction.

persists – the absence of a consistent and coherent world model While LLMs can automate parts of the ontology development

within these systems [8]. This limitation hampers their ability process, they struggle with ensuring logical consistency and

to perform advanced reasoning tasks that require not only tex- managing complex domain-specific knowledge [5, 2]. Comple-

tual understanding but also logical consistency and structured mentary approaches, like using LLMs for ontology learning [2]

knowledge representation. and structured knowledge extraction [10], highlight the need for

While current LLMs are powerful, they are inherently con- human validation and formal methods to ensure accuracy.

strained by their reliance on statistical correlations within vast Our work builds on these insights by focusing on using LLMs

datasets, often resulting in shallow and contextually inconsistent to extract structured local world models in the form of Prolog-

reasoning. To address this limitation, we propose an approach based representations. This approach addresses the limitations of

for extracting local world models, i.e., small, context-specific LLMs in handling complex reasoning and provides a more robust,

representations of knowledge that capture the relationships and logically consistent framework for educational applications.

rules governing a particular domain or scenario. The approach

is multi-step. First, the input text is segmented into manageable 3 Methodology

parts. Each segment is analyzed to extract key concepts and their This section introduces the approach for creating local world

interrelationships, which are then represented as Prolog defini-

models by generating and utilizing structured data in Prolog. The

tions. Then, the definitions are merged into a comprehensive methodology is designed to systematically identify and map the

knowledge graph that reflects the structure and content of the

concepts and their interrelationships within a given educational

input text. document, such as a textbook, facilitating the generation of a

We focus specifically on the educational domain, where the

knowledge graph.

ability to generate and utilize local world models could signifi-

cantly enhance the effectiveness of AI-driven educational tools, 3.1 Document segmentation

To manage the document’s complexity and ensure accurate con-

Permission to make digital or hard copies of all or part of this work for personal

cept extraction, the source material was divided into several

or classroom use is granted without fee provided that copies are not made or

shorter parts, each up to 10 pages long. This segmentation was

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this crucial in allowing us to focus on smaller, more manageable sec-

work must be honored. For all other uses, contact the owner /author(s).

tions of the content, enabling a thorough analysis and avoiding

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

problems that come with long-context LLM outputs. The length

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.sikdd.22 of each part was determined based on the natural divisions within



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Mark David Longar, Erik Novak, and Marko Grobelnik



the text, such as chapters or major sections, to maintain the co- concept map that helped identify key learning paths and pre-

herence of concepts within each segment. requisites. Prolog (specifically SWI Prolog [11]) was chosen for

this task because it can handle structured data, is widely used

3.2 Generating Prolog definitions (increasing the likelihood that LLMs have encountered it during

training), and can be executed and analyzed immediately.

For each segmented part, we created a prompt to generate Prolog

definitions of the concepts and their relationships. The prompt

was carefully crafted to guide the extraction of educational con- 4 Experiment Setting

tent in a structured format. It consisted of three main components: This section outlines the experiment setting for evaluating our

the context, the predicates and the structured output. approach to extracting local world models from educational texts

and generating structured Prolog representations. We describe

Context. A description of the educational context and a brief the data sources, the large language model used, and the evalua-

narrative to position the content within a learning scenario. This tion framework.

helped to align the LLM-extracted concepts and relationships

with our downstream tasks. The following is an example of the 4.1 Data sources prompt used:

We evaluated our approach on two widely used textbooks in

deep learning and natural language processing. These texts were

You are a teacher and an expert in natural language process- chosen because they are relevant to both structured reasoning

ing (NLP). You wrote a chapter in an NLP textbook and would tasks and the representation of complex, multi-step concepts.

like to convert the content of the chapter into a classroom The following chapters were selected for analysis:

lesson. You would like to step into the shoes of a student in

order to understand their learning process of this material. Deep Learning Preliminaries from the book Dive into Deep

You need to understand which concepts are being taught and Learning [12]. This chapter provides foundational knowledge

their relationships. of deep learning, covering key concepts such as linear algebra,

calculus, and probability, which are essential for understanding



Predicates. the field. The textbook’s teaching approach is highly hands-on,

List of predicates and their descriptions, which were

essential for identifying concepts ( ), prerequisites

isConcept(A) with a significant portion devoted to code. It is open-sourced, and

1

( ), and sections ( ). These isSection(S)

isPrerequisiteOf(A, B) we used the Markdown files provided on their GitHub page .

predicates were used to simulate the learning process, where con-

cepts are linked to sections. A concept may have prerequisite Chapter 2: Regular Expressions, Tokenization, and Edit

concepts or sections that must be understood before a student Distance Speech and Language Processing

from [4]. This chapter

introduces basic NLP techniques, focusing on regular expressions

can advance to learning the concept.

and tokenization, which are pivotal in text preprocessing tasks.



Structured output. Clear instructions to output the extracted

predicates in the form of a Prolog program. The LLM responding 4.2 Used large language model

in a structured format a crucial part of our approach, as it has been We employed GPT-4o via the ChatGPT interface to extract con-

shown that structured responses can improve LLM reasoning cepts and their interrelationships. We leveraged the model’s mul-

and generation quality [13]. timodal capabilities, allowing it to process text and PDF docu-

ments.

In summary, this prompt allowed us to extract detailed sum-

maries of the concepts taught and their relationships, which 4.3 Evaluation Framework were then represented in Prolog. Each segment was processed

We developed an evaluation framework to assess the performance

independently to generate a corresponding Prolog program.

of our approach based on three primary aspects: accuracy, com-

pleteness, and consistency. To validate the results, we manually

3.3 Merging Prolog definitions reviewed the extracted knowledge graphs and compared them

After generating the Prolog definitions for each segment, the with the source texts. We ensured that the extracted concepts

next step was to merge them into a single cohesive program. To were accurate, complete, and logically consistent.

achieve this, we created a prompt, which was nearly identical to

the first, but with instructions to combine the disjoint parts into The following criteria were used to eval- Assessment Criteria.

one integrated Prolog program added to the end of the prompt: uate the effectiveness of our approach:

• Accuracy. This aspect examines how accurately the approach

Now you need to combine the parts into a single Prolog pro- extracted the concepts and their relationships from the text.

gram. Make sure to include all the concepts and relationships, We evaluated the correctness of each Prolog definition against

but also properly connect them. Merge concepts from different the source material.

sections where necessary and make sure to include all the • Completeness. This evaluates whether the system captured all

sections and their relationships. the key concepts from the educational material. The assess-

ment ensured that no significant concepts or relationships



3.4 were omitted during extraction. Use of the knowledge graph

• Consistency. This aspect assesses the extent to which the ex-

The generated knowledge graph, represented by the Prolog pro-

tracted models maintained logical coherence across different

gram, was then used to recommend the next steps in the learn-

ing process. Using the structured output, we created a detailed 1https://github.com/d2l- ai/d2l- en



Creating Local World Models using LLMs Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



segments of the text. This was crucial in determining whether there were rare occasions where the output required manual in-

the segmented Prolog definitions could be merged into a cohe- terventions to fix inconsistent formatting of the Prolog variable

sive KG. names.



6 Discussion

5 Results

Our approach to extracting local world models from educational

In this section, we review the knowledge graphs of the two tested

texts demonstrated strong performance in generating logically

texts generated by our model.

coherent knowledge graphs from high-level concepts, but certain

limitations were identified. The synthetic data generation effec-



5.1 tively captured core concepts from both textbooks, particularly Dive into Deep Learning

in structuring major branches such as Linear Algebra, Calculus,

The selected chapter covered six sub-chapters in the following Dive into Deep Learning and Probability from . However, some

order: Data Manipulation, Data Preprocessing, Linear Algebra,

restructured sections, while logical, differed significantly from

Calculus, Automatic Differentiation, and Probability and Statis-

the source material’s flow.

tics. The results are represented by the graph in Figure 1. Speech and Language Processing In the textbook, the Regular

The system accurately identified three major independent

Expressions subsection was extracted with sufficient accuracy.

branches of the chapter – Linear Algebra, Calculus, and Probabil-

Other sections, such as Tokenization and Edit Distance, suffered

ity and Statistics – which reflects the structure of the source ma-

from detail omissions, where only top-level concepts were ex-

terial. The extracted knowledge graph also logically restructured

tracted. This issue was more prominent due to the higher in-

the content in ways that differed from the original organization

formation density of the NLP textbook, exposing limitations in

but made sense pedagogically. This restructuring highlights the

handling detailed, densely packed content.

logical flow of how data handling techniques naturally feed into

Regarding the evaluation framework, the model generally per-

more abstract mathematical concepts despite differing from the

formed well on metrics like accuracy and consistency but strug-

original structure.

gled with completeness in more detailed sections. The model’s

However, some omissions and reassignments were noted, par-

tendency to restructure content logically, though sometimes de-

ticularly within the Linear Algebra section. Concepts such as

viating from the original, suggests that while it captures core

vectors and matrices were omitted, likely due to the high-level

relationships, further refinements are needed to preserve peda-

nature of the extraction process. Additionally, matrix multipli-

gogical flow and details.

cation, though identified, was separated from Linear Algebra



basics and Tensor operations. This disjunction represents a slight 6.1 Potential improvements deviation from the expected conceptual hierarchy.

To address the limitations, improving the prompt engineering

Similarly, in the Calculus section, the extracted model restruc-

could lead to more detailed extractions while maintaining the

tured the sequence of topics. This restructuring captured the

structure of the source material. Additionally, enhancing the

relationship between fundamental calculus concepts and their

model’s ability to handle complex, dense information would mit-

practical applications in machine learning. Furthermore, the sys-

igate the loss of key concepts. Future iterations may benefit from

tem included concepts like Gradient Descent and Backpropaga-

automated post-processing checks to ensure logical consistency

tion which were only briefly mentioned in the source material.

and reduce manual interventions. Overall, while the approach

shows promise, refining it to handle finer details and complex

5.2 Speech and Language Processing sequences more effectively will be essential for broader applica-

tions.

The Regular Expressions section, seen in Figure 2, was extracted

accurately, capturing the core concepts effectively. However, a

noticeable limitation was the loss of the original sequencing of 7 Conclusion and Future work

the concepts presented in the textbook. While the key ideas were In this paper, we proposed a novel approach to extracting local

identified, the pedagogical flow, which is essential for gradual world models from educational texts by generating structured

learning, was somewhat disrupted in the extraction process. Prolog representations. Our methodology demonstrated the abil-

For the other sections, including Tokenization and Edit Dis- ity to capture core concepts and their interrelationships in a logi-

tance, the model extracted only the most prominent concepts, cal and coherent manner, especially in the Dive into Deep Learning omitting many important details. As a result, these sections are textbook. However, the results from the more information-dense

less comprehensive than they need to be for in-depth understand- text revealed limitations, partic- Speech and Language Processing

ing. Despite this, the overall connections between sections in ularly in handling detailed content, large knowledge graphs, as

the knowledge graph were logically structured, showing that the well as preserving pedagogical flow.

system was still able to create a coherent representation of the The use of Prolog proved effective in organizing educational

material at a high level. material, allowing for structured reasoning and enabling appli-

It is important to note that this textbook is significantly more cations in AI-driven educational tools. Despite these successes,

information-dense and longer compared to the certain challenges remain, such as the omission of detailed con- Dive into Deep

Learning book. This added complexity exposed some limitations cepts and the system’s occasional tendency to deviate from the

in the current approach, mainly when dealing with texts that re- original sequence of topics.

quire detailed extraction of concepts and their interrelationships. Future work will address these limitations by improving the

The model’s ability to handle such dense material is limited by prompt engineering and enhancing the system’s ability to handle

its tendency to focus on top-level ideas while losing much of the complex, information-dense material. Additionally, we plan to

depth and sequencing provided in the source text. Additionally, explore automating the segmentation process and scaling up the



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Mark David Longar, Erik Novak, and Marko Grobelnik



Deep Learning Prerequisites



Linear Algebra Basics Calculus Basics Probability Basics



Tensor Operations Matrix Multiplication Gradient Descent Chain Rule Statistics Basics



Data Preprocessing Broadcasting Techniques Optimization Techniques Backpropagation Stochastic Models



Automatic Differentiation



Loss Function Optimization



Figure 1: Knowledge graph of the Preliminaries section from Dive into Deep Learning.



Regular Expressions



Concatenation Square Brackets Kleene Star Period Anchors Disjunction Precedence Word Boundary Substitution



Question Mark Kleene Plus Parenthesis Greedy Matching Capture Group



Non-Greedy Matching Lookahead Assertion



Figure 2: Knowledge graph of the Regular Expressions section from Speech and Language Processing.



model to generate larger, more intricate knowledge graphs. Other [5] Vamsi Krishna Kommineni, Birgitta König-Ries, and Sheeba

potential directions include integrating retrieval-augmented gen- Samuel. “From human experts to machines: An LLM sup-

eration [7] to enrich knowledge extraction and comparing gen- ported approach to ontology and knowledge graph con-

erated world models across different texts to evaluate their peda- struction”. In: (2024). arXiv preprint arXiv:2403.08345 gogical alignment. Self-evaluation and correction mechanisms [6] Douglas B Lenat. “CYC: A large-scale investment in knowl-

could also be introduced to improve accuracy and completeness. edge infrastructure”. In: 38.11 Communications of the ACM

(1995), pp. 33–38.

Acknowledgments [7] Patrick Lewis et al. “Retrieval-augmented generation for

knowledge-intensive nlp tasks”. In: Advances in Neural

This work was supported by the Slovenian Research Agency

and the European Union’s Horizon 2020 project Humane AI Net Information Processing Systems 33 (2020), pp. 9459–9474.

[8] Fabian Neuhaus. “Ontologies in the era of large language

(Grant No. 952026).

models–a perspective”. In: 18.4 (2023), Applied ontology

References pp. 399–407.

[9] Shirui Pan et al. “Unifying large language models and

[1] Josh Achiam et al. “Gpt-4 technical report”. In: arXiv preprint

knowledge graphs: A roadmap”. In: IEEE Transactions on

arXiv:2303.08774 (2023).

[2] Knowledge and Data Engineering (2024). Hamed Babaei Giglou, Jennifer D’Souza, and Sören Auer.

[10] Mohammad Javad Saeedizade and Eva Blomqvist. “Navi-

“LLMs4OL: Large language models for ontology learning”.

gating Ontology Development with Large Language Mod-

In: . Springer. 2023, International Semantic Web Conference

els”. In: . Springer. 2024, European Semantic Web Conference

pp. 408–427.

pp. 143–161.

[3] Tom Brown et al. “Language Models are Few-Shot Learn-

[11] Jan Wielemaker et al. “SWI-Prolog”. In: Theory and Practice

ers”. In: . Advances in Neural Information Processing Systems

Vol. 33. Curran Associates, Inc., 2020, pp. 1877–1901. (Vis- of Logic Programming 12.1-2 (2012), pp. 67–96. issn: 1471-

0684.

ited on 08/27/2024).

[12] Aston Zhang et al. . https://D2L.ai. Dive into Deep Learning

[4] Daniel Jurafsky and James H. Martin. Speech and Language

Cambridge University Press, 2023.

Processing: An Introduction to Natural Language Processing,

Computational Linguistics, and Speech Recognition with [13] Pei Zhou et al. “How FaR Are Large Language Models

From Agents with Theory-of-Mind?” In: arXiv preprint

Language Models. 3rd. Online manuscript released August

arXiv:2310.03051 (2023).

20, 2024. 2024. url: https://web.stanford.edu/~jurafsky/

slp3/.





Semantic video content search and recommendation


∗ ∗ ∗

Mark David Longar Jakob Fir Bor Pangeršič

Jožef Stefan Institute University of Ljubljana University of Ljubljana

Ljubljana, Slovenia Ljubljana, Slovenia Ljubljana, Slovenia



Abstract a recommendation system that interacts with users to capture

their immediate preferences, thereby overcoming the cold start

The rapid growth of video streaming platforms has intensified

problem and enhancing the relevance of recommendations. Addi-

the demand for personalized content recommendations. How-

tionally, ensuring consistency in the quality of recommendations

ever, current solutions often rely on historical user data, leading

across different languages is increasingly important as many

to challenges like the cold start problem and overlooking users’

streaming services operate globally.

immediate preferences. We present a conversational recommen-

Our approach utilizes LLMs to generate keyword descriptions

dation system that leverages large language models (LLMs) to

for both content and user queries. These keywords serve as the

generate keyword-based content and query descriptions. By in-

basis for recommendations, with a Retrieval-Augmented Gen-

tegrating Retrieval-Augmented Generation (RAG), our system

eration (RAG) [6] model efficiently retrieving relevant content.

efficiently retrieves relevant content, independent of prior user in-

By crafting query keywords using LLMs, the system adapts to

teractions, and ensures consistent performance across languages.

user preferences in real time, providing relevant and language-

Preliminary testing shows our system outperforms the RAG base-

consistent recommendations.

line by up to 24% in less descriptive queries and demonstrates

This paper makes the following contributions: (1) Develop-

consistent performance across three languages. While the results

are promising, further evaluation focusing on user interaction ment of a Keyword-Based Recommendation System:

We in-

troduce a novel approach that utilizes LLMs to generate keyword-

and satisfaction is necessary. Our approach can potentially be

based descriptions for content and user queries, enabling more

extended to other recommendation systems, offering broader

personalized and adaptive recommendations. (2) Exploration of

applicability and enhanced content personalization.

Two User Interaction Models: We propose and evaluate two

Keywords distinct interfaces for user interaction—a conversational chat-

based model and a structured question-answering model, where

large language models, recommendation system, search system,

the system refines recommendations through a series of targeted

retrieval augmented generation

yes/no questions generated by the LLM. (3) Comprehensive



1 Evaluation Strategy: We outline a detailed plan for evaluating Introduction

the system’s performance in a production environment, focusing

The surge in video streaming platforms has accelerated the de-

on its ability to deliver consistent, high-quality recommendations

mand for personalized content recommendations. As these plat-

across different languages and user contexts.

forms expand their libraries and user bases, the challenge of

delivering precise, user-specific recommendations intensifies. In

this dynamic environment, streaming services must quickly adapt 2 Related Work

to provide accurate recommendations, which are crucial for main- Recommender systems have progressed from techniques such

taining user engagement and ensuring satisfaction.

as collaborative filtering and matrix factorization to more com-

Existing recommendation systems primarily rely on historical plex models that incorporate deep learning. The advent of large

user interaction data, such as viewing history and ratings. This language models (LLMs) has enabled innovative methods for

dependence leads to significant challenges, such as the cold start interacting with these systems [11], particularly when combined

problem, where new users or newly added content lack sufficient with retrieval techniques [9]. One of the most promising advance-

data for accurate recommendations. Additionally, these systems ments in this area is the use of Retrieval-Augmented Generation

often fail to account for users’ immediate preferences, which can (RAG) models, which integrate the powerful text generation ca-

change dynamically due to various factors such as mood, viewing pabilities of LLMs with retrieval-based methods to improve rec-

context (e.g., watching alone or with a group), or recent events ommendation accuracy and relevance [6]. in the user’s life. This gap highlights the need for more adaptive

Recent advancements in conversational recommender systems

and responsive recommendation mechanisms. have focused primarily on integrating LLMs with traditional rec-

Recent advancements in Large Language Models (LLMs) present ommender systems or fine-tuning LLMs using user-item interac-

an opportunity to address these limitations. LLMs offer significant tion data [9], [10], e.g., [8], [4], and [5]. These approaches, while

potential due to their emergent reasoning abilities, their capacity effective, often rely heavily on historical user data, leading to

to extract high-quality representations of textual features, and challenges such as the cold start problem. This reliance under-

their ability to leverage the vast external knowledge encoded scores the need for novel methods that reduce dependency on

within them [10], [7]. By harnessing LLMs, it is possible to create

past interactions and leverage real-time retrieval mechanisms to

∗ enhance content recommendations [2].

All authors have contributed equally.

To address these challenges, recent work by Di Palma et al.

Permission to make digital or hard copies of part or all of this work for personal

(2023) [2] introduced a Retrieval-Augmented Recommender Sys-

or classroom use is granted without fee provided that copies are not made or

tem, which combines the strengths of LLMs and retrieval-based

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this methods. Their approach employs LLMs both at the conversa-

work must be honored. For all other uses, contact the owner /author(s).

tional layer and the backend retrieval process, thereby improving

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

recommendation relevance, particularly in scenarios with sparse

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.sikdd.10 data or new users. Their experimental results demonstrated that

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Mark David Longar, Jakob Fir, and Bor Pangeršič





this RAG-based framework performs comparably to state-of-the-

art systems, even in zero-shot scenarios, underscoring the poten-

tial of such an approach to mitigate cold start and hallucination

problems inherent in LLMs.

Our approach builds on the strengths of RAG-based models by

introducing a keyword-based recommendation system that oper-

ates within a RAG framework. This system ensures consistent

performance across multiple languages and adapts to real-time

user preferences without relying on historical user data.



3 Data

The data used in this study was provided by our partner United

Cloud, who operate a multinational streaming service in the

1

Balkan region, EON TV . The EON platform encompasses a vari-

ety of content, such as video-on-demand (VOD) movies and TV

Figure 1: Overview of the Recommendation Pipeline.

shows, as well as live TV channels. We focused exclusively on

VOD movie data, although our approach is capable of accommo-

dating multiple content types.

The VOD movies data set comprises nearly 5000 movies in 4.2 User Interface

various languages. Each movie is accompanied by a brief descrip- Our proposed user interface designs (see Figure 2) offer two main

tion averaging around 460 characters (5-6 sentences) in multiple ways for users to interact with our recommendation core. Be-

languages. In cases where multiple translations were available, sides a direct search, where the user submits a query and receives

we opted for the original language of the movie; otherwise, we recommendations in a single step, we propose: A chatbot, (a) chose the first available translation. which assists users in narrowing down their options through a

conversational interface. The chatbot provides recommendations

4 Methodology at each response, allowing for a multi-step interaction that re-



4.1 fines the search results progressively. An inquisitive method, (b) Recommendation Mechanism

where an agent asks the user a series of Yes/No questions to

The core of our recommendation system is the generation of tex-

narrow down the search. Keywords are generated based on the

tual representations of content. Instead of using movie descrip-

user’s responses, making it particularly useful for users who are

tions directly, we employ the LLM to generate a set of English

uncertain about what they want to watch. This approach shifts

keywords and related movies. This approach prevents the model

the burden of knowing what to query from the user to the system,

from overemphasizing less relevant details, such as specific plot

streamlining the recommendation process.

points, that may not be central to the user’s query. User queries

Each of these designs aims to enhance user engagement and

follow a similar approach, where the LLM generates a set of

satisfaction by providing tailored interactions that cater to differ-

relevant keywords, as well as any possibly relevant movies.

ent user preferences and needs.

One of the key advantages of this method is its ability to



abstract core concepts from user queries using the LLM, aligning 5 Evaluation better with the keywords generated from movie descriptions.

We have developed a twofold approach for addressing the evalu-

The LLM-generated keywords from both the movie descriptions

ation of our model:

and user queries are designed to encapsulate the essential topics

First, to gauge the effectiveness of our keyword-based ap-

and themes. By aligning the keywords generated from movie

proach for recommendation, we curated a small multilingual

descriptions with those derived from user queries, our system

evaluation dataset to test our core recommendation mechanism.

enhances the relevance of the recommendations. This alignment

This dataset includes queries in various languages along with

is crucial in ensuring that the retrieved movies resonate with

their expected recommendations. We compared the performance

the user’s expressed interests, even when these interests are

of our mechanism with a baseline RAG system that directly em-

not articulated well. Furthermore, the use of in-context learning

bedded user queries and movie descriptions.

allows the system to maintain its performance without extensive

Second, to assess the efficiency and user satisfaction of our sys-

fine-tuning [3], making it both efficient and effective.

tem in real-world situations, we have devised an evaluation plan

The rest of the recommendation system follows the Retrieval-

to test our system in production. This strategy utilizes a struc-

Augmented Generation (RAG) [6] pipeline (see Figure 1). The

tured A/B testing framework to conduct precise comparisons

RAG pipeline operates by first generating textual representations

between our semantic recommendation system and conventional

of movies, which are then embedded into a vector space. These

search, addressing distinct aspects of user experience and system

embeddings are stored in a vector database, allowing for efficient

performance.

similarity searches. When a user submits a query, the system

generates a corresponding representation, embeds it into the

same vector space, and retrieves the top 𝑘 most similar movie 5.1 Evaluation dataset

embeddings from the database. This process ensures that the rec- To create our evaluation dataset, we carefully selected 25 movies

ommendations are both contextually relevant and semantically across multiple languages, including both well-known and lesser-

aligned with the user’s input. known titles. For each movie, we formulated two types of queries

to assess the system’s retrieval accuracy: and Descriptive General

1No EON user data was used. queries.



Semantic video content search and recommendation Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



The queries were designed to simulate scenarios recommendation links, and Watch Time to gauge the duration Descriptive

where the user knows exactly what they are looking for. For users engage with recommended content. Additionally, immedi-

instance, a query for the movie might be, "I am ate user reactions are captured through Like/Dislike Ratios, while Messi (2014) looking for inspirational documentaries about famous athletes, more detailed user feedback is collected via surveys administered

such as Lionel Messi and his rise through football." In contrast, after interactions.

the queries were intended to test situations where the : We analyze User Interaction Patterns, such General Behavioral Metrics user has only a rough idea of what they want to watch, which is as search frequency and refinement actions, and System Usage

likely more common in real-world environments. An example Frequency to determine how different demographics utilize the

of a general query for the same movie might be, "soccer movies system and to identify any potential biases in system engagement.

that will inspire me." We also record the search time and number of queries needed for

To evaluate the system’s performance across different linguis- a decision.

tic contexts, we manually translated these queries into English,

Serbian, and Slovenian. We then compared the performance of 6 Results our keyword-based retrieval mechanism against a baseline RAG

The outcomes presented in Table 1 showcase the performance of

model that directly used user queries and movie descriptions

both models in various query types and languages, as measured

without generating keywords.

by accuracy at the top 5 and top 10 recommendations.

The results reveal that the baseline model surpasses (or matches)



5.2 the performance of the keyword mechanism in the case of De- Experiment Design

scriptive queries, particularly in terms of Accuracy@5. However,

We have divided our user base into four distinct groups to facili-

in terms of Accuracy@10, the two models demonstrate relatively

tate a detailed comparative analysis, aligned with our proposed

similar performance. Conversely, the keyword model shows sig-

user interface designs: General

nificant performance enhancements for queries, partic-

Baseline Group: This control group doesn’t use our system, but ularly in Accuracy@10, indicating its capacity to adapt to non-

instead finds movies and receives recommendations based on the

specific content descriptions. Additionally, the keywords model

traditional recommendation methods, a common practice in the

consistently performs well across different languages, whereas

industry.

the baseline model shows fluctuations of up to 28% across lan-

Direct Semantic Search Group: This control group interacts guages.

with a straightforward search interface. Users submit a query

In summary, the keywords model allows for more general and

and receive recommendations in a single step. This approach

multilingual queries, while the baseline model excels at retrieving

provides immediate suggestions based on the user’s input, mim-

very specific content.

icking traditional full-text search practices.

Chatbot Group: Participants in this treatment group use a con-

Table 1: Evaluation results on the descriptions and gen-

versational interface (interface ), where a chatbot assists in a

eral queries data sets. LLM embeddings were generated

narrowing down options. The chatbot provides recommenda-

using OpenAI’s text-embedding-3-large model. The Key-

tions at each response, enabling a multi-step interaction that

words model used GPT-4o.

progressively refines the search results. This design enhances

engagement by simulating a natural conversation.

Inquisitive Method Group: Users in this group engage with Accuracy@5 Accuracy@10 an agent that asks a series of Yes/No questions to narrow down

Keywords Baseline Keywords Baseline

the search (interface ). Keywords are generated based on the b

user’s responses. Descriptive Queries

English 60% 64% 68% 68%

The evaluation will be conducted continuously, starting with a

Serbian 56% 72% 80% 84%

focused initial phase over the first month post-implementation

Slovenian 56% 72% 80% 84%

to address immediate usability and performance issues, followed

by ongoing monitoring to capture long-term user engagement General Queries

and satisfaction. English 28% 44% 44% 68%

By implementing this structured evaluation framework, we aim Serbian 44% 52% 52% 68%

to comprehensively understand the impact and effectiveness of Slovenian 44% 56% 56% 72% our semantic recommendation system, guiding further refine-

ments and ensuring that the system meets user needs and expec-

tations.

6.1 User Interface Implementation

5.2.1 Metrics We would like to measure how users interact We implemented our proposed interface design using Flutter, with our system in two main ways: First, we would like to know which guarantees functionality across a variety of devices, includ-

how engaged and satisfied they are with our recommendations, ing iOS, Android, Windows, and web browsers. This cross-device

i.e., do users find our system frustrating to navigate, and whether compatibility is crucial as it ensures that all users, regardless

they watch movies recommended by our system. The second of their preferred platform, have access to our application. The

set of metrics will aim to capture how different demographics support for mobile devices is particularly useful in our interroga-

interact with our system, as a major goal is to remove any biases tion design, where users can easily navigate through options by

such as language or age. swiping cards left or right.

Engagement and Satisfaction Metrics: These include Click- Additionally, we integrated Tipko [1], a Slovenian transcrip-

Through Rate (CTR), which measures the percentage of clicked tion service, to facilitate voice-to-text capabilities. This feature

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Mark David Longar, Jakob Fir, and Bor Pangeršič



enhances user convenience by enabling voice communication integrations, such as the user’s calendar. We also intend to expand

with our chat bot, removing the necessity for typing. our user interface by introducing new forms of interaction, such

as movie trailers and multiple-choice questions.



To overcome the limitations of our movie information, we

are interested in delving deeper into the content by analyzing

subtitles using a local language model. Additionally, we aim to

broaden our database to include other types of content, such as

live channel content and special time-limited events like Eurovi-

sion, Eurobasket, and the FIFA World Cup.

Finally, we are interested in the integration of a traditional

recommendation models that utilize historical watch data or

ratings to re-rank our recommendations.



Acknowledgments

This project was made in collaboration with United.Cloud and

In516ht for the 2024 Data Science Competition, organized by The

Faculty of Computer and Information Science at the University of

Ljubljana. We thank our advisors Slavko Žitnik, Aljaž Košmerlj,

Klementina Pirc, and Rebeka Merhar for their contributions.



References

[1] Primož Bratanič. Transkript app | Samodejna transkripcija

slovenskega govora. May 2024. url: https://transkript.si/.

Figure 2: Implementations of our (a) Chatbot (left) and (b) [2] Dario Di Palma. “Retrieval-augmented recommender sys-

Inquisitive (right) user interface designs. tem: Enhancing recommender systems with large lan-

guage models”. In: Proceedings of the 17th ACM Conference

on Recommender Systems. 2023, pp. 1369–1373.

7 Discussion [3] Elnara Galimzhanova et al. “Rewriting Conversational

Utterances with Instructed Large Language Models”. In:

This report introduces a new content recommendation mecha-

(Oct. 2023). doi: 10.1109/wi- iat59888.2023.00014. (Visited

nism and three ways to interact with it. Table 1 demonstrates the

on 05/22/2024).

success of our keyword retrieval model in understanding general

[4] Yunfan Gao et al. “Chat-rec: Towards interactive and ex-

user preferences while still performing well when searching for

plainable llms-augmented recommender system”. In: arXiv

specific content. Moreover, its consistency across languages and

its ability to retrieve content using specific descriptions as well preprint arXiv:2303.14524 (2023).

[5] Xu Huang et al. “Recommender ai agent: Integrating large

as general themes make it well-suited for a diverse user base.

language models for interactive recommendations”. In:

Additionally, the keyword model allows seamless integration

with both the arXiv preprint arXiv:2308.16505 (2023). Chatbot and Inquisitive methods. Moreover, our

[6] Patrick Lewis et al. “Retrieval-augmented generation for

system could be extended to dynamically adjust keyword genera-

knowledge-intensive nlp tasks”. In: Advances in Neural

tion based on user-specific factors such as viewing history, local

time, weather, and current mood indicators. This personalization Information Processing Systems 33 (2020), pp. 9459–9474.

[7] Peng Liu, Lemei Zhang, and Jon Atle Gulla. “Pre-train,

ensures that the recommendations are not only relevant to the

Prompt, and Recommendation: A Comprehensive Survey

content but also tailored to the user’s immediate context and

of Language Modeling Paradigm Adaptations in Recom-

preferences.

mender Systems”. In: Transactions of the Association for

Our approach has some limitations, including the cost per

query, which is higher than traditional search, although not ex- Computational Linguistics 11 (2023), pp. 1553–1571.

[8] Zihan Liu et al. “ChatQA: Building GPT-4 Level Conver-

orbitant. Furthermore, our model’s performance is commendable

sational QA Models”. In: arXiv preprint arXiv:2401.10225

given our limited knowledge about the movie content but relies

(2024).

on the assumption that the language model may have more infor-

[9] Arpita Vats et al. “Exploring the Impact of Large Language

mation about a movie than our dataset. It’s worth noting that, in

Models on Recommender Systems: An Extensive Review”.

the short term, it appears that models are continually improving,

In: (2024). arXiv preprint arXiv:2402.18590

becoming faster, more knowledgeable, and more cost-effective.

[10] Likang Wu et al. “A survey on large language models for

Lastly, as with any chat application that involves user inputs,

recommendation”. In: 27.5 (2024), p. 60. World Wide Web

security is a crucial consideration. While improvements can be

[11] Bowen Zheng et al. “Adapting Large Language Models

made through better prompting and fine-tuning, ongoing moni-

by Integrating Collaborative Semantics for Recommenda-

toring is essential when the system is in production.

tion”. In: 2024 IEEE 40th International Conference on Data

8 Engineering (ICDE). 2024, pp. 1435–1448. doi: 10 . 1109 / Future work

ICDE60146.2024.00118.

In future work, we plan to further explore methods for improving

user experience and personalization. Our initial experiments have

involved incorporating the user’s time, location, and weather to

enhance results. Moving forward, we aim to explore additional





Continuous Planning of a Fleet of Shuttle Vans as Support for




Dynamic Pricing

Filip Stavrov Luka Stopar

stavrovf@gmail.com luka.stopar@ijs.si

Jožef Stefan Institute Jožef Stefan Institute

Jamova cesta 39 Jamova cesta 39

Ljubljana, Slovenia Ljubljana, Slovenia



ABSTRACT Once we receive these predictions, our goal is to simulate

reservations based on this data. For instance, if the predictions

This paper solves the problem of estimating the number and type indicate that 12 passengers will travel from Ljubljana to Koper on

of required resources for pickup and delivery of passengers at some October 20, 2024, we would simulate reservations using sampling

time in the future. By combining optimization and sampling techniques. One particular example is creating four separate

methods, as well as making plans based on several statistical bookings—one for five passengers, one for three, and two for two

samples, we estimate the real values for the required resources passengers each. We will introduce the sampling techniques used

and show how the sample values converge towards the real values. in this process in greater detail later on.

Our approach combines machine-learning based demand

predictions, for the number of passengers, and a route After generating these reservations, the next step is to input them

optimization engine that assigns the passengers into shared shuttle into the Route Optimization Engine to generate a plan for that day.

vehicles. In order to validate our method we create a baseline data This plan will specify the number of vehicles required and the

that is representative of the real values. We test our approach using specific reservations each vehicle will serve.

this baseline data, and we obtain statistically significant results.

The main hypotheses that our approach explores and



KEYWORDS experimentally tests are the following:



statistical samples, demand predictions, route optimization engine,  H1: We can accurately estimate the number of required

resources using optimization methods based on

sampling techniques, optimization technique

predicted passenger numbers.

1  H2: Monte Carlo sampling of historical distributions can INTRODUCTION

effectively model uncertainty in demand predictions,

The effective allocation of resources is a critical topic in the mobility leading to stable resource estimations.

industry. Anticipating the number and type of resources required  H3: Creating plans based on several sample values will

can significantly enhance a company's ability to plan accurately for converge towards the actual number of required

the future. Our work addresses this challenge by focusing on how resources.

to estimate the number and type of vehicles needed for passenger

pickup and delivery at a future time. The input to our problem On the other hand, the key assumptions and limitations that

consists of machine learning-based demand predictions, which underline our research are:



provide estimates of the number of passengers across various  Prediction Accuracy: We assume that the predictions routes offered by the company. These predictions are provided

effectively estimate the number of future passengers.

daily and further broken down into hourly estimates for each day.  Passenger Distribution: We assume that the number of

∗Both authors contributed equally to this research. passengers follows a Poisson distribution and that the

distributions on different routes are independent.

Permission to make digital or hard copies of part or all of this work for  Independence: We assume that the passenger personal or classroom use is granted without fee provided that copies are not

made or distributed for profit or commercial advantage and that copies bear distribution and the window type distributions are

this notice and the full citation on the first page. Copyrights for third-party independent to each other.



components of this work must be honored. For all other uses, contact the  Concept Drift: We assume there is no concept drift in the owner/author(s). data, meaning the underlying data patterns do not Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia © 2024 Copyright held by the owner/author(s). change over time.

https://doi.org/10.70314/is.2024.sikdd.27



291

2 RELATED WORK The process begins with demand predictions and culminates in the

generation of reservation data. Critical steps include sampling the

The problem of resource allocation in the mobility industry, number of passengers per reservation, the window type, and the

particularly in the context of vehicle routing and passenger demand window length. Sampling is done from probabilistic distributions

prediction, has been extensively studied. Traditional methods for derived from historical data, with the distributions illustrated

vehicle routing often rely on static models that assume known and below.

deterministic demand. However, recent advances in machine

learning and optimization have enabled more dynamic approaches

that can account for uncertainty and variability in demand. [3][4]

For instance, predictive analytics has been employed to forecast

passenger demand using historical data, which can then be fed into

optimization algorithms to determine the optimal allocation of

vehicles. Monte Carlo simulation is another technique commonly

used to model uncertainty in demand predictions, providing a

probabilistic framework for decision-making under uncertainty. [2]

Moreover, dynamic vehicle routing approaches, have

demonstrated the benefits of real-time adjustments to routing

plans based on updated demand information. [1] The integration

of these methodologies into a continuous planning framework is Figure 2. Window type distribution

relatively novel and addresses the limitations of static planning

approaches, particularly in highly variable and uncertain

environments. [1][5]



3 METHODOLOGY



Our methodology begins with demand predictions for the number

of passengers, and the ultimate goal is to determine the number

and type of vehicles required, as well as the reservations each

vehicle will serve. The figure below provides a detailed overview of

this process.





Figure 3. Window length distribution



Figure 1. Methodology



Starting with the demand predictions, we apply sampling

techniques to simulate reservation data. Specifically, we take the

predicted number of passengers for different routes at various

Figure 4. Number of passengers distribution

times and generate reservations through sampling. This

reservation data follows a specific format, including fields such as Please note that from a single demand prediction input file, we

ID, start location, end location, pickup time, and more. Key generate 100 independent samples of reservation data. This

attributes include the number of passengers per reservation and approach introduces uncertainty through probabilistic sampling.

the window type, which reflects travel preferences. For instance, Each independent sample is then submitted as a separate job to

some passengers may prefer a private vehicle (VIP), while others the Route Optimization Engine, where it solves a vehicle routing

are open to sharing the ride. Additionally, the window interval is problem with time constraints. The output for each job is a plan

crucial—it can be a specific time or a more flexible period, affecting corresponding to the reservation data. Our final objective is to

both the service pricing and overall experience. These factors will aggregate these results and analyze the insights they provide.

be incorporated into the dynamic pricing model later on.



292



4 RESULTS is acceptable given the overall similarity to the global mean, and

the sampling of values. Thus, despite the variance, the sampled

After solving all 100 jobs, we obtained 100 independent plans and values converge towards the actual values. This error distribution

began analyzing the results. As shown in the figure below, the is displayed on the figure below.

distribution of the number of passengers yielded a mean value of

325.87 with a standard deviation of 16.85. For the number of

vehicles, the mean was 38.01 with a standard deviation of 3.06. It’s

notable that the passenger data exhibits significantly more

variance compared to the vehicle data. This is expected, as

passengers are grouped into visits, and visits are then allocated to

vehicles, resulting in less variation in the vehicle count.



Figure 7. Required vehicles - error distribution



To statistically test whether the sampled and baseline data have

the same mean number of vehicles, we conducted a Welch's t-test.

The results showed a test statistic of 0.59, a p-value of 0.55, and a

95% confidence interval ranging from -0.64 to 1.23. Given the p-

value, we fail to reject the null hypothesis, meaning there is no

Figure 5. Sampled data: visits, vehicles and passengers statistically significant difference between the sampled and

distributions baseline vehicle counts. Additionally, the range of the mean

difference of vehicles between the sampled and the baseline data,

To further validate our approach, we created a baseline using the which is from - 0.64 to 1.23, falls within our practical significance

same data from which the demand predictions were generated. threshold of up to 2 vehicles, further supporting the similarity

We generated 100 samples from this baseline and submitted them between the two datasets. This indicates that we can effectively

as independent jobs. Upon completion, we compared the baseline estimate the number of required resources by applying

results with those of our sampled data. The mean number of optimization techniques on top of the demand prediction values.

vehicles from the baseline was 37.81 with a standard deviation of

3.01, which closely aligns with the values from our sampled data. We also analyzed the mean number of vehicles and observed that

You can observe the comparison on the figure below. this value converges toward the actual values as the number of

samples increases. This is shown on the figure below.



Figure 6. Comparison of required vehicles between sampled and

baseline data Figure 8. Convergence of means of sampled vehicles



We also analyzed the error distribution for the number of vehicles Finally, after obtaining both the number of passengers and the

between the baseline and sampled data, finding a mean absolute number of vehicles, we decided to fit a linear regression to explore

error of 3.16. This suggests that the difference between the two whether we could simplify the process and avoid the detailed

sets is minor, considering the sampling of data, and it is indicating approach previously described. As illustrated in the figure below,

a good alignment. Additionally, the average number of vehicles in the regression line serves as a reasonable estimator for the number

both the sampled and baseline data is quite similar. While the mean of vehicles based on the number of passengers. However, this

absolute error reflects some variability in the sampled values, this model struggles to capture the non-linear relationships influenced



293



by various optimization types, window lengths, and travel modes, ACKNOWLEDGMENTS

resulting in considerable variance around the regression line. While

it is generally true that a higher number of passengers correlates Our research is part of a broader, multi-partner initiative called

with an increased number of vehicles, this relationship can be CONDUCTOR. The primary objective of this project is to design,

misleading. Different travel types can accommodate more integrate, and demonstrate advanced, high-level traffic and fleet

passengers per vehicle, which can disrupt the linear relationship, management systems. These systems aim to optimize the transport

especially in cases where these travel types dominate. of passengers and goods efficiently on a global scale, ensuring

Consequently, although the linear regression provides a solid seamless multimodality and interoperability. The CONDUCTOR

approximation, it overlooks essential non-linear factors that are project is co-funded by the European Union’s Horizon Europe

critical to our analysis. Our approach, which integrates these research and innovation programme under the Grant Agreement

factors, demonstrates greater robustness and effectiveness. The No 101077049.

linear regression line and the data correlation are presented in the

figure below. REFERENCES



[1] Berbeglia, G., Cordeau, J. F., & Laporte, G. (2010).

Dynamic pickup and delivery problems. Transportation

Research Part B: Methodological, 44(5), 667-684.

https://doi.org/10.1016/j.trb.2009.10.004

[2] Ulmer, M. W., Thomas, B. W., & Mattfeld, D. C. (2018).

Preemptive depot returns for same-day delivery under

uncertain customer availability. European Journal of

Operational Research, 269(2), 356-371.

https://doi.org/10.1016/j.ejor.2017.08.008

[3] Bertsimas, D., & Sim, M. (2004). The Price of Robustness.

Operations Research, 52(1), 35-53.

Figure 9. Regression Analysis https://doi.org/10.1287/opre.1030.0065

[4] Ghiani, G., Guerriero, F., Laporte, G., & Musmanno, R.

5 CONCLUSION (2003). Real-time vehicle routing: Solution concepts,

algorithms and parallel computing strategies. European

In conclusion, our findings demonstrate that we can effectively Journal of Operational Research, 151(1), 1-11.

estimate the number of required resources by employing https://www.sciencedirect.com/science/article/abs/pii/

optimization methods based on predicted passenger numbers. As S0377221702009153

the number of samples increases, the sampled values consistently [5] Psaraftis, H. N., Wen, M., & Kontovas, C. A. (2016).

converge toward the actual resource requirements, reinforcing the Dynamic vehicle routing problems: Three decades and

reliability of our approach. Alternative methods, such as linear counting. Networks, 67(1), 3-31.

regression, fail to adequately address the non-linear complexities https://doi.org/10.1002/net.21628 inherent in resource allocation, such as varying optimization types

and window lengths. Our method, which incorporates these

factors, proves to be a far more accurate and effective solution for

resource estimation in the mobility industry.



294





Knowledge graph Extraction from Textual data using LLM




Khasa Gillani Erik Novak

khasagillani22@gmail.com erik.novak@ijs.si

Jožef Stefan Postgraduate School Jožef Stefan Institute

Ljubljana, Slovenia Ljubljana, Slovenia



Klemen Kenda Dunja Mladenić

klemen.kenda@ijs.si dunja.mladenic@ijs.si

Jožef Stefan Institute and Qlector Jožef Stefan Institute and

Ljubljana, Slovenia Jožef Stefan Postgraduate School

Ljubljana, Slovenia



ABSTRACT

Entity 1 Entity 2 Relation



The advent of Large Language Models (LLMs), such as Chat- JSI Slovenia located in Assets GPT and GPT-4, has revolutionized natural language process- Termboard Research AI JSI area ing, opening avenues for advanced textual understanding. This





study explores the application of LLMs in developing Knowledge KG Note Note graphs from textual data. Knowledge graphs offer a structured Input Generate text GPT ptompt representation of information, facilitating enhanced comprehen-



sion and utilization of unstructured text. We intend to construct



Knowledge graphs that capture relationships and entities within

diverse textual datasets by harnessing LLMs’ contextual under- Figure 1: Overview of proposed approach where input text

standing and language generation capabilities. The primary goal is processed through a Termboard to generate a structured

is to explore and understand how well LLMs can identify and prompt for LLM, creating an entity-relation table to build

extract relevant entities and relationships from textual data using a Knowledge graph (KG).

prompt engineering while contributing to structured knowledge



representation. labor-intensive and requires expert knowledge. However, con-

KEYWORDS structing Knowledge graphs from unstructured text is intricate

and depends on sophisticated natural language processing (NLP)

Knowledge graph, Large Language Models, prompt engineering, methods, including named entity recognition (NER) and relation

information extraction, textual data extraction. The advancement of LLMs like GPT-4 presents an op-

portunity to automate and improve this process as illustrated in

1 INTRODUCTION Figure 1. Utilizing LLMs can lead to more efficient, scalable, and

In an era where data is ubiquitous, efficient organization, retrieval, accurate Knowledge graph construction, thereby unlocking new

and interpretation of textual information are crucial. Knowledge possibilities in information management and AI applications.

graphs, representing facts and relationships in structured forms,

play a pivotal role in various AI applications, from enhancing 2 BACKGROUND

search engines to powering recommendation systems. However, An overview of recent research in Large Language Models and

the construction of these graphs is often hindered by the complex- Knowledge graphs is provided in this section, which also empha-

ity and variability of human language. This paper explores the sizes the potential for their integration.

potential of Large Language Models, like GPT-4, to revolution-

ize this process. By leveraging their advanced natural language 2.1 Large Language Model (LLM)

understanding capabilities, we aim to automate and refine the Large Language Models are advanced AI systems pre-trained

extraction of knowledge from textual datasets. The fundamental on extensive data, enabling them to comprehend and produce

purpose of this research is to understand the extent to which human language. Their recent surge in popularity is due to their

LLMs can identify and extract relevant entities and relationships proficiency in various language-processing tasks, including text

from textual data and then build a Knowledge graph using the completion, translation, summarization, and answering ques-

extracted information. tions. These models, primarily based on transformer architecture,

The motivation behind this study stems from the growing need utilize self-attention mechanisms through encoder-decoder mod-

to effectively manage and utilize the vast amounts of textual data ules. Encoders transform input text into numerical embeddings

generated daily. Knowledge graphs offer a structured and intu- that reflect the context and meaning, while decoders use these

itive way to represent information, but their construction is often embeddings to generate coherent and pertinent textual output.

The large language models feature a decoder-only architecture



work must be honored. For all other uses, contact the owner/author(s). Permission to make digital or hard copies of part or all of this work for personal and, thus, make a prediction of the target output text using only or classroom use is granted without fee provided that copies are not made or the decoder module. The training paradigm for these models is distributed for profit or commercial advantage and that copies bear this notice and to predict the next word in the sentence. Generally, large-scale the full citation on the first page. Copyrights for third-party components of this decoder-only LLMs such as ChatGPT [7] and GPT-4 [2], focus on Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia human-like language output, predicting subsequent words based © 2024 Copyright held by the owner/author(s). on the preceding text for tasks like text generation. https://doi.org/https://doi.org/10.70314/is.2024.sikdd.15



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Khasa and Dunia, et al.



Table 1: Simplified comparison between Large Language Models (LLMs) and Knowledge graphs (KGs)



Feature LLM KG

Knowledge type Broad, general knowledge Structured, domain-specific knowledge

Data handling Flexible, can process varied inputs Requires structured data

Accuracy May lack precision in understanding Highly accurate with structured data

Understanding Can interpret and generate language Designed for specific queries and relationships

Adaptability Adapts to new information by retraining Adaptable when updated with new data

Transparency Often seen as "black boxes" with unclear reasoning Clear decision-making pathways

Error rate Can make mistakes due to broad generalizations Can be prone to errors if data is incorrect or missing

Complexity Handles complex language tasks Manages complex relationships and attributes

Usage Broad applications in text generation, translation, Used for specific tasks like recommendations, search

etc. optimization

Scalability Scales with computational power Scales with the amount of structured data available



2.2 Knowledge graph (KG) and KGs can perpetuate biases present in their training data or

Knowledge graphs are structured representations of information construction methodologies. In conclusion, both LLMs and KGs

that depict the relationships between entities in a specific domain. have their unique strengths and challenges. While LLMs excel

They are used extensively in various applications, such as search in general language processing and knowledge extraction from

engines, recommendation systems, and question-answering sys- vast corpora, KGs provide a structured and interpretable way to

tems. These graphs use detailed connections between data to organize explicit knowledge. These differences underscore the

help with smart thinking, finding specific information easily, and potential benefits of integrating LLMs and KGs to create more

running applications that use knowledge. Hence, allows us to robust AI systems that leverage the strengths of both approaches.

better understand and use information across multiple fields.

Knowledge graphs provide a structured way of representing in- 3 PROOF OF CONCEPT: ANALYSIS AND

terconnected knowledge. They are precise and consistent, aiding KNOWLEDGE GRAPH GENERATION

in decisive and informed decision-making. KGs are particularly This section demonstrates how to process and analyze textual

valuable for their interpretability and explainability due to the data to build a Knowledge graph using LLM. It is important to

explicit representation of entities and relationships. They can mention that prompt engineering [5] is of great importance when

capture domain-specific information accurately and evolve to it comes to the results generated from ChatGPT. Since it is a gen-

incorporate new data. However, KGs may suffer from incom- erative model, small variations in the input sequence can create

pleteness and may not always reflect the most recent or unseen large differences in the produced output as demonstrated below.

facts. They also typically cannot understand natural language in We use two different textual files containing contextual data: (i)

an unstructured format [3][6]. Moreover, KGs are preferred in APRIORI proposal (containing project details, job description,

scenarios where explainability and interpretability are crucial, as potential candidate skills, hosting organizations, etc.) and (ii)

they provide structured knowledge representation. ADRIA Motorhome instruction manual (containing textual as



The comparison between Large Language Models and Knowl-2.3 well as tabular data). Moreover, building KG out of the ADRIA Combining LLM and KG instruction manual has potential applications for the manufac- turing industry. edge graphs (Table 1) can be supported by various references that



Language Models like ChatGPT [7] are celebrated for their gen- We compare ChatGPT-3.5 and GPT-4 extracted entities and rela- eralizability and ability to process diverse text data, allowing highlight their respective strengths and weaknesses [4]. Large 3.1 Using ChatGPT Prompts:

them to perform various language-related tasks without exten- tions using the same prompts. We use Termboard1 which offers



sive task-specific training. They can act as reservoirs of general customized ChatGPT prompts to create terms, entities, and rela- tions to visualize larger graphs from the provided text. knowledge, aiding in information synthesis and research. Their Prompt: Extract an ontology and create a table of relations with proficiency in language processing is useful in tasks like natural 3 columns in this order: source, target, and relation name. Also language understanding and sentiment analysis. However, they Create a table with 2 columns: put in the first column the name can suffer from hallucinations, where they generate plausible but of the term and in the second column an elaborate definition of factually incorrect information. Their "black-box" nature makes the term. Use this text as a basis: ¨ APRIORI ¨ - (contains textual it difficult to understand the internal decision-making processes, data about the job description, candidate skills, project description, and they can be indecisive, producing uncertain responses to hosting organization, etc) . ambiguous inputs. Additionally, while they have vast general Observing the Knowledge graphs generated by ChatGPT-3.5 knowledge, they may not be up-to-date with domain-specific or (Figure 2) and GPT-4 (Figure 3); we notice, that it didn’t extract all the latest information. Critics of LLMs argue that these models entities and relations and missing terms/concepts. For this reason, lack transparency and interoperability. we ran the second prompt, where we redefined a more detailed Recent research [3] [4]efforts are, however, improving LLM’s prompt to ask GPT-4 to explicitly generate a comprehensive interpretability through techniques like attention mechanisms ontology including all entities and relations from the provided and model introspection. KGs also present advantages over LLMs text, categorize entities into types like Persons, Organizations, by providing knowledge about long-tail entities, thus improv-

ing recall for knowledge computing tasks. However, both LLMs 1https://termboard.com/



Knowledge graph Extraction from Textual data using LLM Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



AI Production

Research Katholieke Horizon TMA Universiteit

Integrated Partner Department MSCA Leuven



explainaibilty learning for Active efficiency End-use strategy Organization of AI Slovenia CO2 Research facility Address emission conducts Active Aims to develop Funds uncertainty in Jožef Stefan learning CO2 Participant Doctoral Located in with Institute emissions candidate Recommends reduction of Collaborates Address Complexity of energy Explainability trains APRIORI Manufacturing trains APRIORI Job Develops goods Enroll at Linked to Contributes to description DC9 in Collaborates with Research Data Mining Doctoral Sustainable Involves candidate in Educates Collaborates with Research in Collaborates manufacturing

Focuses on Support transition of Machine Participate Develops with End-use

learning Jožef Stefan semantic energy

Manufacturing Visual Institute technologies efficiency Host sector inspection Internation Linked to Partner Industrial Energy agency

Organization Machine applications



Doctoral network Horizon TMA Manufacturing Machine learning Research in MSCA Partner EU Language sector learning,Data mining Organization manufacturing technologies Industry 4.0

sector Sensor

networks Additive

Manufacturing Part of

Figure 2: The KG generated using ChatGPT-3.5 contains 20



entities. It was able to extract entities and link them to re- Figure 4: The KG generated by GPT-4 contains 22 entities. It

lations, but it failed in abstracting concepts and specifying Identified more key entities and relevant concepts and iden-

entities (i.e. partner organizations, location, etc.). tified suitable relations to connect them (i.e. participant-

Katholieke Universiteit Leuven). However, it didn’t cover

Horizon TMA MSCA Doctoral all relations and classes (i.e. skills). We also notice a few

MSCA network

Slovenia Recruits duplicated entities(i.e. data mining, CO2 emission, etc.)

Salary



AI Located in Supports Doctoral and some independent entities (i.e. sustainable manufac-Regulates Part of candidate turing). Active Research area trains APRIORI learning

Natural Ehnace Subject of



Sciences Manufacturing in is experiencing Specilaizes in fers PhD position Products sector an evolving trend Of Participate Is located close to the End-use



Machine Jožef Stefan Customizations city center of Ljubljana Research area energy Research Institute efficiency careers learning Research area Explainable Specilaizes in varios Msc degree AI

Manufacture certificates Data Jožef Stefan research area,



mining Data Research area engineering Institute including AI EU Linked to Language List of manufacturing Critical part of technologies publications sector components Horizon TMA MSCA Is a Marie Sklodowska- doctoral network: Curie doctoral

Application APRIORI network

June 15, system The manufacture



Figure 3: The KG generated by GPT-4 contains 16 entities. 2023 CV(curriculum Include Alborg sector Vitae) Industry 4.0 universitet , End-use letter of Denmark It was able to identify abstract concepts, and geographic energy sector Additive motivation manufacturing Salary entities that ChatGPT-3.5 doesn’t. Extracted more elabo-Beneficiaries Mobility rules

rated entities with relations. Design 9,B.V, Netherlands Include Temporary works Artificial intelligence CO2

emissions

Include

Materialize NV, Sustainability Include Qlector, MSCA doctoral guidlines and Belgium Slovenia network rates criteria



and concepts, and Geographic Locations, and then identify the Include Katholieke Doctoral Universiteit Leuven, Include Jožef Stefan candidate (DC) relations between these entities. Providing additional information Institute, slovenia Determines the salary for Belgium researchers in MSCA



to GPT-4 resulted in an improved Knowledge graph (Figure 4). doctoral network Identifies critical areas for Internation reducing CO2 emission by energy agency However, ChatGPT-3.5 didn’t produce a quality graph (Figure 5) 2050 Will enhance Europe position in Engineering

compared to Figure 2. sciences



3.2 Python Implementation Figure 5: ChatGPT-3.5 was able to extract a larger number

We use a free, open-source library called spaCY 2 for advanced of entities but it was not successful at abstracting concepts

NLP in Python. We employ the named entity recognition tech- and missing relations. Entities and relations found fre-

nique to identify named entities from a given text using the spaCY quently represented complete sentences rather than con-

model (en-core-web-sm). We used a chunk of textual data from cepts. This occurs because ChatGPT is a conversational

the ADRIA Motorhome manual for experiment purposes. Table 2 model trained on a task to create responses to a given

compares entities, relations, and triplets extracted from the raw prompt and is not particularly trained to recognize en-

texts. The table shows that the number of triplets extracted by tities and relations

algorithms is similar–(Figure 6 and Figure 7). However, the num-



of entities is connected by meaningful relation, leading to fewer or provide additional context for better recognition. Hence re- ber of entities that spaCY extracts are larger but not every pair



triplets. Thus defeating the purpose of creating a Knowledge sults can be improved by pre-processing data into a structured format. Base. When using spaCy for entity extraction, the entities are

typically recognized based on the named entities present in the

text. Named entities are often specific nouns, such as names of 4 EVALUATION

people, organizations, locations, dates, or product names. spaCy When there is no ground truth data available, creating an auto-

might not identify it as a specific entity by default. So to extract mated evaluation metric for a Knowledge graph becomes chal-

specific entities, it might need to customize spaCy’s NER model lenging. In such cases, the evaluation relies on qualitative prin-

ciples to assess the results. Based on the practical framework

2https://spacy.io/models defined in the study [1], the following principles were identified:



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Khasa and Dunia, et al.



Table 2: Knowledge extraction comparison. (ADRIA mo-

Optical equipment Special

manuals Approvals torhome manual dataset)

Provides Original parts

Safety information Required for

information Vehicle data Optical Algorithm Entities Relations Triplets

General safety equipment Adhere to Included in GPT-4 Follow Used in equipment Standard 18 20 20



instructions Describes equip ChatGPT-3.5 24 18 18 Must read Provide/Update Describes Equipment ADRIA Vehicle ADRIA Vehicle state

Saftey chapter Recommended Users Performed on Operates spaCY 22 14 17

Utilize Carry in Service & for

Comply with Repair

Driving on Innovate/Develops Technical ADRIA). We analyzed that extracted entities are duplicated and re-Emergency

Subject to system

public roads equipment

Warranty Imposed on ADRIA design lations have some noise and incomplete information. If you have

obligations team specific patterns or structures in mind that you want to extract

entities and relations based on, you may need to customize the

Figure 6: The KG generated by GPT-4 contains 18 enti- relation extraction logic. Alternatively, more advanced natural

ties using the ADRIA motorhome instruction manual. It language processing techniques or pre-trained models designed

extracted concepts relevant to ADRIA users and vehicle for relation extraction tasks might provide better results. Also,

instructions, their functions, and how they are connected. we analyzed half of the relations-entities extracted by spaCY and

ChatGPT are overlapped.



Provides 5 CONCLUSION

User ADRIA home

Safety

accessories regulations The proposed exploration of using LLMs for Knowledge graph

tyre pressure Instruction

check & tighten extraction holds promise for advancing our understanding of

manuals

Optional

driving license has has must have equipment must comply with how advanced language models can contribute to structured

weight knowledge representation. This paper explores using LLMs to

alter dimension &

has generate Knowledge graphs out of source documents. We uti-

ADRIA Vehicle has Special

Vehicle ensure quality & readiness approvals



nameplates impact check & repair lized ChatGPT-3.5 and GPT-4 models to generate the Knowledge

contains Passengers Graphs for two different textual data and compared the structure Technical

operating system brake system pay attention to manual Service of the KGs. GPT-4 performed better as it successfully identified

gross weight Warranty Affects warranty work more abstract concepts and key entities compared to ChatGPT-

ADRIA design rating obligations 3.5. Therefore, it provides insights into the practical application team doesn't ADRIA dealer

tolerate provides

technical Service & Specialist of LLMs in developing structured knowledge from unstructured assist with

stanstill Repair workshop

textual data, with potential applications in knowledge-based AI

Figure 7: The KG generated by ChatGPT-3.5 contains 24 applications, paving the way for more effective information pro-

entities. Extracted more entities relevant to ADRIA vehi- cessing and utilization. In future studies, we intend to use a more

cles but relations between entities are more generic and formal framework to evaluate the quality of created Knowledge

entities are duplicated. graphs. Such a framework will allow us to efficiently analyze

the quality of KG and provide a standardized method to forecast

missing linkages between concepts and relationships within a

• Triplets should be concise. given domain.

• Contextual information of entities should be captured.

• The Knowledge graph does not contain redundant triples. ACKNOWLEDGEMENTS

• Entities should be densely connected. This research is supported by EU funding HE MSCA Project

• Relations among different types of entities should be in- Apriori (GA: 101073551). The author acknowledges the usage of

cluded. ChatGPT and Grammarly for content paraphrasing, grammar,

• Knowledge graphs should be organized in structured triples and error checking.

for easy processing by machine.

• For tasks specific to a particular domain, it’s essential REFERENCES

that the Knowledge graph is tailored and relevant to that [1] Haihua Chen, Gaohui Cao, Jiangping Chen, and Junhua Ding. 2019. A practi-specific field cal framework for evaluating the quality of knowledge graph. In Knowledge

According to these principles, in our use case, we manually in- Graph and Semantic Computing: Knowledge Computing and Language Under-

standing: 4th China Conference, CCKS 2019, Hangzhou, China, August 24–27,



spected the Knowledge graphs generated above, and we can con- 2019, Revised Selected Papers 4 . Springer, 111–122. [2] R OpenAI. 2023. Gpt-4 technical report. arxiv 2303.08774. View in Article, 2, clude that the ChatGPT-3.5 approach provides a more detailed 13. Knowledge graph without abstract concepts compared to the [3] Jeff Z Pan et al. 2023. Large language models and knowledge graphs: oppor-



GPT-4. However, to create these Knowledge graphs, a few steps tunities and challenges. arXiv preprint arXiv:2308.06374. [4] Shirui Pan, Linhao Luo, Yufei Wang, Chen Chen, Jiapu Wang, and Xindong of refining the answers from ChatGPT are needed. Sometimes Wu. 2024. Unifying large language models and knowledge graphs: a roadmap. the produced output is incorrect and needs to be corrected before IEEE Transactions on Knowledge and Data Engineering . [5] Elvis Saravia. 2022. Prompt engineering guide. (2022). proceeding. When we redefined the prompt, GPT-4 identified [6] Milena Trajanoska, Riste Stojanov, and Dimitar Trajanov. 2023. Enhancing more specific entities, and concepts compared to ChatGPT-3.5. knowledge graph construction using large language models. arXiv preprint

Even though ChatGPT extracted a larger number of entities, it arXiv:2305.04676.

failed to provide abstract concepts and entity-relation. [7] Ce Zhou et al. 2023. A comprehensive survey on pretrained foundation

models: a history from bert to chatgpt. arXiv preprint arXiv:2302.09419.

In the second part of the experiment, we employed the NER

method to extract relations and entities from the given text (i.e.



Solving hard optimization problems of packing, covering, and



tiling via clique search



S´ andor Szab´ o Bogd´ an Zav´ alnij

sszabo7@hotmail.com bogdan@renyi.hu

University of P´ ecs HUN-REN Alfred Renyi Institute of Mathematics

Pecs, Hungary Budapest, Hungary



Abstract numerically hard to solve problem of brick packing popular-

In the paper we propose to convert NP-hard combinato- ized by M. Gardner. We will focus on different approaches

rial optimization problems of packing, covering, and tiling of how to construct an auxiliary graph in order that to

types into maximum or translate this problems into a clique search problem. We 𝑘 -clique problems. The key step is

to come up with a tactically constructed auxiliary graph will try to investigate how these different approaches –

whose maximum or based on packing, covering and tiling– affect the solving 𝑘 -cliques correspond to the sought com-

binatorial structure. As an example, we will consider the time and if they have other consequences as well. First,

problem of packing a given cube by copies of a brick. The we describe the basic problem, then we present theoretical

aim of the paper is two fold to illustrate (i) the modeling discussion of different reformulations, and finally we de-

power and (ii) the feasibility of the clique approach. Since scribe the results of numerical experiments. The emphasis

theoretical tools are not readily available to study the effec- is on the modeling aspect of the computation and not on

tiveness of the solution of the resulting clique problems we reaching new records, as the proposed problem was solved

will carry out carefully conducted numerical experiments. in theoretical manner within months of its formulation.

Here we use it as a prototype of similar problems, and our



Keywords aim to show the versatility of our approach, that is model

a problem by a graph.

mathematical programming, 𝑘-clique problems, combina-

Graphs in this paper will be finite simple graphs. Further

torial optimization

all graphs we use will not have loops or double edges. A

finite simple graph 𝐺 can be described with its set of nodes

1 𝑉 and a subset 𝐸 of the Cartesian product 𝑉 × 𝑉 . The Introduction

subset 𝐸 can be identified by the set of edges of 𝐺.

One can see graphs as a mathematical models that can

Let 𝐺 = (𝑉, 𝐸) be a finite simple graph. A non-empty

describe various fields of interest. Like numbers, functions,

subset 𝐶 of 𝑉 is called a 𝑘-clique if each two distinct nodes

or Linear Programming graph based approach can model

of 𝐶 are adjacent in 𝐺 and in addition 𝐶 has exactly 𝑘

interesting problems and aid us in solving them. Some

elements. If 𝐶 has only one element, then we consider it a

of these approaches are quite straightforward like cliques

1-clique. The 2-cliques of 𝐺 are the edges of 𝐺. A 𝑘-clique

of people in a social interaction graphs or shortest path

𝐶 of 𝐺 is called a maximum clique if 𝐺 does not have

problem in a road map. Other approaches are less obvious

any (𝑘 + 1)-clique. For each finite simple graph 𝐺 there is

but still easily constructed, like conflict graphs in a set of

an integer 𝑘 such that 𝐺 contains a 𝑘-clique but 𝐺 does

codewords where a maximum independent set represents a

not contain any (𝑘 + 1)-clique. This well defined integer

maximum set of suitable error correcting codes [9].

𝑘 is called the clique number of 𝐺. We state two clique

But the approach of modeling and solving various prob-

problems formally.

lems by graphs are more versatile. Namely, we can see



graphs as a language for mathematical programming – if Problem 1. Given a finite simple graph 𝐺 and an inte-

certain combinatorial problems can be solved by construct-

ger 𝑘. Decide if 𝐺 has a 𝑘-clique.

ing a suitable auxiliary graph and finding a maximum or

𝑘-clique of this graph gives the solution. The authors have

Problem 2. Compute the clique number of a given finite

already used this approach in connection with mathemat-

simple graph.

ical conjectures [1], hyper graph coloring [11], subgraph

isomorphism [2], scheduling problems [12], graph coloring

Problem 1 is a decision problem, it is referred as the 𝑘-

problems [13] and protein docking problems in chemistry

clique problem, and it is an NP-complete problem included

[8].

in the original list of 21 NP-complete problems by Karp

Here we would like to give an example, where a hard [7]. Problem 2 is an optimization problem and referred as

combinatorial optimization problem can be solved by this

the maximum clique problem, and as the decision problem

approach. For this we chose a simple to understand but belongs to the NP-complete class it follows that it belongs

to the NP-hard class.

Permission to make digital or hard copies of all or part of this We color the nodes of a finite simple graph 𝐺 with the

work for personal or classroom use is granted without fee provided colors 1, 2, . . . , 𝑘 such that each node receives exactly one that copies are not made or distributed for profit or commercial

advantage and that copies bear this notice and the full citation on color and adjacent nodes never receive the same color. Such

the first page. Copyrights for third-party components of this work a coloring of the nodes of 𝐺 is called a well coloring, a

must be honored. For all other uses, contact the owner/author(s). proper coloring, or a legal coloring (the terminology is not

© 2024 Copyright held by the owner/author(s). unified). The set of nodes of 𝐺 receiving the color 𝑖 is called Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

https://doi.org/https://doi.org/10.70314/is.2024.sikdd.9 the 𝑖-th color class. Clearly, a color class is an independent



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Szab´ o et al.



set of 𝐺, that is, two nodes from a fixed color class are We state five problems related to packings, coverings,

never adjacent. and tilings in a formal manner. Given a finite set 𝑈 and

If the nodes of a finite simple graph can be legally colored its subsets (1).



The reason is that in this situation the nodes of Problem 4. Decide if 𝑈 has a 𝑘-packing using the mem- 𝐺 using 𝑘 colors, then we say that 𝐺 is a 𝑘-partite graph.

form a

bers of the family (1).

union of 𝑘 independent sets and these sets are pair-wise

disjoint. Problem 5. Decide if 𝑈 has a 𝑘-covering using the

In this paper we will focus on the following clique prob- members of the family (1).



lem. Problem 6. Decide if 𝑈 has a 𝑘-tiling using the mem-

Problem 3. Given a finite simple graph 𝐺 whose nodes bers of the family (1).

are legally colored using 𝑘 colors. Decide if 𝐺 has a 𝑘-clique. Problem 7. Compute the packing number of 𝑈 with

Problem 3 is a respect to the family (1). 𝑘-clique problem particularized to case



of 𝑘-partite graphs. This problem is still an NP-complete Compute the covering number of Problem 8. 𝑈 with

problem, as the graph coloring problem can be reduced to respect to the family (1).

such question as shown in [13], and should not be confused

with the problem of complete graphs. Problem 4 can be reduced to Problem 1. We construct a

The problem class we will be focusing on in the present finite simple graph 𝐺. The nodes of 𝐺 are the members of

paper consists of packing, covering, or tiling problems. the family (1). Two distinct nodes 𝐴𝑖 and 𝐴𝑗 are adjacent

Obviously many real world and mathematical problems in 𝐺 whenever 𝐴𝑖 and 𝐴𝑗 are disjoint. A 𝑘-clique in 𝐺

fall into this class, and here we would show some ideas how corresponds to a 𝑘-packing of 𝑈 .

such problems can be modeled by a suitably constructed Problem 5 can be reduced to Problem 3. We sketch the

auxiliary graph where a main points of this reduction. We construct a finite simple 𝑘 -clique search would solve the

original problem. graph 𝐺. The first type of nodes of 𝐺 are ordered pairs

(𝐵, 𝑥), where 𝐵 ∈ {𝐴1 , . . . , 𝐴𝑚}, 1 ≤ 𝑥 ≤ 𝑘. The intuitive

2 meaning of the pair (𝐵, 𝑥) that the subset 𝐵 is the 𝑥-th Packing, covering, and tiling

member of a 𝑘 element family of (1). To the node (𝐵, 𝑥)

First, we describe the problem class in question. Second,

we assign the color 𝑥. Two nodes receiving the same color

we draw up some basic concepts how these problems can

will be non-adjacent in 𝐺. Therefore the first type nodes

be modeled by graphs.

of 𝐺 are legally colored with 𝑘 colors.

Let 𝑈 be a finite ground set and let

We are adding second type nodes to 𝐺. Namely, we are

𝐴 1, . . . , 𝐴𝑚 (1) adding the ordered pairs (𝐴, 𝑢), where 𝐴 ∈ {𝐴1, . . . , 𝐴𝑚},

be subsets of 𝑢 ∈ 𝑈 and in addition 𝑢 ∈ 𝐴 holds. The intuitive meaning 𝑈 . A family of subsets

of the pair (𝐴, 𝑢) is that the element 𝑢 is covered by set

𝐵1 , . . . , 𝐵𝑛 (2) 𝐴. To the node (𝐴, 𝑢) we assign 𝑢 as a color. Two nodes

with {𝐵1, . . . , 𝐵𝑛} ⊆ {𝐴1, . . . , 𝐴𝑚} is called a packing of receiving the same color will not be adjacent in 𝐺. Thus

𝑈 the second type nodes of 𝐺 are legally colored using 𝑡 = |𝑈| if the members of the family (2) are pair-wise disjoint. A

family of subsets (2) is called a covering of colors. Now if we are locating a (𝑘 + 𝑡)-clique in 𝐺, then 𝑈 if the union of

(2) is equal to we select exactly 𝑘 subsets from (1) and each element of 𝑈 . Phrasing it differently, a family of subsets

(2) is a covering of 𝑈 will belong to at least one of these subsets. The missing 𝑈 if each element of 𝑈 belongs to at

least one member of the family (2). If a family of subsets part of the construction, what we left for the reader, is how

(2) is a packing and a covering of the first and second types of nodes are connected by edges. 𝑈 in the same time, then

it is called a tiling of Problem 6 can be reduced to Problem 3. As a tiling is 𝑈 . A tiling of 𝑈 some times referred

as exact covering of a packing and covering at the same time, we can add the 𝑈 .

A packing of packing restrictions, namely not connecting two sets if they 𝑈 is called a 𝑘 -packing if it consists of

𝑘 intersect, to the second type of nodes. On the other hand subsets of 𝑈 . Similarly, a covering of 𝑈 is called a 𝑘 -

covering if it consists of – in case of equal size sets –, we do not need to count the 𝑘 subsets of 𝑈 . Finally, a tiling

of used sets, so we won’t need the first type of nodes, they 𝑈 is called a 𝑘 -tiling if it consist of 𝑘 subsets of 𝑈 . For

a given ground set can be omitted. 𝑈 and for its given subsets (1) there

is an integer The computational difficulties of the 𝑘-packing, 𝑘-covering, 𝑘 such that 𝑈 has a 𝑘 -packing using subsets

of the family (1) but there is no any ( and 𝑘-tiling problems are different. It seems that the cov- 𝑘 + 1)-packing of 𝑈

using members of the family (1). This well defined integer ering problems are the computationally most demanding

𝑘 and the tiling problems are the most manageable. is the packing number of 𝑈 with respect to the family

(1). If the packing number of 𝑈 is equal to 𝑘, then each

𝑘 3 Gardner’s bricks problem-packing of 𝑈 is called maximum packing of 𝑈 .

For a given ground set 𝑈 and for its given subsets (1) We picked Gardner’s problem because it is intuitive and

there is an integer 𝑘 such that 𝑈 has a 𝑘-covering using easy to comprehend among such problems that can be

subsets of the family (1) but there is no any (𝑘 −1)-covering reduced to Problem 3 and so it serves as a good illustration

of 𝑈 using members of the family (1). This well defined of the kind of clique modeling we are dealing with. We do

integer 𝑘 is the covering number of 𝑈 with respect to the not claim any originality in connection with the problem.

family (1). If the covering number of 𝑈 is equal to 𝑘, then We do not prove any new results. Each of the facts we

each 𝑘-covering of 𝑈 is called minimum covering of 𝑈 . use are known from the folklore and we present them only



Solving hard problems via clique search Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



for the reader convenience. The problem was raised by Note that a fixed slab can contain only 0, 2 or 4 Proof.

Foregger in March 1975 [10], popularized by Gardner in unit cubes from any brick of the packing. The point is that

February 1976 [5], and solved by Foregger and Mather in the numbers 0, 2, 4 are all even. Each slab consists of an odd

November 1976 [3]. number of unit cubes. Therefore, each slab must contain an

Let us consider a brick 𝐵 of dimensions 1 × 2 × 4. The odd number of unpacked unit cubes. The number of slabs

brick 𝐵 is a union 8 unit cubes whose edges are parallel to is 7 and so each slabs must contain exactly one unpacked

the coordinate axis. From some reason unknown for us the unit cube. □ brick 𝐵 is referred as canonical brick. Suppose we have a

large supply of congruent copies of We can also form slabs by slicing 𝐶 with planes per- 𝐵 and we want to pack

as many as possible into a 7 pendicular to the second coordinate axes. Each of these × 7 × 7 cube 𝐶 . The cube 𝐶

is a union of 343 unit cubes. Let us divide 343 by 8 with 7 slabs contains exactly one unpacked unit cube. Finally,

remainder. As 343 = (42)(8) + (7), 43 copies of slicing 𝐶 by planes perpendicular to the third axes we get 𝐵 cannot

be packed into that each of these slabs contains exactly one unpacked unit 𝐶 . M. Gardener advanced the question if

42 copies of cubes. These constraints on the uncovered unit cubes are 𝐵 can be placed into 𝐶 . One can place a copy

of independent, but can also be checked independently during 𝐵 into 𝐶 in any possible rotated position as long the

edges of an extended search, and as such can reduce the search 𝐵 are parallel to the coordinate axis. (The answer

to this question is actually: No, one cannot place 42 bricks space well.

into a cube of size 7 × 7 × 7.)

Gardner’s problem can be expressed in terms of comput- 4 Numerical experiments

ing the clique number of a suitable constructed graph 𝐺. Gardner’s brick packing problem can be turned into various

In other words, Gardner’s problem can be reduced to an clique search problems and we carried out numerical exper-

instance of the maximum clique problem. Let us denote the iments with them. We will observe that the same geometric

set of the 343 unit cubes forming 𝐶 by 𝑈 . An 8 elements problem will lead to very different clique search problems.

subset 𝑣 of 𝑈 is a vertex of 𝐺 if the union of the elements When we try to pack 42 congruent copies of the canonical

of 𝑣 is a congruent copy of 𝐵. As it turns out 𝐺 has 1008 brick 𝐵 into the the big cube 𝐶, we get a 𝑘-clique problem.

nodes. Two distinct nodes ′ 𝑣 and 𝑣 of 𝐺 are adjacent in 𝐺 When we notice that the nodes of the auxiliary graph can

if ′ 𝑣 and 𝑣 are disjoint. If 𝐺 contains a (42)-clique, then be legally colored using 42 colors we get a 𝑘-clique prob-

42 congruent copies of 𝐵 can be packed into 𝐶. During lem in a 𝑘-partite graph which is a more tractable search

our numerical experiments a greedy coloring procedure problem. When we try to pack 42 congruent copies of the

provided a legal coloring of the nodes of 𝐺 using 42 colors. brick into the cube 𝐶 together with 7 unit cubes we get

Note that this is just a coincidence, it could’ve happened tiling problem. When we try to pack 42 congruent copies

otherwise. Thus we are facing with a particular case of of the brick into the cube 𝐶 together with 7 unit cubes

the 𝑘-clique problem stated in Problem 3. The nodes of 𝐺 and in addition we distinguish the unit cubes among each

are legally colored with 42 colors and we are looking for a other we get yet another version of the tiling problem.

(42)-clique in 𝐺. Phrasing it differently, we are looking for In the first approach the auxiliary graph 𝐺1 had 1008

a 𝑘-clique in a 𝑘-partite graph, where 𝑘 = 42. vertices. The nodes of 𝐺1 were legally colored using 42

We introduce a coordinate system whose origin coincides colors and we tried to locate a (42)-clique in 𝐺. Note, that

with a corner of the cube 𝐶. although this graph can be colored with 42 colors it was

just a coincidence. There is no theoretical background to



can be packed into Observation 1. If 42 congruent copies of the brick 𝐵 this fact. Of course the expectation was that 𝐺 1 do not 𝐶 , then there is such a packing which have any (42)-clique. contains the congruent copy of 𝐵 whose one corner is the

Let us assume that it is possible to pack 42 congruent

origin. Further the edges of lengths 1, 2, 4 are parallel to copies of the 1 × 2 × 4 canonical brick 𝐵 into the 7 × 7 × 7 the first, second and third coordinate axis, respectively.

cube 𝐶. By Observation 1, we may assume that a brick



Proof. As 343 = (42)(8) + (7) holds, 7 unit cubes of brick coincides with the origin of the coordinate system appear in the packing such that one of the corners of the

𝐶 are not contained by any bricks of the packing. The

and the edges of lengths 1, 2, 4 are along the 1-st, 2-nd,

cube 𝐶 has 8 corners and so at least one of the corners

3-rd coordinate axis. This information can be interpreted

must be contained by a brick. At this point we introduce a

such that there a (42)-clique 𝐶 2 in 𝐺1 which has a specific

coordinate system whose origin is this corner of 𝐶. Then

node. Namely, the vertex 𝑣1 of 𝐺1 that corresponds to the

we introduce the first, second, and third coordinate axis to



satisfy our requirement. □ restrict the graph special corner brick is a node of the of 𝐶2. This suggests to

𝐺 1 to the neighbors of the vertex 𝑣1 to

The cube get a new graph 𝐺2. Then we are looking for a (41)-clique 𝐶 can be sliced into 7 slabs using planes

perpendicular to the first coordinate axes. Each slab is a in 𝐺2. Plainly, the nodes of 𝐺2 are legally colored using 41

colors. This coloring is inherited from the coloring of the

1𝑥7𝑥7 slice of the big cube, that is a union of 49 unit cubes.

The centers of these cubes are in a plane perpendicular to nodes of 𝐺 1 . Since the graph 𝐺2 has fewer vertices than

𝐺1 (actually 960) and we are looking for a smaller clique in

the first coordinate axis. The 7 unit cubes of 𝐶, that are

not contained by any brick of the packing, are referred as 𝐺2 than in 𝐺1. The new clique problem probably requires

less computational effort because the graph is smaller, and

unpacked unit cubes.

because we introduced a symmetry breaking to it.

Observation 2. Two distinct uncovered unit cubes of The problem of packing 42 bricks into a bigger cube can

𝐶 cannot be in the same slab. be viewed as a tiling problem. Namely, we try to tile the



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Szab´ o et al.



7 × 7 × 7 cube 𝐶 by 42 copies of the canonical brick and The results presented here have interesting consequences

7 additional copies of a unit cube. Thus we are facing to and suggest further research problems. First, and as an-

a tiling problem using two different types of tiles and the ticipated, different auxiliary graphs lead to very different

number of the tiles is given. To ensure that we use 42 bricks search space sizes. And although the usual concept in our

we numerate the small cubes as {1, . . . , 7} and ensure in the research is that bigger graphs usually tend to be harder,

graph that each small unit cube is used once, that is we do that is not always the case. Remarkably, numerical results

not connect nodes where the unit square is covered by the indicate that the size of the auxiliary graph alone is not as

same small cube. This tiling problem can also be reduced important as the type of the reformulation. Namely, the

to a clique search problem. We denote the corresponding tiling type auxiliary graphs required less computational

graph 𝐺3. Tiling problems are more manageable compared effort for clique search even if they were not the small-

with packing problems as during the search back-tracking est graphs. Second, there are additional constraints that

can be anticipated earlier. However, the graph associated can be added to some reformulations while they seemingly

with the tiling in our case has more vertices than the graph cannot be incorporated into others. An example to such a

associated with the packing, namely it has 10 465 nodes. constraint is the fact described after the proof of Observa-

Therefore only computations can reveal which approach is tion 1. Namely, that no two distinct uncovered unit cubes

preferable. can appear in the same slab in Gardner’s brick packing

Obviously, in this case we can also fix a brick in the problem. That kind of restriction could be incorporated

corner. This version will be the 𝐺4 graph. into the tiling version of reformulation, and possibly not

In the last clique search equivalent of Gardner’s problem applicable to the packing reformulation. Taking advantage

we construct a graph 𝐺5. In this construction we handle a of the extra constraint made possible to solve the brick

mixed tiling problem but we utilize the extra information packing problem in reasonable time.

that no two distinct unit cube can appear in the same slab. There are other problems that can be solved using similar

By Observation 2, this may be assumed. This is done by approaches as detailed in the paper. Authors could solve

not connecting two nodes associated with unit cubes if smaller instances of the Golomb ruler problem or the Salem-

those unit cubes lay in the same slab. This graph is the Spencer set problem. The results, that lay outside the scope

same size as 𝐺3, as we only delete some edged from it. Also, of the present paper, obtained with those instances open

we can fix a brick in the corner in this case as well, that up even more interesting considerations.

shall be the 𝐺 6 graph.

Once again only numerical experiments can guide us in Acknowledgements

judging the merits of the possible clique search equivalents The present research was funded by National Research,

of the problems. Further, the preconditioning methods per- Development and Innovation Office – NKFIH Fund No.

form differently on the graphs 𝐺1, 𝐺2, 𝐺3, 𝐺4, 𝐺5 , 𝐺6 and SNN-135643.

this adds an extra layer of difficulty to the numerical work.

We used a computer with AMD EPYC 7643 processors, References

C++, and gcc v12.1 with settings-O3 -arch=znver3. [1] K. Corr´ adi and S. Szab´ o, A combinatorial approach for Keller’s

We made all six graphs and performed 𝑘-clique search conjecture. Period. Math. Hungar. Vol. 21, 91–100, 1990.

on them after preconditioning as described in [12, 13]. The [2] M. Depolli, S. Szab´ o and B. Zav´ alnij, An Improved Maximum

Common Induced Subgraph Solver. MATCH Commun. Math.

preconditioning run for 1-2 hours for the bigger graph, and Comput. Chem. 84 pp. 7–28. 2020.

reduced it by half, namely to around 6 000 nodes for [3] T. H. Foregger, and M. Mather, M. E2524. The American 𝐺 , 𝐺 ;

3 4

Mathematical Monthly. Vol. 83, No. 9 (Nov., 1976), pp. 741–

and to around 4 000 for 𝐺5, 𝐺6, that is the graphs where we 742

allow only one small cube in a slab. For the smaller graphs [4] D. Hespe, Ch. Schulz, D. Strash. Scalable Kernelization for

( Maximum Independent Sets. ACM Journal of Experimental 𝐺 , 𝐺 ) the preconditioner runs for a couple seconds but

1 2 Algorithmics. Volume 24, Article No.: 1.16, pp 1–22. 2019.

cannot significally reduce the graph. Three of the six graph [5] M. Gardner, MATHEMATICAL GAMES – Some elegant brick-

could be solved after preconditioning: packing problems, and a new order-7 perfect magic cube. Scien- 𝐺 2 , 𝐺 5 , and 𝐺 6 .

The solution time of tific American. Vol. 234, No. 2 (February 1976), pp. 122–127. 𝐺 (the original graph with fixed

2 [6] M. R. Garey, and D. S. Johnson, Computers and Intractability:

brick in the corner) was 50 days. The solution of 𝐺5 was a A Guide to the Theory of NP-completeness, Freeman, New

bit faster, 29 days. Finally, the graph York, 2003. 𝐺 could be solved

6

[7] R.M. Karp. “Reducibility Among Combinatorial Problems.” In:

more effectively. The running time was 123 484 seconds, Complexity of Computer Computations. New York: Plenum.

that is 34 hours. This clearly show us the importance of pp. 85–103. 1972.

the extra information of slabs. [8] K. Rozman, A. Ghysels, B. Zavalnij, T. Kunej, U. Bren, D.

Janeˇ ziˇ c, and J. Konc, Enhanced Molecular Docking: Novel Al-

gorithm for Identifying Highest Weight k-Cliques in Weighted

General and Protein-Ligand Graphs. JOURNAL OF MOLEC-



5 Conclusions ULAR STRUCTURE. 1304 p. 137639 Paper: 137639. 2024. [9] N. J. A. Sloane. Challenge Problems: Independent Sets in

We detailed several 𝑘-clique search reformulations of a Graphs. https://oeis.org/A265032/a265032.html

certain combinatorial problem in terms of constructing [10] T. H. Foregger. Elementary Problem E2524. The American

Mathematical Monthly. Vol. 82, No. 3 (Mar., 1975), p. 300.

suitable auxiliary graphs. We do not claim, that these [11] S. Szab´ o and B. Zav´ alnij, Reducing hyper graph coloring to

methods result more efficient practical computations than clique search. Discrete Applied Mathematics. 264. pp. 196–207.

other approaches. The point we are trying to make is that 2019.

[12] S. Szab´ o, and B. Zav´ alnij, Clique search in graphs of special

the clique reformulations open up a possibility to use well class and job shop scheduling. Mathematics. 10(5), 697. 2022.

tuned clique solvers, including preconditioning, to handle [13] S. Szab´ o, and B. Zav´ alnij, Graph Coloring via Clique Search

different combinatorial problems in a unified manner as a with Symmetry Breaking. SYMMETRY (BASEL). 14 : 8 Paper:

1574, 16 p. 2022.

general solver.



Zbornik 27. mednarodne multikonference



INFORMACIJSKA DRUŽBA – IS 2024

Zvezek D



Proceedings of the 27th International Multiconference



INFORMATION SOCIETY – IS 2024

Volume D



Miti in resnice o varovanju okolja



Myths and Truths About Environmental Protection



Uredniki / Editors



Rafael Mihalič, Tomaž Ogrin, Erik Margan



http://is.ijs.si



9. oktober 2024 / 9 October 2024

Ljubljana, Slovenia



Multikonferenca Infromacijska družba 2024



Konferenca o mitih in resnicah o varovanju okolja



Otvoritveni nagovor



Information Society Multyconference 2024



Conference on Myths and Truths of Environmental Protection



Opening address



V veliko zadovoljstvo in čast nam je, da vas lahko It is a great pleasure and honour to be able to

ponovno pozdravimo na letošnji konferenci o okolju welcome you all once again on behalf of the editorial

v okviru zdaj že tradicionalne Multikonference board of the Information Society Multiconference,

Informacijska družba. Zahvaljujemo vam, da ste se and to thank you for your submission of articles and

tako številčno odzvali na naš poziv z vašimi presentations at the Conference on Myths and Truth

prispevki in predstavitvami. of Environmental Protection.



This year's main theme is agriculture and forest

Tema letošnje konference je kmetijstvo in

management in the light of the latest European

gozdarstvo v luči nove evropske zakonodaje na teh

legislation on this subject. However, as this is just a

področjih. Pri tem pa gre za mnogo širše probleme v

segment of the much broader problem of both

povezavi z varstvom okolja in podnebnih sprememb,

environmental protection and climate change, we

zato je nujno vključiti v razpravo tudi poročila s teh

have included a number reports on related topics of

drugih področij, posebej energetike. Ravno tako se ni

energy supply and consumption, as well as some

mogoče izogniti razpravi o nekaterih teoretičnih

theoretical basis for assessing the necessity to

vidikih, ki so osnova ukrepov za preprečevanje in

mitigate climate change.

blaženje posledic podnebnih sprememb.



Climate change has become a fashion issue during

Podnebne spremembe so postale modni dodatek the last couple of decades, and the anthropogenic

razpravam, človeški vpliv pa je razglašen za influence on it has been declared the main cause and

poglavitni vzrok in krivca za vse probleme, ki so se a guilt for almost every aspect of problems that have

pojavili v zadnjih desetletjih. recently emerged in the modern society.



Med tem ko spremembe namembnosti zemljišč While land use change certainly affects local climate

zagotovo vplivajo na lokalne podnebne razmere, tega to some extent, the influence of greenhouse gases

pa z enako znanstveno strogostjo ni mogoče trditi za emissions on global climate has been determined to

vpliv ispustov toplogrednih plinov na globalni ravni. a much lower certainty than the scientific rigor

Med tem ko izpuste toplogrednih plinov spremljamo requires. The question of greenhouse gases has

z nesorazmerno veliko pozornostjo, resne okoljske received a disproportionate attention; the question of

škode povzročene z drugimi onesnažujočimi environmental damage by other pollutants has been

dejavniki komaj kdaj omenjamo. discussed only marginally.



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



Med tem ko je človekov vpliv na podnebje še vedno



do določene mere znanstveno vprašljiv, vsaj glede While the science of human influence on climate is

vzrokov ekstremnih vremenskih pojavov, je za still controversial to some extent, at least in regard to

večino medijev in politično elito znanost dorečena in its impact on the frequency and intensity of extreme

je čas za ukrepanje. Podnebna znanost je postala weather events, for the mass media and most of the

močno politizirana. political elite the science is settled and it is now time

for action. Climate science has become highly



Po besedah Rimskega kluba, vplivnega idejnega politicized.

podnebnega združenja ter svetovalnega telesa

Organizacije združenih narodov: In the words of the Club of Rome, a premier

environmental think-tank, consultants to United

"Skupni sovražnik človeštva je človek. Pri iskanju

Nations:

novega nasprotnika, ki naj bi nas poenotil, smo prišli

do spoznanja, da so onesnaževanje, ogrožajoče

"The common enemy of humanity is men. In

podnebne spremembe, pomanjkanje vode in hrane,

searching for a new enemy to unite us, we came up

ter življenja nasploh, ustrezni dejavniki nevarnosti,

with the idea that pollution, the threat of global

ki jih s svojim delovanjem povzroča človek. Le s

warming, water shortages, famine, and the life,

spremembami naravnanosti in obnašanja lahko te

would fit the bill that dangers are caused by human

nevarnosti prebrodimo. Resnični sovražnik je

intervention. It is only by changed attitudes and

človeštvo samo."

behaviours that they can be overcome. The real

enemy is humanity itself."

"Človeštvo je odprlo vrata pekla", je v svoji nedavni



konferenci za novinarje povedal generalni sekretar "Humanity Has Opened the Gates of Hell" is one of

Združenih narodov Antonio Guterres. the recent statements at a press conference by the UN

Secretary General Antonio Guterres.



Po drugi strani pa najnovejše Šesto poročilo IPCC v

poglavju 11 na strani 1520 oznanja: On the other hand, the latest IPCC Sixth Assessment

Report acknowledges in Chapter 11, page 1520:



"Pogostost prihodnjih sicer redkih, toda udarnih "The future occurrence of low-likelihood, high-

dogodkov povezanih s podnebnimi skrajnostmi je na impact events linked to climate extremes is generally

splošno povezana z nizko verjetnostjo, čeprav jih ni associated with low confidence, but cannot be

mogoče izključiti, zlasti ob globalnem segrevanju excluded, especially at global warming levels above

nad 4°C." 4°C."



In še na strani 1608: And on page 1608:



"In summary, the specific relationship between

"Če povzamemo; posamezne povezave med

changes in average and extreme conditions strongly

spremembami povprečnih in skrajnih pogojev so

depends on the variable or phenomenon being

močno odvisni od opazovanih dejavnikov ali

considered. At the local scale, average and extreme

pojavov. Na lokalni ravni so spremembe povprečnih



in skrjnih temperatur površja močno povezane, med while average and extreme precipitation changes are surface temperature changes are strongly related,



padavin pogosto šibka. Oboje, spremembe in average and extreme conditions vary strongly povprečnih in skrajnih pogojev, se močno razlikujejo across different places due to the effect of local and tem ko je povezava sprememb povprečnih in skrajnih often weakly related. For both variables, the changes

med različnimi območji zaradi učinkov lokalnih in

regional processes."

regionalnih procesov."



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



V zanjem času so podnebne spremembe povzročile Lately the question of climate change has become

tudi problem uveljavljana zakonov. Predlagane also an executive problem. Solutions suggested to

rešitve za omilitev domnevnih groženj pogosto mitigate the alleged problems often produce more

ustvarijo več novih problemov, rešujejo pa jih malo new problems, while solving little or nothing. And

ali nič. Predlagana zakonodaja pa pogosto povzroča the proposed legislation is often counterproductive,

ravno nasprotne učinke, kar je postalo očitno zlasti which has become evident mostly in its impact on

pri vplivih na avtomobilsko in drugo težko industrijo, automotive and other heavy industry, as well as on

kot tudi na poljedeljstvo in gospodarjenje z gozdovi. agriculture and forest management.



Probably the greatest problem is a complete lack of

Največji problem pa verjetno leži v popolnem

communication and dialogue between the

pomanjkanju komunikacije in dialoga med vladami

governments and the affected parties, and this needs

in prizadetimi deležniki. Zato je to prvo kar moramo

to change first, if we are to establish real solutions to

spremeniti, če želimo uveljaviti realne rešitve za

real problems.

realne probleme.



Hopefully this conference will be a step in that

Upamo, da bo ta konferenca korak v to smer.

direction.



Konferenčni uredniški odbor

Conference editorial board

Ljubljana, 9. Oktobra, 2024

Ljubljana, October 9th, 2024



PROGRAMSKI ODBOR / PROGRAMME COMMITTEE



Rafael Mihalič



Tomaž Ogrin



Erik Margan



LCA data conforms to Benford’s law



Podatki LCA se ujemajo z Benfordovim zakonom



Bogdan Šinik Aleksandar Tošić

bogdan.sinik@f amnit.upr.si aleksandar.tosic@upr.si

UP FAMNIT UP FAMNIT, InnoRenew CoE

Koper, Slovenia Koper, Slovenia



Abstract se uporabljajo v najpogosteje uporabljenih bazah podatkov LCA.

Naši rezultati preizkusa Ecoinventa, ene najpogosteje uporabljenih

Life cycle assessment (LCA) has been established as the standard

baz podatkov LCA, kažejo, da se podatki LCA močno ujemajo

method for evaluating environmental impacts of products, and

z Benfordovim zakonom. Poleg tega naša analiza vključuje še

processes. However, ISO standards depicting the application of

pet drugih javno dostopnih baz podatkov LCA, ki so prav tako

these standards are tailored more towards LCA practitioners and

skladne, z izjemo Bioenergiedata, kar je verjetno posledica majh-

less on the data acquisition and quality. The data acquisition

nega števila podatkov. Nazadnje smo preizkusili posamezne last-

process lacks robustness, with significant decision-making and

nosti, ki jih ponuja ecoinvent, in ugotovili, da zelo malo stolpcev

quality control entrusted to practitioners. Given the lack of in-

(manj kot 5 %), ki izpolnjujejo kriterije za Benfordovo analizo, ni

centives for industry participants to submit high-quality data,

skladnih. Čeprav so rezultati zanimivi, ti kličejo po podrobnejši

the integrity of LCA databases may be questionable. Moreover, in

analizi v prihodnosti.

some cases, participants may be incentivised to protect their data

from competition. To address these concerns, data is carefully

studied by external experts to verify their credibility. However, it Keywords

is not entirely clear how these experts are chosen, nor how trust LCA, Benford’s Law, Anomaly Detection

is established. In this paper, we apply a well known method Law

of anomalous numbers, commonly referred to as Benford’s law Keywords in order to test the conformity of commonly used LCA databases.

LCA, Benfordov zakon, odkrivanje anomalij

Our results on testing Ecoinvent, one of the most widely used



LCA databases, show that LCA data strongly conforms to Ben- 1 Introduction ford’s law. Moreover, our analysis includes 5 additional publicly

Life Cycle Assessment (LCA) is a systematic methodology for

available LCA databases, which also conformed with the excep-

evaluating the environmental impacts of a product, process, or

tion of Bioenergiedat, which is likely due to the low number of

service throughout its entire life cycle. This encompasses all

observations. Finally, we tested individual properties given by

stages from raw material extraction, production, use, and disposal

Ecoinvent and establish that very few columns (<5%), which pass

or recycling. The main goal of LCA is to identify opportunities

the criteria for Benford’s analysis, are non-conforming. Although

for improving environmental performance and making informed

interesting, these results call for a more fine-grained analysis as

decisions regarding sustainability.

future work.

(1) Goal and Scope Definition Establishing the purpose, bound-:

Abstract aries, and scope of the assessment.

(2) Inventory Analysis (LCI): Compiling an inventory of rel-

Življenjski cikel izdelka (LCA) se je uveljavil kot standardna

evant energy and material inputs and environmental re-

metoda za ocenjevanje okoljskih vplivov izdelkov in procesov.

leases.

Vendar pa so standardi ISO, ki določajo uporabo teh metod, bolj

(3) Impact Assessment (LCIA): Evaluating the potential envi-

prilagojeni strokovnjakom za LCA in manj usmerjeni v pridobi-

ronmental impacts associated with the inputs and releases

vanje kakovost podatkov. Proces pridobivanja podatkov ni zelo

identified in the inventory analysis.

robusten, saj je veliko odločitev in nadzora nad kakovostjo pre-

(4) Interpretation: Analyzing results to make informed de-

puščenih izvajalcem. Glede na pomanjkanje spodbud za udeležence

cisions, identify significant issues, and suggest improve-

iz industrije, da oddajo kakovostne podatke, je lahko integriteta

ments.

baz podatkov LCA vprašljiva. Poleg tega so udeleženci v neka-

terih primerih lahko motivirani, da zaščitijo svoje podatke pred Life Cycle Assessment (LCA) is used in various industries and

konkurenco. Da bi naslovili te pomisleke, podatke natančno sectors to promote sustainable practices, reduce environmental

preučijo zunanji strokovnjaki, da bi preverili njihovo verodosto- footprints, and support regulatory compliance. It is recognized as

jnost. Vendar ni povsem jasno, kako so ti strokovnjaki izbrani a crucial tool for environmental management and policy-making

in kako se vzpostavlja zaupanje. V tem prispevku uporabljamo [6]. Considerable effort has been made over the past decade in

dobro znano metodo “zakon anomalnih števil”, pogosto imeno- an attempt to alleviate early criticism about data quality and

vano Benfordov zakon, da bi preizkusili skladnost podatkov, ki trustworthiness [1]. Over the last few years many different tools

have been made to assess these problems, but there is still not

Permission to make digital or hard copies of all or part of this work for personal

standard that could be used on all LCA datasets [4]. It was shown

or classroom use is granted without fee provided that copies are not made or

that results of LCA for same processes were variable over time

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this which makes it hard to find statistical tool that could easily assess

work must be honored. For all other uses, contact the owner /author(s).

the trustworthiness and reliability of the data. [11] This is the

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

reason why we decided to use Benford’s law as a good indicator.

© 2024 Copyright held by the owner/author(s).

https://doi.org/https://doi.org/10.70314/is.2024.env.1 It is also known as first-digit law and is very often used for fraud



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Bogdan Šinik and Aleksandar Tošić



detection [8]. It was perfect choice for our research since it did not The distribution of digits is presented in Figure 1; the digit 1

require a lot of domain specific knowledge and is not influenced occurs in roughly 30 % of the cases, and the other digits follow in

by variability over time. For this purpose we have decided to a logarithmic curve. It has been shown that this result applies to

analyze Ecoinvent database as it has established itself as one of a wide variety of data sets [15], including electricity bills, street

the most used databases due to the amount of data, and tooling addresses, stock prices, house prices, population numbers, death

provided. More details about databases available in the literature rates, and lengths of rivers. The equation for the distribution of

[10, 17, 9]. the first digits of observed data is given in Equation 1.



2 Benford's Distribution Literature review

We have checked many sources that motivated us to check how

0.3

good Benford’s law would be for detecting inconsistencies in

data. It is estimated that less than 20% of papers related to LCA

conducted any kind of uncertainty analysis. [2]

0.2

Research conducted by Aalto University, Finland [16], com-

pared numbers from five different LCA databases to compare

amount of green house gases produced during the creation of ercentage

P

the buildings. The findings indicate that the databases exhibit 0.1

comparable patterns in the evaluation outcomes, with consistent

disparities in scale across the reference buildings observed across

all databases. Additionally, it was disclosed that there are sig-

0.0

nificant disparities in the numerical values across the databases 1 2 3 4 5 6 7 8 9

at some locations, and these disparities come from several data Digit

fragments.

Early research offered a comparative analysis of 26 sulfate pulp Figure 1: A graphical representation of the Benford’s dis-

mills in Sweden, focusing on the quantified emissions released tribution into the air and water between 1986 and 1993. The analysis re-

vealed significant variations in annual emission variables across

1

a group of companies. The emission parameters for most water 𝑃 𝑑 𝑙 𝑜𝑔 10 ( ) =(𝑑 + 1) − 𝑙𝑜𝑔10(𝑑) = 𝑙𝑜𝑔10 (1 + ) (1)

𝑑

emissions were not influenced by the annual production rates of

We utilized the benford.analysis package [5] in R for our re-

pulp. [11]

search. There were no instances of missing values denoted by

A more recent study statistical analysis of elemental flows in

"NA" in the form, although there were numerous occurrences of

commonly used LCA and LCIA databases and software was anal-

zeros. The Benford function automatically disregarded any zero

ysed [7]. The main conclusions of the study signaled considerable

values, eliminating the necessity for their removal. We needed to

shortcomings in flow clarity, consistency, and extensibility in el-

extract a subset of the data that includes only columns containing

ementary flows. However, no common method of identification

numerical values. The Benford function was modified to accom-

was proposed.

modate negative numbers as well, as the sign was irrelevant for

Moreover, the objectiveness of the evaluators responsible for

our specific investigation. The Ecoinvent dataset comprises 2654

data integrity has been highlighted as crucial to LCA database

columns, of which 2648 are numeric and were utilized in the

quality [13]. While this may be self-evident, a bigger concern

analysis. Within the 2648 columns, we have identified a total of

would be when objectiveness is not guaranteed.

1190 distinct chemical substances. These substances have been

categorized into five classes according to their place of release:

3 Methodology Air, Water, Soil, Natural Resources, and Inventory Indicator.

The first-digit law is an observation about the frequency distribu-

tion of leading digits. It is also known as the Newcomb–Benford 4 Results law or Benford’s law. It has been apparently first discovered by

We have demonstrated that a significant proportion of the columns

polymath Newcomb and published in [14] and later rediscov-

in the Ecoinvent database conform to Benford’s law, specifically

ered by physicist F. Benford and presented in [3]. The Benford’s

2193 out of 2648. The majority of non-conforming cases had a

law [15] defines a fixed probability distribution for leading digits

significant number of missing values, resulting in an insufficient

of any kind of numeric data with the following requirements:

amount of data for analysis. Only 70 columns met the criteria

• Data with values that are formed through a mathematical of having sufficient observations and not conforming. Figure 2

combination of numbers from several distributions. shows proportion of columns that conform and those that do not.

• Data that has a wide variety in the number of figures (e.g., We can see that our R package also divides conformity into four

data with plenty of values in the hundreds, thousands, levels: close conformity, acceptable conformity, and marginally

tens of thousands, etc.) acceptable conformity. For our research we have counted all

• Data set is fairly large, as a rule of a thumb at least 50 – levels of conformity as the same since we can’t expect data to

100 observations [12]. conform perfectly in real life.

• Data is right skewed (i.e., the mean is greater than the In order to show that LCA data generally conforms to Ben-

median), and the distribution has a long right-tail rather ford’s law, in addition to Ecoinvent, which underwent individual

than being symmetric. column testing, other databases were examined as a whole using

• Data has no predefined maximum or minimum value (with the Benford method. These databases are open source and avail-

the exception of a zero minimum). able online, and even though they are not as big and detailed as



LCA data conforms to Benford’s law Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



Database ChiSq ChiSqP MantissaArcTest MantissaArcTestP MADConformity MAD Number of Observations

Ecoinvent 247.684 0 0.033 0 Close conformity 0 49031793

Worldsteel 44.56 0 0.039 0 Close conformity 0.002 73044

OzLCI2019 63.034 0 0.036 0 Close conformity 0.002 178940

Greendelta 29.375 0 0.032 0 Close conformity 0.001 205344

Needs 150.271 0 0.034 0 Close conformity 0.001 808382

Bioenergiedat 29.626 0 0.053 0 Nonconformity 0.018 834

Table 1: Statistical indicators of Benford’s conformity for LCA databases studied.



Conformity of Ecoinvent columns 6 and categorizing items. Needs database was established by the

NEEDS (New Energy Externalities Developments for Sustainabil-

2000 ity) project and contains life cycle inventories of future energy

supplies in Europe. The dataset includes life cycle inventory (LCI)

information related to upcoming transportation services, power,

1500

Close conformity and material supply in the industrial sector. Bioenergiedat was

Acceptable conformity developed as part of the German BioEnergieDat project and was

1000 Marginally acceptable conformity finalized in February 2013. The objective of the project was to

Nonconformity establish supply chains for bioenergy alternatives, with a par-

Number of columns Not suitable data

ticular focus on the German context. The primary focus was

500

on bioenergy derived from wood and wastewood, wheat, and

biowaste.

0

Conforming Nonconforming

ecoinvent database worldsteel database

Conformity

30 30



Figure 2: Conformity of Ecoinvent columns

20 20



ercentage ercentage

P P

Ecoinvent, they also conformed. From the Figure 3 we can see

10 10

that all of these databases conformed with the testing, except

for one that was outdated and lacked sufficient data. By visual

0 0

inspection, the last database appears acceptable, however, from a 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9

Digit Digit

statistical standpoint, it does not meet the required criteria. Table

OzLCI2019 database Environmental Footprint database

1 shows all statistical values produced by our Benford function.

30 30

1

As aforementioned, the Bioenergiedat stands out as the only

non-conforming LCA database due to the low amount of observa-

tions (n=834), the credibility of this statistical test is questionable. 20 20

For other databases, all statistical conformity tests signal strong

ercentage ercentage

P P

conformity with the expected distribution, which results into a

10 10

Close conformity according to commonly used MAD conformity

test. All open source databases were found on OpenLCA Nexus

2 3 0 0

website . Worldsteel presents comprehensive worldwide and 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9

Digit Digit

regional Life Cycle Inventory (LCI) data for 16 different steel needs database bioenergiedat database products, ranging from hot rolled coil to plate, rebar, sections,

30

and coated steels. This study was conducted using the world-

30

steel LCI methodology report and ISO standards 14040 and 14044.

20

It is considered the most extensive and precise LCI dataset for

20

4

steel products worldwide. OzLCI2019 is a free LCA database

ercentage ercentage

P P

created by The Evah Institute in Australia. The database covers 10

10

the supply of goods from the Australasian area, including im-

5

ports, and was created using openLCA. Greendelta is focused on

0 0

providing secondary data for Product Environmental Footprints 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9

Digit Digit

(PEFs) within the openLCA software. The objective is to address



the environmental effects of products, such as carbon emissions, Figure 3: Comparison of first digit distribution between

by creating a standardized European approach for evaluating different databases

1

https://bioenergiedat.de/

2

https://nexus.openlca.org/

3

https://worldsteel.org/

4

http://www.evahnaturepositive.com/

5 6

https://www.greendelta.com/ https://nexus.openlca.org/database/NEEDS



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Bogdan Šinik and Aleksandar Tošić



5 Conclusion and future work (part i): overview and methodology. The International Journal of Life Cycle

Assessment, 21, 1218–1230.

Our analysis has determined that, generally, LCA data should

conform to Benford’s law. The results were consistent with ex-

pectations and demonstrated that this straightforward statistical

approach can be employed to evaluate first assessments of data

reliability. We anticipated that the entire LCA dataset would con-

form, while we acknowledged the possibility that certain minor

components would not. Non-conformity in the results does not

imply the inaccuracy or fabrication of the data; rather, it indicates

the necessity for expert verification. Establishing general confor-

mity is a stepping stone, as the method can be used to reliably test

for database anomalies quickly and efficiently. However, noncon-

formity would still require expert assessment to determine the

reasons. Typically, anomalies might be statistically undetectable

in large number of observations. This is commonly addressed

by sub-sampling the data. By testing individual parameters, we

have shown that for Ecoinvent, almost all are conforming (>95%).

Future work should include other sub-sampling techniques such

as geographic location, compartment (Air, Water, Soil, Natural Re-

sources, and Inventory Indicator), and others. These approaches

might provide a more granular view and possibly be able to detect

anomalies in individual sub-samples, thereby making subsequent

analysis easier and more accurate.



References

[1] Robert U. Ayres. 1995. Life cycle analysis: a critique. Resources, Conservation

and Recycling, 14, 3, 199–223. Life Cycle Management. doi: https://doi.org/1

0.1016/0921- 3449(95)00017- D.

[2] Nicole Bamber, Ian Turner, Vivek Arulnathan, Yang Li, Shiva Zargar Ershadi,

Alyssa Smart, and Nathan Pelletier. 2020. Comparing sources and analysis of

uncertainty in consequential and attributional life cycle assessment: review

of current practice and recommendations. The International Journal of Life

Cycle Assessment, 25, 168–180. doi: 10.1007/s11367- 019- 01663- 1.

[3] Frank Benford. 1938. The law of anomalous numbers. Proceedings of the

American philosophical society, 551–572.

[4] Anna E. Björklund. 2002. Survey of approaches to improve reliability in

lca. , 7, 2, 64–72. doi: The International Journal of Life Cycle Assessment

10.1007/BF02978849.

[5] Carlos Cinelli and Maintainer Carlos Cinelli. 2022. Package ‘benford. analy-

sis’. , Benford analysis for data validation and forensic analytics. Version 0.1

5.

[6] Mary Ann Curran. 2013. Life cycle assessment: a review of the methodology

and its application to sustainability. , Current Opinion in Chemical Engineering

2, 3, 273–277. Energy and environmental engineering / Reaction engineering

and catalysis. doi: https://doi.org/10.1016/j.coche.2013.02.002.

[7] Ashley Edelen, Wesley W Ingwersen, Cristina Rodríguez, Rodrigo AF Al-

varenga, Artur Ribeiro de Almeida, and Gregor Wernet. 2018. Critical review

of elementary flows in lca data. The international journal of life cycle assess-

ment, 23, 1261–1273.

[8] R. M. Fewster. 2009. A simple explanation of benford’s law. The American

Statistician, 63, 1, 26–32. eprint: https://doi.org/10.1198/tast.2009.0005. doi:

10.1198/tast.2009.0005.

[9] Rolf Frischknecht and Gerald Rebitzer. 2005. The ecoinvent database system:

a comprehensive web-based lca database. , 13, Journal of Cleaner Production

13-14, 1337–1343.

[10] Rolf Frischknecht et al. 2005. The ecoinvent database: overview and method-

ological framework (7 pp). , The international journal of life cycle assessment

10, 3–9.

[11] Ole Jørgen Hanssen and Odd Andreas Asbjørnsen. 1996. Statistical proper-

ties of emission data in life cycle assessments. , Journal of Cleaner Production

4, 3, 149–157. doi: https://doi.org/10.1016/S0959- 6526(96)00041- 8.

[12] David A Kenny. 2015. Measuring model fit. (2015).

[13] Alejandro Martínez-Rocamora, Jaime Solís-Guzmán, and Madelyn Marrero.

2016. Lca databases focused on construction materials: a review. Renewable

and Sustainable Energy Reviews, 58, 565–573.

[14] Simon Newcomb. 1881. Note on the Frequency of Use of the Different Digits

in Natural Numbers. , 4, 1-4, 39–40. doi: 10.2307/2369148. Amer. J. Math.

[15] Tommie W Singleton. 2011. IT Audit Basics: Understanding and Applying

Benford’s Law. , 3, 6. Isaca Journal

[16] Atsushi Takano, Stefan Winter, Mark Hughes, and Lauri Linkosalmi. 2014.

Comparison of life cycle assessment databases: a case study on building

assessment. , 79, 20–30. doi: https://doi.org/10.10 Building and Environment

16/j.buildenv.2014.04.025.

[17] Gregor Wernet, Christian Bauer, Bernhard Steubing, Jürgen Reinhard, Emilia

Moreno-Ruiz, and Bo Weidema. 2016. The ecoinvent database version 3





Liquid Neural Network in Modelling Rainfall-Runoff




Henok Teklu† Matjaz Gams Maciej Wielgosz

Applied Artificial Intelligence Applied Artificial Intelligence Applied Artificial Intelligence



Alma mater Eurapae Jamova 39, Slovania,Ljubljana Jozef stefan Alma mater Eurapae

Slovenska ulica 17, 2000



Slovenska ulica 17, 2000 matjaz.gams@ijs.si Maribor, Slovenia henok.teklu@almater.si Maribor, Slovenia wielgosz1@gmail.com



ABSTRACT tako kratkoročnih kot dolgoročnih napovedi, kar nakazuje na

njegov potencial pri izboljšanju hidroloških napovedi in

This study explores the use of Liquid Neural Networks (LNNs) upravljanju z vodnimi viri. to predict runoff for one, three, and six days ahead, highlighting



their superior performance compared to traditional models such KLJUČNE BESEDE



Long Short-Term Memory (LSTM) networks. LNNs leverage a Liquid Neural Networks, napovedovanje odtoka, hidrološko as Artificial Neural Networks (ANNs), Model Trees (MTs), and

dynamic reservoir of neurons, enabling them to capture complex napovedovanje, časovne odvisnosti, strojno učenje, padavinsko-

temporal dependencies inherent in the rainfall-runoff process. odtočno modeliranje

The study employs a case analysis of the Sieve River basin, using

historical hydrological data to train and evaluate the models. The

1 Introduction

results demonstrate that LNNs consistently outperform other

models across all prediction horizons, achieving the lowest Root Rainfall-runoff models are essential tools in hydrology used to

Mean Square Error (RMSE) and Normalized Root Mean Square simulate the transformation of rainfall into runoff, a process

Error (NRMSE) values, and the highest Coefficient of Efficiency critical for water resource management, flood forecasting, and

(COE). This indicates that LNNs are highly effective for both environmental protection. These models help understand and

short-term and long-term hydrological forecasting, offering predict how precipitation translates into river discharge, which is

significant potential for enhancing water resource management vital for designing infrastructure, managing water resources, and

and flood prediction strategies.1 mitigating the impacts of extreme weather events.



KEYWORDS Traditionally, rainfall-runoff models have employed a variety of



Liquid Neural Networks, Runoff Prediction, Hydrological algorithms ranging from empirical and conceptual models to more sophisticated data-driven techniques. Among the data- Forecasting, Temporal Dependencies, Machine Learning, driven approaches, Artificial Neural Networks (ANNs) and Rainfall-Runoff Modeling Model Trees (MTs) have been extensively used due to their

ability to capture complex nonlinear relationships between

POVZETEK

rainfall and runoff [1, 2]. ANNs, inspired by the human brain,

Ta študija raziskuje uporabo Liquid Neural Network (LNN) za consist of interconnected nodes (neurons) that process input data

napovedovanje odtoka vode (runoff) en, tri in šest dni vnaprej. to produce an output. They are highly effective in identifying

LNN je vrsta rekurentne nevronske mreže, ki ohranja dinamično patterns and making predictions based on historical data [3]. On

stanje nevronov, kar omogoča zajemanje kompleksnih časovnih the other hand, Model Trees are decision tree-based models that

vzorcev v podatkih. combine linear regression at the leaves, offering a more

interpretable approach while maintaining good predictive

Raziskava se osredotoča na bazen reke Sieve v Toskani, Italija, performance (Quinlan, 1992).

in vključuje podatke o padavinah, evapotranspiraciji in odtoku v

obdobju treh mesecev. Rezultati kažejo, da LNN dosledno In addition to ANNs and MTs, Long Short-Term Memory

presega tradicionalne metode, kot so Artificial Neural Networks (LSTM) networks and other recurrent neural networks (RNNs)

(ANN), Model Trees (MT) in Long Short-Term Memory have gained popularity for rainfall-runoff modeling. LSTMs are

(LSTM) omrežja, pri vseh napovedovalnih horizontih (1, 3, 6 a type of RNN specifically designed to capture long-term

dni). LNN se je izkazal za najučinkovitejšega pri obvladovanju dependencies in sequential data by addressing the vanishing



1 – Permission to make digital or hard copies of part or all of this work for Information Society 2024, 711 October 2024, Ljubljana, Slovenia personal or classroom use is granted without fee provided that copies are not

made or distributed for profit or commercial advantage and that copies bear © 2024 Copyright held by the owner/author(s). this notice and the full citation on the first page. Copyrights for third-party

components of this work must be honored. For all other uses, contact the

owner/author(s).



gradient problem common in traditional RNNs [4]. LSTMs and Temporal Processing: The temporal processing capability of

similar architectures like Gated Recurrent Units (GRUs) have LNNs allows them to capture and model the evolution of time-

shown promise in modeling time series data due to their ability series data more effectively than static neural networks.

to maintain and update a memory state over long sequences,

making them well-suited for hydrological forecasting [1]. Training: LNNs typically involve training only the output layer,

while the reservoir dynamics are left untrained but fixed, which

While these methods have proven effective for short-term and simplifies the learning process and allows for efficient handling

moderately long-term predictions, their performance tends to of temporal data [7].

degrade over even longer lead times. This limitation is partly due

to the static nature of these models, which may struggle to 3.3 Architecture of Liquid Neural Networks

capture the dynamic temporal dependencies inherent in the

Neurons: Dynamic Neurons: Neurons in LNNs have time-

rainfall-runoff process. Consequently, there is a growing interest

dependent states that evolve based on their interactions with

in exploring more advanced neural network architectures that can

other neurons in the reservoir.

better handle temporal data.



Synapses: Adaptive Synapses: Synapses in LNNs can adapt

based on the input data, allowing the network to learn temporal

2 Modelling

patterns.

This study extends the analysis by employing a Liquid Neural

Network (LNN) to predict runoff one, three, and six days ahead. Network Topologies: Reservoir Computing: LNNs employ a

LNNs are a type of recurrent neural network that maintain a fixed, randomly connected recurrent network (the reservoir) to

dynamic reservoir of internal states, allowing them to capture project input signals into a higher-dimensional space, facilitating

complex temporal patterns in data. Unlike traditional neural the capture of temporal patterns [5].

networks, which rely on static weights and activations, LNNs

leverage a constantly changing network of neurons, making them 3.4 Advantages of Liquid Neural Networks

particularly suitable for modeling temporal data.

Temporal Dynamics: LNNs naturally handle time-series data



The results of this study indicate that the LNN can effectively and dynamic processes, making them well-suited for tasks such as speech recognition, event detection, and time-dependent capture the temporal dependencies in the rainfall-runoff predictions [6]. transformation process. By leveraging the dynamic behavior of

neurons, the LNN can model short-term and medium-term

Energy Efficiency: Due to their dynamic nature, LNNs can be

dependencies and provide accurate predictions for runoff one

more energy-efficient than traditional neural networks, as they

day, three days, and even six days ahead. This promising

maintain a dynamic equilibrium rather than constantly

approach offers significant potential for improving hydrological

recalculating static weights.

forecasting over longer lead times, essential for effective water



resource management, flood prediction, and planning. Biological Plausibility: By mimicking the brain's dynamic

processing of information, LNNs provide insights into biological

neural processes and can be used to study and model neural

3 Liquid Neural Networks and Their

behavior.

Architecture

3.5 Challenges and Future Directions

3.1 Introduction to Liquid Neural Networks

Training Complexity: Training LNNs can be challenging due

Liquid Neural Networks (LNNs) represent an advanced to the complex dynamics of the reservoir. Researchers are

approach to recurrent neural network design. LNNs maintain a exploring various approaches to optimize training and improve

reservoir of dynamic states, enabling them to capture complex performance [7, 8]. temporal dependencies and patterns in data. This feature makes

LNNs particularly well-suited for tasks involving time- Computational Resources: Although LNNs are theoretically

dependent data and dynamic systems [5, 6]. efficient, simulating large-scale LNNs can be computationally

intensive. Advances in neuromorphic hardware aim to address

3.2 Key Concepts and Mechanisms these challenges by providing specialized hardware for efficient



Dynamic Reservoir: LNN simulation [9]. LNNs consist of a reservoir of

interconnected neurons with time-varying states. The reservoir's

dynamic nature allows it to process and retain information over

4 Case Study

varying time scales, making it effective for modeling temporal

dependencies. The study focused on the Sieve River basin, situated in the

Tuscany region of Italy, with a drainage area of 822 km². The

Sieve River, a tributary of the Arno River, extends for 56 km



through predominantly hilly and mountainous terrain. The 1. Data Loader: The training data is loaded in batches using

climate in this basin is temperate and humid. PyTorch's DataLoader to facilitate efficient training and

allow the use of GPUs for acceleration.

For this analysis, three months' worth of hourly data on discharge

(Q), precipitation (R), and evapotranspiration (E) were available, 2. Loss Function and Optimizer: Mean Squared Error

covering December 1959 to February 1960 and comprising 2,160 (MSE) is used as the loss function to measure the difference

data points. This dataset includes a variety of hydrological between the predicted and actual runoff values. The Adam

conditions, with flow rates spanning a wide range. optimizer is chosen for its efficiency in handling sparse

gradients and adaptive learning rate.



5 Result and Discussion 3. Training Loop: The model is trained over several epochs.



This document provides a detailed overview of the In each epoch, the model processes each batch of data,

implementation of a Fluid Neural Network (FNN) for predicting computes the loss, performs backpropagation to calculate



three, and six days ahead using historical data. The model optimizer. The average loss per epoch is monitored to track the training progress. leverages PyTorch for the neural network implementation and rainfall-runoff processes. The aim is to forecast runoff for one, gradients, and updates the model parameters using the

integrates a dynamic adjustment mechanism to optimize the

number of active units in the fluid cells, enhancing its ability to

handle varying input data complexity.

5.4 Model Evaluation and Predictions



5.1 Data Preparation After training, the model is evaluated on the test set to measure

its performance. The test loss is computed to assess how well the

The dataset includes various hydrological parameters such as

model generalizes to unseen data. Additionally, the trained

precipitation, evapotranspiration, and river discharge. The

model is used to make predictions, which are then compared to

features (inputs) and target variables (outputs) are extracted and

actual runoff values to validate the model's accuracy.

normalized to ensure that all variables contribute equally to the



to standardize the features by removing the mean and scaling to 5.5 Results model training process. StandardScaler from Scikit-Learn is used

unit variance. The Fluid Neural Network demonstrated superior performance

in predicting runoff compared to traditional models, particularly

5.2 Model Architecture for longer lead times. For the six-day prediction horizon, the

model showed significant improvements in accuracy, attributed

The core of this approach is the Fluid Neural Network, a variant

to its ability to dynamically adjust to the temporal dependencies

of the LSTM (Long Short-Term Memory) network designed to

in the data. Results are presented in Table 1.

dynamically adjust the number of active units based on the input

variance. This fluid behavior optimizes the model's performance,



especially for time-series data with varying temporal 6 Discussion and Conclusion dependencies.



1. 6.1 Discussion

Fluid LSTM Cell: This cell dynamically adjusts the

number of active units based on the variance of the input The comparison of the models' performance across different

data. High variance input activates more units, allowing the prediction horizons (Day 1, Day 3, and Day 6) reveals significant

model to capture complex patterns, while low variance insights into their efficiency and accuracy.

input activates fewer units to prevent overfitting and reduce

computational load. Artificial Neural Network (ANN):



2. Fluid LSTM Network: The network consists of multiple For short-term predictions (Q₊₁), ANN demonstrates reasonable

layers of Fluid LSTM cells. Each layer processes the accuracy with an RMSE of 5.175 m³/s, an NRMSE of 0.106 m³/s,

sequential data and passes the hidden states to the next layer. and a COE of 0.9886.

Dropout is applied to prevent overfitting.

However, as the prediction horizon extends to Q₊₃ and Q₊₆, the

3. performance of ANN deteriorates, with RMSE values increasing Fully Connected Layer : After processing through the

Fluid LSTM layers, the final hidden state is passed to a fully to 11.353 m³/s and 19.402 m³/s, respectively. Correspondingly,

connected (linear) layer that maps the high-dimensional COE values decrease, indicating reduced model efficiency.

output to the desired number of outputs, which in this case

are the runoff predictions for one, three, and six days ahead. M5 Model tree:



The MT model exhibits superior performance compared to ANN

5.3 Training the Model

for short-term predictions, with an RMSE of 3.612 m³/s, an

The training process involves the following steps: NRMSE of 0.074 m³/s, and a COE of 0.9944 for Q₊₁.



For Q₊₃, the RMSE rises to 12.548 m³/s, and the COE drops to Even for medium-term (Q₊₃) and long-term (Q₊₆) predictions,

0.9331, demonstrating a noticeable decline in performance for LSTM maintains relatively lower RMSE values of 10.500 m³/s

medium-term predictions. and 18.000 m³/s, and higher COE values of 0.9550 and 0.8600,

respectively.

For Q₊₆, the RMSE further increases to 21.547 m³/s, and the COE

falls to 0.8028, indicating significant prediction errors and Liquid Neural Network (LNN):

reduced model reliability for long-term predictions.

The LNN model consistently exhibits the best performance

Long Short-Term Memory (LSTM): across all metrics and prediction horizons. For Q₊₁, it achieves an

RMSE of 2.800 m³/s, an NRMSE of 0.058 m³/s, and a COE of

LSTM models outperform both ANN and MT across all 0.9960.

prediction horizons. For Q₊₁, the RMSE is 3.200 m³/s, the

NRMSE is 0.066 m³/s, and the COE is 0.9952.



Table 1: Results



Prediction ANN: MT: LSTM: LNN:



RMSE NRMSE RMSE NRMSE RMSE NRMSE RMSE NRMSE

COE COE COE COE

(m³/s) (m³/s) (m³/s) (m³/s) (m³/s) (m³/s) (m³/s) (m³/s)

Q₊₁ 5.175 0.106 0.9886 3.612 0.074 0.9944 3.2 0.066 0.9952 2.8 0.058 0.996

Q₊₃ 11.353 0.234 0.9452 12.548 0.258 0.9331 10.5 0.216 0.955 9.5 0.195 0.96

Q₊₆ 19.402 0.399 0.8401 21.547 0.443 0.8028 18 0.37 0.86 16 0.329 0.88



For Q₊₃, the RMSE and NRMSE are 9.500 m³/s and 0.195 m³/s, REFERENCES

respectively, with a COE of 0.9600, demonstrating its robustness [1] Zhang, Y., Vaze, J., & Chiew, F. H. S. (2018). Comparative study of in medium-term predictions. modeling approaches for predicting hydrological responses to climate

change. Water Resources Research, 54(1), 337-354.

For Q₊₆, LNN maintains its superior performance with an RMSE

of 16.000 m³/s, an NRMSE of 0.329 m³/s, and a COE of 0.8800, [2] Berndtsson, R., Bahremand, A., & Singh, V. P. (2019). Advances in

indicating its effectiveness even in long-term predictions. Hydrological Modeling: Application of Soft Computing Techniques.

Water, 11(5), 971.



6.2 Conclusion [3] Kratzert, F., Klotz, D., Herrnegger, M., Sampson, A. K., Hochreiter, S., &



The comparative analysis of the ANN, MT, LSTM, and LNN Nearing, G. S. (2019). Toward Improved Predictions in Ungauged Basins:

Exploiting the Power of Machine Learning. Water Resources Research,

models reveals that the Liquid Neural Network (LNN) 55(12), 11344-11354.

consistently outperforms the other models across all prediction

horizons (Day 1, Day 3, and Day 6). LNN achieves the lowest [4] Hochreiter, S., & Schmidhuber, J. (1997). Long Short-Term Memory.

RMSE and NRMSE values and the highest COE values, Neural Computation, 9(8), 1735-1780.

indicating its superior accuracy and efficiency. [5] Maass, W., Natschlager, T., & Markram, H. (2002). Real-time computing

without stable states: A new framework for neural computation based on

While ANN and MT models demonstrate acceptable perturbations. Neural Computation, 14(11), 2531-2560.

performance for short-term predictions, their accuracy

significantly declines for medium- and long-term predictions. On [6] Jaeger, H. (2021). Reservoir Computing: Model and Tool for Efficient

the other hand, LSTM models show better resilience and Time Series Prediction. Frontiers in Applied Mathematics and Statistics,

7, 50.

maintain relatively lower error rates and higher efficiency than

ANN and MT models. However, LNN models are the most [7] Lukosevicius, M., & Jaeger, H. (2009). Reservoir computing approaches

reliable, providing robust predictions with minimal errors and to recurrent neural network training. Computer Science Review, 3(3), 127-

high efficiency across all tested horizons. 149.



In summary, the LNN model's superior performance across all [8] Bellec, G., Scherr, F., Subramoney, A., Hajek, E., Salaj, D., Legenstein,

R., & Maass, W. (2018). Biologically inspired alternatives to

metrics and prediction horizons underscores its potential as the backpropagation through time for learning in recurrent neural nets. arXiv

most effective model for time series prediction tasks, particularly preprint arXiv:1805.08561.

in scenarios where both short-term and long-term accuracies are

critical. [9] Furber, S. B., Galluppi, F., Temple, S., & Plana, L. A. (2014). The

SpiNNaker project. Proceedings of the IEEE, 102(5), 652-665.



The governance of cultural landscapes in the EU:



A Comparative Analysis of the Nature Restoration Law



and the LULUCF Regulations



Richard J. Schenk



MCC Brussels

Brussels, Belgium



richard.schenk@brussels.mcc.hu



Abstract Land Use-Change and Forestry (LULUCF) Regulation. These two pieces of legislation are chosen due to their relevance for This paper explores how the EU's evolving climate regime cultural landscapes across the EU. The NRL can transform the complex impacts cultural landscapes, focusing on the Nature European landscape anew, and certainly in combination with Restoration Law (NRL) and the revised LULUCF Regulation. other measures [3]. The LULUCF Regulation not only impacts Using Multi-Level Governance (MLG) and regime complex landscapes and forests in its name, but it also plays a crucial role theory, it analyses how these policies, now legally binding, in the overall timetable of the EU in achieving carbon neutrality influence governance across multiple levels. Through case [4], [5]. In the case of the NRL, the study specifically refers to studies on the transformation of the Habitats Directive into the the experiences learnt from the older Habitats Directive in the NRL and the 2023 LULUCF revision, the study highlights the Netherlands. At the same time, the LULUCF is seen as more decisive role of NGOs in integrating climate, biodiversity, and general in the context of alpine and boreal areas since it is a land-use goals. The findings underscore the ongoing challenges newer piece of EU legislation. The analytical framework for the in balancing EU-wide directives with local adaptation needs. analysis is provided by multi-level governance (MLG) theory



Keywords to the two case studies respectively. The contribution ends with and regime complex theory. These two frameworks are applied

EU Climate Policy ; Nature Restoration Law (NRL) ; LULUCF a discussion of the findings.

Regulation : Cultural Landscapes : Multi-Level Governance

(MLG) ; Regime Complex Theory ; Environmental Governance

; NGO Influence ; Biodiversity Conservation ; Sustainable Land 2 The analytical framework: Combining multi-

Use level governance and regime complex theories



1 The study's analytical framework is divided into a macro and The change of EU climate policy instruments

level analysis. On the macro level, MLG deals with the

distribution of authority across different levels of government—

In recent decades, the EU moved from climate policies that are local, regional, national, and international—where multiple

more ’symbolic’ to legally binding targets. At the European level actors and institutions interact to shape policy outcomes. By

at a range of institutions has developed to deal with climate applying MLG, the study explores how the interplay between

policies. Today, these institutions oversee a large toolbox of various governance levels and overlapping regulatory regimes

policy instruments [1]. Knodt even identified three modes of shapes the two case studies [6], [7]. On the meso level, regime

European climate governance, ranging from hard to softer complexes refer to the loosely connected sets of institutions and

governance models [2]. norms that govern overlapping issue areas in global governance.

Changes of this magnitude to EU climate policies naturally Unlike a single, cohesive regime, regime complexes are

impact cultural landscapes in the Union. This contribution sheds characterised by their fragmented and sometimes conflicting

light on how the changes to EU governance shaped two EU regulations. The concept of regime complexes allows for the

policies: the Nature Restoration Law (NRL) and the Land Use, better analysis of the role of NGOs as they advocate for specific

norms, providing expertise, and mobilizing public opinion, often

Permission to make digital or hard copies of part or all of this work for personal or bridging gaps between different governance levels and regimes

classroom use is granted without fee provided that copies are not made or distributed [8], [9]. This look at the meso level is necessary as environmental

for profit or commercial advantage and that copies bear this notice and the full and climate policies display a significant amount of NGO activity

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s). [10].

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.env.3

3 Case Study I: The evolution of the EU

Habitats Directive into the NRL

3.1. Multi-Level Governance offered by the Habitats Directive. The development of the NRL

reflects an MLG approach, where the European Commission,

The Habitats Directive (92/43/EEC), adopted in 1992, is a

influenced by member state experiences, proposed a more

cornerstone of EU biodiversity policy, aiming to protect natural

comprehensive framework. This framework integrates

habitats and species across Europe. It created the Natura 2000

biodiversity restoration into broader land use and agricultural

network, which includes protected areas that member states must

policies, addressing the gaps identified in the previous directive

manage to conserve habitats and species of European

[15].

importance. The Directive operates within an MLG framework,

requiring coordination between the European Commission,

3.2. Regime Complexes

national governments, and local authorities. This necessitates

collaboration across multiple levels of governance to ensure that

conservation objectives are met while considering national and The regime complex surrounding the Habitats Directive is

local contexts. The Habitats Directive is, in this regard, the composed of various overlapping and interconnected regimes,

extension of the global climate policy framework into EU including EU environmental policies (e.g., Habitats and Birds

environmental and climate policies [11]. The Directive operates Directives), agricultural policies (e.g., Common Agricultural

within an MLG framework, requiring coordination between the Policy), legal frameworks, and international biodiversity

European Commission, national governments, and local agreements (e.g., Convention on Biological Diversity).

authorities. Over time, it became evident that the Habitats In the Netherlands, NGOs were instrumental in bringing

Directive, while foundational, had limitations in addressing attention to the nitrogen crisis, particularly its impact on

emerging environmental challenges. These included habitat protected habitats under the Habitats Directive. Organisations

degradation outside Natura 2000 sites and the lack of integration such as Milieudefensie (Friends of the Earth Netherlands) and

with broader land use and agricultural policies. the Dutch Association for Nature Conservation

(Natuurmonumenten) used legal avenues to challenge the Dutch

government's failure to adequately address nitrogen pollution,



leading to significant court rulings that halted construction

projects and forced policy reconsiderations [16], [17], [18]. In

response, the Dutch government revealed a plan to mitigate the

impact of the consequences of the nitrogen crisis with the help of

24 billion EUR [19].

The persistent advocacy and legal actions by NGOs in the

Netherlands were critical in demonstrating the limitations of the

Habitats Directive in addressing complex, cross-sectoral

environmental issues like nitrogen pollution. This helped build

the case for more comprehensive legislation at the EU level. The

Dutch experience provided a clear example how the NGOs could

amplify the issues arising from the Habitats Directive, create

public pressure and cause huge ramification for government and

society. The leading to the proposal of the NRL. It sets ambitious



Figure 1: Conservation Status of habitats at Member quantitative targets in terms of both the areas to restore and the

State level, 2013-18 [12] timeframe [20].



As the Figure 1 highlights, particular issues arose in the Low

Countries. Especially the Netherlands faced significant issues

4 Case Study II: The impact of the EU Green

with nitrogen emissions, largely due to intensive agriculture,

which led to widespread habitat degradation. The crisis Deal on the LULUCF Regulation

highlighted the inadequacies of the existing framework, where

4.1. Multi-Level Governance

national policies were often misaligned with the original

directive [13]. The Netherlands struggled to reconcile its The LULUCF Regulation is primarily an EU-level policy

intensive agricultural sector with the requirements of the Habitats designed to align with the EU's overall climate targets, including

Directive, leading to legal challenges and a halt in construction the European Green Deal and the Fit for 55 package, which aims

projects. In response, the Dutch government had to devise a to reduce net greenhouse gas emissions by at least 55% by 2030

comprehensive compared to 1990. Member states are responsible for nitrogen reduction strategy involving

collaboration across multiple governance levels, from local implementing the LULUCF Regulation. This requires each

municipalities to the national government and EU institutions country to develop national forestry and land-use policies that

[14]. align with the EU’s overarching goals. The 2023 revision

The NRL, proposed by the European Commission in 2022, introduced more stringent national accounting requirements,

can be seen as an evolution of the Habitats Directive, driven by including higher carbon removal targets. The implementation at

the lessons learned from crises like that in the Netherlands. This the regional and local levels involves managing forests,

new law aims to restore degraded ecosystems across the EU, with agricultural lands, and wetlands to maximise carbon

binding targets for habitat restoration beyond the protections sequestration while balancing economic and environmental

objectives. Local authorities and landowners are crucial in for carbon sequestration and broader ecosystem services,

executing these policies, requiring close coordination with including biodiversity and water regulation. This norm has been

national governments. increasingly adopted within the LULUCF framework,

In 2023, the LULUCF Regulation underwent a major revision particularly in the 2023 revision. The organizations lobbied for

to align it with the Green Deal targets. The new LULUCF more stringent monitoring, reporting, and verification (MRV)

Regulation introduced rigorous monitoring, reporting, and processes to ensure that member states' contributions to carbon

verification requirements (MRV). Member states must regularly sequestration are transparent and verifiable. NGOs also pushed

report on their progress toward the targets, and the European for accountability measures to address non-compliance, which

Commission monitors compliance. It set binding targets for each were incorporated into the revised regulation. NGOs are crucial

member state regarding carbon removals in the LULUCF sector, in monitoring the implementation of the LULUCF Regulation

and to ensure compliance, includes provisions for penalties and across the EU. Acting as watchdogs ensures that member states

corrective actions if member states fail to meet their targets. [4], adhere to their commitments and that the interactions between

[5]. This creates a feedback loop where national actions are different regimes remain aligned with overarching

evaluated at the EU level. This marked a major shift in the environmental goals. [22].

vertical integration of LULUCF policies in the EU, including By acting across multiple regimes—climate, biodiversity, and

strengthening the compliance mechanism under the LULUCF agriculture—NGOs have helped to transform the LULUCF

Regulation by imposing a financial penalty of EUR 375 on Regulation into a more comprehensive and coherent policy

Member States for each ton of CO2eq more than the annual instrument.

national target in the given year after 2026.[21].





Figure 3: Comparison of different LULUCF Projections [23]





Figure 2: The newly introduced LULUCF compliance 5 Conclusion



mechanism [21] The evolution of EU climate policy has shifted from symbolic

measures to robust, legally binding targets, exemplified by the

4.2. Regime Complexes

NRL and the 2023 revision of the LULUCF Regulation. These

The LULUCF Regulation is situated at the intersection of several policies illustrate how the EU has strengthened its governance

major international and EU regimes, including climate change framework to address climate change, biodiversity, and land use

(e.g., the Paris Agreement), biodiversity (e.g., the Convention on challenges.

Biological Diversity), and agricultural policy (e.g., the Common Using Multi-Level Governance (MLG) and regime complex

Agricultural Policy - CAP). Each of these regimes has its theories, this study highlights the critical role of coordinated

objectives, norms, and governance structures, which sometimes governance across multiple levels and the influence of NGOs.

overlap or conflict. The NRL, shaped by the Dutch nitrogen crisis and the LULUCF

The 2023 revision of the LULUCF Regulation reflects efforts Regulation, aligned with the European Green Deal, demonstrate

to harmonise these overlapping regimes. It seeks to integrate how integrated approaches and NGO advocacy have transformed

carbon sequestration goals with biodiversity conservation and these policies into powerful tools for achieving EU climate and

sustainable land use, addressing the synergies and tensions ecological goals. Yet, this transformation inherently curbed the

between these regimes. NGOs have been instrumental in shaping ability of member states and local governments to shape their

the LULUCF Regulation by advocating for stronger links environmental and climate policies. Meanwhile, meta-studies

between climate and biodiversity regimes. NGOs often act as highlight the necessity of adapting climate and environmental

bridges between regimes, facilitating communication and policies to local geography, regional preferences, and the needs

coordination among different governance structures. of local communities [24].

NGOs acted as norm entrepreneurs during the revision

process by promoting the idea that land use should be managed References



T. Rayner, K. Szulecki, A. J. Jordan, and S. Oberthür, ‘The global importance [13] B. H. Jacobsen, U. Latacz-Lohmann, H. Luesink, R. Michels, and L.

of EU climate policy: an introduction’, in Handbook on European Union Climate Ståhl, ‘Costs of regulating ammonia emissions from livestock farms near Natura

Change Policy and Politics, T. Rayner, K. Szulecki, A. J. Jordan, and S. Oberthür, 2000 areas - analyses of case farms from Germany, Netherlands and Denmark’, J.

Eds., Edward Elgar Publishing, 2023, pp. 1–23. doi: 10.4337/9781789906981. Environ. Manage., vol. 246, pp. 897–908, Sep. 2019, doi:

[2] M. Knodt, ‘Instruments and modes of governance in EU climate and 10.1016/j.jenvman.2019.05.106.

energy policy: from energy union to the European Green Deal’, in Handbook on [14] A. M. Collet, Dissecting misalignment in agri-environmental policy European Union Climate Change Policy and Politics, T. Rayner, K. Szulecki, A. J. design : The Case of the PAS and the Dutch Nitrogen Crisis. Enschede: University Jordan, and S. Oberthür, Eds., Edward Elgar Publishing, 2023. doi: of Twente, 2023. [Online]. Available: http://essay.utwente.nl/96165/ 10.4337/9781789906981. [15] A. L. R. Jackson, ‘Conservation of areas outside the Natura 2000

[3] A. L. R. Jackson, ‘Introduction’, in Nature Law and Policy in Europe, network: An historical perspective with an eye to the future’, in Nature Law and

1st ed., A. L. R. Jackson, Ed., London: Routledge, 2023, pp. 1–9. doi: Policy in Europe, 1st ed., A. L. R. Jackson, Ed., London: Routledge, 2023, pp. 214– 10.4324/9780429299100. 251. doi: 10.4324/9780429299100.

[4] S. Romppanen, ‘The LULUCF Regulation: the new role of land and [16] C. Backes, ‘“De stikstofcrisis in Nederland is een uniek fenomeen”’,

forests in the EU climate and policy framework’, J. Energy Nat. Resour. Law, vol. Feb. 08, 2023. [Binnenlands Bestuur]. Available: 38, no. 3, pp. 261–287, Jul. 2020, doi: 10.1080/02646811.2020.1756622. https://www.binnenlandsbestuur.nl/ruimte-en-milieu/de-stikstofcrisis-nederland-

[5] S. Romppanen, ‘Targets, timetables and effort sharing as governance een-uniek-fenomeen

tools: emergence, scope and ambition’, in Handbook on European Union Climate [17] A. Jaspers, De stikstoffuik: politici in de ban van de ecolobby. Change Policy and Politics, T. Rayner, K. Szulecki, A. J. Jordan, and S. Oberthür, Amsterdam: Uitgeverij Blauwburgwal, 2023. Eds., Edward Elgar Publishing, 2023, pp. 216–231. doi: 10.4337/9781789906981. [18] K. Mathiesen, ‘The chemist vs. the Dutch farmers’, Politico.eu, Mar.

[6] A. Benz, ‘Multilevel Governance’, in Handbuch Governance, A. Benz, 09, 2023. Accessed: Nov. 11, 2023. [Online]. Available:

S. Lütz, U. Schimank, and G. Simonis, Eds., Wiesbaden: VS Verlag für https://www.politico.eu/article/johan-vollenbroek-netherlands-nitrogen-pollution-

Sozialwissenschaften, 2007, pp. 297–310. doi: 10.1007/978-3-531-90407-8_22. climate-change-farming/

[7] M. Knodt and M. G. Hüttmann, ‘Der Multi-Level Governance-Ansatz’, [19] C. Vallet, ‘In the Netherlands, a drastic plan to reduce nitrogen

in Theorien der Europäischen Integration, H.-J. Bieling and M. Lerch, Eds., emissions angers farmers’. Le Monde, Jul. 14, 2022. [Online]. Available:

Wiesbaden: VS Verlag für Sozialwissenschaften, 2005, pp. 223–247. doi: https://www.lemonde.fr/en/environment/article/2022/07/14/in-the-netherlands-a-

10.1007/978-3-663-11642-4_9. drastic-plan-to-reduce-nitrogen-provokes-farmers-anger_5990080_114.html

[8] L. Gómez-Mera, J.-F. Morin, and T. Van De Graaf, ‘Regime [20] D. Hering et al., ‘Securing success for the Nature Restoration Law’,

Complexes’, in Architectures of Earth System Governance, 1st ed., F. Biermann Science, vol. 382, no. 6676, pp. 1248–1250, Dec. 2023, doi: and R. E. Kim, Eds., Cambridge University Press, 2020, pp. 137–157. doi: 10.1126/science.adk1658. 10.1017/9781108784641.007. [21] H. Böttcher, S. Gores, K. Hennenberg, J. Reise, and A. Graf, ‘Analysis

[9] R. O. Keohane and D. G. Victor, ‘The Regime Complex for Climate of the European Commission proposal for revising the EU LULUCF Regulation’.

Change’, Perspect. Polit., vol. 9, no. 1, pp. 7–23, Mar. 2011, doi: Öko-Institut e.V., Jan. 27, 2022. [Online]. Available: 10.1017/S1537592710004068. https://www.oeko.de/fileadmin/oekodoc/Assumptions_LULUCF_Proposal.pdf

[10] J. J. Schoenefeld, ‘Interest Groups, NGOs or Civil Society [22] A. Savaresi, L. Perugini, and M. V. Chiriacò, ‘Making sense of the

Organisations? The Framing of Non-State Actors in the EU’, Volunt. Int. J. Volunt. LULUCF Regulation: Much ado about nothing?’, Rev. Eur. Comp. Int. Environ. Nonprofit Organ., vol. 32, no. 3, pp. 585–596, Jun. 2021, doi: 10.1007/s11266-020- Law, vol. 29, no. 2, pp. 212–220, Jul. 2020, doi: 10.1111/reel.12332. 00283-w. [23] European Commission, ‘COMMISSION STAFF WORKING

[11] A. Trouwborst, ‘Conserving European Biodiversity in a Changing DOCUMENT IMPACT ASSESSMENT REPORT Accompanying the document

Climate: The Bern Convention, the European Union Birds and Habitats Directives Proposal for a Regulation of the European Parliament and the Council amending

and the Adaptation of Nature to Climate Change’, Rev. Eur. Community Int. Regulations (EU) 2018/841 as regards the scope, simplifying the compliance rules,

Environ. Law, vol. 20, no. 1, pp. 62–77, Apr. 2011, doi: 10.1111/j.1467- setting out the targets of the Member States for 2030 and committing to the

9388.2011.00700.x. collective achievement of climate neutrality by 2035 in the land use, forestry and

[12] European Enviroment Agency, ‘Conservation status of habitats under agriculture sector, and (EU) 2018/1999 as regards improvement in monitoring,

the EU Habitats Directive’. Nov. 18, 2021. [Online]. Available: reporting, tracking of progress and review’. Jul. 17, 2021. [Online]. Available:

https://www.eea.europa.eu/en/analysis/indicators/conservation-status-of-habitats- https://eur-lex.europa.eu/legal-content/EN/ALL/?uri=CELEX:52021SC0609

under?activeAccordion=ecdb3bcf-bbe9-4978-b5cf-0b136399d9f8



Resne napake v oceni toplogrednega učinka ozračja



Serious Errors in the Assessment of the Atmospheric



Greenhouse Effect



Erik Margan

Oddelek za eksperimentalno fiziko osnovnih delcev

Institut Jožef Stefan

Ljubljana, Slovenija





POVZETEK mainly by greenhouse gases emissions. Following the expected

changes the IPCC prepares recommendations to governments to

Toplogredni učinek ozračja je osnovni proces na katerem mitigate the climate changes and reduce their consequences, as

Medvladni odbor za podnebne spremembe (Intergovernmental well as adapt their legislation to implement those measures in



Reports, AR) analizira izhodiščno stanje in napoveduje countries members of the United Nations Organization. The pričakovane podnebne spremembe, ter njihove posledice. Po methodology used by the IPCC for their analyses and predictions Panel for Climate Change, IPCC) v svojih poročilih (Assessment accordance with the Paris Agreement signed by almost all



zadnjih 150 letih pretežno posledica človekovih dejavnosti, errors in the way the greenhouse effect has been presented in the predvsem izpustov toplogrednih plinov v ozračje. Na osnovi IPCC documents, which rises a strong doubt in the correctness njihovih ugotovitvah je segrevanje ozračja, ki ga beležimo v contains a number of errors. Here we show only some serious

pričakovanih sprememb IPCC pripravlja priporočila vladam s of their analysis. Because of those errors the human influence on

katerimi ukrepi bi omilili podnebne spremembe in blažili njihove climate is overestimated, as are their expected consequences,

posledice, ter ustrezno prilagodili zakonodajo za uveljavljanje in therefore the relevance and the necessity to mitigate them is

izvajanje teh ukrepov, čemur so se s svojimi podpisi Pariškega reduced, as well as the urgency to adopt the proposed restrictive

sporazuma zavezale skoraj vse države članice Organizacije legislation. združenih narodov. Metodologija, ki jo uporablja IPCC za svoje



analize in napovedi vsebuje vrsto napak. V tem prispevku se KEYWORDS bomo omejili na nekaj resnih napak v oceni toplogrednega

učinka v dokumentih IPCC, kar vzbuja močan dvom v pravilnost Atmospheric greenhouse effect, climate change, Stefan-

njihovih ugotovitev. Zaradi tega je človeški vpliv na podnebne Boltzmann law, radiative balance, average global air

spremembe in njihove pričakovane posledice močno precenjen, temperature, concentration of greenhouse gases.

kar posledično zmanjšuje nujnost ukrepanja in sprejetja raznih



predlaganih omejevalnih zakonov. 1 Uvod v fizikalno ozadje



KLJUČNE BESEDE Termodinamika je področje fizike, ki opisuje procese prenosa



Toplogredni učinek ozračja, podnebne spremembe, Stefan- toplote med obravnavanimi telesi in njeni zakoni veljajo tako na

Boltzmannov zakon, sevalno ravnovesje, povprečna globalna molekularni kot na makroskopski ravni. Toplotna energija se

temperatura ozračja, vsebnost toplogrednih plinov. manifestira v elektromagnetih interakcijah med obravnavanimi

telesi, kar pomeni, da obstajajo tesne povezave med notranjo

ABSTRACT energijo teles (v obliki kinetične energije, ki se izmenjuje ob

trkih, ali pa kot lastne oscilacije in resonance zaradi vezalne

The atmospheric greenhouse effect is the basis on which the

energije med posameznimi atomi), ter absorbiranim ali

Intergovernmental Panel for Climate Change (IPCC) in their

emitiranim elektromagnetnim sevanjem, katerega energija

Assessment Reports analyse the current climate and predict the

(oziroma pripadajoča frekvenca) mora ustrezati določenemu

expected climate changes and their consequences. In their

molekularnemu rotacijsko-vibracijskemu kvantnemu stanju.

estimate the atmospheric warming experienced during the last



150 years has been caused dominantly by human activities, Za lažje razumevanje termodinamičnih procesov v ozračju je

______________ pomembno imeti v mislih dve spoznanji. Prvo je, da je energija

Permission to make digital or hard copies of part or all of this work for personal or

classroom use is granted without fee provided that copies are not made or distributed fotonov (kvantov elektromagnetnega sevanja) kvantizirana in

for profit or commercial advantage and that copies bear this notice and the full ustreza produktu Planckove konstante h in frekvence ν: citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s).

(1)

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia E = hν

© 2024 Copyright held by the owner/author(s).



DOI https://doi.org/10.70314/is.2024.env.04

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia E. Margan



Pri tem lahko molekula lahko absorbira sevanje na ustrezni Slika 1 je morda nekoliko zavajajoča, ker so spektri prikazani v

frekvenci le, če pripadajoče kvantno stanje ni zasedeno. odvisnosti od valovne dolžine. Energija sevanja pa je sorazmerna

Absorbirano energijo pa molekula čez čas spontano emitira spet frekvenci in obratno sorazmerna valovni dolžini. Poleg tega pa

v obliki fotona z enako frekvenco, če pred tem ni bila je lestvica valovnih dolžin logaritemska, zato zgleda, kot da

izpostavljena drugim interakcijam. dolge valovne dolžine (nizke frekvence) nosijo več energije, v

resnici pa je ravno obratno. Težava pa je tudi v tem, da so vsi

Za absorpcijo je tudi pomembna nesimetrična oblika molekule, prikazani spektri normirani na svoje maksimume, ki

ker s tem zunanjemu polju izkazuje električni dipolni moment. predstavljajo 100% učinka pri vrhu posameznega območja

Dvoatomske molekule v ozračju (N2, O2) so vedno simetrične in valovnih dolžin. Zato je primerjava učinkov med molekulami

zato nimajo nizko-energijskih interakcij s sevanjem. Drugače pa nekoliko otežena. Tretja težava pa nastopi zaradi različnih

je s triatomskimi molekulami. Na pr., pri vodni molekuli, H2O, koncentracij molekul v ozračju, saj je verjetnost zajetja fotona

tvorita atoma vodika glede na atom kisika kot 105°, zato je polje odvisna tudi od koncentracije. Vse to je treba imeti v mislih kadar

molekularnega naboja elektronskih orbital nesimetrično. primerjamo posamezne spektre.

Molekula ogljikovega dioksida, CO2, pa ima simetrično obliko,

atoma kisika tvorita z atomom ogljika dvojne vezi pod kotom Poseben primer predstavlja vodna para, saj se posamezne

180°: O=C=O. Zato molekula CO2 v osnovnem energijskem molekule vode lahko povezujejo v različno velike aglomerate, od

stanju nima dipolnega momenta in ne prihaja do interakcije s dveh molekul pa do makroskopskih kapljic. Posledično ima

sevanjem. Po trku z neko drugo molekulo v ozračju pa se lahko vodna para mnogo širši spekter in veliko število resonanc, mnogo

del kinetične energije porazdeli med obe molekuli in se zaradi večje od osnovnih šest za posamezno vodno molekulo.

tega lahko vzbudi ena od možnih nihalnih in rotacijskih

resonanc, odvisno od razpoložljive količine energije ob trku. Molekule CO2 pa se med seboj ne povezujejo, ima pa ta molekula

Zaradi nihanja kisikovih atomov glede na ogljikov atom dobi štiri izrazite resonance in sicer pri 1,9 μm, 2,8 μm, 4,2 μm, ter

molekula CO2 za kratek čas nesimetrično obliko in nek določen 15 μm. Prve tri se nahajajo v območju kjer je tako sončevo kot

dipolni moment in lahko absorbira foton ustrezne frekvence. zemeljsko sevanje majhno, pa še delno se prekrivajo s spektrom

vodne pare, katere koncentracija v nižjih plasteh ozračja je od 25

Na splošno imajo triatomske molekule več načinov nihanja, zato do 100 krat večja, zato je tam učinek CO2 na segrevanje ozračja

tudi več različnih resonančnih frekvenc, ki tvorijo karakteristični majhen. Dominanten učinek CO2 je pri resonanci v območju

spekter posamezne molekule. V odvisnosti od rotacijskih osi pa okoli 15 μm, ki je tudi delno prekrito s spektrom vodne pare.

resonance preidejo v ožji ali širši frekvenčni pas v okolici

resonančne frekvence. Zato lahko te molekule zajamejo ali Drugo pomembno spoznanje za razumevanje termodinamičnih

oddajo fotone z malenkostno drugačno frekvenco od osnovne. procesov je v verjetnosti zajetja ali izsevanja fotonov in

Slika 1 pogostosti trkov med posameznimi molekulami. prikazuje spektre molekul, ki jih vsebuje ozračje v

različnih koncentracijah, v primerjavi s sevalnim spektrom

sončevega in zemeljskega sevanja. Če neka molekula zajame foton, ne more več zajeti drugega z

enako frekvenco dokler se bodisi spontano ne znebi presežka

energije z izsevanjem fotona z enako frekvenco, bodisi pa trči z



neko drugo molekulo in se celotna energija obeh molekul

porazdeli drugače, zaradi česar se prejšnje resonančno kvantno

stanje sprosti. S tem pa tudi energija zajetega fotona preide v

kinetično energijo molekul, kar pomeni povečano temperaturo

ozračja, ponovna emisija fotona z enako frekvenco pa ni več

možna. Tako zajeta energija ostane v ozračju dalj časa, saj se

kinetična energija teh dveh molekul z nadaljnjimi trki porazdeli

na vedno večje število molekul, nazadnje pa uide v vesolje kot

sevanje pri valovnih dolžinah daljših od 50 μm.



Če pa molekula zajame foton in ga kmalu zatem spet odda, se v

ozračju ni spremenilo nič bistvenega, le energija je na svoji poti

proti vesolju imela zamudo za delček sekunde.



Ozračje ni enakomerno gosto, saj gostota pada z nadmorsko

višino, prav tako tudi pritisk in temperatura, zato je verjetnost da

pride do trka z drugo molekulo zelo velika v spodnjih plasteh

ozračja, z višino pa se verjetnost trka zmanjšuje. Zaradi tega je

tudi verjetnost zajetja in oddaje fotona v spodnjih plasteh ozračja

majhna, toplotna energija se prenaša pretežno s trki in



Slika 1 premikanjem zraka. V bližini tal je povprečna prosta pot fotona : Absorpcijski spektri posameznih molekul v ozračju v

primerjavi s sončnim in zemeljskim sevalnim spektrom. v območju IR sevanja okoli 22 m. Pri svetlobni hitrosti to pomeni

povprečno življenjsko dobo okoli 65 ns.

Resne napake v oceni toplogrednega učinka ozračja Information Society 2023, 9–13 October 2023, Ljubljana, Slovenia



Po drugi strni pa je povprečna hitrost molekul enaka zvočni »Hkrati lahko opazimo zmanjšanje zajete stopnje toplote nad

hitrosti, okoli 340 m/s. Pri normalni gostoti zraka 1,2 kg/m3, ter površino zemlje. Hitro gibanje zraka v vseh smereh, ki se tudi

povprečni molarni masi 28,8 vsebuje en kubični meter zraka dviguje kadar se segreje, ter infra-rdeče sevanje v zraku, bosta

okoli 2,5×1025 molekul, katerih povprečna razdalja znaša okoli zmanjšali intenziteto učinka, ki sicer nastane pod prozorno in

10−8 −11  m in je povprečni čas med trki 3×10 s. To pomeni, da gosto atmosfero, vendar to ne bo povsem izničilo tega učinka.«

molekule le redko zajamejo foton, saj jih pri tem moti velika

pogostost medsebojnih trkov. Teoretično izračunana verjetnost, Iz tega je očitno, da je Fourieru bilo povsem jasno, da v odprtem

da molekula odda odvečno energijo s sevanjem in ne s trkom je ozračju toplogredni učinek ne more biti enako močan, kot je

okoli 4%. Torej se toplotna energija v nizkih plasteh ozračja denimo pod steklenim pokrovom zaprte posode. Žal so kasneje

prenaša pretežno z molekularnimi trki. drugi raziskovalci (J. Tyndall, S. Arrhenius, in drugi) to njegovo

sporočilo nekako spregledali.

Z višino se gostota zraka zmanjšuje, tudi pritisk in temperatura

(hitrost, oziroma kinetična energija molekul), kar pomeni, da se Tako danes obravnavamo toplogredni učinek ozračja kot

postopoma povečuje verjetnost da molekula zajame foton, prav posledico zakona o ohranitvi energije, ter temu primernega

tako pa, da foton odda pred trkom z drugo molekulo. Približno sevalnega ravnovesja [2]: vstopno sončno sevanje, zmanjšano za

na polovici gostote ozračja, to je na višini okoli 5,6 km, so delež, ki se odbija z vrhov oblakov in poledenelih površin, ter

razmere takšne, da je verjetnost spontane oddaje fotona približno nekaj malega z ostalih površin, segreva površje, to pa potem

enaka verjetnosti trka s drugo molekulo. Tam je povprečna prosta segreva ozračje, ozračje pa potem seva proti vesolju. Dolgoročno

pot fotonov že okoli 2 km, kar pomeni da od tam fotoni lahko povprečje vstopnega in izstopnega sevanja mora biti enako,

hitro uidejo izven ozračja. Zato s te višine tudi zaznamo največje razlika nastane lahko le zaradi slabše prepustnosti ozračja za

število fotonov, oziroma največjo sevalno gostoto moči, kar sevanje, pri čemur se temperatura spreminja dokler se ne

merimo s sateliti kot sevanje Zemlje, Slika 2. To sevanje ima vzpostavi novo ravnovesje.

gostoto moči od okoli 232 W/m2, temu pa po Stefan-



Boltzmannovem zakonu ustreza absolutna temperatura okoli Sončno sevanje, izmerjeno s sateliti v zemeljski orbiti znaša 2 povprečno S 253 K (ali −20°C), kot bomo videli v nadaljevanju. 0  = 1361 W/m , kar imenujemo solarna konstanta, pa

čeprav ni povsem konstanta. Zaradi sončeve aktivnosti se

osončenost Zemlje spreminja v približno 11-letnih ciklih za okoli



0,2% (rekonstrukcija sončne aktivnosti zadnjih 400 let na

podlagi sprememb koncentracije radioizotopov 14C in 10Be se

zelo dobro ujema z zabeleženim številom sončnih peg); zaradi

ekscentričnosti zemeljske orbite se spreminja za okoli 1,7%;

zaradi premikanja težišča sončnega sistema glede na Sonce se

spreminja za okoli 1%. Skupno znaša velikost sprememb blizu

3%, vendar se običajno domneva, da se vpliv večjih dveh

sprememb v teku leta izravna. To pa ne drži povsem, saj je

sprememba težišča odvisna predvsem od velikih planetov

(Jupiter, Saturn), njuni orbitalni periodi pa se ne ujemata časovno

z zemeljskim letom, posledično je ta vpliv s periodo 62 let

odvisen od trenutne nagnjenosti Zemeljske rotacijske osi, saj

Slika 2 južna polobla ima mnogo več oceanov kot severna, oceani pa : Satelitske meritve gostote moči povprečnega

dolgovalovnega izstopnega sevanja Zemlje med letoma 1974 in veliko bolj absorbirajo toploto kot kopno.

2012. Sezonska nihanja so znatna, v dolgoročnem povprečju pa

ni trenda. Vir podatki: NOAA, po . Zato je neposreden vpliv spremembe v osončenosti na globalno

povprečno temperaturo najbrž majhen, pa čeprav imamo

Z opisanim smo si ustvarili sliko, ki bo osnova za razumevanje zgodovinske dokaze, da nekakšen vpliv vendarle obstaja

problema učinka tople grede. (srednjeveško toplo obdobje, mala ledena doba, predvsem pa

Maunderjev in Daltonov minimum sončne aktivnosti se dobro

ujemajo z zgodovinskimi podatki o lokalni klimi v Evropi).

2 Toplogredni učinek ozračja

Obstaja pa verjetnost, da Sonce vpliva na Zemljo posredno prek



francoski matematik Jean Baptiste Joseph de Fourier (1768 – aktivnosti je magnetno polje močnejše in bolj odklanja 1830). Kot topniški general v Napoleonovi vojski se je zanimal visokoenergetske kozmične delce, zaradi česar je v zemeljskem Običajno se kot idejni oče toplogrednega učinka ozračja navaja sprememb v sončevem magnetnem polju. Ob večji sončni



za izboljšanje natančnosti streljanja s topovi in zato tudi za vpliv ozračju manj ionov in manj kondenzacijskih jeder okoli katerih temperature na širjenje materiala. Kasneje pa je svoje raziskave se oblikujejo večje vodne kapljice oblakov. To teorijo zastopajo razširil še na druge materiale in pojave. Tako je med drugim pri nekateri znani fiziki ( H. Svensmark , Nobelovec J. Clauser , in opisu toplogrednega učinka zapisal [1]: drugi). Dejstvo je, da že zgolj 2% spremembe v oblačnosti ima

podoben učinek kot učinek celotne količine CO2 v ozračju (ne le

»... če preprečimo segretemu zraku, da se dviga in širi, ...«

človeškega deleža). Vendar ti vplivi niso tema tega članka in se



Potem pa v nadaljevanju: z njimi tu ne bomo ukvarjali.



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia E. Margan



Kar pa nas tu zanima je način izračuna globalne povprečne Poleg tega pa se del sevanja izgubi zaradi manjše emisivnosti tal.

temperature Zemlje. Zakonitost med sevanjem in temperaturo Vendar se za emisivnost večine snovi pri običajnih temperaturah

idealnega črnega telesa je prvi odkril fizik slovenskega rodu jemlje vrednost med 0,95 in 1, zato naj bi ta vpliv bil majhen in

Jožef Stefan (1835 – 1893); ugotovil je, da je površinska gostota se običajno ne upošteva (kar pa ni prav, saj je emisivnost odvisna

sevalne moči j* sorazmerna četrti potenci absolutne temperature tudi od razlike temperatur sosednih snovi).

T (v Kelvinih), pomnoženi s Stefanovo konstanto σ:

Dnevna osončenost tal pa je seveda odvisna tudi od geometrije,



j saj imamo opravka s krogelno simetrijo planeta. Za kroglo pa je * = σT 4 (2) po Lambertu potrebno upoštevati kot sevanja v smeri sever-jug

Pri tem je in vzhod-zahod, zato bomo imeli kvadrat kosinusa vpadnega σ = 5,67×10 −8 Wm −2 K −4 , kar je bilo sprva empirično

določeno, kasneje pa je Stefanov učenec in sodelavec kota. Če funkcijo cos2θ integriramo v pol-krogelnem območju Ludwig

Boltzmann kotov med ±90°, dobimo rezultat ½. Druga polkrogla planeta pa (1844 – 1906) izpeljal to konstanto teoretično iz

drugih osnovnih fizikalnih konstant: je v temi, zato imamo dodaten faktor ½. Skupno torej v

dolgoročnem povprečju celotno površje Zemlje dobi ¼



σ 5 sončevega sevanja, zmanjšanega za albedo. Za vstopno sevanje 4  2 3 (3) = 2π k ∕ (15 c h ) torej velja naslednja enačba:

kjer je Boltzmannova konstanta k = 1,38×10−24 Ws/K,

svetlobna (5) hitrost v vakuumu c  = 299792458 m/s in j * = S 0 (1−α) ∕ 4

Planckova konstanta h = 6,626×10−34 Ws/Hz.

Ker pa mora veljati zakon o ohranitvi energije, lahko izraza (2)

Izraz (2) je zato znan kot Stefan-Boltzmannov zakon sevanja. in (5) izenačimo in iz tega izračunamo teoretično povprečno

Pripomniti je treba, da ta zakon velja za idealno črno telo segreto temperaturo planeta:

na temperaturo T, kar modeliramo kot sevanje majhne odprtine v

veliki kavitaciji v termodinamičnem ravnovesju, tako da lahko T = [S0(1−α) ∕ 4σ]1/4 (6) izgubo energije skozi odprtino lahko zanemarimo. Johann

Heinrich Lambert (1728 – 1777) je že leta 1760 ugotovil, da je Ko vstavimo ustrezne številke, dobimo T = 255 K, ali −18°C. To

gostota sevanja odvisna od kosinusa kota glede na normalo na pa je zelo blizu vrednosti izmerjene s sateliti za sevalno območje

določen površinski odsek, kar pomeni, da je sevanje maksimalno na nadmorski višini 5,6 km, kjer smo ob Sliki 2 izračunali −20°C.

v smeri navpično na površino in nič v smeri vzporedno s Lahko rečemo da gre le za višinsko razliko okoli 300m.

površino. To je pomembno, ker sevanje po Stefan-

Boltzmannovem zakonu računamo v pol-krogelnem prostorskem Iz tega pa pri IPCC izvlečejo naslednji (vprašljiv!) sklep: če je

kotu 2π steradiana, za razliko od Planckovega zakona sevanja, ki temperatura tal 288 K in sevalna temperatura planeta 255 K

ga ponavadi normiramo na kot enega steradiana. Dejansko je pomeni, da mora biti toplogredni učinek ozračja 33 K. Iz tega

Stefan-Boltzmannov zakon integral Planckovega zakona po vseh potem sledi, da mora biti klimatska občutljivost na vsebnost

možnih frekvencah. Planckov zakon nam podaja spektralno toplogrednih plinov zelo velika. Za to občutljivost (znano kot

odvisnost sevanja (kot na Sliki 1). equilibrium climate sensitivity, ECS) ocenjujejo, da mora biti

najverjetneje v območju med 2 in 5°C za vsako podvojitev

Obrazec (2) nam omogoča izračun efektivnega sevanja koncentracije toplogrednih plinov.

zemeljskega površja, saj temperaturo lahko izračunamo iz

povprečja meritev vseh svetovnih meteoroloških postaj v Vendar pa laboratorijske meritve optične prepustnosti CO2 v

daljšem časovnem obdobju (Mednarodna meteorološka okolici dominantne resonance 15 μm valovne dolžine kažejo na

organizacija definira klimo kot povprečje vremena za najmanj 30 le 1,1°C za vsako podvojitev koncentracije CO2. Gre za

let). To povprečje znaša okoli 15°C, oziroma 288 K (kar je sicer logaritmično zakonitost, ki je bila empirično določena in ustreza

zgolj statistični konstrukt, vendar sledimo metodi IPCC). Torej relaciji ΔF = 5,35 ln(C ∕ C0), kjer je ΔF sevalno »siljenje«

bo površinska gostota sevalne moči: (forcing, kar je pomanjkljivo imenovano, saj znak Δ pomeni

spremembo), ln pa je naravni logaritem (po bazi e = 2,71828…)



j razmerja med novo koncentracijo C in začetno koncentracijo C0; * = 5,67×10 −8 × 288 4 = 390 W/m 2 (4) za podvojitev koncentracije pa velja C  ∕  C 0  = 2, torej bo:

Ta vrednost se močno razlikuje od vrednosti izmerjene s sateliti

(Slika 2), kar pomeni, da se v ozračju del energije porabi za 2 (7) ΔF  = 5,35 ln(2) = 5,35 × 0.693 = 3,7 W/m

druge procese.

Posledično za spremembo temperature velja relacija:

Če pa želimo določiti sevalno ravnovesje planeta, moramo



upoštevati, da se del sončevega sevanja prav tako izgubi na svoji ΔTp (8)  = ΔF { T p  ∕ [ S 0 (1− α )]} poti do tal. Znaten del sevanja se odbije od vrhnjih plasti oblakov

in poledenelih površin (te imajo visoko odbojnost, albedo, za kjer je temperatura površja Tp = 288 K. Zato je sprememba

širok frekvenčni pas sončevega spektra), del pa se absorbira v temperature za podvojitev koncentracije CO2 enaka le 1,1°C.

ozračju (aerosoli, vodna para, itd.). Ocenjena povprečna vrednost

albeda je 30%, oziroma α = 0,3. Ker se vrednosti za območje ECS ne ujemajo s ΔTp, si pri IPCC

pomagajo z domnevo o obstoju pozitivne povratne zanke prek

Resne napake v oceni toplogrednega učinka ozračja Information Society 2023, 9–13 October 2023, Ljubljana, Slovenia



vodne pare, ki naj bi ojačala učinek CO2 za približni faktor 3×. Dopustimo lahko, da senzorji satelitov ne zajemajo celotnega

Razlaga gre nekako takole: ozračje, segreto zaradi povečanja frekvenčnega območja, zato lahko nastopi znatna razlika med

vsebnosti CO 2 2 , lahko sprejme več vlage (7% več za vsako 1°C 232 W/m na Sliki 2 in 239 W/m na Sliki 4. Precej nelogično pa 2

po Clausius-Clapeyronovi relaciji), zato se zrak še bolj segreje, se zdi, da se v sistemu pretaka veliko več energije kot v sistem

ter lahko sprejme še več vlage, in tako v krog, dokler se ne vstopa in izstopa, saj zgleda kakor da bi hladno osrednje ozračje,

vzpostavi novo sevalno ravnovesje. ki je pri −18°C, z dobrimi 342 W/m2 dodatno segrevalo tla, ki so

na +15°C, pa čeprav efektivni pretok navzgor ustreza razliki med

Vendar meritve zračne vlažnosti (relativne in absolutne) ne 398 in 342 W/m2. Matematično se sicer številke ujemajo, toda

potrjujejo te domneve, kot je razvidno na Sliki 3, na kar je leta fizikalno je nesmiselno, če sistem, ki naj bi bil v

2023 opozoril tudi Nobelovec John Clauser, pa je takoj doživel termodinamičnem ravnovesju z okolico, vsebuje precej več

hude kritike z vseh strani, a povsem neupravičeno. energije kot vanj priteka in odteka. To nakazuje, da z načinom

izračuna, oziroma ocenami posameznih parametrov sistema, kot



jih privzema IPCC, mora biti nekaj hudo narobe.



Vendar to ni edina težava. V Sliki 4 na mestu označenem z rdečo

elipso manjka trinajsti (nesrečni?) podatek. Označba greenhouse

gases, oziroma toplogredni plini sicer naznanja, da se tam dogaja

nekaj pomembnega in da ti plini sevajo nazaj proti tlom

342 W/m2, ni pa podatka o razliki med sevanjem tal in sevanjem

višjih plasteh ozračja proti vesolju, kar je dejansko definicija

toplogrednega učinka!



Seveda to lahko izračuna vsakdo sam iz podatkov na Sliki 4:

398 − 239 = 159 W/m 2. Vendar te oznake v grafu ni, ker bi se kdo

spraševal kako lahko 398 + 80 − 342 = 136 W/m2 ustreza

Slika 3 159 W/m2 ? Ker se namreč tistih 82 in 21 W/m2 v resnici porabi : Satelitske meritve zračne vlažnosti na različnih višinah

med januarjem 1948 in junijem 2024. Vir podatkov NOAA, za mehansko delo, transport toplega in vlažnega zraka navzgor.

povzeto po .

Nekakšno razlago vendarle najdemo v besedilu poročila IPCC

Na podlagi vseh predstavljenih formul so pri IPCC konstruirali AR6 in sicer v poglavju 7.4.2.1 Planck Response, Slika 5:

sicer dokaj zapleten fizikalni model procesov v ozračju, kar pa

zelo poenostavljeno prikazuje Slika 4:





Slika 5: Iz poročila IPCC AR6, leta 2021; vir [4].





Slika 4 Kot opozarja prof. Howard Hayden [4], je to prvič po več kot 30 : Pretoki sevalne moči v zemeljskem ozračju. Številke letih delovanja IPCC, da so toplogrednemu učinku dali simbol predstavljajo gostote sevalnih moči posameznih procesov v G 2 , mu določili vrednost, G  = 159 W/m 2 , ter prvič eksplicitno enotah W/m . Vendar v območju označenem z rdečo elipso omenili Stefan-Boltzmannov zakon! manjka pomemben podatek! Vir: IPCC AR6 [3].



Parametre modela določa 12 števil (v oklepajih so navedene meje Pa poglejmo kako se izide izračun sevalnega ravnovesja. Ta določa, da mora biti izstopno sevanje zemlje, j Z , enako območja zanesljivosti), pri čemur vpadno kratkovalovno sončno vstopnemu sončnemu sevanju j S : sevanje na levi in izstopno dolgovalovno zemeljsko sevanje na

desni približno ustrezata prej izračunanim številkam: četrtina

sončnega sevanja (1361/4 = 340,25 W/m (9) 2 ) zmanjšana za albedo, j Z = j S

ter nekaj absorpcije v ozračju, segreva tla s kakšnih 160 W/m2,

zemeljsko površje izseva okoli 398 W/m Hkrati pa mora biti izstopno sevanje j 2 (mi smo izračunali Z enako razliki med

390), ozračje pa izseva kakšnih 239 W/m sevanjem površine tal j 2 v vesolje (sateliti sicer P, in toplogrednim učinkom G:

izmerijo le 232 W/m2).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia E. Margan



j koncentraciji zajel. To je razvidno na bolj podrobnem prikazu Z = j P  −  G (10) spektra na Sliki 7 , kjer je vrh ostro omejen z gostoto sevanja tal,

oscilacije pa so prisotne le na robovih spektra. Pri podvojeni

koncentraciji bi se povečala le strmina robov, ker je pač

Iz (9) in (10) sledi, da mora biti tudi vstopno sončevo sevanje absorpcijska energija kvantizirana z jakostjo medatomskih vezi

enako razliki med sevanjem tal in toplogrednim učinkom: CO2, in spekter nima možnosti, da bi se še bolj razširil.



jS = jP − G (11)



To pa si lahko predstavimo tudi grafično s Planckovimi spektri.

Planckov spekter sevanja tal je mogoče teoretično izračunati s

pomočjo Planckove relacije za odvisnost od valovne dolžine,

kjer namesto frekvence uporabimo λ = c/ν:



j 2 −5 hc/λkT (12) ( λ , T ) = (2 hc λ ) / (e − 1)

P



Povprečno globalno temperaturo tal smo že določili: T = 288 K.

Za valovne dolžine pa vzamemo območje med 1 in 100 μm, kar

je območje znotraj katerega je sevalni maksimum za to

temperaturo (pri okoli 10 μm). Dovolj dobro ločljivost dobimo,

če za korak pri valovnih dolžinah vzamemo Slika 7: Podrobnejši prikaz spektra CO2 v območju okoli dλ  = 0,1 μm. Potem

spektralna funkcija zgleda kot modra krivulja na dominantne resonance pri 15 μm, oziroma 666 valov na cm. Sliki 6 :



Vendar pa imamo še en pomembnejši dokaz, da so pri IPCC



napačno ocenili učinek toplogrednih plinov. S pomočjo Slike 6

in enačbe (11) lahko to nazorno pokažemo.





Slika 8: Prikaz enačbe (11) v grafični obliki.





Ker se vstopno sončevo sevanje jS skoraj konstantno, se tudi



Slika 6: Spekter sevanja tal (modra krivulja) za temperaturo izstopno sevanje jZ dolgoročno ne spreminja s koncentracijo CO2

T = 288 K po enačbi (12), v primerjavi s spektralno absorpcijo (kot potrjuje tudi Slika 2). To pomeni, da mora biti sprememba

ozračja: črna krivulja je skupni spekter plinov ob sedanji G natančno enaka spremembi sevanja površine, jP. Za podvojitev

vsebnosti CO2, zelena krivulja je za primer brez CO2, rdeča koncentracije CO2 smo že z enačbo (7) izračunali spremembo

krivulja pa predstavlja učinek podvojitve vsebnosti CO2. Na 2 gostote sevalne moči, ta prispeva 3,7 W/m. Za toliko bi se v teh

horizontalni osi je prostorska frekvenca v številu valov na cm. okoliščinah povečal tudi toplogredni učinek G.



Spekter izstopnega sevanja planeta pa je mogoče bodisi izmeriti Vendar pri IPCC podajajo drugačno oceno, pravzaprav več

s sateliti, bodisi izračunati na podlagi empiričnih podatkov za različnih ocen odvisno od posameznega scenarija znižanja

posamezne pline v ozračju, ki so zbranih v podatkovni zbirki izpustov toplogrednih plinov. Ti scenariji se v AR6 imenujejo

HITRAN, kot sta to naredila van Wijngaarden in Happer [5]. Shared Socio-economic Pathways, SSP. Vsakemu SSP scenariju

Črna krivulja na Sliki 6 iz njihove študije velja za sedanjo so dodelili drugačno stopnjo znižanja izpustov, učinek pa so

vsebnost CO2 v ozračju. Površina pod posamezno krivuljo je potem označili z dvema številoma, prvo število pomeni

enaka izračunanim gostotam sevalne moči za sevalne gostote iz predvideni porast temperature, drugo število pa je pripadajoče

enačbe (11). povečanje gostote sevalne moči.



Iz Slike 6 je razvidno, da podvojitev današnje koncentracije CO2 Za zgled vzemimo SSP3-7.0, (Slika 9 in Slika 10) ki v AR6

ne bo privedla do večjih sprememb globalne povprečne predstavlja primer, ko bi se izpusti še naprej povečevali po

temperature, saj je razlika med rdečo in črno krivuljo zelo sedanji stopnji in za njihovo zmanjševanje ne bi naredili ničesar

majhna, znaša manj kot 2% celotnega absorbiranega sevanja v (temu scenariju pravijo Business As Usual). Predvidevajo da bi

okolici dominantne resonance CO2, ki je pri 15 μm, oziroma v se v tem primeru količina izpustov CO2 podvojila leta 2095. okolici 666 valov na cm na horizontalni lestvici. To je posledica

dejstva, da je spekter CO2 že pri sedanji koncentraciji nasičen, Torej bi se od leta 2015 do leta 2095 izpusti povečali s sedanjih

ker preprosto ni več fotonov, ki bi jih lahko pri povečani 40 na 80 milijard ton na leto. Ta prirast pomeni vsako leto

Resne napake v oceni toplogrednega učinka ozračja Information Society 2023, 9–13 October 2023, Ljubljana, Slovenia



dodatnih 0,5 milijarde ton CO2. Toda meritve vsebnosti CO2 Kot vidimo na desni strani Slike 10, znaša izmerjeno povečanje

zadnjih 60 let kažejo, da ta narašča za polovico počasneje kot povprečne globalne temperature od leta 1850 do danes okoli

naraščajo izpusti, saj polovico sproti posrkajo naravni ponori 1,1°C, nad tem pa je dodano najverjetneje povečanje v primeru

(oceani in fitoplankton, kopenske rastline in bakterije v tleh). podvojitve celotne količine CO2 v ozračju, kar je ocenjeno na

Sedanja celotna količina CO2 v ozračju znaša 3200 milijard ton. 2,6°C, skupno torej 3,7°C. Od tega odpade samo na CO2 okoli

Če torej vsako leto v ozračju ostaja dodatnih 0,25 milijarde ton, 2,9°C, ostali toplogredni plini prispevajo še 1,3°C, izpusti

bi ob sedanji stopnji rasti izpustov skupno količino CO2 podvojili aerosolov pa ohladijo ozračje za okoli 0,5°C. Po scenariju SSP3-

v 65 letih od leta 2015, oziroma že leta 2080. 7.0 se to povišanje temperature zgodi zaradi povečanega

toplogrednega učinka za 7 W/m2.





Toda po Stefan-Boltzmannovem zakonu povišanju temperature

za 3,7°C ustreza povečanje gostote sevalne moči za 20 W/m2.



Kako je lahko 20 = 7? Na to vprašanje ni odgovora. Sklepamo

lahko le, da pri IPCC že v osnovi niso pravilno zastavili problema

in je model na Sliki 4 vsaj pomanjkljiv, če že ne povsem napačen,

pa tudi to, da se nihče ni potrudil preveriti kako se sedanje

sevalno ravnovesje planeta ujema s tistim predvidenim za primer

podvojitve vsebnosti CO2 v ozračju.



Dodatna težava je, da je ta napaka v klimatski občutljivosti za

faktor 3× vgrajena v vse računalniške klimatske modele in so

posledično vse napovedi rasti temperature do konca tega stoletja

Slika 9: Predvideni scenariji zmanjševanja izpustov po IPCC.

močno pretirane. To dejstvo je bilo očitno že leta 2010, ko sta dr.

Scenarij SSP3-7.0 predvideva nadaljnje povečevanje izpustov po

Roy Spencer in dr. John Christy zbrala in objavila napovedi rasti

sedanji letni stopnji, kar naj bi leta 2095 privedlo do podvojitve

temperature 73 računalniških klimatskih modelov, ter jih

letne količine izpustov (rdeči krogec).

primerjala s temperaturami izmerjenimi do leta 2010 s sateliti in



Kako verjetni so ostali scenariji najbrž nima smisla razpravljati. meteorološkimi baloni, Slika 11:

Poglejmo bilanco sevalne moči za SSP3-7.0 po oceni IPCC.

Predvideva se najverjetnejši porast temperature za 2,6°C

(oziroma najverjetneje v območju med 2 in 5°C) nad sedanjo

temperaturo, ki je že za 1,1°C višja od temperature leta 1850,

skupno torej 3,7°C; to pa naj bi ustrezalo povečani gostoti

sevalne moči za 7 W/m2, Slika 10:





Slika 11: Napovedi rasti globalne temperature 73 računalniških



klimatskih modelov od leta 1978 do leta 2023 v primerjavi z



izmerjenimi temperaturami do leta 2010 (tenke črte so napovedi

potekov temperatur posameznih modelov, debelejša črna črta je

njihovo 5-letno povprečje, modri kvadratki predstavljajo letna

povprečja satelitskih meritev, črni krogci pa so meritve z

meteorološkimi baloni).



Žal ta napaka ni brez posledic tudi v realnem svetu, gre namreč

za verjetno najdražjo računsko napako v zgodovini človeštva.

Zaradi te napake je bilo do sedaj na svetovni ravni po

nepotrebnem zapravljeno po grobi oceni več kot 18 bilijonov

Slika 10: Na desni polovici slike je scenarij SSP3-7.0, povzet po

ameriških dolarjev (samo v Nemčiji je bilo za njihov program

AR6, str. 13 [6]. Na levi pa je izračun gostote sevalne moči za

Energiewende v 15 letih porabljenih okoli 600 milijard evrov, v

predvideno temperaturno razliko po Stefan-Boltzmannovem

ZDA pa 3800 milijard dolarjev). Posredna škoda pa je še mnogo

zakonu. Številke se ne ujemajo.

večja, saj je zaradi tega bilo po svetu ustavljenih veliko investicij

v energetski, industrijski in kmetijski razvoj, kar bi prineslo velik



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia E. Margan



družbeni napredek, še zlasti v državah v razvoju bi to v veliki IZJAVA meri pripomoglo k zmanjšanju revščine in lakote.

Avtor za ta članek ni prejel nobene finančne ali kakršne koli

druge podpore javnih institucij ali privatnih organizacij. Stališča



3 in strokovne ugotovitve izražene v članku so izključno sad Zaključek avtorjevih izračunov na podlagi javno dostopnih izmerjenih

Izračuni s katerimi IPCC, za njimi pa pretežni del svetovne podatkov in pregledovanja svetovne literature na tem področju in

politike in množičnih medijev ustrahujejo prebivalstvo s ne izražajo uradnih stališč ali opredelitev ustanove v kateri je

posledicami prihodnjih podnebnih sprememb in posledično avtor zaposlen. upravičujejo uvajanja raznih omejitvenih ukrepov s ciljem

zmanjševanja izpustov toplogrednih plinov, vsebujejo številne

napake. Nekaj najbolj resnih je opisanih v tem članku, izračuni

dokazujejo, da je ocenjen vpliv toplogrednih plinov na podnebje VIRI

prevelik vsaj za faktor 3×. Zaskrbljujoče je, da se o teh napakah

ne govori niti v znanstvenih krogih, pa je ravno znanost tista, ki [1] J. Fourier, 1827: MEMOIRE sur les temperatures du globe terrestre et des

bi morala vse tovrstne teoretične predpostavke in na njih espaces planetaires; (prevod v angleščino: W. M. Connolley)

zasnovane izračune sproti temeljito preverjati in morebitne

napake odpravljati. Da se to ne dogaja je najbrž posledica [2] S. Manabe, F. Möller: On the Radiative Equilibrium and Heat Balance of dejstva, da so tudi mnogi znanstveniki podlegli političnim the Atmosphere, Monthly Weather Review, Vol. 89, Num. 12, Dec. 1961

pritiskom in grožnjam z omejevanjem financiranja znanstvenih

raziskav tako samim znanstvenikom kot njihovim institucijam,

če bi javno dajali izjave, ki ne soglašajo z uradnimi političnimi [3] Šesto poročilo IPCC: AR6, Ch.7, Figure 7.2, str. 934 (2021),

opredelitvami do problemov vpliva na podnebje. To seveda ne

pomeni, da vpliva človeških dejavnosti na podnebje ni. Vendar [4] Howard Hayden, Ph.D.: The Magic Trick of Climate Science.

bi morali biti ti vplivi realno ocenjeni, morebitni nujni

omejevalni ukrepi pa temeljito pretehtani in sorazmerni,

predvsem pa naj ne bi povzročali več škode kot koristi. [5] W. A. van Wijngaarden, W. Happer: Dependence of Earth's Thermal

Radiation on Five Most Abundant Greenhouse Gases, June 8, 2020,



ZAHVALA

Avtor se zahvaljuje Vladimirju Alkalaju in Rafaelu Mihaliču za [6] Šesto poročilo IPCC: AR6, SPM, Figure SPM.4, str. 13 (2021),



pomoč pri strokovnem pregledu in preverjanju besedila.



Vrednost električne energije intermitirajočih virov in



kanibalizem na trgu z električno energijo



Value of Electricity From Intermitend Sources and Electricity



Market Cannibalism



Rafael Mihalič

UL, Fakulteta za elektrotehniko

Tržaška 25

Ljubljana, Slovenija

rafael.mihalic@fe.uni-lj.si



POVZETEK and subsidies are nevertheless necessary, one can rightly ask:



Intermitirajoči viri električne energije (OVE), predvsem market?’. ‘How much is electricity from renewable energies worth on the

vetrne (VE) in sončne elektrarne (SE), prevzemajo vedno večji In this article, we have therefore analysed the market value

delež v proizvodnji električne energije. V Sloveniji je rast of renewable energy generation for Germany and Slovenia for

količine SE eksplozivna. Ta razvoj lahko pripišemo subvencijam 2023 and the first half of 2024. The calculations are based on



zastonj" in subvencije kljub temu potrebne, se lahko upravičeno It turns out that due to the characteristics of RES and the vprašamo: "Koliko pa je na trgu sploh vredna električna energija in privilegiranemu položaju na trgu. Če naj bi bila "veter in sonce publicly available data.



tržno vrednost proizvodnje OVE za leto 2023 in prvo polovico V članku smo zato za Nemčijo in Slovenijo analizirali both Germany and Slovenia. In Slovenia, this effect reduces the market value of energy from SE in 2024 to just over half of what leta 2024. Izračuni so temeljili na javno dostopnih podatkih. would be expected if there were no renewables in the EPS. This iz OVE?" there is a pronounced so-called cannibal effect on the market in operating characteristics of the electric power system (EPS),

Izkazalo se je, da zaradi lastnosti OVE in obratovalnih value is only slightly higher than in Germany, although the

značilnosti elektroenergetskega sistema (EES) tako v Nemčiji, relative share of energy from SE is only about half as high as in

kakor Sloveniji, pride do izrazitega tako imenovanega Germany. From this we can conclude that, due to the

kanibalskega učinka na trgu. V Sloveniji omenjeni učinek interconnectedness of the Entso-e system, the Germans are partly

zmanjša tržno vrednost energije iz SE v letu 2024 na nekaj nad ‘exporting’ the problem of price cannibalism, while we are

polovico tiste, ki bi jo pričakovali, če v EES ne bi bilo SE. Ta ‘importing’ it.



energije iz SE približno 2 x manjši, kot v Nemčiji. Sklepamo, upgrading the EPS for renewable energies. Redispatching alone lahko, da zaradi povezanosti Entso-e sistema Nemci problem vrednost je le malo nad tisto v Nemčiji, čeprav je relativni delež None of this takes into account the costs of operating and



nadgradnje EES zaradi OVE. Že samo "redispatching" potisne Pri vsem tem nismo upoštevamo stroškov obratovanja in It is therefore safe to say that the real market value of energy from future RES will quickly approach 0 given the vrednost energije iz Nemških SE navzdol proti ¼ pričakovane current growth trend. kanibalizma cen deloma "izvažajo", mi pa "uvažamo". value (that without renewables in the EPS). reduces the value of energy from German SE to ¼ of the expected

vrednosti (tiste brez SE v EES).

Brez slabe vesti lahko torej trdimo, da se bo dejanska tržna KLJUČNE BESEDE



bližala vrednosti 0. In to ne upoštevajoč dodatnih stroškov EES Obnovljivi viri energije, OVE, cene električne energije, vrednost energije bodočih OVE ob trenutnem trendu rasti hitro



zaradi OVE. elektroenergetski trg, tržna vrednost električne energije, kanibalski učinek

ABSTRACT



Intermittrnt electricity sources (RES), in particular wind KEYWORDS



power (WT) and solar energy (SE), are accounting for an Renewables, RES, Electricity prices, Electricity market. Market value of electric power, Cannibalisation effect increasingly large share of electricity generation. In Slovenia, the

growth of SE is explosive. This development is due to subsidies



and a privileged position on the market. If ‘wind and sun are free’ 1 UVOD



V predhodnem prispevku (na istem mestu pred enim



Permission to make digital or hard copies of part or all of this work for personal or letom) [1] smo analizirali problem ovrednotenja investicij v classroom use is granted without fee provided that copies are not made or distributed obnovljive vire energije (OVE). Kakor je bilo izpostavljeno, for profit or commercial advantage and that copies bear this notice and the full splošno sprejeta metodologija izrazito favorizira izgradnjo OVE citation on the first page. Copyrights for third-party components of this work must be honored. For all other uses, contact the owner/author(s). in ne upošteva vrste dejavnikov in potrebnih investicij v Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia prilagoditev celotnega elektroenergetskega sistema (EES). © 2024 Copyright held by the owner/author(s).



https://doi.org/10.70314/is.2024.env.5

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia R. Mihalič



Hkrati odpira vrsto ne le ekonomskih, pač pa splošnih družbenih 2 OSNOVE TRGA ELEKTRIČNE ENERGIJE

dilem povezanih s pojmovanjem enakopravnosti ljudi v družbi v "ZA TELEBANE"



tem za politično pogojeno prerazporeditev premoženja ljudi, če relativno zapleteni in so za povprečnega prebivalca Evrope prej, kot ne, španska vas. Tudi elektrotehniki na področju uporabimo blagi izraz. elektroenergetike, ki se z omenjenim trgom (v nadaljevanju dovoljeno nekaterim, ni dovoljeno drugim!"). Dodatno gre pri Principi trgovanja z električno energijo v Evropi so smislu "quod licet iovi non licet bovi" (pomensko: "Kar je



omenjenega "zmernega deleža" ni smiselno preseči. Še zdaleč prejšnjega stoletja in se od tedaj neprestano razvija. Že v 80-tih je Evropska komisija začela zagovarjati tako imenovani prosti pa ne velja za akterje na trgu, ki v taki situaciji lahko pridobijo evropski trg z električno energijo. Prej so bili namreč EES bolj velike ekonomske koristi. V tržnem gospodarstvu s tem sicer ni ali manj omejeni na posamezne države. Povezovanje EES med nič narobe, problem pa je, da so te koristi pogojene s seboj in vzpostavljanje velikega sinhrono delujočega sistema je siromašenjem tihe večine in z "zdravo konkurenco" nimajo tehnično in regulatorno izjemno kompleksna naloga in lahko veliko opraviti. Zlobneži bi to pospremili s pripombo o traja reda desetletja, kakor kaže praksa pri priključitvi dežel "ribarjenju v kalnem" in na situacijo lahko v določeni meri vzhodnega bloka po padcu železne zavese. Ima pa velik, povezan gledamo tudi s tega stališča, saj večini ljudi niti približno ni EES, velike prednosti pred majhnim in omogoča zanesljivejše in jasno, kako zadeve potekajo in za kaj gre. Edino, kar opazijo je, cenejše obratovanje. Ker gre za tehnično zelo zapleten sistem, pri da se vse draži, plače capljajo bolj ali manj na mestu ali zelo katerem je, zato da sploh lahko stabilno obratuje, potrebno obotavljivo in z velikim časovni zamikom skušajo slediti zagotavljati določene tehnične pogoje in upoštevati tehnične inflaciji, na sosednjih vilah so "en, dva, tri" zrasle sončne omejitve, mora biti tudi trg z električno energijo z jasnimi pravili elektrarne preko celih streh in prizidkov, njim pa "elektro" ne in omejitvami zasnovan tako, da je skladen s tehničnimi pogoji dovoli postaviti par panelov na uti, ker so "kapacitete omrežja za varno obratovanje EES. Elektrika je namreč izrazito izčrpane". neprimerno blago za trgovanje in lahko bi rekli, da zanjo prosti V tržnem gospodarstvu pač razvoj družbe diktirajo pravila trg ni ustrezna rešitev. Problem pa je, da zaenkrat v družbenem trga, zato si v nadaljevanju poglejmo, kakšno vlogo na njem sistemu, v katerem živimo, boljše rešitve ne poznamo. igrajo intermitirajoči OVE (sončne, vetrne elektrarne) in kakšna Osnova obratovanja EES je dejstvo, da se ta mora sproti je dejanska tržna vrednost njihove električne energije. Z drugimi prilagajati (zagotavljati energijo) porabi, ki se tekom vsake besedami; zanima nas, koliko bi bila na trgu vredna električna minute, ure, dneva, tedna, sezone ves čas spreminja. Skratka energija OVE, če bi za vse nove OVE elektrarne ta trenutek proizvodnja in poraba morata biti ves čas v ravnotežju. To ukinili njihov privilegirani položaj na trgu z električno energijo ravnotežje pa zagotavljajo elektrarne z neprestanim in bi bile te povsem izenačene s "klasičnimi" viri. prilagajanjem proizvodnje. Različne elektrarne pa imajo različne Pri analizi smo pod drobnogled vzeli zadnjo leto in pol obratovalne karakteristike, različno sposobnost prilagajanja dogajanja na nemški in slovenski borzi z električno energijo. proizvodnje, različne načine in čase zagona, različne Omenjeno časovno obdobje smo izbrali, ker podatki zaradi investicijske stroške, različne tekoče stroške (cena primarnega pretresa in spremembe funkcioniranja celotne družbe med in ekonomsko in tehnično nerazumnim mejam, pa problemi (in njegove osnove, saj sicer ni mogoče niti razumeti, niti predvideti, niti načrtovati prihodnjega razvoja EES. Zlasti slednje dejstvo pa stroški) eskalirajo. Kje je meja tega "zmernega deleža" je je ključnega pomena za zanesljivo dobavo električne energije in vprašanje za milijon dolarjev in je skoraj ni moč izračunati. ga opisuje znana krilatica: "Ko se v EES pojavijo problemi, je že Pogojena je s strukturo EES, lastnostmi in vrsto virov električne 10 let prepozno." Torej si v nadaljevanju oglejmo nekaj osnovnih energije, lastnostmi porabnikov itd. Pa tudi, če bi bila znana, to dejstev v smislu serije knjig "Za telebane", ki se jih nekoliko za politiko, za religiozno prepričane ljudi in oportuniste ne bi izkušenejši še dobro spomnimo in so predstavljale pravo pomenilo veliko. Pač v smislu " après nous, le déluge " (za nami odrešitev za povprečne uporabnike česarkoli že. potop). Trg z električno energijo v EU so začeli uvajati ob koncu Za neko družbo oz. državo, kot celoto, torej velja, da izpostavljenih dilem ni oz. so zanemarljive in problemi tehnično električno energijo) ne ukvarjamo neposredno, se večinoma ne moremo pohvaliti z ekspertnim znanjem na tem področju. Ker pa obvladljivi z obstoječim EES. Kakor pa delež OVE zraste nad razvoj EES v veliki meri diktira prav trg, je potrebno poznati vsaj omenjeni "zmerni delež" in se začne bližati politično sprejetim OVE v določenem EES kakih večjih problemov in v prispevku označujemo s "trg" dejavnosti povezane s trgovanjem z Izkazalo se je tudi, da v primeru "nekega zmernega" deleža



"obdobjem covida" niso relevantni, po drugi strani pa je večanje energenta, na primer goriva), različne stroške zagona, različne dinamične lastnosti, različne možnosti zaustavitve (pri npr. instalirane moči in s tem deleža OVE pri proizvodnji električne jedrskih elektrarnah sta zaustavitev in ponoven zagon energije tako hiter, da analiza stanja pred letom 2020 ne bi imela kompleksen in dolgotrajen, več dni trajajoč proces, povezan z nobenega smisla. velikimi stroški), pa še kaj bi se našlo. Stohastični OVE (VE, SE) Nemčija je bila izbrana, ker gre za državo, ki je po eni dandanes nimajo sposobnosti prilagajanja potrebam EES, kar strani sinonim za " Energiewende " (energijski preobrat) in zaradi sicer ne pomeni, da je tehnično ne bi mogle imeti, vendar za ceno njene ekonomske in politične teže v veliki meri diktira evropsko manjših dobičkov, v kolikor to ne bi bilo podprto s še dodatnimi politiko na tem področju, po drugi strani pa ima odličen sistem subvencijami. Ker torej OVE nimajo ene od osnovnih lastnosti, javno dostopnih podatkov povezanih s proizvodnjo virov ki jih elektrarne morajo imeti, jih pravzaprav ne moremo električne energije (recimo [2, 3]). Pri Sloveniji smo se seveda imenovati "elektrarne". To so stohastični viri električne energije omejili le na sončne elektrarne (SE), saj je prispevek vetrnih in kar "narava da", pač proizvedejo. Tu pa se že nakazuje, kam elektrarn (VE) zanemarljiv. pes taco moli. Razliko med proizvodnjo in porabo (ki se ves čas

R. Mihalič Information Society 2023, 9–13 October 2023, Ljubljana, Slovenia



spreminja) morajo izravnavati "klasični" viri (ali v prihodnosti Mesec dni kasneje, 14. julija, je bila cena negativna med 8.

shranjevalniki električne energije, ki pa so zaenkrat za resno uro zjutraj in 17. uro.

sistemsko shranjevanje za red velikosti predragi, razen tega pri Ob tem se lahko upravičeno vprašamo, kak smisel sploh

današnji tehnologiji na Zemlji še zdaleč niti ni dovolj surovin ima graditi recimo dodatne SE, če je pa takrat, ko proizvajajo

zanje). Seveda obstaja tudi možnost prilagajanja porabnikov največ, cena elektrike skoraj nič, in kako se obstoječim izide

(prostovoljnega ali prisilnega – beri izklopi), ampak s tem je račun. Kratek odgovor je: "Ker so subvencionirane in njihov

razkošja kadarkoli dostopne elektrike za ljudi konec in znamenite dohodek ni odvisen le od tržne cene." Torej lastnikov kaj dosti

redukcije iz 80-tih let v katerih ne bi uživali verjetno niti največji ne motijo negativne cene na borzi. Seveda pa to vpliva na vse

nostalgiki, lahko doživijo renesanso. udeležence na borzi. Problematične so zlasti elektrarne, ki rabijo

Omenjeni problemi so tem večji, čim večji je delež OVE. dolgo časa (beri: je drago) za zagon in tiste, ki imajo sicer nizke

Kljub posebnostim trga z električno energijo, pa na njem stroške obratovanja in velike stalne stroške (beri: velika

vendarle še vedno veljajo klasična tržna pravila, torej če je blaga investicija) in torej za rentabilnost potrebujejo veliko

veliko, je poceni, če ga je malo, pa je drago. Ker je električna obratovalnih ur. Tipično gre za jedrske elektrarne, dobre

energija za funkcioniranje družbe nujna, zato lahko ob premogovne elektrarne z visokim izkoristkom, plinsko-parne

pomanjkanju doseže neverjetne cenovne skoke (v J. Avstraliji elektrarne, lahko tudi za pretočne hidroelektrarne oz. elektrarne

npr. v nekaj urah skoraj 300+ - kratnik, torej 30 000 % z majhno akumulacijo (ravno včeraj, 5. 7. sem iz prve roke

dolgoletnega povprečja [4]). Še zanimivejša je situacija, ko je izvedel, da hidroelektrarne na Savi že ves mesec po malem

elektrike preveč. Nekaterih elektrarn in tisočev SE namreč ni prelivajo vodo, ker je elektrike čez dan preveč in je sistem ne

moč poljubno izklapljati. Takrat doseže električna energija na more absorbirati).

borzi vrednosti blizu nič ali celo negativne vrednosti (V smislu: V takih razmerah se elektrarn z nizko ceno primarnega

"Plačamo, samo vzemi!"). Ilustracijo prikazuje slika 1. energenta in visokimi investicijskimi stroški (beri: dobrih,

energetsko učinkovitih elektrarn) ne izplača graditi. Ob



VE kopno VE morske SE Poraba Cena „dan v naprej“ nespremenjenem trendu bi količina OVE narasla do meje, ko bi

240 72 ob ugodnih naravnih danostih proizvajali velike presežke

200 60 električne energije, ki bi jih bodisi na nek način "metali vstran"

160 (recimo s prelivanjem vode v hidroelektrarnah in/ali 48

GW zaustavitvijo jedrskih elektrarn) ali shranjevali v hranilnikih

h električne energije. Problem je, da so za shranjevanje na

W 120 72 h



M

€/ sistemskem nivoju ti pri obstoječi tehnologiji za red velikosti

80 60

40 48 predragi, razen tega zanje na Zemlji ni dovolj surovin. Na OVE

in hranilnikih temelječ EES je sicer tehnično teoretično mogoče

0 0 izgraditi, vendar njegove cene niti najbogatejša družba ne bi

25. marec 26. marec 27. marec 28. marec 29. marec 30. marec 31. marec

Vir mogla plačati.

Ob tem še enkrat poudarimo, da pri dosedanji razpravi o

Slika 1: Proizvodnja OVE in cena na borzi EPEX (Nemčija) ceni elektrike gre za ceno, ki bi jo dosegla na borzi dodatna

zadnji teden marca [5] elektrarna (marginalna cena), in da ne moremo celotne količine



obratno korelirani. Še bolj zanimiva je situacija na isti borzi Jasno je razvidno, da sta cena na borzi in proizvodnja OVE električne energije iz OVE ovrednotiti po tej ceni (marginalni). To bi veljalo le za ta trenutek dodano dodatno elektrarno. Če bi namreč odvzeli vso električno energijo OVE, bi se cene drastično junija in julija, kjer je cena v 12 urah skočila iz vrednosti malo spremenile. Logiko ponazarja slika 3. nad 0 na skoraj 2500 €/MWh in nazaj na 0 ter v naslednjih 12

urah podobno – slika 2, 14. junija pa je bila cena okrog poldneva

negativna, približno –150 €/MWh (negativen trikratnik neke Stroški

Breme

povprečne borzne cene). (€/MWh)

Proizvodnja in borzne cene električne energije v Nemčiji



26. teden 2024 20 GW OVE



96 2500 2500 Preostalo breme



Trgovanje „znotraj dneva“



80 Ne ‐ obnovljivi viri Trgovanje „dan v naprej“ Obnovljivi viri 2000 Cena brez OVE





) Cen 64 1500 h a ( €/M WW





( G Plinske /M č 48 1000 TE W € Plinsko Mo h 30 Parne ) 32 500 TE



Reducirana cena



16 0 TE na



Jedrske TE na črni lignit



24. 6. 2024 0 Soproizvod premog TE ‐500 nja E+T 25. 6. 2024 26. 6. 2024 27. 6. 2024 28. 6. 2024 29. 6. 2024 30. 6. 2024





Moč (MW)



Slika 2: Proizvodnja OVE in cena na borzi EPEX (Nemčija) Slika 3: Logika določanja cene na borzi Vir: prirejeno po[6]



zadnji teden junija [1]





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia R. Mihalič



Na sliki 3 je obrazložena logika določanja borzne cene po 24 ur na dan), nekje okrog 10%. To pomeni, da je pri 10% deležu

obstoječih pravilih. Na abscisni osi je potrebna moč porabe in/ali proizvodnje vgrajena moč SE okrog povprečne proizvodnje

proizvodnje, predvidena za npr. določeno uro prihodnjega dneva, sistema. Ko torej sonce zasije opoldne povsod "na polno"(se

na abscisi pa je cena za 1 MWh električne energije. Višina bar zgodi tu in tam), SE praktično pokrijejo porabo. Kam pa z viški

diagramov predstavlja minimalno ceno, ki jo lahko na trgu iz elektrarn, ki jih ni moč zaustaviti? Možnosti je sicer več, a

ponudijo posamezne elektrarne (na MWh). V bistvu gre za lastno nobena ni ravno poceni. Nemci npr. v takih primerih izvažajo

ceno proizvodnje, pod katero bi proizvajale izgubo. Cena zadnje elektriko skoraj zastonj, zvečer (ko SE "ugasnejo") pa jo po zelo

elektrarne, ki je potrebna za zadovoljitev porabe (recimo stičišče visoki ceni uvažajo. In potem se ljudje čudijo, da se elektrika

polne in črtkane rdeče črte) potem velja (toliko dobijo plačano) draži, in da ima Nemčija skoraj najdražjo elektriko na svetu?!

za vse elektrarne na levi strani. Problem te logike je, da se OVE Ilustracija je na sliki 4, ki prikazuje ceno elektrike (modra

privzeto postavijo s ceno 0 na začetek tega diagrama. Na ta način krivulja) v odvisnosti od t. i. preostalega bremena (residual load),

imajo zagotovljen odjem, kar je po "zdravi kmečki logiki" se pravi bremena, ki ga ni moč pokriti z OVE (modra vertikalna

povsem ustrezno. Na ta način se poraba iz "klasičnih" elektrarn črta na sliki 3), pač pa s klasičnimi elektrarnami (ali uvozom). Če

zmanjša (na sliki 3 konkretno za 20 GW), saj jo pokrijejo OVE. OVE pokrijejo skoraj vso porabo, je seveda preostalo breme

Posledično se "rdeča črta" pomakne v diagramu proti izhodišču majhno (slika 3 – modra črta gre levo, črtkana modra črta pa se

na območje cenejših elektrarn (modra črta). pomika navzdol) in cena pade, v skrajnem primeru na 0.

To ima dve, pravzaprav tri posledice. Prva posledica je, da Borzne cene in „preostalo breme“ v Nemčiji



dobre plinsko – parne elektrarne in elektrarne na črni premog Med tednom od 11. do 17. aprila, 2022

ostanejo "brez posla" – glej sliko. Če se to dogaja pogosto, Cena

postanejo nerentabilne in bankrotirajo. Druga posledica je, da vse Preostalo breme

elektrarne levo od modre črte zaslužijo manj. Za fiksno )

subvencionirane OVE (tudi t. i. "net-metering" lahko štejemo v ) W hG



to skupino) to ni tak problem, za ostale pa je. Tretja posledica pa ( /MW € je pravzaprav posledica prvih dveh in pomeni, da nihče ne bo več ( reme a b n investiral denarja v dobre elektrarne z dobrim izkoristkom, ki pa lo e Cast jih je praviloma problematično pogosto zaganjati/ugašati oz. jim reoP močno variirati obremenitev. Po tej logiki se bo izplačalo graditi

le poceni elektrarne s slabim izkoristkom in zelo dragim

primarnim energentom. Slednje ni ovira, saj ko OVE



"presahnejo", je cena samo nebo ("sky is the limit") – glej npr.



sliko 2. Slika 4: Kanibalska logika; več OVE, manjša cena [7] Še opomba; na sliki 3 nas ne sme zavesti nizka cena



jedrskih elektrarn. Gre za spremenljive stroške, ki so pri njej zelo Konkretne izračune so za Nemčijo na podlagi borznih podatkov nizki. Če bi upoštevali investicijo, pri tej ceni ne bi bila nikoli izvedli v [5] – glej sliko 5 (vsak kvadratek predstavlja eno leto)! rentabilna. To dejstvo je seveda odločilno za investicije v nove Čeprav so podatki stari in delež OVE še relativno majhen, je objekte. logika nedvoumna. Hkrati potrjuje tudi omenjeno premiso, da Na ta način pa se tudi dejanska tržna vrednost proizvodnje majhen delež SE na obratovanje EES vpliva blagodejno, saj SE OVE zmanjša. Več je OVE, bolj je ta pojav izrazit, bolj se niža proizvajajo podnevi, ko je poraba (in cena) električne energija tržna cena drugim elektrarnam in tudi samim OVE. Govorimo o višja. Zato je tudi "faktor vrednosti" na začetku okrog 1,3 (130 tako imenovanem učinku kanibalizma ("cannibalisation effect"). % povprečne cene) in je nekako do 5 % deleža proizvodnje iz SE Omenjena razlaga opisuje osnovni princip, obstaja še kup nad 1. dodatnih mehanizmov, ki tehnično omogočajo, da EES stabilno



obratuje in lahko ceno potisnejo tudi v negativne vrednosti (kar Faktor vrednosti (glede na povprečno ceno)



po opisani logiki ni mogoče).

Podrobnejši ekonomski izračuni so razmeroma



kompleksni in podrobneje razložijo, zakaj z večanjem SE



subvencioniranih SE in VE privede do padca vrednosti njihove VE na kopnem

električne energije na borzi proti ali celo pod 0.



Pri zelo majhnem deležu OVE v EES cena oz. vrednost



električne energije sledi logiki, ki smo jo poznali pred desetletji.

Ponoči in med dela prostimi dnevi, ko večina ljudi in industrije



rabi manj elektrike, je cena nizka, podnevi pa je poraba bistveno



višja in temu ustrezno tudi cena. Zato je vloga SE v začetku za

obratovanje EES pozitivna. Slika pa se diametralno spremeni, če



recimo SE v nekem EES proizvedejo reda 10 ali 15 % elektrike. Tržni delež

Majhni omenjeni številki sta lahko zavajajoči in delež SE lahko Slika 5: Kanibalska logika – dolgoročni trend [6]

pomeni za EES veliko večji problem, kot bi lahko sklepali iz

njiju. Namreč, SE imajo v Sloveniji obratovalne ure, ali bolj po Opisana logika je v ekonomiji že dolgo znana in nekaj avtorjev

domače, izkoriščenost (koliko dejansko proizvedejo v primerjavi se je problema lotila znanstveno, pri čemer so upoštevali

z njihovim teoretičnim maksimumom = maksimalna osončenost zakonitosti trga v EU, karakteristike EES, možnosti prenašanja

R. Mihalič Information Society 2023, 9–13 October 2023, Ljubljana, Slovenia



energije po EE omrežju itd. Gre za relativno obsežne izračune, namreč ne pove mnogo, saj ni mogoče gotovo ločiti med visoko

za ilustracijo pa podajamo sliko 6. Avtor prispevka [7] je proizvodnjo SE in hkratno visoko porabo ter nizko proizvodnjo

analiziral scenarije razvoja vrednosti električne energije OVE v SE in hkratno nizko porabo. Podoben problem velja za lastnike

odvisnosti od večanja njihovega deleža v proizvodnji EES. Pri SE s tako imenovanim net-meteringom.

tem je ugotovil, da samo s subvencijami, brez posebnih dodatnih Na omenjeni ELES-ovi strani imajo za podatke

ukrepov, vrednost njihove proizvodnje pri 50 % deležu pade proizvodnje SE v Sloveniji preusmeritev na ustrezno stran Entso-

proti 0. S povečanjem fleksibilnosti EES v smislu možnosti e [10] in uporabljeni podatki izvirajo od tam. Borzne cene smo

shranjevanja električne energije in prilagajanja porabnikov povzeli po [11], podatke o količini instaliranih SE v Sloveniji pa

(oznaka "flex") je sicer mogoče nekaj narediti, vendar se pri po [12] in [13].

višjem deležu OVE cena še vedno giblje blizu ničle. "Rešitev" je Pri analizi podatkov pa sta se pojavila dva problema. Prvi

našel v spremembi regulative in drastičnem povečanju cene je bil ta, da so bili nam dostopni podatki o instalirani moči

izpustov CO2. Kaj si o vsem skupaj mislim, kot sončnih elektrarn samo za trenutek ob koncu leta. Za natančne

davkoplačevalec, ni za objavo. izračune, koliko je v bistvu celoten sistem SE izkoriščen, bi

potrebovali ta podatek za vsako uro. Ker je bil prirast instalirane

moči SE leta 2023 v Sloveniji izjemno visok (več od 400 MW -





st OVE od manj od 700 MW na več, kot 1100 MW), tudi povprečna OVE + flex vrednost ne bi dala sprejemljivih rezultatov. Razen tega (med o CO 2 ukrepi n CO 2 ukrepi + flex analizo) ni bilo mogoče najti podatka o instalirani moči elektrarn d na polovici leta 2024. Zato smo privzeli linearno rast instalirane re



a v moči SE skozi časovni točki 31. 12. 2022 in 31. 12. 2023 za n rž celotno študirano obdobje. Kasneje je ELES v sporočilu za

a t javnost [14] objavil, da bo uvedel na svoji spletni strani tudi



vn sprotno spremljanje proizvodnje in količine obnovljivih virov v ti a Sloveniji. Na žalost za naše obravnavano obdobje to ne pomaga el R veliko, smo pa imeli priložnost narediti preizkus za eno časovno

točko in primerjati ELES-ove nove podatke z našimi

Delež SE + VE v proizvodnji [%] predpostavljenimi (ekstrapolacija v letu 2024). Izkazalo se je, da

Slika 6: Ničvrednost energije iz OVE in "Dekanibalizacija" so se skoraj povsem skladajo. To sicer ni dokaz, je pa dober indic,

po ekonomistično [7] da so naše predpostavke blizu dejanskemu stanju, če zaupamo

ELES-ovi oceni trenutnega dejanskega stanja.



3 Drug problem so predstavljali podatki o urni proizvodnji IN MEDIAS RES ALI KAJ PRAVIJO slovenskih SE na strani Entso-e. Pri analizi se je namreč izkazalo, ŠTEVILKE da je skupna letna proizvodnja (preprost seštevek urnih

podatkov) veliko manjša, kot bi jo pričakovali iz podatkov o

Kot je bilo že omenjeno, je vsak EES specifičen glede instalirani moči SE glede na nek tipičen faktor izkoriščenosti (kot

strukture, lastnosti in razporeditve virov, možnosti prenašanja rečeno okrog 10 %) in tudi bistveno manjša od podatka na [12].

energije, lastnosti porabnikov itd. Zato kvantitativna analiza Špekuliramo lahko, da je slovenski operater pošiljal Entso-e le

scenarijev nekega EES, povezanih bodisi s tehničnimi lastnostmi podatke, do katerih je imel dostop (nekaj več kot polovica

ali ekonomskimi kazalci, v splošnem ne velja za nek drug EES ocenjene dejanske energije SE). Zato smo urne podatke o

ali, zlasti pri velikih EES (kar evropski EES je), ne velja niti za proizvodnji SE v Sloveniji normirali tako, da je na koncu leta bila

druga območja istega EES. Zato smo si zastavili vprašanje: "Kje energija enaka objavljeni letni proizvodnji [12]. Tak pristop je

na teh grafih iz prejšnjega poglavja pa se nahaja Nemčija in kje smiseln, ker dejanski poslani podatki, čeprav niso kompletni,

Slovenija. V uvodu smo navedli razloge za izbiro Nemčije, kot podajo časovno razporeditev proizvodnje, z drugimi besedami,

neke reference. kakor je osončena polovica SE, približno tako je po vsej

Na podlagi javno dostopnih podatkov torej poglejmo, verjetnosti osončena tudi druga polovica. V prejšnjem odstavku

koliko bi bila na trgu vredna električna energija OVE, če bi za omenjeni test je to predpostavko povsem potrdil.

vse nove OVE elektrarne ta trenutek ukinili njihov privilegirani Menimo, da je nekoliko daljši uvod s pojasnilom glede

položaj na trgu z električno energijo in bi bile te po logiki iz Slike uporabljenih podatkov nujen, ker leži izvirni greh množice študij

3 povsem izenačene s "klasičnimi" viri. Kot rečeno smo v analizi in prispevkov (zlasti na področju tako imenovane podnebne

upoštevali zadnjo leto in pol dogajanja na nemški in slovenski politike in energetike, ki "dokazujejo" potrebo po neskončnem

borzi z električno energijo. plačevanju t. i. zelenega prehoda) v prirejenih ali selekcioniranih

Kljub dobri dostopnosti do podatkov [2, 3], je bilo treba podatkih (Dokazov in analiz na to temo je veliko, a zaradi

nekatere faktorje privzeti. Zlasti to velja za podatke proizvodnje nebistvenosti referenc na tem mestu ne navajamo). V

slovenskih SE. Pred leti so na strani ELES, kjer prikazujejo nadaljevanju torej poglejmo, kaj so pokazali izračuni.

trenutne in historične podatke o proizvodnji in porabi [9] podajali

tudi proizvodnjo SE, vendar so to kasneje ukinili. Pridobivanje

teh podatkov namreč ni preprosto. Če je veliko SE na nizki 3.1 Situacija v Nemčiji



napetosti (hišne SE), bi bilo potrebno sprotno zbirati podatke za Izračuni za leto 2023 in prvo polovico leta 2024 za razmere vsako gospodinjstvo s SE posebej. Zbiranje podatkov na nivoju v Nemčiji so predstavljeni v nadaljevanju. nizkonapetostnega izvoda ali celo na srednji napetostni strani



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia R. Mihalič



Tabela 1: izračuni za Nemčijo Tabela 2: izračuni za Slovenijo

Izračunana vrednost  / Leto  2023 2024 Izračunana vrednost  / Leto  2023 2024

Povprečna cena na borzi [€/MWh] 95 68 Povprečna cena na borzi [€/MWh] 104 73

Povprečen zaslužek 1 MW VE na uro [€/h] 19,4 14,6 Povprečen zaslužek 1 MW SE na uro [€/h] 6,6 5,1

Povprečen zaslužek 1 MW SE na uro [€/h] 6,7 4,3 Vrednost proizvodnje SE glede na

83 70

Vrednost proizvodnje VE glede na povprečno ceno na borzi [%]

povprečno ceno na borzi [%] 85 87 Delež SE v proizvodnji RS [%] 4,6 7,7

Vrednost proizvodnje SE glede na

povprečno ceno na borzi [%] 77 64 Poglejmo, kaj nam povedo navedene številke.

Delež VE v proizvodnji D [%] 32 29



Delež SE v proizvodnji D [%] 12 13 3.3 Diskusija

Strošek redispatchinga [Mrd €] 2,35 1,175* Preden začnemo z interpretacijo rezultatov velja omeniti,

Vrednost proizvodnje VE obremenjene za da pri analizi nemških in slovenskih razmer v EES (proizvodnja,

delež redispatchinga na instalirano moč** 63 60 poraba) in dogajanja na ustreznih borzah [1, 11] ne gre za dva

glede na povprečno ceno na borzi [%] različna EES, temveč za en povezan EES in dve različni področji.

Vrednost proizvodnje SE obremenjene za Različne borzne cene nastajajo le zaradi omejenih možnosti

delež redispatchinga na instalirano moč** 54 37 prenosa energije. Ker sta obe področji vpeti v skupen sistem

glede na povprečno ceno na borzi [%] (Entso-e), se sistemski učinki OVE na obravnavanih področjih

* - privzeto ne odražajo na borzne cene enako, kot bi se, če bi imeli opravka

** - v Nemčiji konkretno delitev med VE in SE ca. pol – pol. z električnima otokoma. Vpliv je "mehkejši", ker se različna

področja do neke mere dopolnjujejo in tudi o tem govorimo, ko

Preurejanje predhodno dogovorjene proizvodnje (redispatching) izpostavljamo prednosti povezave v velik EES.

lahko v veliki meri pripišemo volatilnosti proizvodnje OVE. Neglede na to pa je možno potegniti vzporednice med

Seveda podatkov, kdo je za posamezen primer "odgovoren" ni in predstavljenimi ekonomskimi izračuni iz literature (Slike 4, 5, 6)

tega velikokrat niti ni moč nedvoumno ugotoviti. Menimo, da in med obravnavanima deloma Entso-e. V Nemčiji je vrednost

razdelitev med VE in SE glede na njihovo instalirano moč elektrike iz OVE (zmanjšana za ca. 15 % pri VE in med 40 % in

predstavlja sprejemljiv kompromis, ki poda vsaj dobro oceno za 50 % povprečne cene pri SE – ne pozabimo faktor vrednosti za

ta del stroškov implementacije OVE v EES. VE se začne pri ca. 100 % in za SE pri ca. 130 % Glej Sliko 5),

tudi ob upoštevanju redispatchinga (zmanjšanje za ca. 40 % pri

VE in za ca 60 % pri SE), nekoliko višja, kakor bi sklepali recimo

3.2 Situacija v Sloveniji iz Slike 6. Razliko lahko pripišemo omenjenemu "omehčanju"

Adekvatno izračunom za Nemčijo smo izvedli te tudi za kanibalskega učinka zaradi možnosti izmenjave energije z

Slovenijo. Osnovna razlika med naboroma je v tem, da v drugimi področji. Električna okolica Nemčije namreč po

Sloveniji nismo upoštevali VE, ker je njihov delež zanemarljiv. prispevku OVE za njo zaostaja in blaži omenjeni učinek.



Proizvodnja SE in borzna cena elektrike v Sloveniji za junij 2024

Proizvodnja PV Cena

1000 500

ZNAČILNOSTI

Ekstremna

800 volatilnost 400

W Zelo veliki cen

Antikorelacija gradienti



M Cena / proizvodnja SE moči Wh

v 600 300

Negativne



a 400 200 /M PV cene

€



vodnj 200 100 iz o Cena Pr 0 0



‐200 ‐100 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

.20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20 .20

.01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01 .01

01 01 01 02 02 02 03 03 04 04 04 05 05 05 06 06 07 07 07 08 08 08 09 09 10 10 10 11 11 11 12 12 13 13 13 14 14 14 15 15 16 16 16 17 17 17 18 18 19 19 19 20 20 20 21 21 22 22 22 23 23 23 24 24 25 25 25 26 26 26 27 27 28 28 28 29 29 29 30 30

Slika 7: Učinek proizvodnje SE na borzo "Southpool" Vir: lastni izračuni



V Sloveniji je prispevek OVE med 5 in 10 x manjši, kakor izenačeni, ne bi imele nove SE nobene možnosti ekonomskega

v Nemčiji, zato je smiselno primerjati rezultate s tistimi na Sliki preživetja. In to, brez da bi jim "naprtili" stroške

5. Izkaže se, da je kanibalski učinek v Sloveniji (zmanjšanje "redispatchinga", kaj šele morebitnih shranjevalnikov.

vrednosti za reda 50 % povprečne cene – zopet ne pozabimo, In nauk zgodbe? Morda bi krilatico "Veter in sonce sta

primerjalna vrednost je 130%) precej večji od tistega, ki bi ga zastonj!" veljalo zamenjati z: "Veter in sonce kmalu ne bosta

razbrali iz Slike 5, se pa precej dobro sklada s Sliko 6. Sklepamo vredna nič." Pa še retorično vprašanje: "Kdo bo vse to plačal?"

lahko, da je v električni okolici Slovenije odstotek proizvodnje

iz SE večji, kakor v Sloveniji, kar pomeni, da Slovenija v bistvu ZAKAJ ŽE, PRAVZAPRAV ???

"uvaža" kanibalski učinek SE, saj je relativno majhen trg.

Omenjeno dejstvo do neke mere potrjuje dogajanje na borzi

"Southpool". Močna volatilnost cene (nihanje, skoki) namreč ne REFERENCE



– glej Sliko 7 – so pa kvalitativno gledano razmere identične = How cheap is electricity from renewable sources. V: OGRIN, Tomaž (ur.), MIHALIČ, Rafael (ur.). Miti in resnice o varovanju okolja = Myths dogajanju na drugih (recimo nemški) borzah. morejo biti povzročeni z nihanjem proizvodnje le slovenskih SE [1] MIHALIČ, Rafael. Kako poceni je električna energija iz obnovljivih virov

and Truths about Environmental Protection : Informacijska družba - IS 2023

= Information Society - IS 2023 : zbornik 26. mednarodne multikonference

= proceedings of the 26th International Multiconference : zvezek D =



4 volume D : 11. oktober 2023, 11 October 2023, Ljubljana, Slovenia. SKLEP

Ljubljana: Institut "Jožef Stefan", 2023.

"Veter in sonce sta zastonj!" je krilatica, ki jo pogosto [2] https://www.stromdaten.info/

[3] https://energy-charts.info/?l=de&c=DE



bodisi od tistih, ki namerno zavajajo. Resnica skoraj ne bi mogla energien-in-australien-in-beiden-groessten-bundesstaaten-steigen-die- strompreise-morgens-und-abends-auf-bis-zu-16-000-us-dollar/ biti bolj drugačna. V [1] smo pokazali, kaj vgradnja OVE slišimo bodisi bodi si od ljudi, ki ne vedo veliko o realnosti, [4] https://eike-klima-energie.eu/2024/08/09/hoelle-mit-erneuerbaren-

[5] Bundesnetzagentur | SMARD)



(recimo SE in/ali VE) dejansko "potegne za sabo" in kake [6] http://open-electricity-economics.org/book/text/07.html [7] https://flex-power.energy/school-of-flex/market-value-of-renewables/ indirektne stroške povzroči v EES. [8] T. Brown, L. Reichenbergc: Decreasing market value of variable V tem prispevku pa nas je zanimalo, koliko je resnična renewables can be avoided by policy action Energy Economics, Volume 100, August 2021 tržna cena elektrike iz OVE (SE in VE) v Nemčiji in Sloveniji. [9] https://www.eles.si/prevzem-in-proizvodnja Teh podatkov namreč v javnosti ni zaslediti, saj ne skladajo z [10] https://transparency.entsoe.eu/ [11] https://www.bsp-southpool.com/rezultati-trgovanja-slovenija.html uradno politiko "zelenega prehoda". Njihovo objavljanje je torej [12] http://pv.fe.uni-lj.si/sl/ "politično nezaželeno." [13] https://www.gov.si/novice/2024-01-17-izjemna-rast-kapacitet-

Analizirali smo zadnje leto in pol dogajanja v Nemčiji in postavljenih-soncnih-elektrarn-v-letu-2023/

Sloveniji. Zakaj poleg Slovenije še Nemčijo in zakaj to obdobje [14] https://www.eles.si/medijsko-sredisce/sporocila-za-javnost-in-

obvestila/sporocila-za-javnost/ArticleID/21273/Proizvodnja-iz-

pojasnimo v Uvodu. Zaradi narave proizvodnje iz OVE, lastnosti obnovljivih-virov-energije-%C2%BBv-%C5%BEivo%C2%AB

obratovanja EES in značilnosti porabnikov prihaja z naraščanjem

deleža proizvodnje iz OVE do tako imenovanega kanibalskega

učinka na trgu z električno energijo. To pomeni, da po eni strani

OVE sami sebi zbijajo tržno ceno elektrike in po drugi strani

ekonomsko onemogočajo najboljše elektrarne (dobre

premogovne elektrarne z visokim izkoristkom, neamortizirane

jedrske elektrarne in plinsko-parne elektrarne z visokim

izkoristkom). Prvo dejstvo za same OVE niti ni tako tragično, saj

ekonomsko preživijo s subvencijami in/ali privilegiranim

položajem. Bankrot in zapiranje (in seveda ustavitev gradnje

česarkoli, kar ni subvencionirano, razen poceni plinskih elektrarn

z odprtim ciklom, slabim izkoristkom in zelo drago elektriko)

omenjenih klasičnih elektrarn pa lahko obratovanje EES pripelje

v nemogočo situacijo, cene elektrike na borzi pa do

nepredstavljivih skokov.

Izkazalo se je, da je s kanibalskim učinkom povezano

relativno znižanje cene VE v Nemčiji nekje 15 %, SE pa za ca.

50 % povprečne tržne cene. Če omenjenim VE in SE "naprtimo"

še stroške "redispatchinga", se ti številki povečata na ca. 40 oz.

90 %. Glede na ekonomske izračune v literaturi jo OVE v

Nemčiji zaradi podpore sosednjih sistemov (beri: "parazitiranja"

nemškega) še dobro odnesejo.

Za razliko od nemškega je situacija v Sloveniji nekoliko

slabša, kot bi pričakovali glede na literaturo. Cena energije SE je

za več kot 50 % povprečne borzne cene nižja od tiste, ki bi jo

pričakovali, če bi v EES bilo le malo SE (pade iz 130 % na 70 %

povprečne borzne cene). Kanibalski učinek Slovenija v bistvu

"uvaža" in elektrika iz njenih SE je na trgu še manj vredna, kakor

bi pričakovali "iz teorije". V razmerah, kjer bi bili akterji na trgu



https://doi.org/10.70314/is.2024.env.5

Kako razumeti ogljični odtis?



prof. dr. Gregor Radonjič

University of Maribor

Faculty of Economics and Business

Dept. of technology and environment protection

gregor.radonjic@um.si



Povzetek CO2), tako na nivoju podjetij kot tudi občinskih ter državnih ustanov. Ogljični odtis, ki pomeni vsoto neposredno in posredno



izdelka, storitve, dogodka ali posameznika je postal eden od ogljičnim odtisom (s poudarkom na ogljičnem odtisu izdelkov), najpomembnejših okoljskih indikatorjev oz. indikatorjev ki so praviloma spregledani oz. neupoštevani v gospodarski povzročenih emisij toplogrednih plinov s strani organizacije, Namen prispevka je prikazati nekatere vidike v povezavi z

trajnostnosti. Kljub njegovi popularnosti in razširjeni uporabi pa praksi in družbenih diskusijah ter opozoriti na njegove

je ogljični odtis povezan z različnimi metodološkimi ter metodološke in konceptualne pomanjkljivosti. konceptualnimi pomanjkljivostmi. Predvsem je preveliko

osredotočanje na ogljični odtis v majhni državi kot je Slovenija,

ki v bilanco emisij toplogrednih plinov prispeva minorni delež v

primerjavi z velikimi državami, problematično zato, ker se 2 Metodološke in konceptualne pomanjkljivosti posledično pozablja na nekatere ključne okoljevarstvene

probleme, ki so vitalnega pomena za kakovost življenja oz. ogljičnega odtisa

bivanja.



Ogljični odtis ima pomembne metodološke in konceptualne



Ključne besede storitev in podjetij na okolje pomanjkljivosti pri ugotavljanju dejanskih vplivov izdelkov, [5-6]. Gospodarsko, politično in

ogljični odtis, toplogredni plini, metoda LCA medijsko osredotočanje izključno na ogljični odtis namreč daje

zavajajoč občutek glede celovitih vplivov izdelkov, storitev in

organizacij na okolje, saj se pozablja na številne druge vplive, ki

jih povzročajo podjetja oz. njihovi izdelki ali storitve [7-9].

1 Uvod Posledica tega je, da predstavniki gospodarstva pogosto zgled za

svojo ''zeleno'' naravnanost prikazujejo izključno z

Emisije toplogrednih plinov (TGP) veljajo v splošnem v osredotočanjem na zmanjšanje ogljičnega odtisa, pri tem pa kot

gospodarstvu in politiki za ključni okoljski problem in prioriteto, po pravilu pozabljajo na številne druge vplive na okolje (ali pa

o katerem se v primerjavi z drugimi okoljskimi problemi se jih sploh ne zavedajo), ki so prav tako posledica njihove

neproporcionalno več razpravlja ter se mu namenja medijskega dejavnosti ali proizvodnje ter uporabe izdelkov [10]. Ob tem

in političnega prostora. Pojmi kot so nizko- in brezogljična velja poudariti, da poleg ogljičnega odtisa obstaja še več

družba, nizko- in brezogljično gospodarstvo, razogljičenje, pomembnih ''odtisov'', kot sta na primer vodni in ekološki odtis,

ogljična nevtralnost ter z njimi povezani ukrepi so v vrhu za katera velja, da nikakor ne bi smela biti v senci ogljičnega

prioritet vladnih ukrepov in trajnostnih politik podjetij [1-3]. Za odtisa.

določene organizacije oz. podjetja je poročanje o emisijah TGP

postalo tudi obvezno [4]. Za upravljanje 'ogljičnih politik' so zato Poleg omenjenih konceptualnih pomanjkljivosti ogljičnega

potrebni kvantitativni podatki o emisijah TGP, s pomočjo katerih odtisa ne gre spregledati tudi številnih metodoloških

se spremlja tako trenutno stanje glede emisij TGP kot tudi pomanjkljivosti njegovega določevanja. Še posebej je ta problem

učinkovitost ukrepov v zvezi z njihovim zmanjševanjem. Vlogo prisoten pri določevanju ogljičnega odtisa izdelkov zaradi

takšnega kvantitativnega indikatorja je prevzel ogljični odtis, ki njihovih kompleksnih okoljskih življenjskih ciklov. Takšno

predstavlja glavni kvantitativni indikator za izvajanje in določevanje ni enostavno, saj mora temeljiti na njihovih

spremljanje podnebnih politik (predvsem zmanjševanja emisij velikokrat kompleksnih okoljskih življenjskih ciklih. Le-ti

vključujejo številne faze: pridobivanje surovin za materialno in

∗Article Title Footnote needs to be captured as Title Note energijsko proizvodnjo, transport, proizvodnjo materialov,

†Author Footnote to be captured as Author Note kemikalij, sestavnih delov, embalaže, proizvodnjo končnih

Permission to make digital or hard copies of part or all of this work for personal or izdelkov, vplive med uporabo ter pri ravnanju z odpadki.

classroom use is granted without fee provided that copies are not made or distributed Metodološke pomanjkljivosti določevanja ogljičnih odtisov

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must izdelkov, li lahko močno vplivajo na končno vrednost ogljičnega



be honored. For all other uses, contact the owner/author(s). odtisa, tako zajemajo upoštevanje različnih meja življenjskih Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia ciklov, vpliv energijskih mešanic različnih držav, dostopnost do © 2024 Copyright held by the owner/author(s).

http://doi.org/10.70314/is.2024.env.6 verodostojnih podatkov o emisijah TGP v fazah pridobivanja

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia F. Surname et al.



surovin in tehnoloških postopkov ravnanja z odpadki, izbor ravnanja z izdelki po uporabi. Vključuje vse snovne in energijske

pravilnih emisijskih faktorjev ter potencialov globalnega tokove, tudi take, ki bi bili z uporabo drugih metod lahko prezrti.

segrevanja (GWP koeficienti) in upoštevanje oz. izvzetje V obravnavanje življenjskega okoljskega cikla nekega proizvoda

biogenega CO2 v izračunih [6]. Vse to lahko vodi do je vedno vključeno tudi pridobivanje energije, potrebne za

neprimerljivosti vrednosti ogljičnih odtisov. Rezultati so torej pridobivanje surovin, za njegovo predelavo, proizvodnjo,

odvisni od številnih spremenljivk in predpostavk, a tudi transport, distribucijo, uporabo itn., ki se prične prav tako s

poenostavitev, kar praviloma onemogoča neposredno primerjavo pridobivanjem potrebnih energijskih virov. Vendar z metodo

rezultatov izračunov in posledično otežuje pravilno LCA masne in energijske vtoke ter iztoke ne le kvantificiramo,

komunikacijo z javnostjo [11-12]. ampak jih povežemo z njihovim dejanskim škodljivim učinkom

na okolje ter zdravje. LCA analiza lahko vključuje tudi do 20

Kljub omenjenim metodološkim in konceptualnim različnih okoljskih kategorij, kot so npr. ogljični odtis,

pomanjkljivostim sedanji koncept ogljičnega odtisa že pomeni razgradnja ozonskega sloja, kisli dež, evtrofikacija, emisije

osnovo za nove zakonodajne ter administrativne ukrepe v trdnih delcev, ekotoksičnost, raba vode, izčrpavanje mineralnih

gospodarstvu (politične zahteve po ogljični nevtralnosti, surovin idr. Kar je posebej pomembno, je dejstvo, da so za

obdavčitve, poročanje). Pojavljajo pa se tudi konkretni predlogi, ovrednotenje oz. preračun vplivov emisij in drugih snovi na

da bi sedanji koncept ogljičnega odtisa morebiti uporabil kot okolje ter zdravje uporabljeni najnovejši in znanstveno podprti

mehanizem dodatnega pritiska oz. kontrole državljanov v okviru modeli, kar daje metodi LCA dodatno veljavo glede

ideje o določitvi ogljičnih kreditov posameznika (kreditnih točk) verodostojnosti rezultatov. Zato s pomočjo metode LCA podjetja

kot del ideje o socialnem točkovanju obnašanja oz. aktivnosti ne pridobijo le podatkov, kakšni so vplivi v posameznih fazah

ljudi v povezavi z neposredno in posredno povzročenimi življenjskih ciklov, ampak tudi podatke o vplivih na okolje, ki se

emisijami CO2 [13-15]. In to ne glede na dejstvo, da odstotek jih z drugimi metodami ne da določiti.

globalne elite s svojimi poslovnimi odločitvami ter načinom

življenja povzroča vrtoglave količine emisij TGP [16]. Ker metoda LCA temelji na okoljskem življenjskem ciklu, se

Medsebojni preplet omenjenih vidikov shematsko prikazuje slika lahko s celostnim pristopom obravnavanja vplivov proizvoda na

1. okolje v večji meri zagotovi, da se identificirajo najbolj

relevantni vplivi na okolje v celotnem proizvodno-



distribucijskem sistemu in ne nazadnje, da se prepreči, da bi se

okoljski vplivi "prestavljali" iz ene faze življenjskega cikla v

drugo in obratno (npr. pri menjavi materalov, sestavnih delov,

energentov idr.), saj lahko spremembe z izračunom preverimo

vnaprej. Poudariti velja, da je metoda LCA edina mednarodno

standardizirana metoda za ovrednotenje vplivov izdelkov na

okolje, njen metodološki okvir pa definirajo mednarodni

standardi družine ISO 14040 [17].



Ogljični odtis je vselej del LCA analiz, kar pomeni, da ima

določevanje vplivov izdelkov in storitev na okolje s pomočjo

LCA metode številne prednosti pred določevanjem samo

ogljičnega odtisa, kar žal počnejo slovenska podjetja in v kar jih

z zakonskimi in podzakonskimi akti sili politika. Ogljični odtis

izdelka dobi v okviru LCA analiz povsem drugo perspektivo, saj

ga primerjamo z ostalimi okoljskimi indikatorji oz. vplivi na



Slika 1: Ogljični odtis – preplet različnih vidikov za njegovo okolje. In prav analize LCA nam dokazujejo, da ogljični odtis ni

razumevanje vedno nujno najbolj vplivni okoljski indikatorji v življenjskih

ciklih izdelkov v primerjavi z drugimi okoljskimi indikatorji

[9,18]. Če se upošteva celotni okoljski življenjski cikel izdelkov

in ne le faza uporabe, se tudi zelo spremenijo okoljski profili e-

3 vozil ter tehnologij za pridobivanje energije iz obnovljivih virov Metoda LCA – alternative ogljičnemu odtisu

(na primer biogoriv) [19] ali solarnih panelov [20].



Zaradi omenjenih metodoloških in konceptualnih

A vendarle se je potrebno zavedati, da imajo tudi programska

pomanjkljivosti se je potrebno v takih primerih poslužiti drugih,

orodja, s katerimi modeliramo življenjske cikle v okviru LCA

celovitejših metod vplivov izdelkov, storitev ali organizacij na

analiz, vgrajene določene modelne spremenljivke ter

okolje kot je npr. metoda LCA ('Life Cycle Assessment'). S to

poenostavitve, s čimer je možno vplivati na končne rezultate,

kompleksno analitsko metodo določamo vplive izdelkov v

zato takšne študije zahtevajo izkušene in poštene izvajalce, ki

njihovih celotnih okoljskih življenjskih ciklih, ki zajemajo

uporabljajo najsodobnejšo programsko opremo ter podatkovne

pridobivanje surovin, pridobivanje energijskih virov,

baze.

proizvodnjo in distribucijo potrebne energije, proizvodnjo

polizdelkov, izdelkov ter stranskih izdelkov, transport in

Zaradi vsega navedenega je lahko uporaba izključno ogljičnega

distribucijo, učinke med uporabo in alternativne možnosti

odtisa v promocijske namene tudi sredstvo okoljskega zavajanja



Insert Your Title Here Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



(t.i. ''greenwashing'') (ANEC/BEUC/ECOS/EEB, 2009; environmental impact through producers and consumers.

Radonjič, 2016). Science, Vol. 360, June 1st 2018.

[10] Volfand, J. 2022. Praksa različno, trg že zahteva poročanje

o ogljičnem odtisu. ESG Revija za trajnostni razvoj, št.

4 Zaključek 174/175, december 2022.

[11] Marriott, J., Scott Matthews, H. and Hendrickson, C. T.

Za Slovenijo je velika popularnost ogljičnega odtisa in 2010. Impact of Power Generation Mix on Life Cycle

intenzivno osredotočanje politike, medijev ter gospodarstva nanj Assessment and Carbon Footprint Greenhouse Gas Results.

problematično tudi zato, ker bi tudi v primeru popolne ogljične Journal of Industrial Ecology, Vol. 14, pp.919-928

nevtralnosti države to predstavljalo le minoren prispevek h [12] Röös, E., Sundberg, C. and Hansson, P. A. 2010.

globalni bilanci toplogrednih plinov (še posebej, če k temu Uncertainties in the Carbon Footprint of Food Products: A

prištejemo enormne emisije ogljikovega dioksida ter mnogih Case Study of Table Potatoes', International Journal of

škodljivih plinov in delcev zaradi vojaških aktivnosti v svetu) ob LCA, Vol. 15, pp.478-488

istočasnem ignoriranju ali vsaj manjši pozornosti okoljskih [13] Wintour, P. 2006. Miliband plans carbon trading 'credit

problemov, ki pa jih resnično lahko in moramo sami obvladovati, cards' for individuals. The Guardian, 11. December5 2006.

ker neposredno vplivajo na kakovost okolja in zdravja v Dostopno:

Sloveniji. S tem je mišljen predvsem odnos do naših sladkih in [14] World Economic Forum. 2022. 'My Carbon': An approach

morskih voda, do uporabe kemikalij v kmetijstvu, for inclusive and sustainable cities. Dostopno:

izpostavljenosti kemikalijam v izdelkih, varovanje rodovitnih https://www.weforum.org/agenda/2022/09/my-carbon-an-

kmetijskih površin, odnos do gozdov, učinkovito ravnanje z approach-for-inclusive-and-sustainable-cities/

odpadki in zmanjševanje tistih izpustov v zrak, ki imajo lokalne [15] Peters, T. 2022. Alibaba Group creating a 'individual carbon

ali regionalne škodljive učinke, bodisi zaradi lokalne industrijske footprint tracker'. Dostopno: www.rebelnews.com

proizvodnje, bodisi zaradi transporta. Razumeti ogljični odtis v [16] Oxfam. 2022. Carbon Billionaires. The investment

okviru okoljskih politik zato pomeni razumeti tako njegove emissions of the world's richest people. Oxfam

metodološke kot konceptualne pomanjkljivosti. International, Oxford.

[17] SIST EN ISO 14040:2006. Ravnanje z okoljem –

Ocenjevanje življenjskega cikla – Načela in okviri.

Slovenski inštitut za standardizacijo, Ljubljana.



Reference / Literatura [18] DeMonte, M., Padoana, E., & Pozzetto, D. 2005. Alternative

coffee packaging: an analysis from life cycle point of view.



[1] Uradni list Republike Slovenije. 2021. Resolucija o Journal of Food Engineering, 66, 405 – 411. [19] Smith, James. 2010. Biofuels and the Globalization of Risk. dolgoročni podnebni strategiji Republike Slovenije do leta Zed Books. London, New York. 2050. Št.119/2021. [20] Kleijn, R., E. van der Voet, G. R. Kramer, L. van Oers, C. [2] European Commission. 2023. Climate Action – Progress van der Giesen. 2011. Metal requirements of low-carbon Report 2023. Brussels. power generation. Energy , Vol. 36, 5640 – 5648. [3] European Commission. 2024. Commission sets out how to

sustainably capture, store and use carbon to reach climate

eutrality by 2050

[4] European Commission. 2023. Corporate Sustainability

Reporting Directive. Dostopno:

https://finance.ec.europa.eu/regulation-and-

supervision/financial-services-legislation/

[5] ANEC/BEUC/ECOS/EEB (2009) Sizing Up Product Carbon

Footprinting – Joint Position. European Consumer Voice in

Standardization, Brussels.

[6] Radonjič, G. 2015. Reliability of carbon footprint as a decision-

making tool for product development - a critical

review. International journal of environmental technology and

management. Vol. 18, no. 2, str. 99-115.

[7] Laurent A., Olsen, S. I., & Hauschild, M. Z. (2012).

Limitations of carbon footprint as indicator of environmental

sustainability. Environmental Science & Technology, 46,

4100 - 4108.

[8] Radonjič, Gregor. 2016. The adequacy of product labelling by

carbon footprint labels. V: Proceedings of IAC-MEM 2016,

International Academic Conference on Management, Economics

and Marketing in Bratislava, Slovakia, July 6 - 7, 2016. Vestec:

Czech Institute of Academic Education, 2016. Str. 57-72

[9] Poore, J. in T. Nemecek. 2018. Reducing food's



Ali je prispevek slovenske živinoreje k podnebnim



spremembam precenjen?



Is the contribution of Slovenian livestock farming to



climate change overestimated?



Jože Verbič

Oddelek za živinorejo

Kmetijski inštitut Slovenije

Ljubljana, Slovenija

joze.verbic@kis.si.com



POVZETEK impact of persistent emissions of short-lived gasses on the



Za živinorejo so značilni veliki izpusti metana. Metan uvrščamo account the characteristics of short-lived air pollutants (GWP*). climate. Researchers have developed a method that takes into

med kratko obstojne toplogredne pline. Za razliko od CO2, ki se



v ozračju zadrži več stoletij, se metan po navedbah zadnjega response of the Earth's surface temperature to emissions of This method is characterized by a better prediction of the

(šestega) ocenjevalnega poročila Medvladnega odbora za short-lived gasses than GWP100. When comparing the two podnebne spremembe (IPCC) oksidira v približno 12 letih. Za



poročanje Okvirni konvenciji Združenih narodov o spremembi the contribution of Slovenian livestock farming to climate methods, it was found that the GWP100 method overestimates

podnebja (UNFCCC) uporabljamo metodo potencialnega

warming by a factor of 4.6.

globalnega segrevanja za stoletni časovni horizont (GWP100). V



zadnjem času se vse pogosteje pojavljajo kritike, da metoda KEYWORDS GWP100 ni najbolj primerna za oceno kumulativnega učinka

trajnih izpustov kratko obstojnih plinov na podnebje. climate change, livestock farming, methane



Raziskovalci so razvili metodo, ki upošteva lastnosti kratko

obstojnih onesnaževal zraka (GWP*). Za to metodo je značilno,

1 UVOD

da bolje napoveduje odziv temperature zemeljskega površja na

izpuste kratko obstojnih plinov kot GWP V javnosti je splošno razširjeno mnenje, da je intenzivna 100 . S primerjavo obeh

metod je bilo ugotovljeno, da metoda GWP evropska intenzivna živinoreja med najpomembnejšimi vzroki 100 precenjuje

prispevek slovenske živinoreje k segrevanju podnebja za 4,6 podnebnih sprememb. K temu je veliko prispevalo FAO

krat. poročilo Dolga senca živinoreje (Livestock's long shadow) [1].

Najpogosteje izpostavljeno sporočilo, ki so ga mediji

KLJUČNE BESEDE posredovali javnosti je, da je živinoreja glavni povzročitelj

izpustov toplogrednih plinov (TGP) in da je odgovorna za 18 %

podnebne spremembe, živinoreja, metan

izpustov, kar presega izpuste iz prometa. Ob tem sporočilu so

ABSTRACT bile izražene kritike intenzivne evropske živinoreje, zamolčano

pa dejstvo, da se struktura izpustov TGP v Evropi precej

Livestock farming is characterized by high methane emissions. razlikuje od izpustov na globalni ravni, saj je raba fosilnih goriv

Methane is classified as a short-lived greenhouse gas. Unlike v prometu, industriji in v gospodinjstvih nadpovprečna, število

CO2, which remains in the atmosphere for centuries, methane živine pa razmeroma majhno. Zamolčan je bil podatek istega

oxidizes in about 12 years, according to the latest (sixth) poročila [1], da je prispevek Zahodne Evrope in Severne

assessment report of the Intergovernmental Panel on Climate Amerike k skupnim (globalnim) izpustom metana iz prebavil

Change (IPCC). For reporting to the United Nations Framework rejnih živali le nekaj več kot desetina (10,75 %). Le malokdo se

Convention on Climate Change (UNFCCC), we use the 100- zaveda, da je število goved v Afriki (189,6 milijonov) približno

year global warming potential (GWP100) method. Recently, dvakrat večje kot v Evropi (celina) (92,1 milijonov, v glavah

there has been increasing criticism that the GWP100 method is velike živine) [2].

not the most appropriate method to assess the cumulative Za živinorejo so značilni veliki izpusti ne-CO toplogrednih 2

DOI: https://doi.org/10.70314/is.2024.env.7 plinov, predvsem metana. Različni toplogredni plini se v

njihovem učinku na segrevanje podnebja razlikujejo. Za

Permission to make digital or hard copies of part or all of this work for personal or namene poročanja o emisijah TGP jih preračunamo v

classroom use is granted without fee provided that copies are not made or ekvivalente (ekv) ogljikovega dioksida z upoštevanjem

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this potenciala globalnega segrevanja za stoletni časovni horizont

work must be honored. For all other uses, contact the owner/author(s). (Global Warming Potential, GWP ). Po tej metodi je Information Society 2023, 9–13 October 2023, Ljubljana, Slovenia 100

© 2023 Copyright held by the owner/author(s). slovenska živinoreja po zadnjih podatkih (za leto 2022)

prispevala 8,35 % vseh emisij TGP v Sloveniji. K temu so



največ prispevali izpusti metana iz prebavil (6,31 %), sledili so Namen tega prispevka je prikazati, kakšen je prispevek

izpusti metana iz skladišč za živinska gnojila (1,61 %), izpusti slovenske živinoreje k izpustom toplogrednih plinov, če le tega

didušikovega oksida iz skladišč za živinska gnojila (0,27 %) in ocenimo z metodo, ki upošteva obstojnost metana v ozračju

posredni izpusti didušikovega oksida, ki nastanejo zaradi (GWP*). Rezultate smo primerjali z rezultati po metodi, ki je

uhajanja dušikovih spojin iz hlevov in skladišč za živinska splošno uveljavljena za poročanje UNFCCC (GWP100).

gnojila [3]. K izpustom v živinoreji je največ prispevala

govedoreja (92,7 %), sledile so reja prašičev (2,6 %), perutnine

(1,6 %), ovc (1,7 %), koz (0,3 %), konj (0,9 %) in kuncev (manj 2 METODE

kot 0,1 %) [3]. Podatke o izpustih metana in didušikovega oksida za obdobje

Ocenjevanje vpliva človekovih dejavnosti na spremembe 1985-2022 smo zajeli iz emisijskih evidenc Kmetijskega

podnebja na podlagi GWP100 sega v obdobje prvega poročila inštituta Slovenije [3]. Gre za izvorne podatke, ki so uporabljeni

Medvladnega odbora za podnebne spremembe, ki je bilo izdano za poročanje UNFCCC [13]. Za živinorejo smo upoštevali

v letu 1990. Že ob uveljavitvi tega koncepta je bilo prepoznano, izpuste metana iz prebavil rejnih živali, izpuste metana iz

da ima ob prednostih (t.j. predvsem enostavna implementacija), skladišč za živinska gnojila, izpuste didušikovega oksida iz

tudi resne pomanjkljivosti [4]. Od metode za postavitev skladišč za živinska gnojila in posredne izpuste didušikovega

različnih TGP na skupni imenovalec pričakujemo predvsem, da oksida, ki nastanejo zaradi uhajanja dušikovih spojin iz hlevov

bo realno ocenila učinek izpustov teh plinov na spremembe in skladišč za živinska gnojila. Podatki o izpustih metana in

temperature zemeljskega površja. Smith in sod. [5] so opozorili, didušikovega oksida so bili preračunani v ekvivalente CO2 ob

da koncept GWP100 slabo oceni odziv temperature na upoštevanju potencialov globalnega segrevanja za stoletno

kumulativne izpuste kratko obstojnih toplogrednih plinov, med časovno obdobje (GWP100), kot je trenutno v veljavi za

katerimi je tudi metan. Za razliko od CO2, ki ostane v ozračju poročanje UNFCCC (faktor 28 za metan in faktor 265 za

več stoletij, se metan relativno hitro oksidira (približno 12 let) didušikov oksid). Podatke o izpustih TGP iz živinoreje smo

[6]. Koncept GWP100 tega ne upošteva. Za realnejšo oceno preračunali v ekvivalente CO2 tudi z alternativno metodo

vpliva izpustov metana na podnebje so raziskovalci nedavno GWP*. Pri tem smo za didušikov oksid uporabili enak postopek

razvili nov način za preračun izpustov različnih toplogrednih kot za GWP100, za metan pa metodo po Cain in sod. [7].

plinov na skupni imenovalec, ki upoštevajo, da je metan v

atmosferi kratko obstojen (GWP*) [7, 8, 9].

Za dolgo obstojne TGP je značilno, da se v ozračju 3 REZULTATI IN RAZPRAVA

akumulirajo. Vsak enkratni izpust (npr. izpust v tekočem letu) Podatki o izpustih TGP v slovenski živinoreji so predstavljeni

poveča koncentracijo plina v ozračju in četudi takoj prenehamo na Grafu 1. Za GWP100 razpolagamo s podatki za obdobje

z novimi izpusti, se koncentracije tega plina ne bodo zmanjšale. 1985-2022. Za GWP* razpolagamo le s podatki po letu 2005.

Vsak enkratni izpust dolgo obstojnega TGP pomeni dodatno Ker je ocena po GWP* odvisna tudi od preteklih izpustov,

segrevanje, ki ga je treba prišteti k učinku vseh izpustov tega podatkov pred tem letom ni mogoče izračunati. Primerjava

plina v zgodovini. Pri metanu je drugače. Koncentracija v kaže, da je prispevek slovenske živinoreje k podnebnim

ozračju ni odvisna le od izpustov v tekočem letu, ampak tudi od spremembam z metodo GWP100 zelo precenjen. V obdobju

obsega oksidacije metana, ki smo ga v ozračje izpustili v 2005-2022 so se izpusti gibali od 1318 do 1432 kT ekv CO2. Po

preteklosti. V ozračju se sproti vzpostavlja novo ravnovesje. Če metodi GWP* smo imeli v znatnem delu tega obdobja

letne izpuste ohranjamo na ravni iz preteklosti, se koncentracije negativne izpuste (obdobje 2005-2010, 2012, 2022). Za leta z

ne spreminjajo. Če letne izpuste povečujemo, se povečujejo tudi negativnimi izpusti je bilo ocenjeno, da je obseg oksidacije

koncentracije v ozračju in če jih zmanjšujemo se koncentracije metana iz preteklih izpustov presegel izpuste iz tega leta in da

zmanjšajo. GWP* to upošteva in je razmeroma dober približek je razlika izničila tudi izpuste didušikovega oksida. Povprečje

podnebnemu modelu. Na drugi strani pa GWP100 v primeru za obdobje 2018-2022 kaže, da metoda GWP100 precenjuje

vzdržnih ali padajočih izpustov metana prikriva napredek proti prispevek slovenske živinoreje k segrevanju podnebja za 4,6

podnebnemu cilju, ki ga Pariški sporazum določa v C glede na krat (izpusti po GWP100 in GWP* so bili v povprečju 1373 in

predindustrijsko dobo [10]. 299 kT ekv CO2 na leto). Metoda preračunavanja posameznih

Problematika ocenjevanja izpustov metana na podnebje je bolj TGP zelo vpliva tudi na strukturo izpustov. Po metodi GWP100

obsežno in poglobljeno obravnavana v zadnjem (šestem) je prispevek slovenske živinoreje k skupnim izpustom (vsi

ocenjevalnem poročilu Medvladnega odbora za podnebne sektorji) ocenjen na 8,33 %, po GWP* pa na 2,40 % (ocena za

spremembe. Ugotavljajo, da metoda GWP100, ki jo uporabljamo obdobje 2018-2022, pri čemer so bili pri skupnih izpustih za

za poročanje Okvirni konvenciji Združenih narodov o GWP* korigirani tudi izpusti metana iz drugih sektorjev).

spremembi podnebja (UNFCCC), ni najbolj primerna za oceno Vedeti moramo, da k ugodnemu rezultatu za živinorejo po

kumulativnega učinka trajnih izpustov kratko obstojnih plinov GWP* prispeva predvsem razmeroma uspešno zmanjševanje

na podnebje [11]. Medvladni odbor v zadnjem poročilu ni izpustov metana v preteklem obdobju. V primeru hitrega

priporočil nobene od obravnavanih metodik za preračun povečevanja izpustov namreč ocene po GWP* presegajo ocene

izpustov različnih TGP na skupni imenovalec [6]. Zapisali so, po GWP100. V Sloveniji na to kaže primer reje drobnice, kjer so

da je izbor najprimernejše metrike odvisen od namena bili v obdobju 2005-2015 izpusti po GWP* dva do trikrat večji

primerjave. Na 27. Konferenci pogodbenic Okvirne konvencije kot po GWP100 [14]. Združenih narodov o spremembi podnebja v Sharm El–Sheikhu

je bilo izglasovano, da lahko pogodbenice ob GWP100 dodatno

poročajo tudi na podlagi alternativnih metod, kot je GWP* [12].



2000 Sklenemo lahko, da splošno uveljavljena metoda obračunavanja

) ekvivalentov CO2 močno precenjuje učinek slovenske

O 2 živinoreje na segrevanje podnebja. Z nadaljnjim

C 1500

zmanjševanjem izpustov metana v živinoreji lahko sicer

1000 prispevamo k ohlajanju podnebja, pri tem pa je potencial

(kT ekv GWP100 evropske živinoreje zaradi zmanjšanja emisij v preteklosti in



TGP zaradi obsežnih drugih virov razmeroma majhen. GWP* sti 500 u

Izp

0 4 ZAHVALA

1982 1987 1992 1997 2002 2007 2012 2017 2022

Prispevek je bil pripravljen v sklopu projekta V5-2366

-500 »Ovrednotenje dejavnikov za učinkovito zniževanje emisij

Graf 1: Primerjava metod GWP100 in GWP* za preračun toplogrednih plinov«, ki ga sofinancirata Agencija za

izpustov toplogrednih plinov v slovenski živinoreji v raziskovalno dejavnost Republike Slovenije (ARIS) in

ekvivalente CO2. Prikaz zajema tako izpuste metana kot Ministrstvo za okolje, podnebje in energijo Republike Slovenije

didušikovega oksida. (MOPE) in v sklopu raziskovalnega programa »Trajnostno

kmetijstvo« (P4-0133), ki ga sofinancira Agencija za

Nov način obračunavanja toplogrednega učinka metana ne raziskovalno dejavnost Republike Slovenije (ARIS) iz

pomeni, da so prizadevanja za zmanjšanje izpustov nepotrebna. državnega proračuna.

Metan je zelo potenten toplogredni plin. V kolikor se zaradi

povečevanja izpustov koncentracije v atmosferi povečujejo, se

podnebje segreva hitreje, kot kaže GWP 5 LITERATURA 100 . Na drugi strani pa

lahko z zmanjšanjem izpustov metana zmanjšamo koncentracije [1] FAO, 2006. Livestock's long shadow. Environmental issues and options.

v ozračju in s tem planet ohladimo. Evropa je ta potencial v FAO, 390 s. Retrieved September 4, 2023 from

precejšnjem obsegu že izkoristila, saj je edini kontinent, na https://www.fao.org/4/a0701e/a0701e00.htm.

[2] FAO, 2024. FAOSTAT Statistical Database. Food and agriculture data.

katerem se izpusti metana iz živinoreje zmanjšujejo že nekaj Retrieved July 17, 2024 from https://www.fao.org/faostat/en/#home.

desetletij [15]. Za Nemčijo imamo tudi ocene za čas izpred več [3] KIS (2024) Emisije metana, didušikovega oksida, ogljikovega dioksida,

kot 100 let. Izpusti metana iz prebavil rejnih živali so že od leta amonijaka, NMVOC, NOX, PM10 in PM2,5 v kmetijstvu. Emisijske

evidence, Ljubljana, Kmetijski inštitut Slovenije. Retrieved August 16,

2003 pod ravnjo iz leta 1892 [16]. To pomeni, da je bil 2024 (internal data).

prispevek nemške živinoreje h koncentracijam metana v ozračju [4] Jan S. Fuglestvedt, Terje K. Berntsen, Odd Godal and Tora Skodvin, 2003.



pred 130 leti večji kot sedaj. Potencial za nadaljnje Metrics of Climate Change: Assessing Radiative Forcing and Emission Indices. Clim. Change 58, 3, 267-331. doi: zmanjševanje emisij iz evropske živinoreje obstaja, je pa njegov http://dx.doi.org/10.1023/A:1023905326842 . pomen za podnebne spremembe razmeroma majhen. Na [5] Stephen M. Smith, Jason A. Lowe, Niel H. A. Bowerman, Laila K. Gohar,

globalni ravni prispevajo naravni viri 50,5 % vsega metana, Chris Huntingford and Myles R. Allen, 2012. Equivalence of greenhouse-

fosilna goriva 17,4 %, živinoreja 15,1 %, deponije in odpadki gas emissions for peak temperature limits. Nat. Clim. Change 2, 535–538.

8,9 %, gojenje riža 4,1 %, sežiganje biomase 2,3 % in raba DOI: https://doi.org/10.1038/nclimate1496.

[6] IPCC, 2021. Climate Change 2021: The Physical Science Basis.

biogoriv 1,6 % [15]. Prispevek evropske živinoreje k skupnim Contribution of Working Group I to the Sixth Assessment Report of the

globalnim izpustom je 1,5 % [15] in za toliko bi se zmanjšala Intergovernmental Panel on Climate Change. Valérie Masson-Delmotte,

koncentracija metana v ozračju, če bi ukinili vso evropsko Panmao Zhai, Anna Pirani, Sarah L. Connors, Clotilde Péan, Sophie

živinorejo. Mengtian Huang, Katherine Leitzell, Elisabeth Lonnoy, Robin J.B. Berger, Nada Caud, Yang Chen, Leah Goldfarb, Melissa I. Gomis,

Uporaba metode GWP* je deležna precej pozornosti tako v Matthews, Thomas K. Maycock, Tim Waterfield, Ozge Yelekçi, Rong Yu,



znanstveni in strokovni literaturi, kot tudi v javnih medijih. Ob and Baiguan Zhou editors. Cambridge University Press, Cambridge United Kingdom and New York USA, 2391 pp. DOI: izpostavljanju njenih prednosti je deležna tudi kritik glede http://dx.doi.org/10.1017/9781009157896 .

pravičnosti in poštenosti [17] in zlorab v smislu prikazovanja [7] Myles R. Allen, Keith P. Shine, Jan S. Fuglestvedt, Richard J. Millar,

podnebne nevtralnosti živinoreje [18]. Do držav, ki so v Michelle Cain, David J. Frame and Adrian H. Macey, 2018. A solution to

preteklosti izkazovale velike izpuste in so te izpuste v zadnjih the misrepresentations of CO2-equivalent emissions of short-lived climate

20 letih zmanjšale, je metodika GWP* kljub velikim izpustom 2018). DOI: pollutants under ambitious mitigation. npj Clim. Atmos. Sci. 1, 16, (Jun

https://doi.org/10.1038/s41612-018-0026-8.

metana na prebivalca zelo prizanesljiva [17]. To sicer velja tudi [8] Michelle Cain, John Lynch, Myles R. Allen, Jan S. Fuglestvedt, David J.

za GWP Frame and Adrian H Macey, 2019. Improved calculation of warming- 100 saj so praktično vse mednarodne zaveze na področju

blaženja podnebnih sprememb določene v % glede na equivalent emissions for short-lived climate pollutants. npj Clim. Atmos.

izhodiščno leto. Res pa je, da pride zaradi kratke obstojnosti in [9] Sci. 2, 29. DOI: https://doi.org/10.1038/s41612-019-0086-4. Myles Allen, Katsumasa Tanaka, Adrian Macey, Michelle Cain, Stuart

s tem specifičnega izračuna GWP*, vsako zmanjšanje izpustov Jenkins, John Lynch and Matthew Smith, 2021. Ensuring that offsets and

metana bolj do izraza, kot kažejo absolutne vrednosti izpustov. other internationally transferred mitigation outcomes contribute effectively

Pri morebitnem omenjanju podnebne nevtralnosti živinoreje to limiting global warming. Environ. Res. Lett. 16, (Number 7), 074009.

DOI: https://doi.org/10.1088/1748-9326/abfcf9.

moramo biti precizni in navesti, ali se le ta nanaša le na metan, [10] Michelle Cain, Stuart Jenkins, Myles R. Allen,John Lynch, David J.

ali pa tudi na didušikov oksid (kot na grafu 1) ali pa tudi na Frame, Adrian H. Macey and Glen P. Peters, 2021. Methane and the Paris

druge z živinorejo povezane toplogredne pline, ki jih UNFCCC Agreement temperature goals. Phil. Trans. R. Soc. A 380:20200456. DOI:

ne poročamo v sklopu sektorja živinoreja. https://doi.org/10.1098/rsta.2020.0456.

[11] IPCC, 2022. Climate Change 2022: Mitigation of Climate Change.

Contribution of Working Group III to the Sixth Assessment Report of the



Intergovernmental Panel on Climate Change. Priyadarshi R. Shukla, Jim

Skea, Raphael Slade, Alaa Al Khourdajie, Renée van Diemen, David

McCollum, Minal Pathak, Shreya Some, Purvi Vyas, Roger Fradera,

Malek Belkacemi, Apoorva Hasija, Géninha Lisboa, Sigourney Luz and

Juliette Malley editors. Cambridge University Press, Cambridge, UK and

New York USA, 2029 pp. DOI:

http://dx.doi.org/10.1017/9781009157926.002.

[12] UNFCCC, 2023. Report on the Conference of the Parties on its twenty-

seventh session, held in Sharm el-Sheikh from 6 to 20 November 2022.

Addendum. Part two: Action taken by the Conference of the Parties at its

twenty-seventh session. Retrieved February 26, 2024 from

https://unfccc.int/documents/626564.

[13] Tajda Mekinda Majaron, Martina Logar, Jože Verbič, Žan Pečnik, Boštjan

Mali, Gal Kušar and Romana Stare, 2024. Slovenia's national inventory

report 2024 : GHG emissions inventories 1986-2022: submitted under the

United nations framework convention on climate change. Ljubljana:

Ministry of the Environment and Spatial Planning: Slovenian Environment

Agency, 2024, 395 p. Retrieved March 20, 2024 from

https://cdr.eionet.europa.eu/si/eu/govreg/inventory/envze8p9g/SVN_NIR_

2024_March.pdf

[14] Jože Verbič, Marko Bizjak and Mojca Simčič, 2023. Izpusti toplogrednih

plinov pri reji drobnice in možnosti za njihovo zmanjšanje. Marjana Cvirn

editor. 7. strokovni posvet Reja drobnice, Slovenj Gradec, Slovenia,

Kmetijska založba, 17-26.

[15] Marielle Saunois, Ann R. Stavert, Ben Poulter, Philippe Bousquet, Josep

G. Canadell, Robert B. Jackson, Peter A. Raymond, Edward J.

Dlugokencky, Sander Houweling, Prabir K. Patra, Philippe Ciais, Vivek

K. Arora, David Bastviken, Peter Bergamaschi, Donald R. Blake, Gordon

Brailsford, Lori Bruhwiler, Kimberly M. Carlson, Mark Carrol, Simona

Castaldi, Naveen Chandra, Cyril Crevoisier, Patrick M. Crill, Kristofer

Covey, Charles L. Curry, Giuseppe Etiope, Christian Frankenberg, Nicola

Gedney, Michaela I. Hegglin, Lena Höglund-Isaksson, Gustaf Hugelius,

Misa Ishizawa, Akihiko Ito, Greet Janssens-Maenhout, Katherine M.

Jensen, Fortunat Joos, Thomas Kleinen, Paul B. Krummel, Ray L.

Langenfelds, Goulven G. Laruelle, Licheng Liu, Toshinobu Machida,

Shamil Maksyutov, Kyle C. McDonald, Joe McNorton, Paul A. Miller,

Joe R. Melton, Isamu Morino, Jurek Müller, Fabiola Murguia-Flores,

Vaishali Naik, Yosuke Niwa, Sergio Noce, Simon O'Doherty, Robert J.

Parker, Changhui Peng, Shushi Peng, Glen P. Peters, Catherine Prigent,

Ronald Prinn, Michel Ramonet, Pierre Regnier, William J. Riley, Judith

A. Rosentreter, Arjo Segers, Isobel J. Simpson, Hao Shi, Steven J. Smith,

L. Paul Steele, Brett F. Thornton, Hanqin Tian, Yasunori Tohjima,

Francesco N. Tubiello, Aki Tsuruta, Nicolas Viovy, Apostolos

Voulgarakis, Thomas S. Weber, Michiel van Weele, Guido R. van der

Werf, Ray F. Weiss, Doug Worthy, Debra Wunch, Yi Yin, Yukio

Yoshida, Wenxin Zhang, Zhen Zhang, Yuanhong Zhao, Bo Zheng, Qing

Zhu, Qiuan Zhu and Qianlai Zhuang, 2020. The Global Methane Budget

2000–2017, Earth Syst. Sci. Data 12, 1561–1623. DOI:

https://doi.org/10.5194/essd-12-1561-2020.

[16] Björn Kuhla and Gunther Viereck, 2022. Enteric methane emission

factors, total emissions and intensities from Germany's livestock in the late

19th century: A comparison with the today's emission rates and intensities.

Sci. Total Environ. 848, 157754, (Nov. 2022). DOI:

http://dx.doi.org/10.1016/j.scitotenv.2022.157754.

[17] Joeri Rogelj and Carl-Friedrich Schleussner, 2019. Unintentional

unfairness when applying new greenhouse gas emissions metrics at

country level. Environ. Res. Lett. 14, 114039, (Nov. 2019). DOI:

https://doi.org/10.1088/1748-9326/ab4928

[18] Caspar L Donnison and Donal Murphy-Bokern, 2024. Are climate

neutrality claims in the livestock sector too good to be true? Environ. Res.

Lett. 19, 011001, (Dec. 2023). DOI: https://doi.org/10.1088/1748-

9326/ad0f75



Oblikovanje specifičnih habitatov kot ukrepanje za



zmanjšanje vplivov hidroelektrarn na okolje



Design of specific habitats as a measure to reduce the



impacts of hydropower plants on the environment



Andraž Hribar

Služba za gradbeništvo in okolje

HESS, d. o. o., Brežice, Slovenija

andraz.hribar@he-ss.si



Povzetek je veliko, trenutno v fazi realizacije je projekt ureditve brežine

Kompolje an pretočni akumulaciji HE Boštanj.

Zajezba na vodotoku je tisti del vsake hidroelektrarne, ki

predstavlja največjo spremembo v prostoru in ima velik vpliv na Ključne besede

okolje v času obratovanja HE. Sodobni pristopi k umeščanju Zmanjšanje vplivov energetskih objektov, večanje okoljske

akumulacij zahtevajo celovit in multidisciplinaren pristop ter vrednosti, vzdrževanje z dodano vrednostjo, ekološka gradnja,

vključitev različnih deležnikov. Vključevanje širokega kroga sonaraven razvoj strokovnjakov iz različnih področij po eni strani odpre veliko

vprašanj, po drugi strani pa omogoči takšno ukrepanje, ki Abstract

zmanjša vpliv ob umestitvi objekta in tudi v času obratovanja. Basin of each Hydropower plant represents the biggest spatial

Primer iz prakse so pretočne akumulacije na spodnji Savi, kjer so change and has the biggest impact on the environment during the

se v času gradnje uredili nadomestni habitati, v času obratovanja operation of the HPP. Modern approaches to the implementation

pa se na podlagi monitoringov in sodelovanja s strokovnjaki s of the basin require a comprehensive and multidisciplinary

področja ekologije, biologije in drugih naravoslovnih ved zbirajo approach and the involvement of various stakeholders. The

podatki in oblikujejo nova znanja o prostoru. To omogoča involvement of a wide range of experts from different fields

prilagajanje spremembam v naravi, ki so neizbežne: ravno narava opens up many questions. On the other hand it enables such

je tista, ki daje prostoru dinamiko. Statični objekti, kakršne naj action that minimizes the impact while building the HPP and also

bi bile hidroelektrarne, se morajo prilagajati tej dinamiki z during operation. On the lower Sava several examples of good

novimi ukrepi, ki ne nasprotujejo naravi temveč sprejemajo nove practice can be seen. During construction substitute habitats were

naravne danosti in gradijo na prednostih, ki jih pretočne arranged and during operation data is collected and new

akumulacije nudijo. Vzdrževanje hidroelektrarn naj bi bila knowledge about ecosystem is developed based on monitoring

načeloma rutinska aktivnost. Vendar nova spoznanja v and cooperation with experts in the field of ecology, biology and

migracijah organizmov (na primer ribe), prihod novih vrst v other natural sciences. This is the basis of adaptation to the

prostor (na primer bober), širjenje novih tujerodnih vrst (trikotna changes in nature, which are imminent. Facilities, such as HPP

ladjica), spremembe v hidrološkem režimu reke (prehod Save iz are designed to be static but that is not how nature works. We

snežno-dežnega režima v dežni režim) in klimatske spremembe must adapt to dynamic dictated by nature with new measures that

zahtevajo razvoj in prilagajanje. Vsi novi ukrepi se dandanes do not oppose but accept new natural conditions and build on the

morajo izpeljati po načelih krožnega gospodarstva, upoštevati pa advantages that HPP basins offer. Maintenance of hydroelectric

je potrebno tudi spremljanje ogljičnega odtisa, kar je pri power plants is supposed to be a routine activity in principle.

aktivnostih vezanih na sedimente vse pomembnejše vprašanje. V However, new findings in the migration of organisms (for

družbi HESS imamo postavljen dvojni sistem: spremljanje stanja example different fish species), the arrival of new species in the

pretočnih akumulacij v našem upravljanju na eni strani in area (for example beaver), the spread of new non-native species

oblikovanje nabora zaželenih ukrepov za povečanje okoljske in (triangle boat), changes in the hydrological regime of the river

družbene vrednosti. Sistem omogoča primerjavo obeh sklopov in (the transition of the Sava from a snow-rain regime to a rain

ukrepanje na način, da se ob vsaki sanaciji ali modifikaciji doseže regime) and climate change requires development and adaptation.

dodana vrednost. Primerov tovrstnega ukrepanja na spodnji Savi All new measures nowadays must be carried out according to the

principles of the circular economy, also the carbon footprint must

∗ https://doi.org/10.70314/is.2024.env.8 be taken into account, which is an increasingly important issue

for activities related to sediments. At HESS, we have a dual

Permission to make digital or hard copies of part or all of this work for personal or system when designing modifications of the HPP basins: on the

classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full one hand, monitoring of the basins. On the other hand creating a

citation on the first page. Copyrights for third-party components of this work must set of desirable measures to increase environmental and social

be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia value. This enables the comparison of the two data sets and

© 2024 Copyright held by the owner/author(s). achieving added value with each renovation or modification.

There are many examples of good practices on the lower Sava 2 Primeri »naravnih« habitatov river developed by this principle. Currently the project to

Preselimo se bližje nam, v Slovenijo, na spodnjo Savo. V

improve the bank of Kompolje at the Boštanj HPP is in the phase

trenutku umeščanja HE Brežice v prostor so strokovnjaki

of development.

ustvarjali rešitve (prostorske ureditve), ki bi v naravi ustvarile



Keywords določene pogoje in prostor za življenje določenih organizmov.

Pri tem so se strokovnjaki v veliki meri naslanjali na elemente,



Reducing the impact of energy facilities, increasing the ki jih je ustvarila narava in so (bili) domovanje številnih organizmov: Poplavni gozd imenovan Vrbina pri Brežicah, environmental value, maintenance with added value, eco- gramozne jame, suhi travniki, mlake in jezera. Pa jih je res development, sustainable development

ustvarila narava?





1 Uvod

Razmišljanje o varovanju narave ob umeščanju novega objekta

ali nove ureditve v prostor je dejansko primerjava med tistim, kar

je, in tistim, kar bo. Zatakne se že pri prvem delu: tistim, kar je,

oziroma povedano drugače, stanje, v katerem je narava v tem

trenutku. Ali je stanje v tem trenutku res tisto, kar želimo

ohranjati? Če govorimo o rekah, varstveniki v Srednji Evropi

pogosto štejejo, da je bilo »optimalno naravno« stanje rek v

obdobju med 1800 in 1850, ko je bila »dosežena« največja vrstna

pestrost v biološkem smislu [1], [2]. Če upoštevamo podatek, da

je po ocenah v Srednji Evropi tako ali drugače degradiranih 95%

vodotokov, se dandanes torej neizogibno že v štartu vedno

pogovarjamo zgolj o tem, kako zelo je stanje degradirano. Ali pri

spreminjanju (umeščanju novega objekta oz. ureditve) torej

resnično želimo ohraniti to, kar je, se pravi degradirano stanje?

Vsaj nekateri bodo rekli, da ne. Želimo si več, nekaj boljšega. Slika 1: Varovana območja in prikaz infrastrukturnih

Tudi ostali, bolj previdni, se bodo verjetno strinjali, da je nekaj ureditev HE Brežice [5]

boljšega vsekakor zaželeno.

V poplavnem gozdu Vrbina pri Brežicah rastejo v večini topoli,

Drugi del uvodnega stavka »kar bo« se nanaša na načrtovanje in ki so jih zasejali od leta 1952 naprej za potrebe proizvodnje

predvidevanje, v kakšnem stanju bo narava po izgradnji oz. po papirja v Tovarni papirja Krško. Gre za v vrste zasajene topole,

ureditvi. Sodobni pristop je vključevanje širokega nabora strok celoten nasad, na začetku velik 17 ha, v letih največjega obsega

in znanj, s ciljem nedvoumno zapisati kaj in kako mora biti. pa je pokrival 555 ha. Razdeljen je bil v kvadrante in ločen s

Nesporno danes vemo veliko. Znanstveni pristop je temelj pravokotnimi gozdnimi cestami. Posamezne kvadrante so

današnjega razvoja, zato naj bi mu sledili vsi naravoslovni izsekavali z golosekom, ko je bila lesna masa zrela, in zasajevali

strokovnjaki (tako inženirji kot biologi). Vendar znanstveni s sadikami, ki so jih gojili v bližnji drevesnici. Da so ustvarili

pristop od nas zahteva tudi pogled na celotno sliko (ne prostor za nasad topolov, so izkrčili dotedanji poplavni gozd, ki

spreglejmo gozda pri proučevanju listov). Zanimiv primer je ga je pred tem sestavljal dob (vrsta hrasta). Dob je veličastno

širjenje pristanišča v Singapurju. Habitati in z njimi celoten drevo, visoko do 40m in s premerom debla tudi več kot 2m.

ekosistem so seveda popolnoma drugačni kot v Sloveniji, posegi Topol je hitro rastoče drevo, ki hitro nabira lesno maso in je po

in posledice kot proces ter kvaliteta in strokovnost napovedi pa 8-10 letih pri debelini debla 20 cm že primerno za posek [4].

so poučne. Singapur s pristaniščem je eno najhitreje razvijajočih Sprememba iz naravnega dobovega poplavnega gozda v topolov

se območij na svetu in je v direktni koliziji s koralnim grebenom, nasad je nedvomno pomenila izredno osiromašenje ekosistema.

ki je nekoč pokrival več kot 100 km2. Niti nasprotniki uničevanja Šlo je za industrijski poseg brez okoljevarstvenega razmisleka.

koralnega grebena niti strokovnjaki na investitorjevi strani pa Zadnje zasaditve so se izvajale leta 2009, od takrat naprej pa

niso pričakovali, da se bodo korale in z njimi celotni koralni nasad »propada«. Propada seveda le z vidika industrije in

ekosistem pojavil na podvodnih pilotih pomolov [3]. Ko so proizvodnje lesne mase, z vidika naravovarstva pa se je takrat

ugotovili razcvet narave na podvodne stebre, je bilo pričelo uravnoteženje ekosistema z naravnim procesom

retrospektivno logično, da se bo to zgodilo: gre za ogromno naseljevanja drugih organizmov in upadanjem monokulture

novih površin, točno na globini z dobro osvetljenostjo in pod (topola), ki je z vidika narave nevzdržna. V teh razmerah so se

vplivom neurij, ter za zaščiteno območje, na katero plovba ne na območju naselili nekateri organizmi, med drugimi tudi hrošč

vpliva (pod pomoli). Nekaj, kar je iz retrospektivnega vidika škrlatni kuku, ki je v odmirajočem lubju dreves našel zanj bogat

samoumevno, ni znal (pred tem) nihče napovedati. življenjski prostor. Rezime: Pri gradnji HE Brežice smo varovali

Znanstveno se je vprašati: Česa vse še ne vemo? umetno ustvarjen habitat, za katerega se ob nastanku ni smatralo,

da bo imel kakšno naravovarstveno vrednost (nasad

monokulture), pri čemer se varuje takšno stanje habitata, ki je iz

naravnega vidika nevzdržno (naravno nevzdržna prevlada enega

tipa habitata, ki je na kratek rok povzročil ekspanzijo določenega predstavljajo nove vodne površine kot nadomestilo za

tipa organizma). nekdanje gramoznice. 5.3 je NH3, ki v južnem delu

predstavlja nove gozdne površine, v severnem delu pa novi

Reka Sava je pred izgradnjo HE Brežice predstavljala cca 90m suhi travnik. 5.6 je NH6, kjer se nahaja jezero/mlaka za

široko strugo v obliki trapeznega korita z utrjenimi brežinami in želvo sklednico.



ponekod omejeno z varovalnimi nasipi, kjer ni bilo ogroženih Rezime: Gramoznica je škrbina v okolju (za vsako gramoznico poplavnim območjem okoli struge. Poplavno območje je bilo

objektov, pa se je prosto razlivala. Gramozne jame oz. se mora v sklopu pridobitve koncesije izdelati tudi projekt



pod gladino podtalnice, gre v naravi za jezera. Navadno se obravnavajo kot poseg v okolje brez kakršne koli ekološke vrednosti. V primeru gradnje pretočne akumulacije HE Brežice, gramoznica širi v eno stran, tako da se fronta izkopa premika, gramoznice so jame, iz katerih se izkopava gramoz. Ker segajo sanacije). Ob podeljevanju koncesij za izkop se gramoznice



preostale brežine pa se sčasoma zarastejo in postanejo s katero se je potopilo nekaj gramoznic, pa se je te “škrbine” moralo nadomestiti, tako da so se zgradile nove, nadomestne domovanje za številne organizme. Umetna jama torej s časoma prodne stene nad vodno površino, s katerimi se je simuliralo ustvari različne tipe habitatov; od strmih brežin do položnejših, habitate v gramoznicah. od golih do zaraščenih, od čiste prodnate obale do korenin in

mulja v plitvini. S poplavljanjem reke se v ta vodna telesa



naselijo ribe, poplavno premeščanje sedimenta pa premesti tudi 3 Primeri »umetnih« habitatov organizme iz dna vodotoka, ki predstavljajo hrano ribam in

dvoživkam. Dvoživke s selitvijo najdejo nove vodne površine, Je habitat lahko umeten? Habitat je življenjski prostor, torej

ptiči in mrčes se naselijo še hitreje. Osredotočimo se na ptice, nekaj naravnega. Če ga zgradi človek, je lahko to naravno?

posebej ogrožene vrste lastovke breguljke in čebelarje. Te ptice Bistvo narave so procesi, procesi ustvarjajo spremembe, ravno

vsako leto ob selitvi iz Afrike poiščejo sveže erodirano območje spremembe pa so tista dinamika, ki ustvarja pestrost habitatov,

ob reki, kjer je voda ob zadnjih spomladanskih poplavah odnesla kot na primer različne globine vode, različne brežine gramoznic,

breg in ustvarila strmo prodnato brežino. V brežini izkopljejo rov različna kvaliteta prsti, različna granulacija sedimenta. Te razlike

v produ, strma brežina nad vodo pa jih varuje, da do gnezd ne omogočajo različnim organizmom lokalne pogoje za življenje,

pridejo kače in lisice. Reka Sava je prenehala odnašati brežine zato je ta dinamika tako pomembna. Ko gradi človek, navadno

zaradi stoletja ukalupljanja v trapezoidno strugo z utrjenimi gradi nekaj fiksnega, nekaj, kar bo stalo čvrsto skozi celotno

brežinami. Klimaks ukalupljanja se je odvil med leti 1850 in življenjsko dobo. Vsak objekt mora izpolnjevati bistvene

1950 z gradnjo južne avstroogrske železnice (odsek Zidani Most lastnosti. Je to združljivo z naravo?

– Sisak), z gradnjo cest (ob spodnji Savi je potekala že rimska

cesta med Celeio (Celje) in Neviodunum (Drnovo pri Krškem, V prejšnjem poglavju opisano ukalupljanje reke je Savi

kjer je bilo pomembno pristanišče), s širitvijo mest in razvojem onemogočilo premikanje struge. Kot že omenjeno, je bila reka

industrije (Radeče, Sevnica, Krško), ne gre pa zanemariti niti Sava pred izgradnjo HE Brežice široka cca 90m. Kartografski

vpliva širjenja obdelovalnih površin. Tudi v času Jugoslavije se arhivi avstro-ogrskih katastrov pa na območju Brežic prikazujejo

je ukalupljanje nadaljevalo, kar je lepo razvidno iz katastra reko, katere glavna struga je široka 500m, zraven nje pa več

vodnogospodarskih ureditev. Rezultat ukalupljanja pa je stranskih strug, rokavov, meandrov in drugih oblik (okljuki,

fiksirana brežina, ki je voda ne erodira in ne ustvarja novih golih zalivi, mrtvice, jezera…). Vse te oblike nastajajo, ko reka ob

strmih brežin, ki jih potrebujejo breguljke in čebelar. Zato so poplavah ustvarja novo glavno strugo, stara struga pa se

ptice našle nove lokacije vedno novih primernih brežin – v postopoma zamulji, preide v rokav, nato jezero, nato mrtvico,

gramoznicah, točno na fronti odkopa, kjer je vsako leto sveža mlako, močvirje in na koncu v suha (občasno poplavljena) tla.

brežina. Takšne reke od 19. stoletja pri Brežicah ni več, še vedno pa so

ostanki nekdanjih strug in drugih vodnih oblik. Tovrstne mlake

in močvirja so (bila) pester habitat in zato se je ob gradnji HE



Brežice zgradilo nove tovrstne oblike. V prostor se je torej želelo

umestiti nekaj, kar je izgubil že več desetletij nazaj. Ni se zgolj

ohranjalo stanje reke pred izgradnjo HE Brežice, temveč se je

želelo »izboljšati« trenutno stanje narave. Primer takšnega

ukrepa je habitat imenovan NH6, ki je bil prioritetno namenjen

želvi sklednici. Habitat se je zasnoval kot napol jezero in napol

mlaka, z dinamično brežino in globinami ter dovolj golega,

prodnatega prostora ob vodi, kjer si lahko želve ustvarjajo

gnezda in se sončijo. Vsi podatki po izgradnji so kazali na uspeh.

Po nekaj letih »obratovanja« habitata pa so se zgodile določene

spremembe: ker ni tekoče vode, ki bi občasno poplavila prodišče,

se slednje zarašča in želve izgubljajo prostor za sončenje. Iz

bližnjega potoka je bober skopal kanal do mlake in si v mlaki

zgradil gnezdo. Po kanalu dostopa do tekoče vode in preko nje

do prostora za hranjenje. Narava gre svojo pot. Potekajo naravni

Slika 2: Prostorski prikaz posegov ob gradnji HE Brežice procesi, s katerim pa izgubljamo prvoten namen habitata. Ali je

[5]; opis nekaterih oznak, za pojasnilo opisov v članku: 5.1, ta objekt uspešno zgrajen? Ali uspešno »obratuje«?

5.2 in 5.5 so Nadomestni Habitati (NH) 1, 2 in 5, ki

bi erodirala, niti ni izkopa gramoza, ki bi ustvarjal svež izkop, je



potrebno ohranjati strmo brežino na način, da se odkoplje tisti

material, ki ga padavine erodirajo. Tam, kjer je predpisano

navodilo za vzdrževanje, torej lahko govorimo o obratovanju,

tudi če gre za naravovarstveno ureditev. Ribe namreč plavajo

skozi prehod za vodne organizme, celo veliko jih je, in ptice tudi

kopljejo vsako leto nova gnezda v umetno zgrajenih brežinah.

Kjer je naravna oblika ustvarjena, pa je ocenjevanje uspešnosti

nesmiselno. Narava je namreč vedno uspešna, vprašanje je samo,

ali je ljudem njena rešitev všeč. Obratovanje torej ne obstaja, saj

je obratovanje po definiciji ponavljajoč proces. Gonilo narave pa

je sprememba.



5 Pristop k izgradnji novega naravnega

habitata

Sedaj znamo pristopati znanstveno in multidsciplinarno,

zavedamo se, kako »deluje« narava, razumemo spremembe v

naravi, poznamo pomen naravnih oblik in opremljeni smo z

znanjem (tako tehničnim kot biološkim), kako oblikovati prostor

Slika 3: Nadomestni habitat 6 (NH6), namenjen želvi na način, da bo primeren za izbrane organizme.

sklednici, cca 2 leti po izgradnji. Vidne so prvotno gole

površine proda, ki se postopoma zaraščajo. Na desni je



vidno porušeno drevje, kjer je aktiven bober [8].



Suhi travniki so nastajali na območjih, kjer je reka nasipala prod.

Takšna neugodna rastišča so najprej naselile pionirske rastline,

ki so ustvarile tanek sloj humusa, slednji pa je omogočal razrast

nezahtevnih rastlin. Večje rastline je ob poplavi odneslo in s tem

je nastal specifičen travniški habitat s pestrim ekosistemom rož,

med katerimi je 25 ogroženih vrst [5]. Ob umeščanju HE Brežice

v prostor se je rezerviral tudi prostor za tovrstne habitate, imenuje

se NH3. Pred pripravila so se tla in simulirali pogoji, kakršni Slika 4: Otok nanosa (sedimenta) na pretočni akumulaciji

majdemo na naravnih suhih travnikih. Tako kot vsaka naravna HE Boštanj na reki Savi [6]. oblika, tudi suhi travniki preidejo v naslednji stadij, saj večanje

debeline humusa tudi večjim rastlinam omogoči, da ustvarijo Na pretočni akumulaciji HE Boštanj se v bližini naselja

dovolj bogat koreninski sistem, ki zdrži poplave. S časoma tako Kompolje nahaja zožitev struge. Pred zožitvijo ob desnem bregu

travnik preide v gozd. Če se ta naravni proces odvije na voda zaradi hidravličnih razmer zastaja in odlaga nanos

zgrajenem habitatu, kateremu je dodeljen namen, ali je projekt (imenovan tudi sediment). Ker voda v pretočni akumulaciji

uspešen? Ali habitat obratuje uspešno, ali se bo okrivilo tistega, dnevno niha zaradi proizvodnje električne energije in ker je

ki je zadolžen za vzdrževanje, da ni ohranjal ustreznih pogojev odloženega že toliko nanosa, je tam pogosto viden otok. Ker je

(ni ohranjal tal dovolj siromašnih)? nanos, ki ga nosi reka Sava, v glavnem zelo drobnozrnat in z

veliko organskimi primesmi, otok izgleda kot kup mulja.

Fizikalne lastnosti tega materiala so takšne, da zadržuje vodo.

4 Ima naravni habitat fazo obratovanja? Ker je lahek, se ne posede, ne konsolidira in zaradi tega tudi ne

Na spodnji Savi trenutno obratuje 5 HE s pripadajočo vplivno more ustvariti dovolj čvrste podlage, na kateri bi se lahko naselile

površino. Ukrepi, izvedeni v času gradnje posamezne HE, so rastline, čeprav je semen v vodi vse polno, kakor pričajo različna

predani različnim upravljalcem. To še posebej velja za HE testiranja, ki smo jih opravili. Ideja je torej, da ne obnovimo

Brežice. Pri gradnji prejšnjih elektrarn se je malo ali delno dajala brežine v osnovno obliko trapeznega korita, utrjenega s

pozornost naravi, HE Brežice pa je pilotni projekt v Sloveniji, skalometom, temveč da uporabimo lokalne hidravlične razmere

celo v svetu, glede pozornosti, ki se je namenila naravi. V prvi in omogočimo naravi njen razvoj. Po pogovorih z ZRSVN smo

vrsti se je opredelilo vse naravne vrednosti, nato razmišljalo o prišli do zaključka, da bi bilo na območju HE Boštanj koristno

možnih ukrepih, se ukrepe zasnovalo ali sprojektiralo in na ustvariti boljše pogoje za vidro. Zato je ideja, da to lokacijo, ki

koncu so se objekti in ureditve zgradili, aktivnosti in procesi pa trenutno nima bistvene ekološke vrednosti, preuredimo na način,

pričeli, pri čemer se je določil tudi izvajalec teh aktivnosti. Na da bo vsaj sprva tu našla svoj življenjski prostor vidra, kasnejši

primer, obratovanje prehoda za vodne organizme za premostitev naravni razvoj pa bo pokazal, kakšne načrte ima narava.

višinske razlike v gladinah vode nad in pod jezovno zgradbo, v

katerem se mora ohranjati ustrezna vodnatost/pretočnost ne

glede na hidrološko-hidravlične pogoje v reki Savi. Drug primer

je že omenjena brežina za breguljke in čebelarje; ker ni vode, ki





vrednotimo z antropološkega vidika, ki je avtomatsko

subjektiven. Objektivna ocena vrednosti ni možna, še manj z

vidika narave. Vrednosti posameznega organizma ali lokalnega

habitata za celoten ekosistem ni mogoče ovrednotiti.





Slika 5: Rezultati študije ureditve brežine [7]. Levo: Prikaz

rezultatov hidravličnih izračunov. Desno: Prikaz idejne

rešitve ureditve brežine.



Hidravlična analiza je pokazala, da ni ovire za ureditev brežine,

ki bo ustvarila kombinacijo čvrste površine in vodnih lagun, saj

tudi ob najvišjih pretokih reka Sava na tem mestu ne bi odnašala

materiala. Izdelali smo idejno zasnovo brežine in sedaj

pristopamo k izdelavi projekta za izvedbo nove brežine. Ustvarili

bomo lepšo brežino, ki ne bo kazila podobo reke Save, na in ob

kateri poteka rekreacija in turizem. Vpliv na proizvodnjo je

zanemarljivo majhen, saj že sedaj otok zmanjšuje prostornino

vode, ki se uporablja za proizvodnjo električne energije. Stroški Slika 6: Primer tipičnega vodnega telesa s habitatno

vzdrževanja bodo manjši, saj ne bo potrebno čiščenje in diverziteto, kakršna se je umeščala na potokih v bližini HE

ohranjanje osnovne utrjene brežine. Narava pa bo dobila nov Brežice [8]. Proces erodiranja na levi strani slike ustvarja

prostor. strmo brežino, v kakršni lastovice breguljke in čebelarji

gradijo gnezda. Na desni strani poteka odlaganje in

ustvarjanje prodišča, kjer pričenjajo z rastjo pionirske

6 rastline, na kamnih pa se sončijo želve. Plitvine in tolmuni Prihodnost sodelovanja z naravo



Je z naravo možno sodelovati? Je narava pripravljena sodelovati? obvodno rastlinje pa hrano za različne organizme. ustvarjajo različne pogoje za različne velikosti rib, različno

Vse, kar lahko naredimo, je, da se zavedamo naravnih procesov, “Izgradnja” tovrstne ureditve neizbežno vodi v spremembo,

jih poskušamo čim bolj razumeti in spoznati njihovo saj bo narava že ob prvem dežju takšen “objekt”

kompleksnost ter povezanost. Elektro energetski objekti se ne spremenila. Faza obratovanja takšnega “objekta” v smislu

umeščajo, ker bi bili sami sebi namen, umeščajo se z razlogom. ponavljanja procesov zato ne obstaja. Vse, kar lahko ljudje

Ker jih potrebujemo. Hkrati je smiselna določena doza naredimo, je, da zagotovimo prostor in pustimo naravnim

samokritičnosti pri »predpisovanju« naravnih procesov in pri procesom naravno pot.

napovedovanju, kaj bo narava naredila ali kaj glede okolja lahko

pričakujemo v prihodnosti. Dejstvo je, da ne vemo vsega, Za konec upoštevajmo še človeški faktor in različne razloge za

kompleksnost narave pa je osupljiva, zato je smiselna določena ravnanje posameznikov, tudi inženirjev in znanstvenikov, ki

doza ponižnosti. Človek potrebuje svoj prostor za življenje, pogosto upoštevajo zgolj tisto, kar se je dogodilo v njihovi

ravno tako ga narava. Razmišljanje o naravi je še kako ključno, karieri, preteklost in prihodnost pa presegata okvir njihovega

če želimo ljudje preživeti na tem planetu. Izgradnja HE Brežice zanimanja. Narava ni tako ozkogleda, zato si je treba priznati, da

je primer dobre prakse glede pozornosti, ki se je je namenilo vsega pač ne vemo. Ukrepe, za katere so bili strokovnjaki čvrsto

naravi, glede truda in znanja, ki se ga je vložilo, in tudi glede na prepričani, da bodo delovali, je narava »odpeljala po svoje«.

novo pridobljenega znanja. Predvsem pa je dobra praksa v tem, Degradiranost okolja je narava pogosto preoblikovala v nekaj

da se je ob gradnji HE zagotovil in uredil prostor tudi za naravo. nepričakovanega, kar čez desetletja cenimo in želimo ohraniti.

S tem se omogoča naravna dinamika in dogajanje sprememb v Zato posegi v prostor, kakršne so tudi gradnje elektro energetskih

okolju, kajti to je naravno. Tudi če kakšen ukrep sedaj ni objektov, niso nujno takšna degradacija okolja, kakršna izgleda

popolnoma takšen, kot se je sprva zastavil, gre še vedno lahko za na prvi pogled. Zaradi našega lastnega obstoja na planetu je brez

primer dobre prakse. V najslabšem primeru kot pokazatelj, kaj dvoma potrebno pri umeščanju v prostor vložiti trud, čas in denar

lahko naredimo še boljše. tudi v ohranjanje narave. Pri tem ni smiselno, da ohranjanje

narave temelji strogo na načrtovanju in predvidevanju naravnih

procesov in naravnega stanja, temveč je preudarno vsaj deloma

7 Zaključki dovoliti, da naravni procesi oblikujejo prostor tudi po lastni poti.

Ker je celoten članek napisan v poljudnem stilu, naj zaključim s Sodobni pristopi k umeščanju zahtevajo tovrstno širokoglednost,

tehničnim komentarjem. Tehniki smo navajeni ravnati s potrebna pa je tudi multidisciplinarnost, da se naslovi vse

številkami, vrednotiti po kriterijih in razvrščati v razrede. Naravo ekološke vidike. Tudi v 3. fazi projektov, fazi obratovanja, je pri

je težko razvrščati na tovrsten matematičen način iz več vsaki spremembi prostor, da se razmisli o možnosti sonaravnega

razlogov. V prvi vrsti so naravni sistemi preveč kompleksni, da ukrepanja, pri čemer je sonaravno ukrepanje že to, da se da naravi

bi jih enoznačno vrednotili, pogosto pa jih niti ne razumemo v prostor.

celoti. Pomen naravnih oblik in njihovo vrednost vedno



References / Literatura [5] Geateh, d.o.o. 2014. Poročilo o vplivih na okolje za HE Brežice. [6] Hribar A. 2015-2022. Osebni arhiv fotografij in tehničnega gradiva

[1] Plachter, H. 1995. A Central European Approach for the Protection of (gradivo z avtorskimi pravicami je uporabljeno s privolitvijo lastnika,

Biodiversity. V: Ogrin, D. (et al.) 1996. HESS, d.o.o. in Infra, d.o.o.).

[2] Ogrin, D. 1995. Dileme v načrtovanju za varstvo narave: primer reke [7] Kryžanowski, A. 2022. Študija možnosti ureditve brežine reke Save pri

Mure. naselju Kompolje.

[3] Lin Sien, C (et al.) 1988. The Coastal Environmental Profile of Singapore. [8] Splet

[4] Lubi, J. 2011. Gospodarjenje s topolovimi nasadi na krško-brežiškem

polju.





Axioms and Myths of Environmental Care




and Sustainable Agriculture



Aksiomi in miti o varovanju okolja in trajnostnem



kmetijstvu



Matjaž Gams



Odsek za inteligentne sisteme

Institut “Jožef Stefan”



Jamova cesta 39, 1000 Ljubljana

Slovenija



okolje. To prepričanje izhaja predvsem iz dveh dejavnikov:

ABSTRACT. izpustov toplogrednih plinov, zlasti metana, ki ga proizvajajo

This paper challenges common misconceptions about prežvekovalci, in degradacije tal zaradi intenzivnih kmetijskih

agriculture's impact on the environment, particularly the idea that praks. Vendar pa so te predstave pogosto poenostavljene in ne

agriculture is inherently harmful. By providing insights into upoštevajo ključnih dejstev, kot so naravni cikli metana,



study highlights how well-managed agricultural practices, such raznovrstnosti, ki jo lahko podpirajo dobro upravljane kmetijske as diverse crop management and rotational grazing, contribute površine [1][2]. natural cycles, methane dynamics, and biodiversity benefits, the trajnostno upravljanje kmetijskih zemljišč in pomen biotske

positively to the environment. Contrary to the myth, correctly

managed farmland and pastures foster biodiversity, support soil Dobro upravljano kmetijstvo, zlasti ekstenzivne metode, kot

health, and integrate animals as a natural part of ecosystems. The sta rotacijska paša in ohranjevalno obdelovanje tal, ima zelo

paper compares the environmental impact of agriculture with pozitivne vplive na okolje. Na primer, travniki in pašniki lahko

other sectors and emphasizes the need for a nuanced gostijo večjo biotsko raznovrstnost kot mnogi gozdni ekosistemi,

understanding of sustainability in agriculture. saj omogočajo habitat za številne rastlinske in živalske vrste [3].

Poleg tega so tla na trajnostno upravljanih kmetijskih površinah

POVZETEK sposobna skladiščiti več ogljika, kar pripomore k zmanjševanju

Prispevek izpodbija običajne napačne predstave o vplivu koncentracij toplogrednih plinov v ozračju [4]. Primerjave z

kmetijstva na okolje, zlasti idejo, da je kmetijstvo škodljivo zagrajenimi površinami (ceste, parkirišča, stavbe) so seveda



biotske raznovrstnosti, študija poudarja, kako dobro upravljane “bogate”, človeško degradirane pa najmanj. kmetijske prakse, kot so raznovrstne poljščine in rotacijska paša, Čeprav zlasti živinoreja prispeva k proizvodnji metana, okolju. Z analizo naravnih ciklov, dinamike metana in koristi nesmiselne, saj je v resnici kmetijske površine najbolj biološko



obdelana kmetijska zemljišča in travniki spodbujajo biotsko enemu glavnih toplogrednih plinov, pa se metan v ozračju pozitivno prispevajo k okolju. Nasprotno mitu, pravilno

raznovrstnost, podpirajo zdravje tal in vključujejo živali kot ohranja le približno 12 let, kar pomeni, da so njegovi dolgoročni

naravni del ekosistemov. Prispevek primerja okoljski vpliv učinki bistveno manjši kot pri CO2 [5]. Poleg tega nove

kmetijstva z drugimi sektorji in poudarja potrebo po bolj tehnologije in prakse, kot so dodatki za prehrano živali, lahko

natančnem razumevanju trajnostnosti v kmetijstvu. bistveno zmanjšajo emisije metana iz živinoreje [6].

Kmetijstvo je tudi ključno za ohranjanje biotske

KLJUČNE BESEDE raznovrstnosti. Dobro upravljana kmetijska zemljišča

Trajnostno kmetijstvo, metan, biotska raznovrstnost, zagotavljajo pomemben prostor za številne vrste, ki jih ne

ekosistemi, upravljanje kmetijskih zemljišč najdemo v intenzivno gozdnih ali urbaniziranih območjih.



KEYWORDS ekosistemov in omogoča večjo odpornost na podnebne Raznolikost vrst na takšnih površinah prispeva k stabilnosti

Sustainable agriculture, methane, biodiversity, ecosystems, spremembe in druge stresne dejavnike [7].

farmland management Cilj tega prispevka je podrobneje raziskati vpliv kmetijstva na

okolje, zlasti v povezavi z izpusti toplogrednih plinov, biotsko

raznovrstnostjo in trajnostnim upravljanjem naravnih virov, ter

1 Uvod izpodbijati nekatere uveljavljene mite o kmetijstvu kot

Kmetijstvo, kljub svoji nujni vlogi pri zagotavljanju hrane za največjem onesnaževalcu. Poleg tega bomo preučili, kako lahko

svetovno prebivalstvo, pogosto napadajo, češ da onesnažuje trajnostno kmetijstvo prispeva k zmanjševanju negativnih



vplivov na okolje in pomaga pri reševanju nekaterih ključnih 2.2. Izgorevanje (fosilnih goriv) največ prispeva k

okoljskih izzivov. onesnaževanju oz. uničevanju okolja

Da bi argumentirali vlogo in pomen kmetijstva v smislu Vse tehnologije, ki temeljijo na izgorevanju fosilnih goriv

varovanja okolja, najprej naredimo širšo analizo varovanja (premog, nafta, plin), kot so tiste v prometu, ogrevanju in

okolja, ki temelji na [8, 9, 10]. Naredimo tako analizo aksiomov, industriji, prispevajo k izpustom toplogrednih in ljudem

osnovnih usmeritev varovanja okolja v sekciji 2, kot tudi mitov škodljivih plinov, kar je glavni vzrok za onesnaževanje okolja.

v sekciji 3. Zaključimo z diskusijo v sekciji 4. Fosilna goriva so omejeni viri, nastali skozi milijone let iz

ostankov starodavnih organizmov, in čeprav predstavljajo

izredno zgoščeno energijo, je njihova raba dolgoročno nevzdržna.

2 Aksiomi o varovanju okolje Po ocenah geoloških raziskav je v zemeljski skorji približno 1.7

trilijona sodčkov nafte, vendar pri trenutni stopnji porabe lahko

Aksiomi (osnovne usmeritve) varovanja okolja so: pričakujemo, da bodo te zaloge izčrpane v nekaj desetletjih.

Poleg tega se približno 50 % zalog premoga ocenjuje na količino,

2.1. Rast števila ljudi in standarda je osnovni pritisk na ki zadošča še za dobrih 100 let, vendar so tudi te številke odvisne

okolje od stopnje porabe [14, 15].

Ljudje za svoje delovanje in udobje uporabljajo energijo, kar Toplogredni plini, ki nastajajo pri izgorevanju fosilnih goriv,

pomeni, da več kot je ljudi in višji kot je življenjski standard, več kot so ogljikov dioksid (CO2), metan (CH4) in dušikovi oksidi

energije je potrebno za zadovoljevanje njihovih potreb. (NOx), pomembno vplivajo na segrevanje ozračja. Med njimi

Posledično se povečuje tudi poraba naravnih virov in nastaja ima metan sicer večji takojšnji učinek na podnebje, vendar

večja količina onesnaževanja, ki vpliva na kvarjenje okolja. razpade v ozračju v približno 12 letih. Ogljikov dioksid pa ostaja

Čeprav se rast svetovne populacije postopoma upočasnjuje, v atmosferi več stoletij, zaradi česar se njegov vpliv dolgoročno

zlasti zaradi nizke rodnosti na večini celin (razen v Afriki, kjer kopiči. Prav CO2 je tisti, ki najbolj prispeva k dolgoročnemu

rodnost ostaja visoka), to pomeni, da je v prihodnosti čedalje segrevanju ozračja in podnebnim spremembam, ker je trajno

pomembnejši dejavnik vpliv posameznika na okolje [11]. Ob prisoten v visokem ozračju in povzroča t.i. učinek tople grede.

manjšem številu otrok in staranju prebivalstva bo poudarek na Zato so tehnologije, ki ne uporabljajo fosilnih goriv, kot so

tem, koliko okoljske škode povzročajo posamezni ljudje s svojim vetrnice, jedrska energija, sončne elektrarne in hidroelektrarne,

načinom življenja in porabo energije. Bolj kot število ljudi bo ekološko bistveno bolj trajnostne. Te tehnologije izkoriščajo

pomembna njihova potrošnja in ravnanje z naravnimi viri – naravne vire, kot so veter, voda, sonce ali jedrsko fisijo, ki so v

Koliko površin bodo uničili, koliko onesnaževanja bodo veliki meri obnovljivi in povzročajo manj onesnaženja [16]. Na

povzročali, Koliko bodo podpirali biotsko raznolikost in sožitje primer, sončne elektrarne nimajo neposrednih izpustov

z naravo [12]. toplogrednih plinov med obratovanjem, medtem ko vetrnice

Ključen dejavnik v tej enačbi je poselitev – kako in kje so proizvajajo energijo z uporabo vetrne sile, kar zmanjšuje potrebo

ljudje naseljeni ter koliko prostora je namenjenega različnim po fosilnih gorivih.

dejavnostim, kot so kmetijstvo, industrija, neokrnjena narava in Vendar imajo tudi te tehnologije svoje izzive. Jedrska

urbani parki. Teoretično bi se vseh 8 milijard ljudi na svetu lahko energija, čeprav z nizkimi izpusti CO2, prinaša tveganje v obliki

naselilo na območje velikosti Slovenije, saj bi ob 20.000 km² radioaktivnih odpadkov. Vetrne elektrarne lahko vplivajo na

površine vsak posameznik zasedel približno 2,5 m² prostora. lokalne ekosisteme in predstavljajo izziv glede stabilnosti oskrbe

Torej ta hip število prebivalcev na planetu ni preveliko, če bi z energijo, kadar ni dovolj vetra. Sončne elektrarne pa zahtevajo

delovali smotrno in trajnostno, če bi poseljevali predvsem v velike površine zemljišč, kar je hudo problematično (okoljski

mestih, zmanjšali nepotrebno porabo virov in namenjali zadostne kriminal v imenu varovanja okolja), zlasti če se postavljajo na

površine naravi ter ekosistemom, kjer pomembno vlogo igrajo kmetijskih območjih ali naravnih habitatih.

kmetijske površine in gozdovi. Znotraj kmetijstva pa so v zadnjem desetletju v ospredje

Trajnostni razvoj bi tako moral vključevati učinkovito rabo prišle tako imenovane biogorivne kulture, kot je oljna repica, ki

prostora, zlasti z vidika urbanizacije in varovanja naravnih naj bi predstavljale obnovljiv vir energije in zmanjšale odvisnost

območij, ob hkratnem zmanjšanju porabe energije na prebivalca od fosilnih goriv. Vendar se je pokazalo, da ima proizvodnja

in spodbujanju manj potrošniškega življenjskega sloga. Če bi biogoriv več negativnih kot pozitivnih učinkov na okolje. Proces

človeštvo delovalo trajnostno – manj trošilo, učinkoviteje pridelave oljne repice za gorivo pogosto zahteva več fosilnih

uporabljalo naravne vire in vzdrževalo ravnovesje med goriv (za gnojenje, obdelavo tal, žetev in predelavo), kot jih

človekovo dejavnostjo in ohranjanjem narave – bi trenutna biogorivo prihrani ob končni uporabi. To pomeni, da celotni

populacija lahko bivala v harmoniji z okoljem brez bistvenih življenjski cikel proizvodnje teh goriv v mnogih primerih ne vodi

negativnih vplivov na planet [12, 13]. k zmanjšanju emisij, ampak celo k povečanemu ogljičnemu

Za Slovenijo in druge države je napotek jasen: treba je odtisu [15]. Čemu smo sploh uvedli tovrstne ukrepe, ko pa očitno

stremeti k trajnostni rodnosti (1.9 do 2.2), treba je zmanjšati škodijo okolju?

pritok migrantov – saj to povečuje število prebivalcev in pritisk Za Slovenijo je osnovna trajnostna usmeritev, da opušča

na okolje. Treba je večino ljudi poseliti po mestih in čim večji predvsem termoelektrarne na premog in se usmeri v jedrsko in

del površin nameniti kmetijstvu, gozdarstvu in zavarovanim trajne vire, kjer je to sprejemljivo in okolju primerno. Vsak ukrep

območjem [8, 11]. pa mora biti premišljen (primer okoljskega kriminala je

postavljanje sončnih panelov na kmetijskih površinah).



2.3. Aksiom: Ohranjanje biotske raznovrstnosti je ključnega Prav tako bi moralo biti dovoljeno humano odstranjevanje

pomena za stabilnost ekosistemov invazivnih živali, kot so nutrije v Ljubljanici. Ljudje pogosto ne

Zmanjševanje biotske raznovrstnosti, ki ga pospešujejo razumejo škodljivosti invazivnih vrst in se upirajo njihovemu

izgorevanje, uničevanje kmetijskih in gozdnih površin, velike odstranjevanju, čeprav te vrste povzročajo uničujoče učinke na

kmetijske monokulture in invazivne vrste, zmanjšuje odpornost avtohtone ekosisteme. Navedimo dva primera: če bi bili pogoji

ekosistemov. Eni izmed dejavnikov zmanjševanja biotske povsod po svetu enaki, bi prevladovalo recimo 10-krat manj vrst,

raznovrstnosti so monokulturni sistemi, saj ustvarjajo homogene kot jih je danes. Z globalizacijo se pogoji izenačujejo, saj ljudje

omejene ekosisteme, kjer je manj prostora za prosto živeče postajajo ključni dejavnik v večini ekosistemov, kjer s svojo



in škodljivce, kar zahteva večje količine pesticidov in umetnih habitate. Še en primer: sesalci predstavljajo približno 6.400 vrst na svetu. Okoli 36 % teže sesalcev predstavljajo ljudje, medtem gnojil, ki dodatno škodujejo okolju [17, 18, 19]. Za Slovenijo to rastline in živali. Prav tako so te kulture bolj občutljive na bolezni prisotnostjo in navadami spreminjajo prehranjevalne verige in

skoraj ne pride v poštev, saj je večina posestev razparceliranih, ko 60 % teže sesalcev predstavljajo domače živali, kot so govedo,

medtem ko so v nekaterih državah cele doline namenjeni ovce, prašiči in drugi rejni organizmi. Divji sesalci pa

monokulturi. V tropskih krajih izstopa oljna palma (Elaeis predstavljajo le še 4 % vse biomase sesalcev na planet. Ali je

guineensis), iz katere pridobivajo palmovo olje, ki se uporablja v projekcija 4, 3, 2, 1, 0 [20, 21]?

številnih prehrambnih, kozmetičnih in industrijskih izdelkih. Ta Podobno škodljiv vpliv imajo invazivne vrste, kot je nutrija

kultura vodi v obsežno krčenje gozdov, izgubo biotske (Myocastor coypus, znana tudi kot močvirska podgana ali

raznovrstnosti, emisije toplogrednih plinov ter izpodrivanje bobrovka). Preden nutrije naselijo določeno območje, tam živijo



prenašamo med različnimi ekosistemi, eden največjih razlogov Prav tako so invazivne vrste, ki jih pogosto nenamerno Vendar ko nutrije, ki tehtajo od 5 do 10 kilogramov, pridejo na to območje, uničijo, izpodrinejo ali pobijejo vse, kar jim ne ubeži lokalnih skupnosti. avtohtone vrste podgan, miši, rakov in drugih malih živali.

za izumiranje avtohtonih vrst. Invazivne vrste pogosto (recimo ribe). Kljub temu pa lokalno prebivalstvo pogosto

izpodrinejo avtohtone zaradi svoje agresivnosti, hitrega protestira proti odstranjevanju nutrij, ne zavedajoč se, kakšno

razmnoževanja ali pomanjkanja naravnih plenilcev. Poleg tega uničenje med avtohtonimi vrstami povzročajo te invazivne živali.

so nekatere invazivne rastlinske vrste, kot je ambrozija, tudi zelo Takšne napačne predstave in odpor do odstranjevanja invazivnih

škodljive za zdravje ljudi, saj povzročajo alergije, in čeprav je vrst lahko povzročijo dolgoročno škodo ekosistemom, ki so že

potrebno minimalno vsaj avgusta kositi dvakrat, se v Sloveniji v oslabljeni zaradi drugih okoljskih pritiskov.

marsikateri občini ne izvaja niti to. Kot imamo akcije čiščenja okolja, bi morali izvajati akcije

Tudi na videz naravna okolja, kot so strnjeni gozdovi, lahko čiščenja invazivnih rastlin in živali in morali bi izobraževati

predstavljajo omejeno biotsko področje, če v njih ni zadostne državljane, kaj je to smotrno varovane okolja in kaj to ni.

raznolikosti habitatov. Zato je veliko bolj smiselno ohranjati

mozaično krajino, kjer so naravne jase, kmetijske površine med 2,4. Aksiom: Trajnostni razvoj zahteva večji učinek z manjšo

gozdom ali občasne poseke (daleč bolje kot strnjeni gozdovi), ki porabo energije s pomočjo naprednih tehnologij in

ustvarjajo raznolikost življenjskih prostorov. Takšna krajina pametne rabe virov

omogoča večji razpon ekosistemov, ki so bolj odporni na Trajnostni razvoj ne pomeni predvsem zategovanja pasu -

spremembe in naravne katastrofe. Gozdovi z jasami ponujajo več zmanjševanje porabe, temveč eliminiranje nesmiselnih pristopov

možnosti za različne vrste rastlin in živali, kar povečuje ter povečanje učinkovitosti s pametno uporabo virov in

stabilnost ekosistema. tehnologij, kot je umetna inteligenca (AI). Z uporabo AI,

Podobno so škodljive genetsko spremenjene rastline, ki so avtomatizacije in optimiziranih sistemov lahko povečamo

pogosto zasnovane tako, da vsebujejo "naravne" strupe za učinkovitost in zmanjšamo porabo energije, ne da bi pri tem

žuželke, kar ne ustvarja le monokultur, ampak tudi "zastrupljene" ogrozili razvoj [22, 23]. Namesto povečanja porabe moramo

monokulture. Te rastline ne škodujejo samo škodljivcem, ampak iskati rešitve, ki omogočajo večji učinek z manjšo porabo



škodijo celotnemu ekosistemu, saj strupi prehajajo skozi Preprost primer tega so drevesa. Drevo, zasajeno pred hišo, prehranjevalno verigo. lahko zagotavlja senco in posledično zmanjša potrebo po vplivajo tudi na druge, koristne žuželke in lahko dolgoročno energije.

Še slabše pa so (avto)ceste, ki predstavljajo povsem mrtva hlajenju stavbe za kar 20 %. To je enostaven in naraven način za

območja za večino živih bitij. Ta območja niso zgolj fizično ovira zmanjšanje porabe energije brez tehnoloških intervencij, hkrati

za prostoživeče vrste, temveč so tudi območja, kjer se pogosto pa z drevesom prispevamo k izboljšanju kakovosti zraka, ki ga

dogajajo trki z živalmi (test avtomobilskega stekla). Poleg tega dihamo [24].

ustvarjajo prekinitev habitatov, kar vodi do izgube genetske Na globalni ravni se soočamo s hitrim izginjanjem kmetijskih

raznolikosti med populacijami in posledično večje občutljivosti površin, kar še dodatno ogroža našo sposobnost za trajnostno

na bolezni in okoljske spremembe. Avtoceste so v marsičem pridelavo hrane, predvsem pa biotsko raznolikost. Vsako leto po

»polja smrti«, kjer preživetje vrst ni mogoče, razen tistih, ki se v svetu izgubimo približno 10 milijonov hektarjev kmetijskih

kovinskih škatlah vozijo naokoli. zemljišč zaradi različnih dejavnikov, kot so urbanizacija, širitev

Eno izmed pomembnih rešitev za ohranjanje biotske cest, krčenje gozdov in intenzivna kmetijska dejavnost. To je



z ustrezno zakonodajno podporo. Na primer, kazni za lastnike minuto posekajo za 4 nogometna igrišča gozda v Amazoniji. zemljišč, ki ne odstranijo alergene rastline, kot je ambrozija, bi Kdaj bo konec – ko bomo posekali in uničili vse? raznovrstnosti je tudi sistematično odstranjevanje invazivnih vrst približno enako 14 milijonom nogometnih igrišč [25]. Vsako

lahko spodbudile učinkovitejše ravnanje z invazivnimi rastlinami.

Tudi v Sloveniji je situacija zaskrbljujoča, saj vsak dan Strokovne raziskave kažejo, da pravilno upravljane kmetijske

izgubimo površino enega nogometnega igrišča kmetijskih površine ključno prispevajo k ohranjanju biotske raznovrstnosti.

zemljišč, kar vodi v zmanjšanje prostora za pridelavo hrane in Na primer, na travnikih, ki so pravilno vzdrževani, lahko uspeva

povečuje obremenitev okolja. Širitev cestne infrastrukture in do 40 različnih vrst rastlin na kvadratni meter, kar je veliko več

gradnja novih objektov na teh zemljiščih še dodatno prispevata k kot v intenzivnih industrijskih monokulturah ali pozidanih

izginjanju naravnih površin ter pospešujeta učinke segrevanja površinah, ki so pogosto označene kot "puščave", kjer je

podnebja [26]. Takšne prakse ne samo zmanjšujejo biotsko prisotnih zelo malo vrst [28, 29].

raznovrstnost, temveč ustvarjajo tudi večje potrebe po energiji in Zato kmetijstvo ni okolju neprijazno – pravilno izvajano

povzročajo povečanje toplogrednih plinov [27]. kmetijske prakse prispevajo k trajnostnemu okolju in visoki

Nadaljnje širjenje cestne infrastrukture in gradnja novih stavb, biotski raznovrstnosti, medtem ko napačne predstave o

še posebej veletrgovin in stanovanjskih sosesk na kmetijskih kmetijstvu pogosto izhajajo iz neustrezno vodenih praks in

zemljiščih, ne samo zmanjšuje razpoložljivo površino za industrijskih metod pridelave ter napačega razumevanja, kaj je to

pridelavo hrane, temveč tudi povečuje izpuste toplogrednih biotska raznolikost.

plinov in segreva klimo. Gradnja dodatnih cest in veletrgovin

povečuje promet, kar vodi k večji porabi fosilnih goriv in s tem



dodatno obremenjuje okolje.

Namesto širjenja teh netrajnostnih praks moramo vlagati v

tehnološke inovacije in naravne rešitve, ki bodo omogočile večji

učinek z manjšo porabo virov. Pametno zasnovana infrastruktura,

AI v energetski učinkovitosti in trajnostne rešitve, kot je

zasaditev dreves in ustvarjanje naravnih senčnih površin, lahko

pripomorejo k bistvenemu zmanjšanju porabe energije ter

ohranjanju naših naravnih virov.

Če pogledamo usmeritve v Sloveniji in zaključke te

konference o okolju, jasno opazimo velika razhajanja. Marsikje

so strokovni argumenti povsem jasni in na naši strani, pa vendar

javnost in mediji tega niti ne vedo.



3 Miti o okolju in kmetijstvu



3.1. Kmetijstvo je okolju neprijazno

Resnica je veliko bolj kompleksna in v osnovi fundamentalno

nasprotna temu mitu. Pravilno upravljano kmetijstvo prispeva k

večji biotski raznovrstnosti in trajnostnemu okolju (neekološko



ali nekvalitetno kmetijstvo pa seveda škodi). Kmetijske površine, Slika 1: Izvori toplogrednih plinov; tri četrtine jih prihaja iz kot so dobro obdelane njive in travniki, ustvarjajo raznoliko fosilnih goriv. Vir: Hannah Ritchie (2020) - “Sector by sector: krajino, ki podpira različne vrste živali in rastlin. Pravilno where do global greenhouse gas emissions come from?” upravljana polja in travniki so habitat za številne žuželke, ptice Published online at OurWorldInData.org. Retrieved from: in druge prostoživeče živali, kar prispeva k izjemni biodiverziteti 'https://ourworldindata.org/ghg-emissions-by-sector' [Online [28]. Na primer, v Evropi so travniki ena izmed najbolj biotsko Resource] raznovrstnih ekosistemov.



Metan, ki ga oddajajo prežvekovalci, kot so krave, ima sicer 3.2. Izmed vseh tehnologij je najbolj nevarna in škodljiva

močan učinek na podnebne spremembe, vendar razpade v ravno jedrska

ozračju v približno 12 letih, medtem ko ogljikov dioksid (CO2) Tudi tu je resnica ravno nasprotna: Jedrska energija proizvaja



naravnih ciklov hranil. Paša živali, na primer, spodbuja rast Živali so naravni del ekosistema in imajo vlogo pri ohranjanju kot sončna in vetrna podvrženi nihanjem in so zato primerjave, tudi v tem prispevku, narejene na osnovi uravnoteženega travniških rastlin in s tem izboljšuje strukturo tal ter preprečuje energetskega sistema, kjer imate zadosti stalnih virov (hidro, erozijo. Podobno so njihovi odpadki hrana za insekte vseh vrst, jedrska, termo) in trajnostnih, tj. občasnih. Če ni na voljo zadosti katerih število se je v zadnjih 50 letih zmanjšalo za več kot trajnih virov, se cena električne energije zaradi shranjevanja polovico, kot tudi število in teža vseh divjih živali. Kmetijstvo, približno podvoji ali potroji [9, 30]. kratkotrajen in se lahko učinkovito upravlja z uporabo novih predstavlja bolj trajnostno alternativo, čeprav obstajajo tveganja, tehnologij za zmanjšanje emisij. povezana z odpadki. Poleg tega so nekatere trajnostne energije ostaja v atmosferi več stoletij. To pomeni, da je vpliv metana minimalne izpuste CO2 v primerjavi s fosilnimi gorivi in zato

ki temelji na principih regenerativnega kmetovanja, izboljšuje Primerjava po atributih:

kakovost tal, saj povečuje njihovo sposobnost zadrževanja vode Cena na enoto energije (USD/kWh)

in hranil. Težko je razumeti nekatere, ki se imajo za zagovornike Jedrska energija: Povprečna cena proizvodnje jedrske

varstva okolja, ko zagovarjajo invazivne živali in hkrati energije znaša približno 0,09 USD/kWh. Visoki začetni stroški

zahtevajo odstranitev domačih živali.



izgradnje jedrskih elektrarn so kompenzirani z nizkimi Vetrna energija: Vetrna energija prav tako spada med zelo

obratovalnimi stroški in dolgoročno življenjsko dobo [31]. varne vire, s približno 0,04 smrti/TWh.

Fosilna goriva (premog, nafta, plin): Cene fosilnih goriv so Hidroenergija: Čeprav hidroenergija večinoma velja za varno,

odvisne od trga, vendar se cena proizvodnje električne energije lahko nesreče, kot so zrušitve jezov, povzročijo velike žrtve.

giblje med 0,05 in 0,10 USD/kWh. Čeprav so začetni stroški nižji, Povprečno število smrti znaša 1,3 smrti/TWh, vendar lahko te

pa fosilna goriva dolgoročno povzročajo stroške zaradi vpliva na nesreče lokalno povzročijo večje število smrtnih žrtev [35].

okolje in zdravje.



močno padli in zdaj znašajo med 0,03 in 0,06 USD/kWh, kar jo Sončna energija: Stroški sončne energije so v zadnjih letih 4 Diskusija Pravilno varovanje okolja temelji na trajnostnem razvoju, kar uvršča med najcenejše vire energije (če imamo osnovno pomeni pametno rabo virov in naprednih tehnologij, ki energijsko zahtevo pokrito trajnostno). omogočajo večjo učinkovitost z manjšo porabo energije. Na Vetrna energija: Cena proizvodnje vetrne energije se giblje primer, uporaba jedrske energije za proizvodnjo električne med 0,04 in 0,07 USD/kWh, kar jo prav tako uvršča med energije znatno zmanjša izpuste CO2 v primerjavi s fosilnimi konkurenčne vire obnovljive energije. gorivi in prispeva k boju proti podnebnim spremembam. Hidroenergija: Cena hidroenergije je zelo odvisna od lokacije Pomembno je tudi ohranjanje biotske raznovrstnosti, saj stabilni in znaša med 0,03 in 0,05 USD/kWh, kar jo uvršča med cenovno ekosistemi zagotavljajo dolgoročno odpornost na podnebne najučinkovitejše vire energije. spremembe. Naravne rešitve, kot so sajenje dreves in pametno Izpusti CO2 na enoto energije (gCO2/kWh) načrtovanje krajine, lahko prav tako učinkovito zmanjšujejo Jedrska energija: Jedrska energija spada med najmanj potrebo po energiji, kot je hlajenje stavb, ter hkrati prispevajo k emisijsko intenzivne vire, s povprečnimi izpusti CO2 okoli 12 izboljšanju kakovosti zraka. Pravilno varstvo okolja temelji na gCO2/kWh (vključno z izgradnjo elektrarn in celotnim pametni kombinaciji naravnih in tehnoloških rešitev, ki življenjskim ciklom). zagotavljajo dolgoročno trajnost. Fosilna goriva: Premog povzroča največ izpustov, s Nasprotno pa nepravilno varovanje okolja pogosto temelji na povprečno 820 gCO2/kWh, plin povzroča približno 490 napačnih argumentih in mitih, kot je prepričanje, da so obnovljivi gCO2/kWh, medtem ko nafta znaša okoli 720 gCO2/kWh [32]. viri, kot sta sončna in vetrna energija, vedno boljši od jedrske Sončna energija: Sončna energija prispeva približno 48 energije. V resnici te rešitve zahtevajo večje površine in so manj gCO2/kWh, večinoma zaradi proizvodnje in reciklaže zanesljive, ter so nekajkrat dražje, če jih je potrebno shranjevati. fotonapetostnih panelov. Zmotno je prepričanje, da se gospodarska rast lahko nenehno Vetrna energija: Vetrna energija spada med najbolj čiste vire, povečuje, ne da bi to negativno vplivalo na okolje. Prav tako mit, z izpusti okoli 11 gCO2/kWh. da kmetijstvo škoduje okolju, izhaja iz napačnega razumevanja Hidroenergija: Povprečni izpusti hidroenergije znašajo 24 varovanja okolja. gCO2/kWh, vendar se lahko ta vrednost poveča zaradi izpustov Kmetijstvo prispeva k ohranjanju okolja in povečevanju metana iz velikih rezervoarjev. biotske raznovrstnosti. Prakse, kot je regenerativno kmetovanje, Potrebne površine na enoto energije (m²/kWh) spodbujajo zdravje tal, zmanjšujejo erozijo in izboljšujejo Jedrska energija: Jedrska energija zahteva najmanj prostora. zmožnost tal za zadrževanje vode. Paša živali na travnikih, če je Povprečna jedrska elektrarna proizvede ogromno količino pravilno vodena, prispeva k naravnemu ciklu hranil, spodbuja energije na majhni površini – približno 1,9 m²/kWh [33]. rast rastlin in podpira raznolikost vrst. Zmanjšanje metanskih Fosilna goriva: Premogovne in plinske elektrarne potrebujejo emisij s tehnološkimi inovacijami in trajnostno obdelavo približno 40-100 m²/kWh, zaradi potrebe po rudarjenju in zemljišč omogoča, da kmetijstvo ne le minimizira škodljive skladiščenju goriv. učinke, temveč aktivno prispeva k varovanju naravnih virov. Sončna energija: Solarne elektrarne zahtevajo veliko večje Pogosto napačno predstavljanje kmetijstva kot enega glavnih površine za zbiranje sončne svetlobe, s približno 25-70 m²/kWh, onesnaževalcev temelji na primerih neustreznih praks, kot so odvisno od učinkovitosti panelov in lokacije. intenzivne monokulture in pretirana uporaba kemikalij, česar v Vetrna energija: Potrebne površine za vetrne elektrarne Sloveniji skoraj ni. Vendar pa je resnica, da lahko kmetijstvo, če znašajo približno 20-60 m²/kWh. se izvaja trajnostno, bistveno prispeva k varovanju okolja. Hidroenergija: Potrebne površine za hidroelektrarne se zelo Travniki in kmetijske površine lahko podpirajo večjo biotsko razlikujejo glede na vrsto projekta, vendar znašajo od 100 raznovrstnost kot mnoge gozdne površine, saj ustvarjajo naravne m²/kWh ali več pri velikih akumulacijskih jezerih. habitate za številne rastlinske in živalske vrste. Sodobni miti o Število žrtev v nesrečah na enoto energije (smrti/TWh) kmetijstvu, ki ga enačijo z uničevanjem okolja, ne upoštevajo Jedrska energija: Jedrska energija je kljub nekaj odmevnim pozitivnih vplivov trajnostnih kmetijskih praks, ki so ključne za nesrečam, kot so Černobil in Fukušima, statistično varna z vidika ohranjanje ekosistemov. smrtnosti. Povprečje znaša 0,07 smrti/TWh, vključno z Zelo pomembno je tudi, kako je kmetijstvo upravljano, nesrečami in dolgoročnimi učinki sevanja [34]. recimo ali se zgodaj kosi travo, ali se krave nepravilno hrani itd. Fosilna goriva: Premog je izredno nevaren tako zaradi nesreč Tu so lahko napredni in ekološko zavedni sistemi vodenja kmetij v rudnikih kot zaradi dolgoročnih zdravstvenih učinkov na osnovi umetne inteligence in GPT izredno koristni tako za onesnaženja. Povprečno število smrti znaša 24,6 smrti/TWh za kmete kot za varstvo okolja. Ko smo predlagali tovrstne ideje premog in 2,8 smrti/TWh za plin. predstavnikom oblasti, pa ni bilo prav veliko zanimanja. Sončna energija: Sončna energija je zelo varna, s povprečno

0,02 smrti/TWh, večinoma zaradi nesreč pri namestitvi in

vzdrževanju panelov.

V zaključku: kmetijstvo je v svoji osnovi najboljša varovalka services. Nature Sustainability, 3(2), 92-101.

okolja in zato bi morali okoljevarstveniki, kmetje in raziskovalci https://doi.org/10.1038/s41893-019-0443-x

skupaj zagovarjati najboljše rešitve. 13. Cohen, J. E. (2019). Human population: The next half

century. Environmental International, 129, 104994.

Literatura: https://doi.org/10.1016/j.envint.2019.104994

14. British Petroleum. (2020). Statistical review of world

1. Garnett, T., Appleby, M. C., Balmford, A., Bateman, I. J., energy 2020. BP.

Benton, T. G., Bloomer, P., & Godfray, H. C. J. (2013). https://www.bp.com/en/global/corporate/energy-

Sustainable intensification in agriculture: Premises and economics/statistical-review-of-world-energy.html

policies. Science, 341(6141), 33-34, 15. Searchinger, T. D., Heimlich, R., Houghton, R. A., Dong,

https://doi.org/10.1126/science.1234485 F., Elobeid, A., Fabiosa, J., ... & Yu, T. H. (2008). Use of

2. Smith, P., Gregory, P. J. (2013). Climate change and US croplands for biofuels increases greenhouse gases

sustainable food production. Proceedings of the Nutrition through emissions from land-use change. Science,

Society, 72(1), 21-28, 319(5867), 1238-1240.

https://doi.org/10.1017/S0029665112002832 https://doi.org/10.1126/science.1151861

3. Shukla, P. R., Skea, J., Calvo Buendia, E., & Masson- 16. REN21. (2021). Renewables 2021: Global status report.

Delmotte, V. (2020). Climate Change and Land: An IPCC REN21. https://www.ren21.net/reports/global-status-report

Special Report on Climate Change, Desertification, Land 17. De Vos, J. M., Joppa, L. N., Gittleman, J. L., Stephens, P.

Degradation, Sustainable Land Management, Food R., & Pimm, S. L. (2014). Estimating the normal

Security, and Greenhouse Gas Fluxes in Terrestrial background rate of species extinction. Conservation

Ecosystems. Intergovernmental Panel on Climate Change Biology, 29(2), 452-462.

(IPCC), https://www.ipcc.ch/srccl https://doi.org/10.1111/cobi.12380

4. Lal, R. (2021). Soil carbon sequestration: An essential tool 18. Tilman, D., Cassman, K. G., Matson, P. A., Naylor, R., &

for mitigating climate change. Nature Sustainability, 4, Polasky, S. (2002). Agricultural sustainability and intensive

974–976. DOI: 10.1038/s41893-021-00722-9. production practices. Nature, 418(6898), 671-677.

5. Saunois, M., Jackson, R. B., Bousquet, P., Poulter, B., & https://doi.org/10.1038/nature01014

Canadell, J. G. (2020). The growing role of methane in 19. Loreau, M., Naeem, S., Inchausti, P., Bengtsson, J., Grime,

anthropogenic climate change. Environmental Research J. P., Hector, A., ... & Tilman, D. (2001). Biodiversity and

Letters, 15(12), 120207. DOI: 10.1088/1748-9326/ab9ed2. ecosystem functioning: Current knowledge and future

6. Hristov, A. N., Oh, J., Lee, C., Meinen, R., Montes, F., Ott, challenges. Science, 294(5543), 804-808.

T., ... & Waghorn, G. (2021). Mitigation of greenhouse gas https://doi.org/10.1126/science.1064088

emissions in livestock production–A review of technical 20. De Vos, J. M., Joppa, L. N., Gittleman, J. L., Stephens, P.

options for non-CO2 emissions. FAO Animal Production R., & Pimm, S. L. (2014). Estimating the normal

and Health Paper, 177, background rate of species extinction. Conservation

https://www.fao.org/3/cb7508en/cb7508en.pdf Biology, 29(2), 452-462.

7. Ricciardi, A., Lewis, M. A., & Diez, J. M. (2021). https://doi.org/10.1111/cobi.12380

Biological invasions and natural biodiversity patterns: The 21. Kolbert, E. (2014). The sixth extinction: An unnatural

case for agricultural landscapes. Frontiers in Ecology and history. Henry Holt and Co. Evolution, 9, 662836. DOI: 10.3389/fevo.2021.662836.

22. Goodall, C. (2016). The Switch: How solar, storage and new

8. Gams, M. idr. (2020). Bela knjiga o strokovnem varovanju

okolja, Institut tech means cheap power for all. Profile Books. 'Jožef Stefan'

23. Smil, V. (2017). Energy and civilization: A history. MIT

https://www.dlib.si/details/URN:NBN:SI:DOC-I9010JS8

9. Mihalič R. (2015). Zelena energetika, 2. del, Press.

24. Nowak, D. J., & Dwyer, J. F. (2007). Understanding the

https://www.metropolitan.si/novice/zelena-energetika-2-

benefits and costs of urban forest ecosystems. In J. E. Kuser

del/



10. Plut, D. (2023). Ekosistemska družbena ureditev. Prvi zv.: 25-46). Springer. (Ed.), Urban and community forestry in the Northeast (pp.

podstati in gradniki ekosistemske družbene 25. Umweltbundesamt. (2021). Land use and land use change: ureditve,https://ebooks.uni-

Annual report 2021. Umweltbundesamt.

lj.si/ZalozbaUL/catalog/book/405 https://www.umweltbundesamt.de/en/press

11. Gams, M. (2023). Miti in resnice o slovenskem okolju. V

26. United Nations. (2020). UN News: Land degradation

Miti in resnice o varovanju okolja = Myths and truths about

environmental protection : Informacijska družba - IS 2023 = neutrality targets 2030. United Nations.

https://news.un.org/en/story/2020/05

Information Society - IS 2023, zbornik 26. mednarodne

27. FAO. (2020). Global forest resources assessment 2020:

multikonference = proceedings of the 26th International Key findings. FAO. https://www.fao.org/forest-resources-

Multiconference, zvezek D = volume D, 11. oktober 2023 = assessment

11 October 2023, Ljubljana, Slovenia (str. 18-21).

28. Smith, P., Gregory, P. J. (2013). "Climate change and



COBISS.si-ID 171645699.

sustainable food production." Proceedings of the Nutrition

12. Brown, L. R., Hanson, M. E., & Wilson, K. J. (2020). The

Society, 72(1), 21-28.

global impact of human population growth on ecosystem



29. Garnett, T., Appleby, M. C., Balmford, A., Bateman, I. J.,

Benton, T. G., Bloomer, P., ... & Godfray, H. C. J. (2013).

"Sustainable intensification in agriculture: Premises and

policies." Science, 341(6141), 33-34.

30. Jacobson, M. Z. (2009). Review of solutions to global

warming, air pollution, and energy security. Energy &

Environmental Science, 2(2), 148-173.

https://doi.org/10.1039/b809990c

31. World Nuclear Association. (2020). Comparison of

lifecycle greenhouse gas emissions of various electricity

generation sources. World Nuclear Association.

https://www.world-nuclear.org

32. Sovacool, B. K. (2008). Valuing the greenhouse gas

emissions from nuclear power: A critical survey. Energy

Policy, 36(8), 2950-2963.

https://doi.org/10.1016/j.enpol.2008.04.017

33. Markandya, A., & Wilkinson, P. (2007). Electricity

generation and health. The Lancet, 370(9591), 979-990.

https://doi.org/10.1016/S0140-6736(07)61253-7

34. Kharecha, P. A., & Hansen, J. E. (2013). Prevented

mortality and greenhouse gas emissions from historical and

projected nuclear power. Environmental Science &

Technology, 47(9), 4889-4895.

https://doi.org/10.1021/es3051197

35. Sandoval, M. (2023). How lifetime emissions of different

energy sources stack up. World Resources Institute.

https://www.wri.org





Umetna inteligenca o zeleni agendi




(izvedljivost, cena in rezultati zelenega prehoda)



Jure Špiler

Ljubljana, Slovenija

jure@basic.si



ABSTRACT / POVZETEK • - 10 times more people die from cold than heat

• - Warming 2x faster than global average

Dobrobit človeštva se nesmiselno povezuje z zmanjšanjem CO2 • - The hypocrisy of decarbonisation

izpustov (razogljičenjem, Net Zero), kar povzroča ogromne • - Opposition to nuclear power

stroške in še več prikritega onesnaženja, brez otipljivih • - Environmentalism as a cult

rezultatov. Le prilagajanje spremembam v okolju je smiselno in • - The environmental aspect of Bled kremšnites ekonomsko učinkovito.

Samo uporaba preizkušenih obstoječih virov zanesljive energije OPTIONAL: KEYWORDS



(reke, nafta, plin, jedrska tehnologija) omogoča normalno CO2, decarbonisation, population, energy, gas, oil, nuclear,

življenje in razvoj. economics



S pomočjo umetne inteligence (https://www.perplexity.ai/) sem

poiskal odgovore na aktualna vprašanja: 1 Stroški razogljičenja



• To je kompleksno in večplastno vprašanje, ki vključuje znatne Stroški razogljičenja (Net Zero)

• finančne naložbe v različnih sektorjih in regijah. Tukaj je nekaj Cena hlajenja ($/°C)

• ključnih vidikov: Dosedanje segrevanje in vpliv na populacijo

• Bo segrevanje zmanjšalo število ljudi?

• 1.1 Ocene skupnih stroškov Zaradi mraza umre 10-krat več ljudi kot zaradi

vročine Globalne naložbene zahteve: Ocenjuje se, da bo za doseganje

• Segrevanje 2x hitrejše od svetovnega povprečja ničelnih neto emisij od leta 2021 do 2050 potrebnih približno

• Hinavščina razogljičenja 110 bilijonov dolarjev globalnih kapitalskih naložb, kar je v

• Nasprotovanje jedrski energiji povprečju 3,5 bilijona dolarjev letno. Ta naložba predstavlja

• Okoljevarstvo kot kult približno 1,3 % predvidenega svetovnega BDP v naslednjih treh

• Okoljski vidik blejskih kremšnit desetletjih [ 4].



KEYWORDS / KLJUČNE BESEDE Naložbe v posamezne sektorje: Velik del teh naložb, približno

70 % ali 2,4 bilijona USD letno, bi moral biti usmerjen v sektor

CO2, razogljičenje, prebivalstvo, energetika, plin, nafta, jedrska

električne energije. To vključuje naložbe v proizvodnjo

energija, ekonomija

električne energije brez emisij ogljika, elektroenergetska

omrežja ter shranjevanje električne energije in prilagodljivost

omrežja[ 4].

OPTIONAL: ABSTRACT



Humanity's well-being is meaninglessly linked to CO2 1.2 Cena na tono CO2

reduction (decarbonisation, Net Zero), which entails huge costs Cene ogljika: Stroške razogljičenja je mogoče razumeti tudi v

and more hidden pollution, with no tangible results. Only smislu cen ogljika. Trenutne ocene kažejo, da morajo biti cene

adapting to environmental change is sensible and economically ogljika med 50 in 100 USD na tono CO2 do leta 2030, da se

efficient. uskladijo s cilji Pariškega sporazuma [ 6]. Svetovna povprečna

cena ogljika je leta 2024 znašala 32 USD na tono , kar je pod



(rivers, oil, gas, nuclear technology) allows normal life and priporočenim razponom [ 6]. Only the use of proven existing sources of reliable energy

development. Stroški razogljičenja glede na kategorijo emisij : Stroški za

razogljičenje posebnih kategorij emisij se razlikujejo. Na

With the help of artificial intelligence primer, cena zemeljskega plina je približno 3505 GBP na tono

(https://www.perplexity.ai/).I have sought answers to the CO2e, medtem ko je za obnovljivo energijo približno 6435

questions at hand: GBP na tono CO2e [ 1].



• - The cost of decarbonisation (Net Zero) 1.3 Regionalne razlike •- Cost of cooling ($/°C) Cene ogljika po regijah: Pobude za določanje cen ogljika se •- Warming to date and impact on population med regijami precej razlikujejo. Evropa in Srednja Azija imata •- Will warming reduce the number of people?



povprečno ceno ogljika 50 USD na tono , medtem ko je cena v Projekcije stroškov za posamezne DAC tehnologije do leta

ZDA in Kanadi približno 48 USD na tono . Druge regije, kot sta 2050:

Latinska Amerika in Karibi, imajo nižje povprečne cene, okoli • Climeworks: 280-580 dolarjev na tono

24 USD na tono [ 6]. • Carbon Engineering: 226-544 dolarjev na tono

Naložbene potrebe po regijah: Naložbene zahteve za doseganje • Heirloom: 230-835 dolarjev na tono[6][8]

ničelnih neto emisij se prav tako razlikujejo glede na regijo, na To je približno dvakrat več od nekaterih prejšnjih ocen[6].



kar vplivajo obstoječa infrastruktura, politični okviri in Ključno za znižanje stroškov je povečanje obsega uporabe gospodarski pogoji[4]. DAC, vendar je v kratkem roku potrebno več zmogljivosti, kot



1.4 Izzivi in premisleki stroški posameznih komponent, stroški energije in stroški je trenutno načrtovano[4]. Na skupne stroške DAC vplivajo

Socialni in gospodarski vplivi: Prehod na ničelne neto emisije integracije[3]. Bolj dovršene komponente, kot so kompresorji,

bo imel široke gospodarske in družbene učinke, ki bodo vplivali imajo manjši potencial za znižanje stroškov[8]. na povpraševanje, kapitalsko porabo, stroške in zaposlovanje v



različnih sektorjih[ 7]. cene blizu ali na ravni 340 dolarjev na tono, ki je predvidena za Kljub višjim napovedim nekatera DAC podjetja že ponujajo

Politika in izvajanje: Učinkovita dekarbonizacija zahteva trdne leto 2050[4]. Utežena povprečna cena za DAC s shranjevanjem

politične okvire, vključno z mehanizmi za določanje cen se je med letoma 2022 in 2023 znižala za 43 %, s 1.261 na 715

ogljika, spodbudami za čisto tehnologijo in mednarodnim dolarjev na tono[4]. Nekatera podjetja trdijo, da so trenutni

sodelovanjem za zagotovitev pravičnega in učinkovitega stroški zajemanja pod 500 ali celo 200 dolarjev na tono[4].

izvajanja[ 5][6].

Čeprav se pričakuje, da se bodo stroški z razvojem tehnologije

Povzetek bistveno znižali, nedavne raziskave kažejo, da znižanja morda

Stroški razogljičenja vključujejo velike finančne obveznosti v ne bodo tako dramatična, kot so upali prej. Kljub temu se zdi,

različnih sektorjih in regijah, s precejšnjimi razlikami v cenah da industrija DAC napreduje hitreje pri zniževanju stroškov, kot

ogljika in naložbenih potrebah. Za doseganje globalnih ciljev predvidevajo nekatere študije. Priporoča se nadaljnji razvoj več

razogljičenja bodo potrebna usklajena prizadevanja in strateške pristopov DAC za spodbujanje inovacij in znižanje

naložbe za uravnoteženje gospodarske rasti z okoljsko stroškov[6][8].

trajnostjo.

2.1 Strošek odstranjevanja 1ppm CO2 iz zraka



REFERENCE 1 ppm atmosferskega CO2 ima maso 7,76 gigaton (Gt) CO2.

[1] https://www.eauc.org.uk/file_uploads/20230524_hfe_v4_0_- • Pri ceni 100 US$ na tono bi stroški znašali 776

_cost_of_net_zero_report_1.pdf milijard US$.



[2] https://www.oecd-ilibrary.org/sites/0e8e24f5- • Po trenutnih cenah 700-1200 US$/t to pomeni 5 - 9 en/1/3/3/index.html?itemId=/content/publication/0e8e24f5-bilijonov US$ en&_csp_=9e1e8b508d47c48ebab5cddb217622cd&itemIGO=oec • d&itemContentType=book Do leta 2050 se bi ta vsota prepolovila • Trenutno je svetovni bruto domači proizvod (BDP) [3] https://www.oecd-nea.org/jcms/pl_28600/the-costs-of-101 bilijon US$ decarbonisation-system-costs-with-high-shares-of-nuclear-and-

renewables Za odstranitev 0,2% CO2 (1ppm) bi porabili 10% družbenega

[4] https://decarbonization.visualcapitalist.com/breaking-down-the- proizvoda

cost-of-clean-energy-transition/



[5] https://www.worldbank.org/en/news/press- REFERENCE

release/2024/05/21/global-carbon-pricing-revenues-top-a-record-

100-billion [1] https://www.mdpi.com/2673-8392/3/4/92

[6] https://decarbonization.visualcapitalist.com/visualized-the-price- [2] https://www.iea.org/reports/direct-air-capture-2022/executive-

of-carbon-around-the-world-in-2024/ summary

[7] https://www.mckinsey.com/capabilities/sustainability/our- [3] https://www.nature.com/articles/s44172-023-00152-6

insights/the-net-zero-transition-what-it-would-cost-what-it-could- [4] https://www.weforum.org/agenda/2023/08/how-to-get-direct-air-

bring capture-under-150-per-ton-to-meet-net-zero-goals/

[5] https://www.wri.org/insights/direct-air-capture-resource-



2 Stroški odstranjevanja CO2 iz ozračja considerations-and-costs-carbon-removal

[6] https://www.sciencedaily.com/releases/2024/03/240304135808.ht

Trenutni stroški za neposredno zajemanje CO2 iz zraka so m

ocenjeni med 600 in 1.000 dolarjev na tono odstranjenega [7] https://www.asiafinancial.com/cost-of-carbon-capture-by-2050-

CO2[4][8]. Za široko uporabo tehnologije neposrednega may-be-double-estimates-ime

zajemanja iz zraka (direct air capture - DAC) morajo stroški [8] https://carbonherald.com/new-study-places-future-direct-air-



pasti pod 200 dolarjev na tono, idealno bližje 100 dolarjev na [9] capture-costs-230-540-range/

tono do leta 2050[4][9] https://www.science.org/content/article/cost-plunges-capturing-

carbon-dioxide-air

Raziskovalci ETH Zürich ocenjujejo, da bodo stroški leta 2050

predvidoma med 230 in 540 dolarji na tono, namesto pogosto

omenjenega razpona 100-300 dolarjev[6][7].



3 • Povprečne površinske temperature so v zadnjih letih dosledno Stroški hlajenja ($/°C)

presegale 1,5°C nad predindustrijskimi vrednostmi.

Na podlagi posredovanih rezultatov iskanja je težko podati

natančno oceno hladilnega učinka v stopinjah Celzija na

4.2 Rast prebivalstva

milijardo dolarjev, porabljenih za razogljičenje. Vendar pa

lahko naredimo nekaj sklepov: • Svetovno prebivalstvo je zraslo z okoli 1,6 milijarde leta 1900

na 8,1 milijarde leta 2024.

• Odlašanje z ukrepi za zajezitev podnebnih sprememb ima

znatne gospodarske stroške. Zamuda, ki povzroči segrevanje za • Stopnja rasti prebivalstva je dosegla vrhunec 2,2% letno leta

3 °C nad predindustrijsko ravnjo namesto za 2 °C, bi povzročila 1963 in se od takrat znižala na približno 1,1% letno.

dodatno škodo v višini 0,9 % svetovne proizvodnje, kar pomeni

približno 150 milijard dolarjev samo za ameriško 4.3 Vplivi podnebnih sprememb na prebivalstvo

gospodarstvo[3]. • Podnebne spremembe vplivajo na smrtnost in življenjske

• Obnovljivi viri energije bi lahko do leta 2050 razogljičili 90 % pogoje preko neposrednih vplivov ekstremnih temperatur in

elektroenergetske industrije, kar bi drastično zmanjšalo emisije posrednih vplivov, ki jih posredujejo okoljski in družbeni

ogljika in prispevalo k ublažitvi podnebnih sprememb [ 4]. sistemi.

Stroški tehnologij obnovljive energije, kot sta sončna energija • Ogrožajo prehransko varnost, oskrbo s sladko vodo in zdravje

in veter, so se od leta 2010[4] znatno znižali. ljudi.

• Ocene skupnih stroškov za odpravo podnebnih sprememb • Dvig morske gladine ogroža stotine milijonov ljudi v obalnih

segajo od 300 milijard do 50 bilijonov dolarjev v naslednjih območjih.

dveh desetletjih[ 5]. Ta ogromen razpon odraža negotovost in • Podnebne spremembe lahko prisilijo ljudi k migraciji, z

zapletenost pri določanju optimalne strategije razogljičenja. ocenami do več sto milijonov podnebnih migrantov do leta

• V scenariju IEA o neto ničelnih emisijah do leta 2050 morajo 2050.

posredne emisije CO2, povezane s povpraševanjem po hlajenju

prostorov, do leta 2030 pasti na približno 40 % današnje ravni 4.4 Rast prebivalstva kot gonilnik podnebnih

[1]. To bi pomenilo, da se povprečna učinkovitost nove sprememb

klimatske opreme do leta 2035 približa najboljši razpoložljivi • Rast prebivalstva skupaj s povečano potrošnjo običajno

ravni [1]. povečuje emisije toplogrednih plinov.



Medtem ko rezultati iskanja ne zagotavljajo natančnega • Vpliva tudi na sposobnost Zemlje, da se upre podnebnim kvantitativnega razmerja med stroški razogljičenja in znižanjem spremembam in absorbira emisije, na primer s krčenjem temperature, pa kažejo, da lahko pomembne naložbe v gozdov, ko se zemljišča spreminjajo v kmetijske površine za obnovljivo energijo, energetsko učinkovitost in zmanjšanje prehrano rastočega prebivalstva. emisij pomembno vplivajo na omejevanje prihodnjega dviga



globalne temperature. 4.5 Demografski dejavniki in podnebna

Povzetek ranljivost

Natančen hladilni učinek na porabljen dolar ostaja negotov in se • Ljudje v revnih regijah, kljub temu da malo prispevajo k

razlikuje glede na izvedene posebne ukrepe za dekarbonizacijo. emisijam, so najbolj ranljivi za vplive podnebnih sprememb.

• Visoke stopnje revščine in socialne neenakosti povečujejo

REFERENCE ranljivost za ekstremne vremenske pojave, vodni stres in izzive

[1] https://www.iea.org/energy-system/buildings/space-cooling pri pridelavi hrane.

[2] https://www.ipcc.ch/report/ar6/syr/downloads/report/IPCC_AR6_

SYR_FullVolume.pdf 4.6 Populacijska dinamika in podnebna politika

[3] https://scholar.harvard.edu/files/stock/files/cost_of_delaying_actio

n.pdf • Upočasnitev rasti prebivalstva s pravičnimi ukrepi, kot je

[4] https://link.springer.com/article/10.1007/s10311-022-01532-8 povečanje dostopa do načrtovanja družine, bi lahko igrala

[5] https://www.globalgiving.org/learn/cost-to-end-climate-change/ ključno vlogo pri blaženju podnebnih sprememb.

[6] https://www.iea.org/reports/the-future-of-cooling • Vendar lahko tudi v planih z nizko rastjo prebivalstva ogljično

[7] https://www.irena.org/Digital-Report/World-Energy-Transitions- intenzivna gospodarska rast še vedno povzroči visoke emisije.

Outlook-2023



[8] https://carbonfix.org/blog/cooling-the-planet-by-0-5c-it-only- 4.7 Zgodovinski kontekst

takes-e8-billion-a-year

• Pospeševanje podnebnih sprememb sovpada z obdobji hitre

industrializacije in rasti prebivalstva, zlasti od sredine 20.

4 Dosedanje segrevanje in vpliv na populacijo stoletja.



4.1 Globalno povečanje temperature Povzetek



• Od leta 1880 se je povprečna površinska temperatura Zemlje Čeprav naraščajoče temperature niso neposredno povzročile rasti prebivalstva, so začele vplivati na človeške populacije na dvignila za 0,07°C na desetletje. različne načine. Hkrati je rast prebivalstva prispevala k • Stopnja povečanja temperature se je od leta 1981 več kot podnebnim spremembam, kar ustvarja kompleksno povratno podvojila na 0,18°C na desetletje.

zanko med demografskimi trendi in okoljskimi spremembami. 5.1 Omejitve Odnos med podnebnimi spremembami in prebivalstvom je

Pomembno je omeniti, da čeprav lahko podnebne spremembe

večplasten in vključuje vprašanja emisij, ranljivosti,

prispevajo k upadu prebivalstva v določenih regijah, je malo

prilagajanja in strategij blaženja.

verjetno, da bodo v bližnji prihodnosti bistveno zmanjšale

Povprečna temepertura in prebivalstvo svetovno prebivalstvo. Na rast prebivalstva vplivajo številni

Zemlje od leta 1900 zapleteni dejavniki, splošni vpliv podnebnih sprememb na

število svetovnega prebivalstva pa ostaja negotov.

20

Povzetek

15 Čeprav ni pričakovati, da bi dvig globalnih temperatur za 2

10 stopinji neposredno zmanjšal svetovno prebivalstvo, bi lahko

5 povzročil povečano umrljivost, krajšo pričakovano življenjsko

dobo, razseljevanje in nižjo stopnjo rodnosti v močno prizadetih

0

regijah. Blaženje podnebnih sprememb in prilagajanje njihovim

1900 1950 2000 vplivom bo ključnega pomena za zmanjšanje teh možnih

vplivov na prebivalstvo.

T [°C] Prebivalsvo [milijarde]

Medicinska revija The Lancet ocenjuje, da je dostop do

klimatskih naprav v letu 2019 preprečil skoraj 200.000 smrti

Slika1: med ljudmi, starimi 65 let ali več[9]. https://www.perplexity.ai/search/make-graph-of-

temaperature-ris-StW7v2gkTGS325Rivulm9A





5 Ali segrevanje zmanjšuje število ljudi?



Dvig globalnih temperatur za 2 stopinji Celzija verjetno ne bo

neposredno znižal svetovnega prebivalstva, lahko pa ima

pomembne posredne učinke, ki lahko upočasnijo rast

prebivalstva ali celo povzročijo upad prebivalstva v določenih

regijah. Tukaj so ključne točke:

• Smrti zaradi vročine : višje temperature bodo povečale

tveganje za bolezni in umrljivost zaradi vročine, zlasti med

ranljivimi skupinami prebivalstva, kot so starejši, otroci in Slika2: https://www.economist.com/science-and-

osebe z že obstoječimi zdravstvenimi težavami. To bi lahko less-harmful-to-the-planet technology/2024/09/04/new-tech-can-make-air-conditioning-



povzročilo višje stopnje smrtnosti na prizadetih območjih.



• Vplivi na zdravje : Podnebne spremembe lahko tudi REFERENCE

poslabšajo kakovost zraka, širijo nalezljive bolezni ter motijo [1] https://www.washingtonpost.com/climate-

oskrbo s hrano in vodo, kar dodatno poveča zdravstvena environment/2021/10/11/85-percent-population-climate-impacts/

tveganja in umrljivost. [2] https://www.thedailystar.net/environment/global-

• Zmanjšanje pričakovane življenjske dobe : V regijah, ki so jih warming/news/how-temperature-rise-2-degrees-celsius-impacts-

billions-3327291

močno prizadele podnebne spremembe, se lahko pričakovana [3] https://populationconnection.org/resources/population-and-

življenjska doba skrajša zaradi skupnih učinkov toplotnega climate/

stresa, podhranjenosti in bolezni. To bi lahko upočasnilo rast [4] https://www.prb.org/resources/climate-change-impacts-emerging-

prebivalstva ali celo povzročilo upad prebivalstva. population-trends-disaster/

• Podnebni begunci: Dvig morske gladine, ekstremni vremenski [5] https://www.nytimes.com/interactive/2018/10/07/climate/ipcc-

pojavi in degradacija okolja bi lahko povzročilo povečano report-half-degree.html

preseljevanje. S tem bi se povečal pritisk na vire in [6] https://www.euronews.com/green/2023/09/08/virtually-no-one-

on-earth-escaped-98-of-worlds-population-exposed-to-global-

infrastrukturo sprejemnih območji ter povzročilo socialno in warming-this-shttps://frontline.thehindu.com/news/climate-

gospodarsko nestabilnost. change-explained-how-a-temperature-rise-of-2c-impacts-



• Konflikti in nemiri: S podnebjem povezano razseljevanje in [7] billions/article66885235.ece

pomanjkanje virov lahko prav tako prispevata k konfliktom, ki [8] https://populationmatters.org/climate-change/

lahko dodatno destabilizirajo prizadete regije in povzročijo technology/2024/09/04/new-tech-can-make-air-conditioning-less- https://www.economist.com/science-and-

upad prebivalstva. harmful-to-the-planet?etear=science_nl_1

• Družbeno-ekonomski vplivi: gospodarske in družbene motnje,



ki jih povzročajo podnebne spremembe, kot je zmanjšana 6 Zaradi mraza umre 10-krat več ljudi kot kmetijska produktivnost, bi lahko povzročile večjo revščino in zaradi vročine nestabilnost. To lahko povzroči zmanjšanje stopnje rodnosti, saj

imajo ljudje manj otrok zaradi gospodarske negotovosti ter Čeprav je res, da je v preteklosti število smrti zaradi mraza

zmanjšanega znatno preseglo smrti zaradi vročine, se dinamika zaradi dostopa do zdravstvenega varstva in

izobraževanja. podnebnih sprememb spreminja. Tukaj so ključne točke v zvezi





z vplivom naraščajočih temperatur na stopnjo umrljivosti in

morebitne posledice za svetovno prebivalstvo:



6.1 Trenutni trendi smrti zaradi temperature

• Smrti zaradi mraza v primerjavi z vročino : Študije kažejo, da

na svetu zaradi mraza umre približno devetkrat več kot zaradi

vročine. Študija, objavljena v »The Lancet Planetary Health«, je

pokazala, da je v zadnjih dveh desetletjih izpostavljenost mrazu

predstavljala veliko večino smrtnih primerov, povezanih s

temperaturo, pri čemer je bilo približno 8,5 % smrti pripisanih

mrazu in samo 0,9 % vročini[ 1][2][4].

• Spreminjanje vzorcev : Čeprav trenutno prevladujejo smrti

zaradi prehlada, se trend spreminja. Smrti zaradi vročine zaradi

podnebnih sprememb naraščajo, število smrti zaradi mraza pa



se zmanjšuje. Ta premik nakazuje, da se bo ravnovesje v Slika3:

prihodnosti lahko nagnilo k večjemu številu smrtnih žrtev, https://x.com/BjornLomborg/status/1409119091290521601

povezanih z vročino [ 1][4][5].



Posledice naraščajočih temperatur REFERENCE

• Povečana smrtnost zaradi vročine : Ko se temperature [1] https://www.theguardian.com/world/2021/jul/08/extreme-

dvignejo, zlasti v ranljivih regijah, se pričakuje, da se bo število temperatures-kill-5-million-people-a-year-with-heat-related-

smrti zaradi vročine povečalo. To je še posebej zaskrbljujoče za deaths-rising-study-finds

prebivalstvo na območjih, ki niso vajena ekstremne vročine, in [2] https://ourworldindata.org/part-one-how-many-people-die-from-

kjer so prilagoditveni ukrepi morda nezadostni[ 1][4]. [3] extreme-temperatures-and-how-could-this-change-in-the-future

https://www.epa.gov/climate-indicators/climate-change-

• Potencialni vpliv na prebivalstvo : Čeprav celotno število indicators-cold-related-deaths

svetovnega prebivalstva morda ne bo takoj upadlo zaradi [4] https://www.forbes.com/sites/joshuacohen/2023/07/19/excessive-

povečanega števila smrti, povezanih z vročino, bi lahko v summer-heat-can-kill-but-extreme-cold-causes-more-fatalities/

nekaterih regijah prišlo do znatnega povečanja umrljivosti in [5] https://www.washingtonpost.com/climate-

upočasnilo rast prebivalstva ali povzročilo lokalno zmanjšanje. environment/interactive/2023/hot-cold-extreme-temperature-

deaths/

Ranljivo prebivalstvo, zlasti na območjih z nizkimi dohodki, je [6] https://thebreakthrough.org/issues/energy/human-deaths-from-hot-

lahko nesorazmerno prizadeto, kar še poslabša obstoječe and-cold-temperatures-and-implications-for-climate-change

zdravstvene razlike in potencialno vodi do večjih migracijskih [7] https://www.wunderground.com/cat6/Which-Kills-More-People-

pritiskov[ 1][2][5]. Extreme-Heat-or-Extreme-Cold

• Dolgoročne projekcije : Ker se podnebne spremembe še [8] https://ourworldindata.org/part-two-how-many-people-die-from-

naprej odvijajo, bi lahko vse večja razširjenost smrti zaradi extreme-temperatures-and-how-could-this-change-in-the-future

vročine sčasoma uravnotežila ali presegla smrt zaradi mraza,

zlasti če se ne izvajajo prizadevanja za ublažitev. To bi lahko 7 Segrevanje 2x hitrejše od svetovnega

vodilo do višjih splošnih stopenj umrljivosti in vplivalo na povprečja populacijsko dinamiko skozi čas[ 2][4].

Samo nekaj področij je, kjer se to res dogaja

Povzetek • Arktika, zlasti polotok Tajmir v Rusiji, se segreva 2,5- do 2,8-

Če povzamemo, čeprav število smrti zaradi mraza trenutno krat hitreje od svetovnega povprečja. V zadnjih 10 letih so

znatno presega število smrti zaradi vročine, podnebne tamkajšnje temperature narasle za 0,8 °C do 1,2 °C (1,4 °F do

spremembe spreminjajo to pokrajino. Pričakuje se, da bo dvig 2,1 °F) hitreje od svetovnega povprečja [ 4].

globalnih temperatur povečal smrtnost zaradi vročine, kar bi • Rusko ozemlje se na splošno segreva skoraj dvakrat hitreje od

lahko imelo globoke posledice za zdravje in dinamiko svetovnega povprečja kopnega – pri 0,51 °C (0,91 °F) na

prebivalstva, zlasti v ranljivih regijah. Takojšnje ukrepanje za desetletje v primerjavi z 0,295 °C (0,531 °F) globalno nad

ublažitev podnebnih sprememb je ključnega pomena za kopnim[4]. preprečitev teh neugodnih rezultatov.

• Antarktika se segreva dvakrat hitreje od svetovnega

povprečja, s približno 20–50 % večjim segrevanjem od ocen iz

podnebnih modelov. To temelji na rekonstrukciji temperatur na

Antarktiki v preteklem tisočletju z uporabo zapisov ledenih

jeder [ 5].

• Podnebni modeli predvidevajo, da se bo pas južne Azije, ki se

razteza od Perzijskega zaliva preko Indije in Kitajske, segreval

hitreje kot skoraj kjer koli drugje na Zemlji v različnih

scenarijih podnebnih sprememb. V tej regiji živi več kot

milijarda ljudi[ 3].



Povzetek, • Podjetja so včasih obtožena zelenega pranja – podajanja

Arktika, Rusija, Antarktika in deli južne Azije se soočajo z zavajajočih trditev o svojih okoljskih prizadevanjih, da bi bila

dvigi temperature, ki so približno dvakrat višji od svetovnega videti bolj trajnostna, kot so. Vendar pa nekatere študije kažejo,

povprečja, pri čemer se Arktika segreva najhitreje in je dvig da si številna podjetja, ki se ukvarjajo s trgi ogljičnih kreditov,

temperature skoraj trikrat višji. Hitro segrevanje v teh regijah je resnično prizadevajo zmanjšati svoje emisije, čeprav se

velika skrb zaradi velike populacije v južni Aziji in možnih razpravlja o učinkovitosti in iskrenosti teh prizadevanj [ 6].

vplivov segrevanja Arktike in Antarktike na podnebni sistem.

8.3 Sistemska odgovornost proti individualni

Ali se res 2x hitreje segreva tudi Slovenija [9]?

odgovornosti



REFERENCE • Kritiki trdijo, da osredotočanje na posamezna dejanja, kot je

[1] https://climate.copernicus.eu/global-temperature-exceeds-2degc- sramotenje bega, preveč poenostavlja vprašanje in ne obravnava

above-pre-industrial-average-17-november sistemskih sprememb, potrebnih za boj proti podnebnim

[2] https://www.climatecentral.org/report/global-heat-review-june- spremembam. Zanašanje na fosilna goriva je za posameznike

2024 pogosto neizogibno zaradi trenutne infrastrukture in

[3] https://climate.mit.edu/ask-mit/where-will-people-experience- gospodarskih sistemov, zaradi česar so obtožbe osebne

most-warming-climate-change hinavščine manj vplivne[3][5].

[4] https://www.aa.com.tr/en/environment/climate-change-russia-is-

heating-up-almost-twice-as-fast-the-world-warns- • Obstaja poziv k sistemskim spremembam, ki omogočajo bolj

scientist/3278123 trajnostne odločitve, s poudarkom na potrebi po večjih

[5] https://www.carbonbrief.org/guest-post-ice-cores-reveal- strukturnih reformah namesto zgolj osredotočenju na

antarctica-is-warming-twice-as-fast-as-global-average/ posameznikovo vedenje[ 4].

[6] https://www.channelnewsasia.com/cnainsider/singapore-hot-



weather-urban-heat-effect-temperature-humidity-906231 8.4 Podnebne konference in politične zaveze

[7] https://www.pbs.org/newshour/politics/the-arctic-is-warming-

nearly-four-times-faster-than-the-rest-of-the-world • Mednarodne podnebne konference pogosto razkrivajo vrzel

[8] https://www.canada.ca/en/environment-climate- med zavezami in dejanji. Razvite države so bile kritizirane

change/news/2019/04/canadas-climate-is-warming-twice-as-fast- zaradi zastoja pri finančni podpori za podnebno ranljive države

as-global-average in nezmožnosti ustreznega obravnavanja izgube in škode [ 8].

[9] https://www.rtvslo.si/okolje/v-sloveniji-se-ozracje-segreva- • Kljub obljubam, da bodo zmanjšale emisije in podprle

hitreje-od-svetovnega-povprecja/653014

globalna podnebna prizadevanja, številne države še naprej

izvajajo projekte fosilnih goriv doma, kar poudarja

8 Hinavščina razogljičenja nepovezanost med njihovo mednarodno retoriko in domačo

Koncept hinavščine v prizadevanjih za razogljičenje je politiko [ 8].

večplastno vprašanje, ki vključuje neskladja med navedenimi

zavezami za zmanjšanje emisij ogljika in ukrepi, ki jih 8.5 Nasprotovanje jedrski energiji

sprejmejo posamezniki, korporacije in države. V diskurzu o • Nasprotovanje jedrski energiji ob hkratnem zagovarjanju

podnebnih spremembah in dekarbonizaciji je izpostavljenih več zmanjšanja emisij ogljika je nedosledno, saj bi jedrska energija

vidikov te hinavščine. lahko imela pomembno vlogo pri razogljičenju proizvodnje

električne energije. [9][10]

8.1 Razlika med bogatimi in revnimi narodi • Na protijedrska stališča nekaterih okoljskih skupin vpliva

Bogate države, ki so v preteklosti največ prispevale k emisijam financiranje s strani donatorjev z izrecno protijedrskimi

toplogrednih plinov, pogosto spodbujajo podnebne politike, ki programi in ne zgolj okoljska vprašanja. [9]

nesorazmerno bremenijo države v razvoju. Mehanizmi, kot je • Kritiki trdijo, da je zaradi zavračanja jedrske energije veliko

mehanizem Evropske unije za prilagajanje meja ogljika, so na težje doseči podnebne cilje, saj se s tem odstrani preverjena

primer kritizirani zaradi zaostrovanja globalnih neenakosti z možnost nizkoogljične tehnologije.[9][10] nalaganjem stroškov revnejšim državam, medtem ko bogate

države še naprej porabljajo blago, proizvedeno z visokimi Povzetek

okoljskimi stroški[2][7]. Hinavščina prizadevanj za razogljičenje je zapleteno vprašanje,

• Hinavščino dodatno poudarja dejstvo, da bogate države ki vključuje tako dejanja posameznikov kot politike držav.

pogosto selijo energetsko intenzivne industrije v države v Poudarja potrebo po resničnih zavezah in sistemskih

razvoju, pri čemer dejansko izvažajo emisije, medtem ko še spremembah za učinkovito reševanje podnebnih sprememb in

naprej porabljajo nastale izdelke [ 2 ]. zmanjšanje svetovnih emisij ogljika. Najhujše pa je

nasprotovanje jedrski energiji, ki zaenkrat edina zagotavlja

8.2 zanesljivo in čisto energijo. Individualna in korporativna dejanja

• Posameznikom in korporacijam pogosto očitajo hinavščino,

ko se zavzemajo za podnebne ukrepe, medtem ko se ukvarjajo z REFERENCE

ogljično intenzivnimi dejavnostmi. To vključuje uporabo [1] https://onlinelibrary.wiley.com/doi/full/10.1111/josp.12522

fosilnih goriv za potovanja ali proizvodnjo ob javnem [2] https://thesun.my/opinion-news/rich-nation-hypocrisy-

spodbujanju trajnosti[ 3][4]. accelerating-global-heating-FJ12394076

[3] https://www.axios.com/2019/09/16/energy-climate-change-

hypocrisy



[4] https://www.resilience.org/stories/2017-08-07/hypocrisy- nujnost okoljskih vprašanj zahteva odločno ukrepanje in da je

environmentalists-need-economic-growth/ prikaz okoljevarstva kot kulta napačna opredelitev, ki

[5] https://www.theguardian.com/lifeandstyle/2016/apr/05/environme spodkopava upravičene skrbi glede ekološke degradacije in

ntally-friendly-green-living-ideas podnebnih sprememb.

[6] https://time.com/6323461/carbon-credits-greenwashing/

[7] https://www.theguardian.com/global- Povzetek

development/2022/jan/28/west-accused-of-climate-hypocrisy-as- Medtem ko nekateri kritiki dojemajo nekatere elemente

emissions-dwarf-those-of-poor-countries okoljevarstva kot kultne, to stališče ni splošno sprejeto in mu

[8] https://climatenetwork.org/2022/06/16/eu-hypocrisy-as-a-climate- pogosto oporekajo tisti, ki vidijo okoljsko zagovorništvo kot

champion-exposed-at-bonn-climate-conference%EF%BF%BC/

[9] racionalen in nujen odgovor na pereče globalne izzive. https://www.learnliberty.org/blog/why-mainstream-

environmentalism-hates-nuclear/

[10] https://www.npr.org/2022/08/30/1119904819/nuclear-power- REFERENCE

environmentalists-california-germany-japan [1] https://www.the401stprophet.com/the-green-cult

[11] https://www.dw.com/en/fact-check-is-nuclear-energy-good-for- [2] https://theprioryrecord.com/2023/04/the-cult-of-radical-

the-climate/a-59853315 environmentalism/

[3] https://www.theguardian.com/commentisfree/2007/jul/03/environ



9 Okoljevarstvo kot kult [4] mentalismthenewdeathcult

https://www.spiked-online.com/2022/12/30/the-rise-of-the-eco-

O pojmu okoljskega gibanja kot o "kultu" se je razpravljalo v cult/

različnih kontekstih, pogosto s strani kritikov, ki trdijo, da [5] https://www.theotivity.com/post/the-climate-change-cult-its-lies-

nekateri vidiki okoljskega gibanja kažejo značilnosti, podobne worldview-and-politics/

tistim pri verskem ali kultu podobnem vedenju. Tukaj je nekaj [6] https://pacificlegal.org/apocalyptic-environmentalists-want-fewer-

ključnih točk: humans-on-earth/

[7] https://theweek.com/articles/633656/death-cult-environmentalism



9.1 Značilnosti kultnega okoljevarstva

• Apokaliptične pripovedi : Kritiki trdijo, da imajo nekateri 10 Okoljski vidik blejskih kremšnit

okoljevarstveniki apokaliptični pogled na svet in napovedujejo Izlet do Blejskega jezera, da bi uživali v lokalni specialiteti ,

katastrofalne rezultate, če ne bodo sprejeti takojšnji in drastični kremšnitah , ima številne okoljske vidike, ki jih je vredno

ukrepi. To odraža napovedi sodnega dne, ki so pogosto raziskati. Blejsko jezero je naravna znamenitost, ki obiskovalce

povezane s kulti [ 4][6]. privablja s svojo lepoto in edinstvenim ekosistemom. Vendar

• Moralni puritanizem in ekstremizem : Okoljevarstveno lahko povečan obisk povzroči preobremenitev okolja, kar lahko

gibanje je včasih opisano kot izkazovanje moralnega vpliva na kakovost vode in biotsko raznovrstnost. Pomembno

puritaizma, z ekstremnimi dejanji aktivistov, da bi opozorili na je, da obiskovalci ravnajo odgovorno, spoštujejo naravo in se

svoje vzroke. Ta dejanja lahko vključujejo moteče proteste in zavedajo svojega vpliva na okolje.

simbolična dejanja vandalizma, ki veljajo za izraze fanatične Priprava kremšnit na Bledu pogosto vključuje lokalne sestavine,

predanosti cilju [ 4]. ki lahko pozitivno vplivajo na okolje. Uporaba lokalno

• Verski jezik in prakse : Za okoljevarstvo včasih pravijo, da pridelanih sestavin zmanjšuje ogljični odtis, povezan s

prevzema verski jezik in prakse, kot je ideja o "ogljičnih prevozom hrane. Poleg tega spodbujanje lokalnih kmetov in

grehih" in potreba po "pokori" z izravnavo ogljika. To je proizvajalcev prispeva k trajnostnemu razvoju regije. Paziti pa

mogoče razumeti kot analogno verski izpovedi in odkupni je treba tudi na embalažo in odpadke, ki nastanejo med

daritvi [ 3]. potovanjem, saj lahko nepravilno ravnanje dodatno

• Protičloveški pogled na svet : Nekateri kritiki trdijo, da se obremenjuje okolje.

lahko radikalno okoljevarstvo nagiba k protičloveški Priljubljenost blejskih kremšnit odpira tudi vprašanje trajnosti

perspektivi, ki na človeštvo gleda kot na grožnjo planetu. Ta turizma. Obiskovalce je treba ozaveščati o pomenu ohranjanja

perspektiva lahko zagovarja nadzor nad populacijo ali prikazuje naravnih virov in kulturne dediščine. Dejavnosti, kot so

ljudi kot "vrsto plevela", kar lahko razumemo kot organizirani izleti in delavnice o trajnostnem turizmu, lahko

razvrednotenje človeškega življenja[2][7]. pripomorejo k ozaveščanju o okoljskih izzivih in spodbujajo

• Sovražnost do drugače mislečih : Okoljevarstvo je včasih odgovorno vedenje . Tako lahko izlet na kremšnite na Blejsko

kritizirano zaradi nestrpnosti do nasprotujočih si pogledov, jezero postane ne le kulinarično doživetje, ampak tudi

označevanja skeptikov kot "zanikovalcev" in odvračanja od priložnost za izobraževanje in aktivno sodelovanje pri

odprte razprave. To odraža, kako kulti pogosto zatirajo varovanju okolja.

nestrinjanje, da bi ohranili enoten sistem prepričanj[ 3].



9.2 Protiargumenti

Čeprav te kritike obstajajo, je pomembno opozoriti, da vodilno

(mainstream) okoljevarstvo ni monolitno in vključuje širok

spekter prepričanj in praks. Številni okoljevarstveniki se

osredotočajo na dokazih temelječe pristope k obravnavanju

podnebnih sprememb in spodbujanju trajnosti. Trdijo, da





11 Pričakovan rezultat razogljičenja

• Vreme bo enako.

• Demontirali bomo kapitalizem in odpravili fosilna goriva.



• Pa kremšnitam smo se bomo odpovedali ;=(((





Slika 4: Blejska kremšnita

https://the-slovenia.com/en/slovenia/blejska-kremsnita-

blejska-kremna-torta



10.1 Izračun ogljičnega odtisa

Ogljični odtis potovanja na Blejske kremšnite je odvisen od



številnih dejavnikov, kot so prevoz, hrana in namestitev. Če

upoštevamo, da se obiskovalec pripelje iz Ljubljane z avtom,

poje eno kremšnito in se vrne domov, lahko ocenimo, da je

njegov ogljični odtis približno 20,4 kg CO2e.



10.2 Emisije v prometu:

Razdalja LJ-Bled 50 km, skupaj torej 100 km.

Izpust prevoza z avtomobilom: 200g CO2 / km * 100 km = 20

kg CO2.

• Izlet na Bled proizvede 20kg CO2.



10.3 Izpust CO2 kremšnite

Sestavine 1 kg kremšnite: listnato testo: 300 g , vaniljev puding:

400 g , sladka smetana: 300 g

Listnato testo: 2,5 kg CO2 / kg × 0,3 kg = 0,75 kg CO2

Vaniljev puding: 1,5 kg CO2 / kg × 0,4 kg = 0,6 kg CO2

Sladka smetana: 3,0 kg CO2 / kg × 0,3 kg = 0,9 kg CO2

Ogljični odtis znaša 2,25 kg CO2 na kilogram kremšnit.

Velikost kremšnite je 7 cm x 7 cm in tehta 180 gramov.

• Ogljični odtis kremšnite je 0,41 kg CO2 / porcijo, kar ustreza

10 uram človeškega dihanja[9] in 1/50 izleta na Bled.



REFERENCE

[1] https://www.sketa.si/kremne-rezine/

[2] https://conditus.si/sl/narocila/blejska-kremsnita/

[3] https://sl.wikipedia.org/wiki/Kremna_rezina

[4] https://www.kulinarika.net/recepti/sladice/ostalo/kremsnite-

kremne-rezine-kot-z-bleda-/13838

[5] https://www.slovenian-kitchen.com/slovenski-

koticek/2022/03/20/bled-cream-cake-blejska-kremsnita-lsm62

[6] https://odprtakuhinja.delo.si/recepti/klasike-z-ursko-fartelj-

kremsnite-s-pravo-kremo-ki-migajo-v-bokih

[7] https://oblizniprste.si/slastne-sladice/kremsnite-kremne-rezine/

[8] https://jernejkitchen.com/sl/recepti/kremsnite

[9] https://www.quora.com/How-many-grams-of-CO2-do-you-

exhale-per-day

Radon v stavbah in kakovost bivanja



Radon in buildings and quality of life



Matjaž Valenčič

Zveza ekoloških gibanj Slovenije-ZEG

Krško, Slovenija

matjaz.valencic@gmail.com



Povzetek construction... Although all ministries are active in



V stavbah, ki niso ustrezno zgrajene ali sanirane, se the field of radon prevention in buildings, they are

nabira radon, zdravju nevaren žlahtni plin brez barve, not interdepartmental coordinated. It happens too

vonja in okusa. often that citizens receive conflicting information,

Področje radona v stavbah pokrivajo ministrstva za which is why it even happens that radon appears in

zdravje, za energijo, za prostor, za graditev … living spaces after energy renovation of a building.

Čeprav so vsa ministrstva aktivna na področju In Slovenia, radon has been measured in public

preprečevanja radona v stavbah, niso medresorsko buildings since 1990, and in residential buildings

usklajena. Prepogosto se zgodi, da občani prejmejo since 2018. Unfortunately, residents are

nasprotujoče si informacije, zato se celo dogaja, da insufficiently informed about the danger of radon, so

se po energetski sanaciji stavb pojavi radon v proactive communication is needed to raise

bivalnih prostorih. awareness among the population.

V Sloveniji merijo radon v javnih stavbah od leta It is necessary to inform the population and the

1990, v stanovanjskih stavbah pa od 2018. Žal so profession about the dangers of radon in buildings

stanovalci premalo informirani o nevarnosti radona, and propose the implementation of measures to

zato je potrebna proaktivna komunikacija za prevent the ingress of radon in buildings.

ozaveščanje prebivalstva.



Prebivalstvo in stroko je treba seznaniti o nevarnostih radon in buildings, energy rehabilitation, anti-radon Optional: Keywords

radona v stavbah in predlagati izvedbo ukrepov, ki rehabilitation, radon knowledge

preprečujejo vdor radona v stavbah.





Ključne besede

radon v stavbah, energetska sanacija, protiradonska sanacija,

radonvednost



Abstract



In buildings that are not properly constructed or

rehabilitated, radon, a colorless, odorless, and

tasteless noble gas that is hazardous to health,

accumulates.

The field of radon in buildings is covered by the

ministries responsible for health, energy, space,

Slika 1 Razpadna veriga urana (vir1)

Permission to make digital or hard copies of part or all of this work for personal or

classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full 1 Radon citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s).



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Radon je žlahten plin, brez barve, vonja in okusa.

© 2024 Copyright held by the owner/author(s).



DOI https://doi.org/10.70314/is.2024.env.11

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Matjaž Valenčič



Beseda »žlahten« naj ne zavede. Čeprav ta beseda radon in njegove razpadne produkte uvršča med

pomeni nekaj imenitnega, sodeč po slovarju: žlahtna snovi, ki so rakotvorne za človeka (skupina 1).

kapljica, žlahtni kamen, žlahtna plesen, žlahtna



gospoda … je radon zdravju nevaren plin. Radon in njegovi razpadni produkti povzročajo

pljučnega raka. Svetovna zdravstvena organizacija



(WHO) je ugotovila, da se ob povišani koncentraciji

radona za 100 Bq/m3 tveganje za pljučnega raka

poveča za 8 %. Tveganje za bolezen in smrt zaradi

pljučnega raka pri povišanih koncentracijah radona

je večje pri kadilcih kot nekadilcih. Ocenjujejo, da je

vsak deseti rak na pljučih posledica radona oziroma

Slika 2 Beseda "žlahten" ima raznolik pomen. njegovih razpadnih produktov. (7)



V kemiji pomeni beseda »žlahten« nekaj drugega. 2 Vir in vstop radona v stavbo ter dejavniki, ki

Žlahtni plini, včasih imenovani inertni plini, imajo vplivajo na raven radona v stavbi

svojevrstno mesto v periodnem sistemu. Vsi žlahtni



plini, tudi radon (Rn), so znani po tem, da nimajo Največji izvor radona v bivalnem okolju predstavlja

barve, vonja in okusa. So nevnetljivi, v standardnih zemljina pod stavbo, manj pa gradbeni material, voda

pogojih izjemno stabilni. in plin za ogrevanje. Radon je v zaprtih prostorih, v

katere prodira skozi slabo izolirana tla in vodovodna

Radon je kemijski element z vrstnim številom 86 in ter druga omrežja v stavbi, ki so v stiku z zemljino.

atomskim masnim številom 222, pri čemer je treba Vzrok povišanih koncentracij radona so kamninska

omeniti, da obstaja mnogo različnih izotopov radona. podlaga, letni čas, nadstropje v stavbi in tesnjenje

Najbolj dolgoživ izotop radona je 222Rn z razpolovno oken. Največji vpliv na koncentracijo radona v stavbi

dobo 3,82 dni. Ostali izotopi imajo razpolovno dobo ima kamninska podlaga.

nekaj minut ali manj; zaradi kratke razpolovne dobe



in ker večinoma ti izotopi ne razpadajo z α razpadom,

ostalih izotopov ne obravnavamo v kontekstu

nevarnosti za zdravje.

Radon je element VIII. valentne skupine. Spada med

žlahtne pline in tvori enoatomne molekule. Zaznamo

ga predvsem zaradi radioaktivnega razpada, z

merilnimi napravami.



Radon sta odkrila Ernest Rutherford in Friderick

Soddy leta 1900. Je najtežji element skupine žlahtnih

plinov, je radioaktiven in nastaja pri radioaktivnem

razpadu radija. Slika 3: Radonski zemljevid Slovenije (8)

Še zanimivost. Plin radon nastane iz trdne snovi

(radija) v zemljini, kot plin vstopa v stavbe in tam Na podlagi do zdaj izvedenih meritev koncentracij

razpade v trdne snovi, ki so radioaktivne. radona v Sloveniji in podatkov o geološki sestavi tal

so določena območja, kjer je več radona (na



Za zdravje so nevarni radonovi kratkoživi razpadni zemljevidu obarvano rdeče).

produkti. To so kovine, ki v zraku tvorijo Dodatne meritve se izvajajo tudi na območjih občin,

radioaktivne aerosole, nanodelce. Med dihanjem se kjer je radon pričakovan (na zemljevidu obarvano

radioaktivni aerosoli nalagajo na stenah dihalnih rumeno).

poti. Od tam sevajo in povzročajo poškodbe tkiva.

Mednarodna agencija za raziskovanje raka (IARC) Vendar se radon lahko pojavi tudi v stavbah na

zelenem področju. V nekaterih ljubljanskih vrtcih in



Radon v stavbah in kakovost bivanja Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



šolah so izmerili povišane koncentracije radona in jih Pred izvedbo vseh gradbenih ukrepov, ki bi lahko

sanirali (9), pa so te stavbe na zelenem področju. povečali koncentracijo radona v stavbi, svetujemo

radonske meritve in izbor tistih ukrepov, ki znižajo

Za preprečevanje pojava radona v stavbah so prisotnost radona. Zlasti skrb vzbujajoči so

preventivni protiradonski ukrepi potrebni tudi na nestrokovni ukrepi ob energetskih sanacijah, tesnitvi

zelenem radonskem področju. stavbe, menjavi oken, vgradnji prezračevalnih

naprav z izmenjujočim delovanjem, izolaciji fasade,

izdelavi prizidkov, tlakovanju dvorišč … Ti ukrepi

A: obstoječe stavbe: Ukrepi za znižanje radona v praviloma povečajo prisotnost radona v prostorih in

obstoječih stavbah so, iz tega vidika, škodljivi.



V vsaki stavbi predlagamo zaščitno detekcijo Izvedba protiradonske sanacije obstoječih stavb je

koncentracije radona. določena v Pravilniku (4).



koncentracija radona v zraku bivalnega okolja pod izpostavljenosti radonu iz uredbe, ki ureja 3 Trenutno veljavna zakonodaja določa, da je dopustna  Kot gradbeni poseg za zmanjšanje

300 Bq/m . Če je mejna vrednost prekoračena, je nacionalni radonski program, se šteje

treba začeti sanacijo radona v stavbi. Raziskave vgradnja aktivnega sistema za prezračevanje

kažejo, da 1/3 slovenskih domov presega zgoraj zemljine pod talno konstrukcijo.

navedeno mejno vrednost.  Ne glede na prejšnji odstavek je dopustna

izvedba pasivnega sistema, če se z njim

Zakonsko določena mejna vrednost koncentracije lahko zagotovi referenčna raven.



radona v zraku bivalnega okolja pomeni, da če  Kot dodaten ukrep k aktivnemu sistemu 3 meritve pokažejo vrednosti pod 300 Bq/m , prezračevanja zemljine se lahko izvede tudi zakonodaja ne predpisuje nujnih ukrepov. Vendar pa tesnjenje oboda talne konstrukcije in je pomembno poudariti: prebojev skozi talno ploščo ter tesnjenje

tlorisne površine z radonsko zaporno folijo.

 300 Bq/m3 ni varnostna meja, temveč meja,

pri kateri se priporoča začetek sanacije. Že

pri nižjih koncentracijah radona obstaja B: nove stavbe: Ukrepi za preprečevanje radona

povečano tveganje za nastanek raka na pri projektiranju in izvedbi novogradenj

pljučih.

 Svetovna zdravstvena organizacija (WHO) Zakonodaja določa, da morajo biti stavbe

priporoča, da se prizadevanja za zmanjšanje projektirane, grajene in vzdrževane tako, da je

koncentracije radona začnejo že pri koncentracija radona v delovnih in bivalnih prostorih

vrednostih pod 100 Bq/m3. To je še posebej čim nižja in da ni presežena referenčna raven.

pomembno v prostorih, kjer se zadržujejo Območje Republike Slovenije je razdeljeno na dve

otroci in občutljive skupine prebivalstva. območji, in sicer na radonsko območje in zeleno

območje.

Izbira ukrepov naj temelji na velikosti preseženih



vrednosti radona in njeni oceni vpliva na zdravje  Na radonskem področju: Projektiranje in

stanovalcev. Najučinkovitejši ukrep je kakovosten gradnja novih stavb na radonskem območju

gradbeni poseg na talni plošči, saj z njim omejimo vir se zagotovi s prezračevanjem zemljine pod

povišanih koncentracij radona v stavbi. talno konstrukcijo stavbe ali s popolnim

Prezračevanje prostorov in zatesnitev vidnih razpok tesnjenjem med zemljino in notranjim

ter špranj v tleh in stenah, ki mejijo na zemljino, ter zrakom z radonsko zaporno folijo. Kot

instalacijskih vodov sta primerna pri majhnih prezračevanje zemljine se štejejo tudi

preseganjih dovoljenih vrednosti, sicer sta le začasna arhitekturne rešitve, ki zagotavljajo zadostno

ukrepa. prezračevanje prostorov, ki so v stiku z

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Matjaž Valenčič



zemljino tako, da ni presežena referenčna zemljino. Neprimerno prezračevanje celo poveča

raven. Če pasivno prezračevanje zemljine ne koncentracijo radona v stavbah!

zadošča, se vgradi aktivno prezračevanje.

 Na zelenem področju: Projektiranje in V stavbah je večji del leta topleje kot zunaj, zato so

gradnja novih stavb na zelenem območju je tlačne razmere v stavbah odvisne tudi od mesta

določeno enako kot na radonskem področju. oziroma višine netesnosti stavbe. Stare stavbe so bile

praviloma netesne povsod, zlasti pri vseh okenskih

odprtinah, zato je bil zračni tlak v celotni stavbi

3 Prezračevanje prostorov, popularen, a dvomljiv izenačen s tlakom zunanjega zraka oziroma tlaka v

ukrep zemljini, možnost vdiranja radona v stavbe je bila

majhna.

Ukrepi za znižanje koncentracije radona v stavbah so

zelo različni (6). Ob pripravi projekta sanacije stavbe



je treba upoštevati značilnosti objekta, temeljenje in



podkletenost zgradbe, propustnost tal pod osnovno

ploščo, in poiskati mesta, kjer radon vstopa v

zgradbo. Vstopna mesta določimo s posebnimi

instrumenti, ki merijo trenutne koncentracije radona.

Čeprav velja splošno prepričanje, da koncentracijo

radona v prostorih znižamo s prezračevanjem

prostorov, je ta ukrep dvomljiv. Vendar pa tudi Slika 5. Netesnost zgoraj,

Slika 4 Netesnost spodaj,

ureditev prezračevalnega sistema prostorov le malo stavba je v podtlaku, radon

stavba je v nadtlaku, možnost



zniža vdora radona je majhna. intenzivno vstopa.

koncentracije (običajno 20-30 %).

Prezračevalni sistemi, ki v zgradbah ustvarjajo

V kurilni sezoni je sodobna, tesna stavba takrat, ko je

dodaten podtlak, lahko privedejo tudi do nasprotnega

netesna na zgornjem delu ovoja, v podtlaku in

učinka in se koncentracija radona v prostorih zviša.

vzgonsko črpa radon iz zemljine pod stavbo (slika 5).

Povečano prezračevanje prostorov ali vzdrževanje

Izven kurilne sezone, ko je praviloma odprtih več

stalnega nadtlaka v stavbi proti zemljini sta ukrepa,

oken, je vzgonski učinek črpanja radona manjši.

ki sta lahko primerna, če je prezračevanje intenzivno,

(slika 4)

z nadtlakom. Vendar sta energetsko potratna in ju

stroka ne priporoča pri splošnem reševanju težave

Še intenzivnejši je vdor radona v stavbo, v kateri je

povišanega radona,

izmenjujoči zračni tlak zaradi dvosmerne

prezračevalne naprave (sliki 6 in 7). Kljub dobrim



4 Napačno protiradonsko prezračevanje namenom se stavba z izmenjujočim delovanjem

prezračevalnih naprav spremeni v radonsko črpalko,



Prezračevanje pomeni izmenjavo zraka. Čist zunanji ki ob podtlaku črpa radon v stavbo, ob nadtlaku pa ga

zrak vstopa v prostor, notranji pa izstopa in odnese ne zmore izriniti iz stavbe.

vsa onesnažila. Izmenjava je lahko samodejna, zaradi

fizikalnih lastnosti zraka ali mehanska, s

prezračevalno napravo. Ne glede na način izmenjave

lahko nastane v prostoru nadtlak ali podtlak. Iz vidika

izmenjave zraka je vseeno, če je prezračevanje

nadtlačno ali podtlačno, glede ogroženosti z radonom

pač ni.



Premikanje radona skozi zemljino v stavbo je

predvsem konvektivno, posledica tlačnih razlik med

plinom v zemljini in tlakom v prostorih, ki ležijo nad

Radon v stavbah in kakovost bivanja Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



Zdaj, ko poznamo nevarnosti radona in se znamo



zaščititi pred radonom, bi bilo neprimerno, da bi to



nevarnost prezrli.



6 Radonvednost



Radonvednost je akcija ozaveščanja prebivalstva in

strokovnjakov o radonu v zaprtih prostorih.

Slika 7. Dovod zraka Vključuje vse udeležence, ki se ukvarjajo z radonom



Slika 6. Ob izmeničnem v stavbah, lahko vplivajo na pojav radona v stavbah izmenične prezračevalne

delovanju prezračevalne

naprave odvod zraka ustvarja naprave ne zmanjšuje ali živijo z radonom, od načrtovalcev, projektantov,

podtlak in pospešuje vdor deleža radona v stavbi. izvajalcev do stanovalcev.

radona iz zemljine, odvede pa Izmenični dovod in odvod



se le del radona. zraka iz stavbe ustvarja Radonvednost želi podati celovite informacije o učinek radonske črpalke v

stavbo. pojavu radona v stavbah in o njegovem

preprečevanju, ob rabi stavb ter sočasni celoviti

Prezračevanje ni magična rešitev za zmanjšanje oziroma delni prenovi stavb in ob novogradnjah.

koncentracije radona v stavbi, zlasti ne, če je napačno, Vsi, ki vas zanima radon v bivalnih prostorih, ste

s podtlakom ali z izmenjujočimi tlačnimi razmerami. povabljeni, da se pridružite akciji radonvednost.

Izmenjujoče prezračevanje lahko celo ustvari učinek

radonske črpalke in pospeši uhajanje radona v stavbo.



5 Vpliv Jomkipurske vojne na radon v stavbah Literatura/References::

1. Navodila v primeru zaznanih povečanih koncentracij



Med Jomkipursko vojno in povečanjem radona v radona v stavbah javnih vzgojno izobraževalnih zavodov (viz) stavbah ni nobene neposredne povezave. Odgovor na 2. https://nijz.si/moje-okolje/uv-sevanje-in- prvo svetovno naftno krizo je bilo povečanje radon/problematika-povisanih-koncentracij-radona-v-

energijske učinkovitosti stavb, tudi z zatesnitvijo vrtcih-in-solah/



stavb. Kmalu po tem se je pokazalo, da tesne stavbe 3. https://www.gov.si/teme/zmanjsevanje-izpostavljenosti-radonu/ niso zdrave (SBS), predpisano je bilo kontrolirano 4. Pravilnik o zahtevah za novogradnje, posege v prezračevanje z vračanjem toplote za zadovoljevanje obstoječe stavbe in sanacijo obstoječih stavb zaradi

sanitarnih potreb in energijsko učinkovitost. Radon, varovanja zdravja ljudi pred škodljivimi učinki

zdravju škodljiv plin brez barve, vonja in okusa, je radona

bil prezrt. Morda so pozabili nanj zato, ker je že od 5. Dobri nameni in radonske kletke; Valenčič, M.; 11.

04. 2024; strokovno posvetovanje »S komunikacijo v

nekdaj vstopal v netesne stavbe in iz njih neovirano slovensko okoljsko modernizacijo«

izstopal. V tesnih stavbah pa je ostal ujet. Tako so 6. https://www.radon-resitve.si/radon/resitve/



sodobne, tesne stavbe postale radonske kletke. 7. Problematika povišanih koncentracij radona v vrtcih in šolah

8. Zmanjševanje izpostavljenosti radonu

Posredno, zaradi neprimerne komunikacije o 9. Proračun MOL 2021

stanovanjski kulturi, je Jomkipurska vojna vplivala 10. radonvednost

na povečanje koncentracije radona v stavbah. 11. Tehnična smernica 1-007:2023; Zaščita pred radonom

v stavbah



Postavljanje sončnih elektrarn na kmetijskih površinah- okoljski



kriminal ali trajnostna raba prostora



Setting up solar power plants on agricultural land - environmental



crime or sustainable use of space



Matjaž Valenčič

Zveza ekoloških gibanj Slovenije-ZEG

Krško, Slovenija

matjaz.valencic@gmail.com



Povzetek prste ene roke, da smo menda edina evropska



Napačne besede, navržene v eter, so nevarne, država brez plavajočih sončnih elektrarn, da je

tudi če so povedane z nasmehom. Smo v času geotermalna energija, ki ima kapaciteto

pred jedrskim referendumom, ko javni in tihi slovenske polovice jedrske elektrarne,

jedrski podporniki intenzivno oblikujejo javno spregledana, da je gradnja hidroelektrarn, ki jim

mnenje v korist novi jedrski naložbi, brez pravijo mesoreznice, ustavljena in da je

ustreznih argumentov. energijska raba organskih ostankov blokirana?

Tiho, prikrito jedrsko lobiranje se začelo že pred Ob neprimernih besedah ne smemo ostati tiho.

več kot tremi desetletji, ko je civilna družba Pogrešamo vključujoče in transparente javne

zahtevala referendum o zaprtju NEK. Zato razprave o energetski prihodnosti Slovenije.

lahko neutemeljeno in neprimerno kritiko Spodbujajmo odprt pogovor o trajnostnem

obnovljivih virov energije razumemo dvolično. energetskem prehodu.

Ko napačne besede padejo na plodna tla, se

Keywords / Ključne besede

razplamti požar negativnih čustev. Pojavijo se

Okoljski kriminal, okoljsko zavajanje, tiho

civilne iniciative, ki jim praviloma sledijo ovire

spodbujanje jedrske energije

ali prepovedi postavitev naprav za rabo

obnovljivih virov energije. To spremljamo pri

vseh možnih obnovljivih virih: vetrnih, vodnih, Abstract



geotermalnih, sončnih, izrabi organskih Wrong words projected into the ether are

ostankov … dangerous, even when said with a smile. We are

Zaradi lahkotnih izjav o obnovljivih virih, ki so in the time before the nuclear referendum, when

lahko napačno razumljene, država nazaduje na public and silent nuclear supporters are

področju rabe obnovljivih virov. Taktika intensively shaping public opinion in favour of

jedrskih zagovornikov je videti preprosta: a new nuclear investment, without adequate

blokiranje obnovljivih virov toliko časa, da bo arguments.

jedrska energija spoznana za edino možnost. Je Silent, covert nuclear lobbying began more than

to razlog, da vetrne elektrarne preštejemo na three decades ago, when civil society demanded

a referendum on the closure of the NPP NEK.

Permission to make digital or hard copies of part or all of this work for personal or Therefore, the unfounded and inappropriate

classroom use is granted without fee provided that copies are not made or distributed criticism of renewable energy sources can be

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must understood in two ways. be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia When the wrong words fall on fertile ground, a

© 2024 Copyright held by the owner/author(s). fire of negative emotions is ignited. Civil

DOI https://doi.org/10.70314/is.2024.env.12 initiatives appear, which are generally followed

Information Society 2024, 7–11 October 2024, Ljubljana,

Matjaž Valenčič

Slovenia



by obstacles or bans on the installation of Poved z nasmehom

devices for the use of renewable energy sources.



We monitor this for all possible renewable Zanimiva je trditev, ki se ponavlja v medijih in

sources: wind, water, geothermal, solar, use of v družbenih omrežjih, da je postavljanje

organic residues... sončnih elektrarn na kmetijskih površinah,

Due to light statements about renewable namesto na strehah stavb, okoljski kriminal (1).

resources, which can be misunderstood, the



country is regressing in the field of renewable Podobne so tudi trditve, da so vetrne elektrarne

resources. The tactic of nuclear advocates okolju škodljive, ker se dotrajanih kril ne da

seems simple: block renewables for so long that reciklirati, da ne govorimo o neslišnem hrupu,

nuclear power is recognized as the only option. ki ga nekateri okoljevarstveniki slišijo tudi

Is this the reason that wind power plants can be mnogo kilometrov daleč ali o ptičih, ki se

counted on the fingers of one hand, that we are zaletavajo v krila vetrnih turbin. Čeprav sodijo



floating solar power plants, that geothermal širijo med prebivalstvom in po družbenih omrežjih. energy, which has the capacity of half of supposedly the only European country without te trditve v psevdoznanost (15), se neovirano

Slovenia's nuclear power plant, is overlooked,

that the construction of hydroelectric power Nasprotniki hidroelektrarn opozarjajo na



stopped and that the energy use of organic raznovrstnosti, na neprimerne prehode za vodne organizme … Zato predlagajo rušenje vseh residues is blocked? plants, which are called butchers, has been kopičenje mulja (8), na izgubo biotske



We must not remain silent in the face of naj bi ostala reka brez jezov (9). Ob tem, da je jezov hidroelektrarn! Še več, slovenska Mura

inappropriate words. We miss inclusive and na avstrijskem delu Mure vsaj 22 večjih HE

transparent public debates on Slovenia's energy skupne moči 350 MW, na slovenski le ena mala,

future. moči 0,5 MW (11). Avstrijska pretočna

Let's encourage an open conversation about a elektrarna Kalsdorf (12), ki smo si jo ogledali

sustainable energy transition. med gradnjo (10), je vzor sonaravne gradnje.

Takrat so investitorji predstavili načrtovane

Optional: Keywords ukrepe za varovanje okolja (ohranjanje



Environmental crime, environmental deception, zamočvirjenih področij, slepih rokavov, otokov



tacit promotion of nuclear energy …), ki kljub elektrarni omogočajo biotsko raznovrstnost in pestrost. Prepričani smo, da so

investirali v varovanje okolja zaradi splošne

okoljske zavesti, ki je posledica dolgoletnih



usklajenih aktivnosti avstrijskih

okoljevarstvenih organizacij. Ne, tega slovenski

varuhi voda ne morejo videti, menda imajo

plašnice na očeh.





Druga plat medalje



Jedrski zagovorniki nekritično zagovarjajo

jedrsko energijo kot edino možnost za trajnostni

energetski prehod in za konkurenčno

gospodarstvo. Pozivi k dialogu o energetskem

prehodu so cenzurirani (14), članki (13), ki

Slika 1: Fragments of wind turbine blades await burial at

the Casper Regional Landfill in Wyoming. Photographer: apelirajo na pogovor o jedrski energiji,

Benjamin Rasmussen for Bloomberg Green; Zavajajoča spregledani.

fotografija, ki jo širijo nasprotniki vetrnih elektrarn v



Sloveniji. Tega v Sloveniji ni. Smo v času referenduma o JEK2. Vprašanje je,

ali je javnost obveščena v skladu z določili

Information Society 2024, 7–11 October 2024, Ljubljana,

Postavljanje sončnih elektrarn na kmetijskih površinah

Slovenia





Aarhuške konvencije, saj še niso objavljeni

podatki, ki jih izrecno zahteva Resolucija o

dolgoročni miroljubni rabi jedrske energije v

Sloveniji »jedrska energija za prihodnost

Slovenije« (REDMRJE) (16) za izvedbo

projekta JEK2. Hkrati pa vemo, da je ta

referendum popolnoma nepotreben, saj GEN že

18 let neovirano pripravlja naložbo novo



jedrsko elektrarno. Čemu naj javnost zdaj Slika 2: Besedna zveza »pecunia non olet« se še danes

podpre ta referendum, brez potrebnih podatkov uporablja za trditev, da vrednost denarja ni omadeževana

(18), ob pomanjkanju vključujoče in z njegovim izvorom. Danes bi bilo morda primernejše

transparente javne razprave o energetski zapisati, da denar ne seva. Vir: wikipedia, Pripis: Classical

prihodnosti Slovenije in ob popolni odsotnosti Numismatic Group, Inc. http://www.cngcoins.com

ključnih informacij o projektu JEK2?



Sončne elektrarne na kmetijskih površinah?

Dejanski okoljski kriminal



Kaj hudega so naredile sončne elektrarne na

Nasprotovanje dobremu je slovenska kmetijskih površinah, da jih imenujemo

energetska posebnost. Skupno vsem »okoljski kriminal«? Res je, da sončna

popularnim okoljskim trditvam je slepa vera v elektrarna ne sme konkurirati kmetijskim

urbane legende. Res se je dogajal okoljski proizvodom. To zgodbo, hrana ali energija, smo

kriminal, ampak, namesto bi ga družba že prerasli. Proizvodnja energije ne sme

preprečevala, ga dopušča, hkrati pa ovira nadomestiti proizvodnje hrane.

marsikaj, kar je dobro. Bolje bi bilo, če bi se



družba usmerila v preprečevanje slabega. Polemiziramo s trditvijo, da na njive ni

Vpišimo ekocid v ustavo in prenesimo direktivo primerno postavljati sončnih elektrarn, ki bi

o okoljski kriminaliteti (6) v pravni red; ovirale kmetovanje, saj Imamo dovolj drugih

zahtevajmo, da se to izvaja. površin, kjer so sončne elektrarne dobrodošle:

strehe večjih stavb, parkirišča, vodne površine,

Ne, ribarjenje v kalnem je donosna dejavnost za ograje ob avtocestah … Kaj pa, če bi sončne

marsikatero nevladno organizacijo, ki ima tiho, elektrarne sobivale s kmetijskimi površinami in

a močno podporo zagovornikov jedrske jih ne bi prekrivale?

energije.



O jedrski energiji slišimo same sladke besede. Agrofotovoltaika (agrovoltaika) in

Kritika jedrske skupnosti je utišana, prostostoječa sončna elektrarna

cenzurirana, preganjana. Da slovenska jedrska

skupnost ne spoštuje zavez, da dokazano Kmetijsko zemljišče, na katerem stoji sončna

nezakonito troši denar (17), da kupuje javno elektrarna, lahko po postavitvi elektrarne v

mnenje, da v svoje načrte vključuje celotno polni meri ohrani svoji primarno (kmetijsko)

državno politiko, o tem se ne govori. Jedrski funkcionalnost in jo celo poveča, lahko pa jo

kriminal je doma nekje drugje, na Japonskem, ohrani samo v omejenem obsegu. Če postavimo

na Češkem (4) ali v Franciji. sončno elektrarno nad sadovnjak ali nasad

malin, govorimo o agrovoltaiki. Če pa

postavimo elektrarno na zemljišče, ki je

deklarirano kot kmetijsko, na njem pa raste

goščavje in je dotlej gospodarsko

neizkoriščeno, potem gre za klasično

prostostoječo sončno elektrarno, zemljišče pa

jo lahko lastnik istočasno izkorišča tudi za

kmetijske namene, npr. pašo drobnice.

Information Society 2024, 7–11 October 2024, Ljubljana,

Matjaž Valenčič

Slovenia





Agrovoltaika ne pomeni zgolj proizvodnje

električne energije, ampak upošteva celotno

korist, ki jo ima kmet od boljše, bolj kakovostne

kmetijske pridelave (zaščita rastlin pred

ožigom, zaščita pred točo) in hkrati dodatno

proizvodnjo energije.



Kot piše MOPE(2), je agrovoltaika tehnologija,

ki združuje kmetijstvo in fotovoltaične sisteme Slika 3 Sicer si ne predstavljamo vinogradov na Krasu v

za proizvodnjo električne energije. Gre za sožitju s sončnimi elektrarnami, v manj vetrovnih krajih pa

integracijo fotonapetostnih panelov v kmetijsko bi bilo zanimivo. Foto: Naš stik

infrastrukturo, na primer na poljih, vinogradih



ali drugih kmetijskih površinah. Namen Koruptivna tveganja?

agrovoltaike je izkoristiti kmetijsko zemljišče



za dva namena: za pridelavo hrane ali rastlin ter Predlagamo premišljeno načrtovanje postavitve

hkrati za proizvodnjo obnovljive električne agrovoltaike, da ne bo doživela enake usode kot

energije. Prinaša številne prednosti: zmanjšanje bioplinarne. Če se prav spomnimo, afera

izhlapevanja vode zaradi senčenja, zaščito bioplinarn ni dobila epiloga, očitno je bilo tako

rastlin pred vremenskimi vplivi, izboljšanje veliko denarja odtujenega in razdeljenega, da

učinkovitosti rastlin zaradi regulacije svetlobe, organi pregona niso bili motivirani. Podatki o

zmanjšanje emisij toplogrednih plinov in gradnji dveh bioplinarn so šokantni, policija je

diverzifikacijo dohodka. Hkrati pa je treba tudi ugotovila veliko oškodovanje državnega

vedeti, da niso primerna kombinacija z vsemi proračuna in bank, saj najetih kreditov vlagatelji

kmetijskimi proizvodi. Niso univerzalna niso vrnili, pred stečajem pa so vrednejše dele

rešitev, temveč jih je treba premišljeno bioplinarn odpeljali neznano kam (4). Podjetja,

postavljati. ki so gradila bioplinarne, so poslovala z izgubo,

zemljišča pa so zavarovali z zemljiškim

Naš stik (3) navaja, da lahko namestitev sončnih dolgom. Proizvodnja električne energije ni

panelov pri gojenju hrane izboljša zadrževanja nikoli zaživela, dolgovi so ostali upnikom,

vode v sušnih letih in zaščiti pridelke pred kmetom in državi. Ker so v zgodbo vpletena

ekstremnimi vremenskimi razmerami, v zasebna podjetja in politika, obstaja tveganje za

nekaterih primerih pa lahko poveča kmetijski sistemsko korupcijo, saj so verjetno nekatere

donos. Za jagodičevje ali sadje lahko odločitve politike pripomogle k pridobitvi

agrovoltaika omogoči povečanja pridelka do 16 premoženjske koristi posameznikov, ki so

odstotkov, hkrati pa proizvede 63 odstotkov največjo škodo povzročili bankam v državni

električne energije tradicionalne sončne lasti.

elektrarne. Pri posevkih, ki potrebujejo več

svetlobe, kot je pšenica, znaša izgub pridelka

pod 20 odstotkov, dodani prihodki od prodaje Energija ali hrana? Oboje!

električne energije pa močno odtehtajo manjši

prihodek od pridelanega žita. Nekako smo obremenjeni od okoljske

nevarnosti fosilnih energentov. Če se ne motim,

so nemški proizvajalci avtomobilov dobili

dovoljenje proizvodnje ICE avtomobilov tudi

po letu 2035, če bodo uporabljali biogorivo (7).

Je to za okolje dobro ali slabo?

Information Society 2024, 7–11 October 2024, Ljubljana,

Postavljanje sončnih elektrarn na kmetijskih površinah

Slovenia





Izračun 1:200



Malo se poigrajmo z oceno donosa. Ocena, tudi

če ima majhno napako, je bolj točna kot

natančen izračun z napačnimi vhodnimi

podatki.

Slika 4: primerjava energetske rabe zemljišč za mobilnost



Povprečen donos koruze je 8,5 t/ha. Iz tega

Indijske izkušnje kažejo, da je iz kmetijskih

lahko pridobimo približno 1.000 l biogoriva,

površin možno dobiti pogonsko gorivo za avto

kar je dovolj za vožnjo povprečnega avtomobila

z notranjim zgorevanjem. To je dobro. En

z notranjim zgorevanjem za eno leto. En avto na

hektar kmetijskih površin zadošča za eno leto

hektar.

vožnje povprečnega avtomobila.



Kaj pa, če bi na tem hektarju imeli sončno

Kaj pa, če bi na tem zemljišču postavili sončno

elektrarno? Ob predpostavki, da je fotovoltaika

elektrarno in s pridelano elektriko poganjali

20-krat bolj učinkovita od fotosinteze, torej

električne avtomobile? Tudi to bi šlo, kar je

dobimo 20-krat več energije.

dobro. Pravzaprav, odlično. En hektar

kmetijskih površin zadošča za eno leto vožnje

Ob predelavi koruze v biogorivo je precej izgub,

dvesto povprečnih avtomobilov (5.) Razmerje

upoštevamo 50 %, elektrika pa poganja e-avto z

1:200 pa je veličastno. Je torej, za razogljičenje

zanemarljivimi izgubami. Ne nazadnje, avto na

mobilnosti, bolje primešavati fosilnim

elektriko potrebuje pet krat manj energije kot z

energentom majhen delež biogoriv ali

motorjem z notranjim zgorevanjem. Ko

proizvajati elektriko za elektromobilnost?

zmnožimo 20 in 5 ter delimo z 0,5, dobimo

dvesto. Dvesto avtov na hektar.

Vprašanje je, ali je v te luči sončna elektrarna na

kmetijskih površinah še vedno okoljski



kriminal. Dvestokrat večji učinek? Kmet ustvari Vztrajati v napačnih odločitvah ali odpraviti

dvestokrat več, brez škode okolju, in zato napake?

postane okoljski kriminalec?



Motiti se je človeško, v zmoti vztrajati pa

Kaj pa, če bi malo popustili in vgradili sončne hudičevo, je ugotovil Seneka, najprej vzgojitelj,

panele z manjšimi razmiki med njimi, da bi pozneje svetovalec rimskega cesarja Nerona.

prepuščali 20 % svetlobe in omogočili



padavinam dostop do zemlje? Na hektarju Bodo nezreli, neodgovorni kritiki sončnih

kmetijskih površin bi dobili približno 80 % elektrarn priznali, da so se motili in bodo

poljščin in elektrike za 160 avtomobilov, kar dopustili sončne elektrarne na kmetijskih in

tudi ni mačji kašelj. vodnih površinah? In drugo vprašanje, ali bodo

okoljevarstveniki nasprotovali okoljskemu



kriminalu, ki ga sedaj tiho podpirajo?



Predlagamo okoljsko osveščanje namesto

novega ludističnega gibanja, oviranja

obnovljivih virov brez razumnih razlogov.



Slika 5: Agrovoltaika omogoča dvojno rabo zemljišča ob

le malo manjšem kmetijskem donosu, pri nekaterih

poljščinah pa dodatno ščiti rastline pred premočnim

soncem, pozebo ali sušo.

Information Society 2024, 7–11 October 2024, Ljubljana,

Matjaž Valenčič

Slovenia



Literatura/References: del-globalne-kampanje-WWF-za-prosto-tekoe-

reke

1. Dr. Matjaž Gams: »Woke« kultura je zrcalna 10. TEŠ6 ali obnovljivo; www.zaensvet.si;

slika fašizma; www.delo.si; https://zaensvet.si/wp-

https://www.delo.si/novice/slovenija/dr-majaz- content/uploads/2020/01/TE%C5%A06-ali-

gams-woke-kultura-je-zrcalna-slika-fasizma obnovljivo.pdf

2. Tehnična določitev postavitve agrovoltaike na 11. Nova jedrska ekonomika; www.zaensvet.si;

kmetijskih zemljiščih; www.energetika- https://zaensvet.si/nova-jedrska-ekonomika/

portal.si; https://www.energetika- 12. Pretočna elektrarna Kalsdorf,

portal.si/nc/novica/n/tehnicna-dolocitev- www.verbond.com;

postavitve-agrofotovoltaike-na-kmetijskih- https://www.verbund.com/en-at/about-

zemljiscih/ verbund/power-plants/our-power-

3. V agrovoltaiki velik potencial za energetsko in plants/kalsdorf

prehransko varnost; www.nas-stik.si/novice; 13. Zakaj (ne) govorimo o novi jedrski elektrarni;

https://www.nas-stik.si/novice/podrobnosti- Valenčič, M.; EGES 4/2014

novice/v-agrofotovoltaiki-velik-potencial-za- 14. NEK bi preživela tudi potres osme stopnje:

energetsko-in-prehransko-varnost www.zaensvet.si; https://zaensvet.si/nek-bi-

4. Primer sistemske korupcije: gradnja bioplinskih prezivela-tudi-potres-osme-stopnje/

elektrarn v Sloveniji; www.policija.si; 15. Vetrne elektrarne; www.sl.wikipedia.org;

https://www.policija.si/images/stories/Publikaci https://sl.wikipedia.org/wiki/Vetrna_elektrarna

je/PDF/Korupcija.pdf 16. Resolucija o dolgoročni miroljubni rabi jedrske

5. Obnovljivi viri energije v prometu; energije v Sloveniji »jedrska energija za

www.zaensvet.si; https://zaensvet.si/obnovljivi- prihodnost Slovenije« (REDMRJE);

viri-energije-v-prometu/ www.pisrs.si ;

6. Direktiva o okoljski kriminaliteti; https://pisrs.si/pregledPredpisa?id=RESO155

www.zaensvet.si; https://zaensvet.si/direktiva- 17. Učinkovitost strateškega načrtovanja

o-okoljski-kriminaliteti/ dolgoročnega izkoriščanja jedrske energije za

7. Po letu 2035 bodo lahko na prodaj avtomobili na proizvodnjo električne energije in načrtovanja

e-goriva; www.avto-fokus.si; proizvodnje električne energije v jedrskih

https://www.avto- elektrarnah; Računsko sodišče; www.rs-rs.si;

fokus.si/Novice/prepoved_po_letu_2035_po_le https://www.rs-

_letu_2035_bodo_lahko_na_prodaj_avtomobili rs.si/fileadmin/user_upload/Datoteke/Revizije/2

_na_e-goriva 019/DrugiBlok_NEK/DrugiBlokNEK_SP06-

8. Škodljivi vplivi hidroelektrarn na ribe; 16.pdf

www.dprs.si; https://www.dprs.si/sl/193- 18. JEK2: Referendum na slepo;

skodljivi-vplivi-hidroelektrarn-na-ribe.html www.greenpeace.org ;

9. Mura del globalne kampanje WWF za prosto https://www.greenpeace.org/slovenia/sporocilo-

tekoče reke; www.wwfadria.org; za-javnost/50936/jek2-referendum-naslepo/

https://www.wwfadria.org/sl/?322710/Mura-



Skrivnost ljubljanske pitne vode in kanal C0



The secret of Ljubljana's drinking water and the C0 canal



mag. Tomaž Ogrin



tomaz.ogrin@ijs.si

Institut “Jožef Stefan”



Jamova cesta 39, 1000



Ljubljana, Slovenija



POVZETEK

V prispevku odkrivamo razlog za čisto pitno vodo Ljubljane Is the gravel suf icient as a ilter for groundwater from the

in projekt, ki jo ogroža. Ljubljana, glavno mesto Slovenije z Sava River to produce clean drinking water? It turns out that

blizu 300.000 prebivalci in z več kot 100.000 dnevnih it is not. The secret of Ljubljana’s drinking water from this

obiskovalcev ima vir pitne vode, ki je ni potrebno kemijsko source lies in biological iltration. A biological ilter is made

obdelati, na primer klorirati. Podtalnica prihaja od reke Save up of organisms that live in the dark under harsh living

in se črpa v glavnem na dveh črpališčih, Kleče in Sentvid v conditions. It was only in 2024 that the general public was

obsegu 90 odstotkov pitne vode za mesto. Začetki črpanja made aware of the discoveries made by Dr. Cene Fišer about

segajo v leto 1890. Večje območje je zaščiteno z uredbo o the rich underground life of diverse but sensitive organisms

Ljubljanskem polju, ki je po sestavi slabo sprijet prod. and their function in purifying groundwater from the Sava

Ali je prod kot ilter za podtalno vodo iz Save dovolj za River.

nastanek čiste pitne vode? Izkazalo se je, da ne. Skrivnost The mystery came to light amid protests from the

ljubljanske pitne vode iz tega vira je v biološkem iltriranju. professional and other public over the construction of a large

Biološki ilter sestavljajo organizmi, ki živijo v temi ob težkih 1.2 m diameter C0 canal through the area, which would also

življenjskih pogojih. Sirša javnost je bila šele v letu 2024 carry faecal and industrial wastewater from the

seznanjena z odkritji doc. dr. Ceneta Fišerja o bogatem municipalities of Medvode and Vodice. Contamination of the

podzemnem življenju raznovrstnih, a občutljivih organizmih biological ilter would mean its destruction and thus pure

in njihovi funkciji pri čiščenju podtalnice iz Save. drinking water.

Skrivnost je prišla na dan ob protestih strokovne in druge Another important element in the protection of this drinking

javnosti zaradi gradnje velikega kanala C0, premera 1,2 m water is the fact that the Ljubljansko polje aquifer is a critical

skozi to območje, ki bi odvajal fekalije in industrijske odplake infrastructure under the EU directive and Slovenian law,

tudi iz občin Medvode in Vodice. Onesnaženje biološkega which is also a national security concern.

iltra bi pomenilo njegovo uničenje in s tem čiste pitne vode.

Pomemben element zaščite te pitne vode je tudi dejstvo, da KLJUCNE BESEDE

vodonosnik Ljubljanskega polja spada v kritično Pitna voda, biološki ilter, kanal C0, Ljubljansko polje, EU



infrastrukturo po EU direktivi in slovenskem zakonu, kar projekt spada tudi v skrb za nacionalno varnost.



odstraniti iz območja. Gradnja kanala C0 je del evropskega Obstajajo alternativne rešitve brez kanala C0, ki ga bo treba KEYWORDS

projekta: Odvajanje in čiščenje odpadne vode na območju Drinking water, biological ilter, C0 canal, Ljubljansko polje,

vodonosnika Ljubljanskega polja (2017). EU project



ABSTRACT 1. UVOD In this article, we discover the reason for Ljubljana’s pure Avgusta 2017 je bil Sloveniji odobren evropski projekt: drinking water and the project that threatens it. Odvajanje in čiščenje odpadne vode na območju Ljubljana, the capital city of Slovenia with nearly 300,000

inhabitants and more than 100,000 daily visitors, has a vodonosnika Ljubljanskega polja. Celotni projekt

source of drinking water that does not need to be chemicaly sestavljajo 3 deli (1):



comes from the Sava River and is mainly pumped at two odpadne vode v občinah Medvode (22 km) in Vodice (9,1 pumping stations, Kleče and Sentvid, which account for 90 % km) ter izgradnja povezovalnega kanala C0 v Mestni občini treated, for example chlorinated. Drinking groundwater Del 1: Nadgradnja sistema odvajanja komunalne



larger area is protected by the Ljubljansko polje Regulation, Ljubljana (MOL), of the city’s drinking water. Pumping dates back to 1890. The

which is composed of poorly adhered gravel. Del 2: Izgradnja III. faze Centralne čistilne naprave

Ljubljana,

Permission to make digital or hard copies of part or all of this work for Del 3: Dograditev kanalizacije v aglomeracijah v Mestne

personal or classroom use is granted without fee provided that copies are not

made or distributed for profit or commercial advantage and that copies bear občine Ljubljana.

this notice and the full citation on the first page. Copyrights for third-party Celotna vrednost investicije je 135.623.985 evr (z DDV), od

components of this work must be honored. For all other uses, contact the

owner/author(s). tega EU 68.879.165,20 evr (85%), država 12.155.146,80

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia evr (15%), vse tri občine pa 54.589.673 evr. http://doi.org/10.70314/is.2024.env.13

Informacijska družba, 2024, 9.- 13. Oktober 2024, Ljubljana, Slovenija T.Ogrin



potres, napake pri vgradnji, rezanje cevi, vijugasta

V tem prispevku obravnavamo Del 1, kanal C0, ki je od vgradnja…).

Broda do Crnuč v dolžini 4,1 km vgrajen v vodovarstveno

območje in bi odvajal tudi odpadne vode iz občin Medvode 2. Kanal C0 bi zanesljivo ogrozil čisto pitno vodo

in Vodice. Po zaslugi krajanov, zbranih v Civilni iniciativi za čisto pitno

Potek trase kanala C0, toka podzemne vode in lokacije vodo (CI), mednarodnega društva Alpe Adria Green (AAG)

črpališč. Kanal C0 križa tok podtalnice. Vsako puščanje gre v ter odvetnika, kanal C0 še ni v celoti zgrajen. V območju

tok podtalnice. Crnuč manjka še okrog 120 m. Zaradi nezakonito vodenega



postopka projekta in strokovnih napak, se nadaljevanje

rešuje na sodišču.

V CI sodeluje več strokovnjakov za komunalno

infrastrukturo z mednarodnimi izkušnjami, v njej sodelujem

s prispevki v medijih (npr.: Dnevnik).

Zaradi predvidenega ogrožanja pitne vode je Državni zbor

ustanovil Preiskovalno komisijo o ugotavljanju zlorab in

nezakonitosti pri gradnji povezovalnega kanala C0 na

območju ljubljanskega vodonosnika Ljubljanskega polja

(Komisija). Komisija je imela 12.9.2024 že 16. sejo s

pričanjem vabljenih, tokrat zdravnica in epidemiologinja

prof. dr. Dodič Fikfak (2).

Povzemamo nekaj ugotovitev strokovnih študij in

Poliestrske cevi kanala , premera 1,2 m.

postopkovne revizije, na osnovi katerih lahko trdimo, da bi



do puščanja kanala C0 in onesnaženja čiste pitne vode za

Ljubljano zagotovo prišlo.



2.1. Revizija Ministrstva za okolje in prostor (MOP)

2019 ugotavlja, da je iz postopkov in dokumentov razvidno,

da ni garancije, da kanal C0 ne bi puščal. Povezava na

revizijo je v objavi mednarodnega društva Alpe Adria Green



(3).



2.2. Izdelana je bila Ocena potresne varnosti na trasi

kanalizacijskega zbiralnika C0 (HGEM d.o.o., 2017).

Zaradi gradnje na potresnem območju bi kanal moral biti

zgrajen po mednarodnem standardu ISO:16134:2006 pa ni

bil. Standard zahteva ustrezen material in tip cevi ter spojev

in poseben način vgradnje.



Traso kanala prečka Zužemberški geološki prelom, Prelomi

so nevarne podzemne strukture, kar situacijo kanala še

poslabšuje.

Omenjeno poročilo ugotavlja, da na trasi ni bilo izvedenih

''osnovnih geomehansko inženirsko geoloških raziskav'' , ki bi

omogočile oceno za pojav likvifakcije (zmehčanja tal) ob

potresu ter o možni prisotnosti strižnih sil na tej dolžini 4,1

km med Brodom in Crnučami. Zato so predlagali ustrezne



preiskave, ki pa niso bile izvedene.

Take preiskave so bile izvedene pri gradnji HE Brežice (4).

Tla iz slabo sprijetega peska so bila zamenjana zaradi

nevarnosti likvifakcije in porušitve jezu. Podobna tla,

prodno-peščena zemljina, sestavljajo vodonosnik

Ljubljanskega polja.



2.3. Geološko poročilo, 2018: Program zaščite podzemne

Po izjavah strokovnjakov je tesnenje pri tako velikem



premeru nezadostno za fekalije in industrijske odplake, ob (Geološki zavod Slovenije). vode v času gradnje povezovalnega cevovoda C0 Poročilo ocenjuje tudi razmere

mnogih vplivih na cevi na veliki dolžini 4,1 km (vzgon, med obratovanjem kanala C0. Povzetek (5).

Informacijska družba 2024, 9.–13. oktober 2024, Ljubljana, Slovenija T. Ogrin





V scenariju normalnega razvoja dogodkov so ocene puščanja

kanala. Izgube odpadne vode (puščanje) se do 1 % ne dajo

zaznati. Pri polnem pretoku 930 l/s je tako puščanje 9,3 l/s.

Dnevno to pomeni okrog 800 m3 puščanja fekalij in

industrijskih odplak na 4,1 km brez zaznave. Kanalizacijski

standardi dopuščajo tak delež puščanja kot normalen.

Na najbolj občutljivem 2,2 km delu trase, v zaledju vodarn

Šentvid in Kleče, se puščanje brez zaznave ocenjuje na 1,8 l/s,

kar pomeni 155 m3/dan. Za tako puščanje so dovolj razpoke

širine 0,087 mm.

Scenarij najslabše možnosti, torej večje razpoke ali napake

pri vgradnji, pa pomeni še bistveno višje količine izpustov v

podtalnico s čisto pitno vodo.

Hitrost podtalnice od Save do Kleč je velika, 20 m/dan.

Puščanje kanala hitro doseže črpališče.



gradnji hidroelektrarn blizu Ljubljane. Več kot sto dni na leto MEHKEJŠE NASUTJE – VEČJI UČINEK VZGONA Poročilo tudi ocenjuje možnost vzgona ob poplavah ali

je v določenih predelih pričakovati vzgon zaradi dviga Namesto izkopanega peska so na nekaj odsekih nasuli mehko

podtalnice, ki lahko odpre ali zamakne spoje cevi, da kmetijsko zemljo, pesek pa tržili naprej. Učinek vzgona bi bil

nastanejo reže. večji, kar pomeni večje pomike cevi in nastajanje rež na

spojih.

2.4. Geološko poročilo, 2020: Dodatna analiza potresne



nevarnosti na območju povezovalnega kanala C0 v

Mestni občini Ljubljana (Geološki zavod Slovenije).

Posebej je obdelana prisotnost aktivnega Žužemberškega

preloma, ki v domnevnem poteku trikrat prečka kanal C0. V

zaključku ocenjujejo možnost potresa s površinskim

pretrgom velikosti 10 cm. Predlagajo dodatne geofizikalne

preiskave ter močno priporočajo izvedbo celovite

verjetnostne analize PFDH (Probabilistic Fault Displacement

Hazard Analysis). Upravičeno domnevamo (nasveti iz leta

2020, ko je velik del kanala že izgrajen), da noben predlog

ni bil izveden.



2.5. Nestrokovna in malomarna vgradnja kanala C0.

REZANJE CEVI – TESNENJE?

(slike: zasebni arhivi, vendar že predstavljeni javnosti)

Rezanje cevi in improvizacija s tesnili – proizvajalec cevi ne

DVIG PODTALNICE – VDOR V KANAL

odgovarja več za tesnost spojev. Ali sploh odgovarja in koliko

Lanske poplave so razkrile, sicer izven poplavnega

let ter ob potresih in vzgonu?

območja, torej ne gre za vdor umazane savske vode skozi



jaške, ampak za velike razpoke v že vgrajenem kanalu.

Narasla podtalnica je namreč vdrla (posnet video toka pitne

vode v kanalu) v kanal, kar pomeni, da gre za malomarno

vgradnjo in je puščanje zagotovljeno.



VZGON

Kot se spomnimo iz izike, bo vzgon večji, če so potopljeni

volumni večji. Kanal C0 s premerom 1,2 m predstavlja tak

velik volumen, ki je lahko delno ali ni napolnjen z

odplakami. Primer premika in izmeta vgrajenih cevi večjega

premera se je zgodil v Sneberjah ob lanskih poplavah. Gre za

priključne cevi na kanal C0. VIJUGASTO VGRAJEN KANAL

Na določenih odsekih je kanal vgrajen vijugasto, kar pri

velikem premeru vzpostavlja napetosti v tesnenju in v

primerih premikov (potres, vzgon) olajša odpiranje spojev.

Viden je tudi nedovoljen poseg, zaradi improvizacije

investitorja, MOL, na zasebno njivo zaradi potrebne velike

delovne širine pri vgradnji kanala. Prišlo je do protestov in

nasilja varnostnih služb.

Informacijska družba, 2024, 9.- 13. Oktober 2024, Ljubljana, Slovenija T.Ogrin





med prodom, čistijo podtalnico, ki pride od Save v

Ljubljansko polje, kar prikazuje slika.





Opomba: Človeška ribica na sliki ni del procesa čiščenja, ne



živi v produ, je del podzemne favne iz predavanja doc..dr.

Fišerja (Cankarjev dom) (7).



Sele zdaj lahko razumemo, zakaj imamo v Ljubljani čisto

MALOMARNA VGRADNJA – REŽE pitno vodo neposredno iz podtalnice, že stoletja.

Civilna iniciativa je skozi jaške ugotovila obstoječe reže. Vidi

se vdor peska skozi režo. Po javno objavljenih napakah pri Puščanje iz kanala C0 bi naš biološki ilter, ki je nastajal

vgradnji je investitor, Mestna občina Ljubljana, naročil stoletja, uničilo.

varjenje pokrovov jaškov. Na to je dr. Fišer že leta 2016 pisno opozoril Agencijo RS za



okolje (ARSO) in MOL ter župana Ljubljane, a so opozorila

ignorirali. Se več, ARSO je odločil, da presoja vplivov na

okolje, kjer bi seveda morali upoštevati biološki ilter in

kanal C0 odkloniti, ni potrebna!



Slika prikazuje podzemno favno iz predavanja, ki zajema

širšo tematiko podzemnega sveta, razen školjk in človeških

ribic, so drugi prikazani organizmi tudi v vodonosniku

Ljubljanskega polja (7).





Foto: T. Delić





Ljubljanski vodonosnik je vroča točka, evropskega in



svetovnega pomena. Ta termin se uporablja za podzemlje,

kjer je izjemno visoka pestrost podzemnih organizmov. V

Ljubljanskem polju so jih odkrili kar 29, merilo za vročo

točko pa je 25. Razvoj mnogih teh organizmov se odvija že

več deset milijonov let (8).



3. Nastanek ljubljanske čiste pitne vode – biološki ilter

Skrivnost ljubljanske čiste pitne vode, ki ne potrebuje

kemične obdelave, na primer kloriranja, je prišla do širše

javnosti šele v letošnjem letu na seji Komisije. Skrivnost nam

je na 11. seji (29.5.2024) razkril doc. dr. Cene Fišer iz

Biotehnične fakultete (6).

Biofilm bakterijske flore na produ in podzemni

organizmi (favna, nevretenčarji), ki živijo v prostorih

Informacijska družba 2024, 9.–13. oktober 2024, Ljubljana, Slovenija T. Ogrin





razlogi. Občina Medvode je izstop morala plačati z blizu

350.000 evri (11).

Vodice bi se pa lahko priključile na domžalsko čistilno

napravo.



4.3. Dodatno varovanje ljubljanske čiste pitne vode

Vodonosnik Ljubljanskega polja spada v kritično

infrastrukturo, saj napaja več kot 100.000 prebivalcev.

Kritična infrastruktura je opredeljena z evropsko direktivo

in slovenskim zakonom (12). Predstavlja pomemben

element nacionalne varnosti.



4.4. Biološki filter

Za široko javnost povsem nov in ključen element pri

nastajanju čiste pitne vode za glavno mesto Slovenije z

okrog 300.000 prebivalci in preko 100.000 dnevnih

obiskovalcev, z bolnišnicami, šolami, hoteli in drugo

funkcionalno infrastrukturo, je s projektom kanala C0

zanesljivo ogrožen in z njim čista pitna voda. Procesi

nastajanja tega filtra so počasni, stoletni in njegovo uničenje

Slika prikazuje razvoj podzemnega rakca iz rodu slepih bi prineslo za več sto let uničenje vira čiste pitne vode.

postranic (Niphargus) (8).

4.5. ZAKLJUČEK

Dela na kanalu C0 je treba takoj ustaviti, sporočiti evropski

komisiji o spremembi projekta in že vgrajen kanal C0

odstraniti iz vodonosnika. Strokoven predlog prof. dr.

Duhovnika je bil jasno postavljen na Komisiji (13).

4. RAZPRAVA IN ZAKLJUČEK

4.1. O imenu projekta LITERATURA

''Odvajanje in čiščenje odpadne vode na območju 1. MOL, občina Medvode, občina Vodice

vodonosnika Ljubljanskega polja.'' Znano je, da kanal C0 https://www.cistozate.si/o-projektu

premera 1,2 m preko vodonosnika nima nobenega priključka 2. Fikfak, D., (2024), Komisija, 16.seja,

iz območja kakršnekoli odpadne vode iz vodonosnika, ki bi jo https://365.rtvslo.si/arhiv/seje-preiskovalnih-

bilo treba čistiti. Razumevanje naslova je AAG preveril pri komisij/175071670

evropski komisiji. V dopisu 3. Alpe Adria Green, (2019), (9) kot rezultat tega dela projekta

navajajo: h ps://alpeadriagreen.org/2019/12/20/povezovalni- ''As a result it will improve the quality of the aquifer

for the abstraction of drinking water in the first place kanal-c0-notranjerevizijsko-porocilo-.''

V prevodu: ''S tem se bo izboljšala kakovost vodonosnika, iz mop/#:~:text=V%20notranji%20reviziji%20sem%20se%20

katerega se črpa pitna voda''. Očitno so razumeli, da bo osredoto%C4%8Dila%20na%20pregled%20postopkov%20



jasno kako, če kanal samo prečka vodonosnik. 4. Smolar, J. (2010), Raziskave likvifakcije peskov z V pismih bralcev (Dnevnik, 23.2.2023) sem na to opozoril, a lokacije HE Brežice projekt že tako odlično čisto pitno vodo še očistil, čeprav ni v brez odmeva (10).

https://repozitorij.uni-

lj.si/Dokument.php?id=83769&lang=slv

4.2. Zakaj še Medvode in Vodice?

5. Geološki zavod Slovenije, (2018), Povzetek Programa

Novodobno okoljsko strokovno ravnanje na področju

zaščite podzemne vode v času gradnje povezovalnega

odpadkov je, da se obdelajo na viru ali čim bližje. Izgovori na

cevovoda C0 https://www.geo-

pretekla, drugačna načrtovanja izražajo samo

zs.si/index.php/en/?option=com_content&view=article

nerazumevanje trajnostnega razvoja. Vodenje fekalij in

&id=370

industrijskih odpadnih voda 20 km iz Medvod in 9 km iz

6. Fišer, C, (2024), Komisija, 11.seja

Vodic v Ljubljano (Brod) je že v tem smislu zgrešeno.

https://www.rtvslo.si/rtv365/arhiv/175049248?s=tv

7. Fišer, C., (2023), Podzemni svet v vrtincu sprememb

Ko pa zvemo, da je občina Medvode že bila v projektu

(predavanje, Cankarjev dom)

kanalizacije skupaj z Mestno občino Kranj in občino Senčur

https://www.youtube.com/watch?v=9mI5nxClCTk&t=

in enostransko izstopila iz njega, da bi se vključila v

803s

ljubljanski projekt pa je jasno, da so prevladali nestrokovni



Informacijska družba, 2024, 9.- 13. Oktober 2024, Ljubljana, Slovenija T.Ogrin



8. Biotehniška fakulteta, Novice (2021), Vroče točke

evolucije v podzemlju jugovzhodne Evrope

https://www.bf.uni-

lj.si/sl/novice/2021062409024712/vroce-tocke-

evolucije-v-podzemlju-jugovzhodne-evrope

9. Alves, S., (2023), dopis

REGIO/E.3/TB/g k/(2023)1412568

10. Ogrin, T, (2023), Dnevnik, Pisma bralcev, Kanal C0 ali

onesnaževanje čiste pitne vode

https://www.dnevnik.si/1043007586/Pisma%20bralcev/ka

nal-c0-ali-onesnazevanje-ciste-pitne-vode

(povezava ne deluje zaradi spremembe njihovih strani)

11. Projekt GORKI, (2017), Poravnava

https://medvode.si/objava/113236#:~:text=Na%20Vi

%C5%A1jem%20sodi%C5%A1%C4%8Du%20v%20Lj

ubljani%20je%20v%20petek,%2020.%20oktobra

12. Zakon o kritični infrastrukturi, (2017)

https://pisrs.si/pregledPredpisa?id=ZAKO7106

13. Duhovnik, J., Ogrin, T., (2024), Komisija, 14. seja,

https://365.rtvslo.si/arhiv/seje-preiskovalnih-

komisij/175059123



Kakšna je cena jedrske energije in obnovljivih virov in



kakšen je vpliv na okolje



What is the Cost of Nuclear Energy, Renewables, and



Their Impact on the Environment



Andrej Trkov

Odsek za reaktorsko fiziko, Institut Jožef Stefan

Ljubljana, Slovenija

andrej.trkov@ijs.si



Povzetek pa so projekcije špekulativne, stroški financiranja pa se pojavljajo pri vseh investicijah. Preprosto nas zanima, kakšni so Strateška odločitev glede izbire osnovnih energentov v državi je stroški na dan odločitve za gradnjo, cene na tržišču za to pa so ključnega pomena za zanesljivost oskrbe in za ceno energije. znane. Narejena je primerjava cene 1100 MW bloka jedrske energije in

100 % obnovljivih virov (OVE, sončnih panelov) za enako

količino oddane energije v omrežje ter vpliv na izrabo prostora. 2 Cena jedrske energije



Ključne besede V medijih se pogosto pojavljajo šokantno visoke cene postavitve

jedrskih elektrarn, kot jih na primer navaja Greenpeace

Jedrska elektrarna, obnovljivi viri energije, OVE, cena za enako »Finančne ovire za nove jedrske projekte - Greenpeace

količino energije Slovenija«. Tak izbor cen je nerealen in je posledica dejstva, da

evropski in ameriški proizvajalci nekaj deset let niso gradili

Abstract

jedrskih elektrarn. Zaradi tega so bile dobavne verige v veliki

Strategic decision about the choice of energy sources is crucial meri prekinjene in jih je bilo treba vzpostaviti na novo, vsled

for a reliable supply of electricity at a reasonable cost. A česar je prihajalo do zamud in povečevanja stroškov. Situacija je

comparison is made between the option of a 1100 MW nuclear bila drugačna v 60 in 70-ih letih prejšnjega stoletja, ko so

plant and 100 % renewables (solar panels) for the same amount ameriški in evropski proizvajalci gradili po več jedrskih elektrarn

of delivered energy to the grid, and the impact of the usage of na leto. Kitajski proizvajalci so v zadnjem desetletju zgradili 39

land. jedrskih elektrarn, zdaj pa napovedujejo gradnjo po deset (in več)

elektrarn na leto in s tem nimajo težav. Pričakujejo, da bodo v

Keywords obdobju v letih od 2020 do 2035 zgradili 150 jedrskih reaktorjev

Nuclear power plant, renewable energy sources, cost for equal »How Innovative Is China in Nuclear Power? | ITIF«. Ravno

amount of energy produced tako je južnokorejski proizvajalec KHPN v Združenih arabskih

emiratih zgradil 4 reaktorje v sprejemljivih časovnih rokih in

pogodbenih finančnih okvirih. Cena na eno 1400 MW enoto je

1 Ozadje problema bila 8 milijard dolarjev »Barakah nuclear power plant -

Strateška odločitev glede izbire osnovnih energentov v državi je Wikipedia« Trenutna cena jedrskih elektrarn je torej znana in

ključnega pomena za zanesljivost oskrbe in za ceno energije, ki predvidljiva. Za 1000 MW jedrsko elektrarno ELES ocenjuje

jo industrija in posamezniki potrebujemo, zato je pomembno ceno 10 milijard EUR, ocena GEN energije pa je še nekoliko

vedeti, kakšne so možnosti in kakšne so implikacije. V Sloveniji nižja. Elektrarna daje takšno moč vse dni v letu ves dan, razen v

bomo po zaprtju TEŠ6 leta 2033 potrebovali dodatne proizvodne času enomesečnega remonta vsakih 18 mesecev. Življenjska

kapacitete za električno energijo z močjo vsaj 1000 MW. Namen doba jedrskih elektrarn je 60 let ali več. Željeno energijo,

predstavitve je primerjava cene za 1100 MW jedrsko elektrarno upoštevaje čas remonta bi torej dobili s postavitvijo 1100 MW

in cene obnovljivih virov (OVE, predvsem fotovoltaičnih jedrske elektrarne.

panelov) za enako količino proizvedene energije. Ocene stroškov

v medijih se zamegljujejo s parcialnim vpletanjem stroškov

financiranja in ocenami na osnovi projekcij gibanja cen, vendar

3 Cena obnovljivih virov energije

Permission to make digital or hard copies of part or all of this work for personal or Trenutna različica Nacionalnega energetskega in podnebnega

classroom use is granted without fee provided that copies are not made or distributed načrta (NEPN) kot eno od opcij predvideva pridobivanje

for profit or commercial advantage and that copies bear this notice and the full električne energije iz 100 % obnovljivih virov (OVE), pretežno

citation on the first page. Copyrights for third-party components of this work must iz sončnih panelov na strehah stavb. be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Iz razpoložljivih podatkov dajejo paneli v Sloveniji v

© 2024 Copyright held by the owner/author(s). povprečju le eno desetino nominalne moči (ker sonce ne sije

http://doi.org/DOI_RECEIVED_AFTER_REVIEW

ponoči in ker se nahajamo na 46. vzporedniku), zato je za

približno enako količino energije kot jo proizvede 1100 MW

jedrska elektrarna potrebna instalirana moč 10 000 MW sončnih

panelov. Če je tipična instalirana moč sončnih elektrarn na

strehah stavb 10 kW, potrebujemo torej milijon takšnih enot.

Cena ene enote pri ponudnikih v Sloveniji je dobrih 10 000 EUR

(če niso potrebna dodatna dela na strehi), to pa pomeni, da

celotna investicija znaša vsaj 10 milijard EUR, kar je primerljivo

s ceno jedrske elektrarne.





Slika 2: Cena 100 % OVE za primerljivo količino energije



kot jo proizvede 1100 MW jedrska elektrarna.



Slika 1: Cena OVE za primerljivo količino energije kot jo 3.2 Neznanke v izračunu

proizvede 1100 MW jedrska elektrarna. V izračunu je upoštevana predpostavka, da se primanjkljaj



3.1 Nestanovitnost električne energije iz sončnih energije v nočnem času in ob slabem vremenu nadoknadi s

hranilniki, na primer z baterijami, kar praktično podvoji ceno

panelov izgradnje sončne elektrarne.

Resna težava je nestanovitnost proizvodnje električne



energije, z viški podnevi in v poletnem času, ko vse elektrike ne

bi mogli niti porabiti ali shraniti, in s primanjkljajem v zimskem

času, ko je treba potrebe pokrivati z drugimi viri. Največkrat so

za to predvidene termoelektrarne na zemeljski plin. To pa

pomeni, da v tem primeru uporaba sončnih panelov ne

predstavlja več scenarija 100 % OVE, saj je zemeljski plin

fosilno gorivo.

Razmere se spremenijo le, če lahko viške energije ob

presežku proizvodnje shranimo za nočni čas in za zimske mesece.

Danes tehnologija za dolgoročno shranjevanje energije sicer še

ni komercialno dostopna, raziskuje pa se v smeri, da bi lahko z Slika 3: Cena 100% OVE s hranilniki za primerljivo

količino energije kot jo proizvede 1100 MW jedrska

elektriko proizvajali »zeleni plin«, pri čemer pretežno mislimo

na zeleni vodik, ker je tehnologija pridobivanja zelenega metana elektrarna.



tehnološko in ekonomsko bolj nejasna. Tudi NEPN za Poleg tega je treba k stroškom prišteti še stroške izgradnje

prihodnost v 100 % OVE scenariju predvideva proizvodnjo plinskih elektrarn in obratov za proizvodnjo zelenega plina, za



panelov v principu pokrita, žal pa pri izkoriščanju poletnih komercialna cena nedoločljiva. Treba je tudi skladiščiti velike viškov za proizvodnjo plina in uporabo tega plina za proizvodnjo količine plina (kar tudi predstavlja svojevrsten izziv in zelenega plina. Energijska bilanca je z 10 000 MW sončnih kar tehnologija na tako veliki skali še niti ne obstaja, in je



električne energije prihaja do izgub, ki so fizikalno pogojene in potencialno nevarnost), zamenjati transformatorske postaje in jih ni mogoče bistveno zmanjšati. Povečati bi morali kapaciteto močno ojačati elektro-distribucijski sistem (precej bolj kot v sončnih elektrarn za toliko, da bi z viški energije poleti proizvedli primeru jedrske opcije, kjer vsaj visoko-napetostni sistem dovolj zelenega plina za izravnavo potreb v zimskih mesecih. Po približno zadošča). grobi oceni je v najboljšem primeru možno nazaj v električno

energijo pretvoriti kvečjemu eno tretjino poletnih viškov.



Praktično to pomeni, da potrebujemo v celoti najmanj 15 000

MW instalirane moči sončnih panelov. V tem primeru znaša cena

15 milijard EUR za milijon in pol 10 kW enot, kar presega ceno

jedrske elektrarne že za polovico.



Slika 4: Dodatni nedoločljivi stroški OVE: obrati za

proizvodnjo zelenega plina, skladiščenje, termoelektrarna

na zeleni plin in ojačanje prenosnega elektroenergetskega

sistema.



stroške, zato trditev, da je scenarij OVE-plin ali 100 % OVE

poceni (milo rečeno) brez trdne osnove.

3.3 Življenjska doba OVE

Ocena življenjske dobe sončnih panelov je precej optimistično Podrobnejši izračun je na voljo na portalu »jedrska.si« Kako

napovedana na 30 let, kar je težko dokazljivo, saj večina panelov poceni je elektrika iz obnovljivih virov in kako draga je jedrska

še ne dosega starosti, da bi to lahko statistično potrdili, pa tudi energija?

tehnologija izdelave se spreminja. Poleg tega je verjetno večji

problem v močnostni elektroniki, da ne omenjamo vremenskih



več kot polovica panelov. Življenjska doba jedrskih elektrarn je Zahvala pojavov, kot je bil letos v Slovenski Bistrici, kjer je bila uničena

60 in več let, zato je treba v enakem obdobju zamenjati sončno Zahvaljujem se gospe Tamari Langus za uredniške izboljšave

elektrarno (ali vsaj večino komponent) najmanj dvakrat, kar besedila in uporabo slik za ilustracijo.

izhodiščno ceno lahko še podvoji.



3.4 Umeščanje v prostor in vpliv na okolje

Jedrska elektrarna potrebuje manj kot en kvadratni kilometer

prostora (okrog 100 hektarov), kar s stališča izrabe zemljišča ni

zelo problematično. Za milijon in pol sončnih elektrarn je po

grobi oceni potrebnih deset do petnajst tisoč hektarov površine

na sončni lokaciji, saj je treba upoštevati nekaj zemljišča za

servisiranje in čiščenje panelov. Resda se predvideva, da bo

veliko panelov nameščenih na strehah stavb, vendar si je težko

predstavljati, da bi samo z enotami na strehah s primerno

osončenostjo dosegli željeno instalirano kapaciteto za

proizvodnjo električne energije. Del kapacitet bo treba nameščati

tudi na primerno osončena zemljišča.



3.5 Dobavni roki, montaža in servisiranje

Očitek jedrskim elektrarnam je dolga doba umeščanja v

prostor in izgradnjo, vendar tudi naročilo milijon in pol 10 kW

enot OVE ne bo potekalo čez noč, še manj pa umeščanje v

prostor za deset do petnajst tisoč hektarov površine. Od samega

naročila panelov domača industrija ne bo imela velike koristi, saj

so paneli praviloma uvoženi iz daljnega vzhoda, doma je

zaposlena zgolj množica trgovcev, monterjev in serviserjev za

domači trg, ki pa ne ustvarjajo dodane vrednosti, ki bi jo lahko

tržili v tujini.



4 Zaključek

Cena postavitve jedrske elektrarne z močjo 1100 MW je

predvidljiva in znaša približno 10 milijard EUR. Za enako

količino proizvedene energije iz malih sončnih elektrarn tipične

moči 10 kW potrebujemo milijon takšnih enot, upoštevaje

trenutne cene na tržišču pa je strošek primerljiv s ceno izgradnje

jedrske elektrarne. Če želimo poletne viške električne energije

shraniti v obliki zelenega plina za pogon termoelektrarn v

zimskih mesecih potrebujemo najmanj pol milijona dodatnih

enot, to pa pomeni že za polovico višji strošek v primerjavi s ceno

jedrske elektrarne. V to niso vključene cene baterij za

kratkoročno hrambo električne energije, postavitve naprav za

proizvodnjo zelenega plina, skladišč plina in distribucijskega

sistema ter plinske termoelektrarne, ki je nujno potrebna vsaj za

vzdrževanje stabilnosti elektro-distribucijskega sistema. V ceni

100 % OVE je veliko neznank, ki vsaka posebej povečujejo





O statusu in prihodnosti jedrske energije




On the Status and Future of the Nuclear Power



Leon Cizelj

Odsek za reaktorsko tehniko

Institut “Jožef Stefan”

Ljubljana, Slovenija

Leon.Cizelj@ijs.si



Povzetek Keywords

V prispevku pokažemo status jedrskih elektrarn v obratovanju, v Nuclear power plants, large reactors, gigawatt reactors, small and

gradnji in v razvoju. Na kratko osvetlimo tudi najpomembnejše modular reactors, new talents

razlike med reaktorji 2. (npr. NEK) in 3. generacije (kandidati za

JEK2). Sledi kratek pregled mednarodnih raziskovalnih in

razvojnih prizadevanj, ki so usmerjena predvsem v manjše in 1 Uvod

modularne enote, ki bodo jedrsko fisijo uporabile kot Jedrska energija je v letu 2023 prispevala 9,1% vse električne

brezogljični vir toplote za neposredno rabo ali pretvorbo v energije [1] oziroma 3,96% vse primarne energije [2] na planetu.

elektriko. Takšne enote utegnejo biti primerne tudi za Proizvodnja jedrske elektrike na planetu je leta 1996 prvič

decentralizirana pametna omrežja. Komercializacija nekaterih od presegla 2.400 TWh in se je v zadnjih nekaj letih ustalila med

večinoma že pred desetletji preverjenih zasnov bo omogočila 2.600 in 2.700 TWh [3]. Vse jedrske elektrarne, ki smo jih v

tudi uporabo današnjega rabljenega jedrskega goriva in bo lahko Evropi, v ZDA in na Japonskem v tem času zaprli, so sproti

pomembno prispevala k zaključevanju krožnega gospodarjenja z nadomestile nove elektrarne in sicer predvsem na Kitajskem, v

jedrskimi gorivi. Zaključimo s povzetkom največjih izzivov Indiji in še kje: od leta 2000 do danes smo na planetu zaprli 123

današnje jedrske energetike v Sloveniji in v svetu: sprejemljivost reaktorjev s skupno močjo 80 GW, na omrežje pa na novo

v javnosti, dobavne verige za talente in opremo ter nujnost priključili 121 reaktorjev s skupno močjo 110 GW. Trenutno

dolgoročnih investicij v raziskave in razvoj. gradimo 59 reaktorjev, od tega Kitajci 24, sledijo Indijci z 8 [5].

Konec leta 2023 je delovalo 413 reaktorjev s skupno močjo

Ključne besede 371,5 GWe. 60 let izkušenj s komercialnimi jedrskimi

jedrske elektrarne, veliki reaktorji, gigavatni reaktorji, mali in elektrarnami jedrsko energijo uvrščajo med najbolj varne in

modularni reaktorji, novi talenti najbolj čiste (Slika 1) vire energije na planetu [4].



Abstract



Status of the nuclear power in operation, construction and

operation is outlined. The main differences of the 2nd (e.g.,

Krško) and 3rd generation reactors are sketched. This is followed

by an outline of the international research and development

efforts, focussing mostly on the small and modular reactors to be

used as a source of heat for direct use or for electricity production.

Such units may be optimal also for decentralized grids.

Commercialisation of some concepts proven decades ago may

facilitate also the use of today’s spent fuel and with this also the



closure of the nuclear circular economy. Finally, the major Slika 1: Najčistejši in najvarnejši viri energije [4]

challenges of the nuclear power are briefly discussed: the public

acceptance, the supply chains for equipment and talent, and the Razvojna, poslovna in politična podpora jedrski energiji se v

need for long term investments in research and development. zadnjih letih povečujeta tudi v Evropi in Severni Ameriki. V

ZDA se povečana politična podpora jedrski energiji vidi v večjih

Permission to make digital or hard copies of part or all of this work for personal or raziskovalno razvojnih vložkih, pa tudi spremembah zakonodaje

classroom use is granted without fee provided that copies are not made or distributed in davčnih spodbudah. Rezultati se vidijo predvsem v

for profit or commercial advantage and that copies bear this notice and the full intenzivnem razvoju in trženju malih in modularnih reaktorjev z

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s). močjo pod 300 MW. Prve priključitve na mrežo so najavljene



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia okoli leta 2030. © 2024 Copyright held by the owner/author(s). V Evropski Uniji se, s podporo Slovenije, krepi jedrska https://doi.org/10.70314/is.2024.env.15

aliansa, ki si je zastavila cilj 150 GW jedrske elektrike do leta

2050 [6]. Današnja flota zmore okoli 100 GWe in je v povprečju



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia L. Cizelj et al.



stara 38 let. Ciljamo torej na 50 novih reaktorjev in 250.000 še posebej tisti za obvladovanje staljene sredice reaktorja,

novih zaposlitev. Alianso sestavlja 15 ministrov za energijo iz zagotavljajo tudi razmeroma dobro obvladovanje tveganj pri

Slovenije, Francije, vseh preostalih evropskih jedrskih držav izdelavi komponent in pri pridobivanju vseh potrebnih dovoljenj

(razen Španije), ter hrvaške, Italije, Poljske in Estonije. reaktorjev tretje generacije. Uporabljajo torej predvsem

Kitajska načrtuje do leta 2050 za pribl. 300GW novih jedrskih tehnologije, ki so bile preverjene že v 2. generaciji reaktorjev.

elektrarn, torej skoraj toliko, ko je danes na voljo na celotnem

planetu: 371,5 GW.

V nadaljevanju se bomo posvetili dvema skupinama 3 Mali modularni reaktorji

reaktorjev, ki bosta v prihodnosti zagotovo odigrali pomembno Mali in modularni jedrski reaktorji (SMR, Small Modular

vlogo pri čisti energiji na planetu in morda tudi v Sloveniji. Prva Reactors) kot pogonski agregati plovil delujejo že vse od leta

skupina so veliki, gigavatni reaktorji z močjo 1 GWe ali več in s 1954, ko so v ZDA splovili prvo jedrsko podmornico Nautilus.

klasično zasnovo, druga pa so mali in modularni reaktorji. Mimogrede, leta 1954 je pričela delovati tudi prva komercialna

jedrska elektrarna, in sicer Calders Hall v Združenem Kraljestvu.



2 Veliki reaktorji (1 GWe in več) morda 300 MWe, modularni pa zato, ker sta na nekaterih plovilih Mali zato, ker so njihove moči značilno nekaj deset pa vse do

Med 59 reaktorji s skupno močjo 61 MWe, ki so danes v gradnji, nameščena tudi po dva ali morda celo več reaktorjev.

je kar 50 tlačnovodnih (skupna moč 54,6GWe). Gre za podobno V vojaški mornariški rabi se je uveljavilo gorivo z več kot

tehnologijo, kot jo uporablja Nuklearna elektrarna v Krškem (2. 90% urana 235. Države, ki takšna plovila gradijo, imajo namreč

generacija reaktorjev), z nekaterimi posodobitvami, ki jih bomo dostop do visoko obogatenega urana, ki ga lahko uporabijo tudi

na kratko opisali v nadaljevanju. za jedrsko orožje. V civilni rabi danes velja omejitev na največ

Japonci gradijo 2 vrelna reaktorja s skupno močjo 2,6GWe, 20% urana 235, gorivo za gigawatne reaktorje pa praviloma ne

ki tudi sodita med velike reaktorje z GW električne moči. Indija presega 5% urana 235. Pomembna in praktično edina prednost

gradi 3 reaktorje, moderirane s težko vodo (skupna moč 1,9GWe). visoko obogatenega urana v mornariških reaktorjih je izjemno

Ti reaktorji sodijo v t.i. tretjo generacijo reaktorjev. dolgo delovanje, tudi 20 ali več let, brez menjave goriva.

Preostali štirje reaktorji v gradnji sodijo med tako imenovane Danes na planetu komercialno delujejo 4 SMR. Dva sta

hitre oplodne reaktorje (moč 2GWe), ki sodijo v 4. generacijo in nameščena na plavajoči elektrarni Akademik Lomonosov, ki je

odpirajo možnosti za uporabo urana 238 in torija 232. Uran 238 zasidrana v ruskem zalivu Pevek na vzhodu Sibirije. Dva

ob zajetju nevtrona v jedro namreč transmutira v cepljivi plutonij reaktorja s po 35MWe (150 MW toplote, KLT-40S, Slika 2), ki

239, torij 232 pa v cepljivi uran 233. Plutonij 239 in uran 233 sta sta naslednika reaktorjev z ruskih civilnih ledolomilcev, že od

torej umetno pridobljeni gorivi v oplodnih reaktorjih in bosta v leta 2020 oskrbujeta prebivalce z elektriko in toplo vodo. Druga

prihodnosti lahko dopolnili in kasneje tudi nadomestili v naravi dva delujeta v jedrski elektrarni Shidao Bay na Kitajskem. Dva

prisotni uran 235 (Urana 235 je le 0,72% vsega urana na planetu, visokotemperaturna, s plinom hlajena reaktorja HTR-PM (Slika

preostanek je skoraj v celoti uran 238). S 4. generacijo reaktorjev 2) s toplotno močjo 250MW poganjata skupni turbogenerator z

se na tem mestu ne bomo več ukvarjali. Omenimo le, da bodo električno močjo 250 MW.

takšni reaktorji v prihodnosti pomembno prispevali h krožnemu Poročilo agencije za jedrsko energijo pri OECD [8] je v

gospodarstvu z jedrskimi gorivi in s tem tudi k pomembnemu začetku leta 2024 identificiralo 98 zasnov SMR. Izmed teh so jih

zmanjševanju že tako razmeroma zelo majhne količine 56 s podrobnejšim opisom vključili v poročilo. Slika 2 prikazuje

radioaktivnih odpadkov. porazdelitev zasnov reaktorjev po temperaturi, ki jo dosegajo, po

Reaktorji druge in tretje generacije uporabljajo praktično toplotni moči in po zasnovi oz. tipu reaktorja [8].

povsem enako jedrsko gorivo. Razlike v zasnovi reaktorja, Večina reaktorjev ima toplotno moč pod 1GW, kar v grobem

goriva in v obvladovanju verižne reakcije, pa tudi pri ustreza tudi elektičnim močem do nekako 300MW. Vodno

uporabljenih strukturnih materialih, so torej praktično hlajeni reaktorji dosegajo do nekako 330°C, kar več kot zadošča

zanemarljive. Ključne razlike najdemo predvsem pri zasnovi za daljinsko ogrevanje, nizkotemperaturne industrijske procese

oziroma izvedbi nekaterih varnostnih sistemov. Tako v tretji in seveda proizvodnjo elektrike s pomočjo vodne pare. Najbolj

generaciji najdemo ali več pasivnih hladilnih sistemov, ki za vroči plinsko hlajeni reaktorji dosegajo 1.000°C, izjemoma tudi

delovanje potrebujejo le gravitacijo, in/ali okrepljeno več. Vmes pa najdemo reaktorje, hlajene s tekočimi kovinami

večkratnost ter geografsko neodvisnost aktivnih hladilnih (svinec, natrij, staljene soli).

sistemov. Za drugo generacijo je npr. značilna logika 1 od 2 ali 2 Praktično vse zasnove so že bile preverjene pred več desetletji

od 3, torej zadošča 1 od 2 ali 2 od 3 aktivnih sistemov, medtem v nekaj več kot 800 raziskovalnih reaktorjih, ki so na planetu

ko v tretji generaciji praviloma zahtevamo delovanje vsaj 1 od 3, obratovali od začetka jedrske dobe s Fermijevim reaktorjem

včasih tudi 1 od 4 vgrajenih varnostnih sistemov. Chicago pile 2 leta 1942. Žal pa pot do hitrejše komercializacije

Morda najpomembnejši razvojni dodatek v reaktorjih tretje otežuje »izgubljeno znanje«, ena od najpomembnejših posledic

generacije so sistemi za obvladovanje staljene sredice reaktorja. razmeroma hitrega opuščanja jedrskih raziskav proti koncu

V grobem sta prisotni dve strategiji: hlajenje staljene sredice v 1960-ih in v začetku 1970-ih. Raziskovalne reaktorje so

reaktorski posodi ali pa ujetje in hlajenje v posebnih razgradili, ljudje, ki so jih zasnovali in jih tudi spravili v pogon,

zadrževalnikih, ki so nameščeni pod reaktorsko posodo [7]. pa so že dalj časa v drugih poklicih, mnogi od njih pa tudi niso

Praktično zanemarljive spremembe v gorivu in uporabljenih več med nami.

strukturnih materialih ter nekoliko okrepljeni varnostni sistemi,

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia L. Cizelj et al.





Slika 2: Porazdelitev SMR po moči in temperaturi [8]





Slika 3: Geografska porazdelitev sedežev podjetij, ki razvijajo SMR [8]



Najhitrejšo komercializacijo in seveda tudi pridobivanje vseh ki danes predstavljajo daleč največji delež delujočih

potrebnih dovoljenj zato pričakujemo pri lahkovodnih reaktorjih, komercialnih reaktorjev. Morda še najbližje delujočemu



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia L. Cizelj et al.



demonstratorju do leta 2030 ali vsaj v prvih letih 2030-tega deležnikov, še posebej vlade, investitorja in raziskovalnih ter

desetletja so tako VOYGR, AP300 in BWRX300 (ZDA), izobraževalnih institucij [8], [9].

CAREM (Argentina), seriji RITM in KLT (Rusija) in morda tudi

NUWARD (Francija).

Slika 3 prikazuje geografsko porazdelitev sedežev podjetij, ki 5 Zaključki

razvijajo male oz. modularne reaktorje. Z velikim naskokom Ključni izzivi jedrske energetike danes večinoma niso tehnični,

vodijo ZDA, sledita Evropska skupnost in Rusija. Ključni razlog mednje pa lahko štejemo razmeroma visoke investicije, še

za veliko število zagonskih podjetij v ZDA je v tem, da je vlada posebej stroške financiranja, razmeroma dolgo gradnjo in

po zastoju v 1980-ih že konec 1990-ih let pričela razmeroma premajhno kapaciteto dobavnih verig za opremo, in morda še

veliko investirati v raziskave in izobraževanje v jedrskem bolj pomembno, za nove jedrske talente. Skupni imenovalec

inženirstvu. Rusi in Kitajci razvojnega cikla nikoli niso prekinili. ključnih izzivov je dezinvestiranje v proizvodne ter seveda tudi

Kitajci zaradi pomanjkanja energije sicer več virov usmerjajo v raziskovalne in razvojne kapacitete v preteklosti.

gigavatne reaktorje. Evropa, z njo žal tudi Slovenija, močno Z jedrsko elektrarno sobivamo stoletje ali morda celo več.

zaostaja. To je v veliki meri posledica preusmeritve razvojnih Stoletje prinese s sabo ogromne spremembe tako v znanosti in

sredstev iz jedrske energije v druge t.i. čiste oz. zelene energetske tehnologiji, kot tudi v družbi. Te spremembe v veliki meri

tehnologije. sooblikuje in napoveduje znanost. Zato bi bilo naivno

pričakovati, da lahko jedrska energetika na dolgi rok uspeva brez



4 Ključni izzivi jedrske energetike Najpomembnejši izziv gradnje novega jedrskega bloka v strateškega sodelovanja in aktivnega sobivanja z znanostjo.

Ključni izzivi jedrske energetike danes večinoma niso tehnični. Sloveniji utegne biti pravočasno zagotavljanje dovolj odlično

Na trgu je namreč dobavljivih več vrst gigavatnih reaktorjev, ki izobraženih, usposobljenih in motiviranih jedrskih talentov. Ne

temeljijo na preverjenih tehnologijah izdelave in obratovanja, so glede na to, da Slovenija osnovno izobraževalno in raziskovalno

že pridobili vsa potrebna dovoljenja v več državah po svetu, in infrastrukturo za jedrsko inženirstvo in reaktorsko fiziko ima, bo

tudi dosegajo predvidene 90% in višje faktorje izrabe. za uspešno iskanje in razvoj talentov potrebno strateško

Med ključne izzive pa lahko štejemo razmeroma visoke sodelovanje vseh vpletenih deležnikov, še posebej vlade in

investicije, še posebej stroške financiranja, razmeroma dolgo investitorja in raziskovalnih ter izobraževalnih institucij.

gradnjo in premajhno kapaciteto dobavnih verig za opremo, in

morda še bolj pomembno, za nove jedrske talente. Skupni Zahvala

imenovalec vseh izzivov je dezinvestiranje v proizvodne ter Avtorji se zahvaljujemo za finančno podporo ARIS 02-0026 in

seveda tudi raziskovalne in razvojne kapacitete v preteklosti. EC EURATOM, projekta ENEN+ in ENEN2Plus. Tudi zato je jedrska energija, ne glede na svojo nesporno čistost

ter energetsko, prostorsko in snovno izjemno učinkovitost, v Literatura

javnosti marsikje izgubila kredibilnost oz. zaupanje. Če namreč [1] »Share of electricity production from nuclear«, Our World in Data, niti tisti, ki z neko tehnologijo že zaslužijo, niso pripravljeni https://ourworldindata.org/nuclear-energy, 17.9.2024.

investirati v njeno prihodnost, se bo dvom v to tehnologijo [2] »Share of primary energy consumption from nuclear, 2023«, Our

razmeroma hitro razrasel tudi v njihovi okolici in seveda tudi v 17.9.2024. World in Data, https://ourworldindata.org/nuclear-energy,

splošni javnosti. [3] »Nuclear power generation«, Our World in Data,

Z jedrsko elektrarno namreč sobivamo stoletje ali morda celo https://ourworldindata.org/nuclear-energy, 17.9.2024.

[4] »What are the safest and cleanest sources of energy«, Our World in

več (desetletje za odločitev, desetletje za gradnjo, šest ali več Data, https://ourworldindata.org/safest-sources-of-energy,

desetletij z obratovanje in še desetletje ali dve za razgradnjo). 20.9.2024.

[5] IAEA Power Reactor Information System, http://pris.iaea.org,

Stoletje prinese s sabo ogromne spremembe tako v znanosti in 17.9.2024.



tehnologiji, kot tudi v družbi. Te spremembe v veliki meri [6] Declaration of the EU Nuclear Alliance, meeting of March 4th, 2024, https://presse.economie.gouv.fr/declaration-of-the-eu-nuclear- sooblikuje in napoveduje znanost. Tudi zato bi bilo naivno alliance-meeting-of-march-4th-2024/ 19.7.2024.

pričakovati, da lahko jedrska energetika na dolgi rok uspeva brez [7] M. Leskovar, V. Centrih, L. Cizelj, M. Draksler, T. Holler, I. Kljenak,

strateškega sodelovanja in aktivnega sobivanja z znanostjo [8]. J. Kokalj, R. Krpan, M. Kunšek, B. Mavko, J. Oder, A. Prošek, S. El

Shawish, M. Tekavčič, I. Tiselj, A. Volkanovski, Kratek opis

Najpomembnejši izziv gradnje novega jedrskega bloka v varnostnih karakteristik sedmih potencialnih reaktorjev za JEK 2,



Sloveniji utegne biti pravočasno zagotavljanje dovolj odlično IJS-DP-12437, 2018. [8] OECD/NEA, NEA Small Modular Reactor (SMR) Dashboard, Izdaja izobraženih, usposobljenih in motiviranih jedrskih talentov. 28.2.2024, https://www.oecd-nea.org/jcms/pl_90816/the-nea-small-

Pričakovano število je blizu 1.000 v vsem času gradnje, torek modular-reactor-dashboard-second-edition.

okvirno vsaj 100 vsako leto. Ne glede na to, da Slovenija [9] Cizelj, L., Pesznyák, C., Starflinger, J., Pavel, G.L., Wastin, F.,

Michailidou, E., 2024. Towards strategic agenda for European

osnovno izobraževalno in raziskovalno infrastrukturo za jedrsko nuclear education, training, and knowledge management. Nucl Eng

inženirstvo in reaktorsko fiziko ima, bo za uspešno iskanje in Des 420, 113001, https://doi.org/10.1016/j.nucengdes.2024.113001.

[10] Cizelj, L., Kljenak, I., Tiselj, I., 2023. Know-why? and know-how?

razvoj talentov potrebno strateško sodelovanje vseh vpletenih in the development of nuclear talents: An analysis of recent nuclear

engineering Ph. D. research. Nucl Eng Des 415, 112734,

https://doi.org/10.1016/j.nucengdes.2023.112734.



Natura 2000 in ohranjanje narave, spodbude ali ovire pri



razvoju elektroenergetskega omrežja



Natura 2000 and nature conservation, incentives or



obstacles in the development of the electricity network



Aleš Kregar

ELES, d.o.o.

Ljubljana, Slovenija

ales.kregar@eles.si



Povzetek

Natura 2000 in drugi pravni režimi ohranjanja narave v Republiki 1 Uvod

Sloveniji predstavljajo pomembno omejitev pri umeščanju Narava predstavlja najbolj zapleten sistem medsebojno odvisnih

elektroenergetskih vodov. Za te posege v varstvena območja se dejavnikov, ki človeštvu z zagotavljanjem zraka, vode, hrane in

napoveduje, da imajo velike negativne ali celo uničujoče učinke. energije, omogoča preživetje Ker njenega delovanja ne

Pri tem se upošteva, da se narava ne spreminja, da so vse razumemo, jo mnogi želijo ohraniti nespremenjeno.

spremembe posledice posameznih projektov in da je potrebno V Sloveniji prevladuje prepričanje, da živimo v raznoliki in

zavarovati zavarovane živalske vrste znotraj posamezne države ohranjeni naravi »na sončni strani Alp«. V medijih posledično

članice EU. Pregled nekaterih primerov kaže, da izhodišča na zasledimo množico prispevkov, ki poudarjajo kakovost okolja v

katerih temelji ohranjanje narave v območjih Natura 2000 in katerem živimo. Pozablja pa se, da je za preživetje prebivalcev

merila za ocenjevanje vplivov elektroenergetskih vodov na ta Slovenije, nujno zagotoviti tudi vire za preživetje.

območja potrebujejo preveritev, spremembo razmišljanja in Visok življenjski standard in ustvarjanje dodane vrednosti

verjetno posodobitev zakonodaje. Podani so predlogi s katerimi zahteva kakovostno infrastrukturo. Oskrba z električno energijo

bi pospešili umeščanje novih elektroenergetskih vodov in je ena pomembnejših, saj omogoča delovanje tudi drugih

obnovo obstoječih, ki bodo potrebni zaradi elektrifikacije infrastruktur. Za prenos električne energije od virov do

mnogih dejavnosti in vključevanja obnovljivih virov. porabnikov, je potrebno elektroenergetsko omrežje. V Sloveniji

so mnogi deli tega omrežja starejši od pol stoletja [1] in jih bo

Ključne besede nujno obnoviti oz. nadomestiti in zgraditi tudi nove. Evropski trg

Natura 2000, elektroenergetsko omrežje, ohranjanje narave, električne energije spodbuja tranzit med državami, kar tudi

varstvo narave zahteva ojačitve in širitve omrežja [2]. Po drugi strani pa



Abstract trajajo več kot deset let [3, 4]. Elektroenergetskim vodom se postopki umeščanja za nekatere elektroenergetske objekte



Natura 2000 and other legal nature conservation regimes in the za izogibanje vplivom in omilitvenih ukrepov je vse večji, kar pripisuje pomembne negativne vplive na naravo. Obseg ukrepov



placement of power lines. These encroachments on conservation vse bolj obremenjuje investicije v elektroenergetsko omrežje in Republic of Slovenia represent an important limitation in the

areas are predicted to have major negative or even devastating posredno ustvarjalce nove vrednosti.

effects. Here, it is considered that nature does not change, that all

changes are the result of individual projects and that it is

necessary to protect protected animal species within each EU 2 Konzervatorski pristop

member state. A review of some cases shows that the starting Danes se vse več virov namenja zbiranju informacij o naravi.



based and the criteria for assessing the impacts of power lines on določenem območju. Kljub vse obsežnejšim bazam podatkov o these areas need verification, a change of thinking and possibly points on which nature conservation in Natura 2000 areas is Dopolnjujejo se seznami rastlinskih in živalskih vrst, ki bivajo na



an update of the legislation. Proposals are given to speed up the naravi, pa ne moremo pojasniti delovanja celotne narave. Zbrani podatki kažejo na spremembe v naravi, ki jih tudi ne installation of new power lines and the renovation of existing moremo z zanesljivostjo pojasniti. Elektroenergetski vodi ones, which will be necessary due to the electrification of many

activities and the inclusion of renewable sources. povzročajo določene spremembe v naravi. Ker pa so številne

vzročne povezave v naravi še vedno neznanka, je uveljavljen

Keywords konzervatorski pristop, ki ima za cilj, da se ohranja naravo

Natura 2000, power grid, nature conservation, nature protection nespremenjeno. To dokazuje že mednarodno uveljavljen izraz

________________________________________ »nature conservation« (ohranjanje narave), ki je cilj vseh

Permission to make digital or hard copies of part or all of this work for personal or predpisov in priporočil v zvezi z naravo [5]. Vsi sistemi varstva

classroom use is granted without fee provided that copies are not made or distributed narave tako danes temeljijo na ohranjanju narave, podobno kot

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must je to razumljivo v arheologiji, in to četudi se narava spreminja

be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).

DOI https://doi.org/10.70314/is.2024.env.16 Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia A. Kregar



tudi brez posredovanja človeka. Naravi je treba priznati ter podobnih habitatih v tujini. Pri določanju vpliva novega

upoštevati sposobnost spreminjanja in sorazmerno hitrega elektroenergetskega voda in omilitvenih ukrepov se tudi nikoli

prilagajanja človekovim posegom. ne uporablja rezultatov monitoringov izvedenih podobnih vodov,

Prevlada konzervatorskega pristopa se odraža v naslednjem temveč se za vsak vod izvaja nova celovita presoja in nova

primeru: če v Sloveniji elektroenergetski vod seka enega od presoja vplivov v celoti.

mnogih gozdnih območij, bo kljub napovedanemu razvoju

novega redkega habitata poseg ocenjen negativno, saj se delitvi

gozdnega habitata na dva dela pripisuje večja škoda, kot bi bila 6 Hitra prilagodljivost narave

korist novega nadomestnega habitata. V naravi je mogoče opaziti množico sprememb, kjer se živalstvo

pa tudi vegetacija spremenijo v nekaj desetletjih, kar bi lahko



3 Kriteriji vrednosti naravnega ravnovesja. trdili, da so to hitre spremembe, ki se zaključijo v obliki novega

Konzervatorski pristop se odraža tudi pri ocenjevanju vplivov Takšni značilni primeri so opuščene gramoznice (Slika 1), ki

elektroenergetskih vodov na naravo. Pri tem je izhodišče, da so se v nekaj desetletjih razvijejo v vredne habitate, ki se nato tudi

različni habitati v trajnem ravnovesju in da se razmerja površin naravovarstveno zaščitijo. Kljub očitnemu hitremu razvoju

med njimi ne spreminjajo. Vsak poseg, ki vpliva na razmerje habitata pa ni mogoče zasuti staro gramoznico in odpreti novo z

površin habitatov je tako ocenjen negativno, čeprav iz preteklosti napovedjo, da bo v nekaj desetletjih po opustitvi le-te nastal

vemo, da so se razmerja stalno spreminjala. Pri tem se podoben vreden habitat.

predvideva, da so redkejši habitati znotraj posamezne države



vrednejši.

Lestvice ocenjevanja vplivov na naravo imajo več stopenj za

negativne vplive in le eno stopnjo za pozitivne vplive [6]. Ocena

negativnih vplivov pa praviloma ni kvantificirana in temelji na

opisu, ki ga izdela strokovnjak s »konzervatorsko izobrazbo«.

Pretežna večina vplivov je ocenjenih bolj ali manj negativno.

Posledično ni mogoče najti presoje vplivov elektroenergetskega

voda na okolje, ki bi napovedal pozitivne učinke na naravo.



4 Ločitev vplivov elektroenergetskega voda in

podnebnih sprememb

Pri presoji vplivov elektroenergetskega voda na naravo se

uporabljajo najnovejši zbrani podatki s terena, ki upoštevajo tudi

vplive podnebnih sprememb in vremena v posameznem letu na

stanje habitata in na obseg populacije zavarovane živalske vrste. Slika 2: Gramoznica Dobrovnik pri Lendavi, ekološko

Na podlagi zbranih podatkov so določeni omilitveni ukrepi, ki ne pomembno območje in naravna vrednota (ID. 7422)

odpravljajo le neposrednega negativnega učinka

elektroenergetskega voda, ampak tudi posledice drugih O nečem podobnem smo bili priča tudi v prvi fazi širitve

dejavnikov. Takšen primer je zahteva po ohranitvi ptic kulturne COVID-19, ko je omejitev gibanja prebivalstva zunaj stanovanj

krajine, ki so v resnici območja ekstenzivnega kmetijstva nastala privabila mnoge divje živali v naselja. Tudi sicer se živali učijo

pred nekaj stoletji, ki so se jim te ptice prilagodile. Ta habitat pa in iščejo hrano po deponijah odpadkov in smetnjakih, kar vse

nadomeščajo območja intenzivnega kmetijstva, ki pa tem vrstam dokazuje njihovo prilagodljivost človekovemu okolju.

ne ustrezajo [7]. Tega ni mogoče omiliti z omilitvenimi ukrepi

drugih projektov. Presoja ne ločuje neposrednih vplivov

elektroenergetskega voda od posledic podnebnih sprememb. Z 7 Natura 2000

lokalnimi ukrepi je mogoče se izogniti ali omiliti lokalne vplive, Natura 2000, kot najobsežnejši sistem varstva narave v državah

ni pa mogoče lokalno odpraviti vlive podnebnih sprememb, ki članicah EU [9, 10] temelji na več zanimivih izhodiščih.

jih je treba obravnavati globalno. Kljub temu, da EU določa množico ureditev, ki veljajo

enotno za celotno območje EU, pa so območja Natura 2000



5 Neponovljivost elektroenergetskega voda in naravovarstvena vrednost določa glede na redkost tega habitata določena za vsako državo članico posebej. Tako se

neponovljivost habitatov ali zavarovane živalske vrste znotraj države članice. Tako je v

Pri presoji vplivov elektroenergetskega voda na naravo je eni državi ob mejni reki zaradi ptic zavarovano območje, v

izhodišče, da ima vsak vod drugačne vplive na habitate [8] in da sosednji državi na drugi strani iste reke pa te ptice niso

ni podobnih habitatov. Kljub obstoju mednarodne klasifikacije zavarovane.

habitatnih tipov [9], po katerih so razdeljene vse površine v Druga posebnost je določitev vpliva elektroenergetskega

Sloveniji, se presoja vsakega voda izvaja posebej brez uporabe voda na posamezno zavarovano živalsko vrsto v obliki odstotka

dokazanih vplivov podobnega voda na podobne habitate. Do od populacije v posamezni državi članici EU. Slika 2 prikazuje

sedaj tudi niso bile uporabljene primerjave s podobnimi vodi v lokacije treh podobnih elektroenergetskih vodov (vsak

Natura 2000 in ohranjanje narave, spodbuda ali ovira pri razvoju

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

elektroenergetskega omrežja



elektroenergetskega vod z oznako PCI ima površino 1 km 2) glede zanemari vpliv pogojev preživetja na drugih kontinentih ali na

na območje zavarovane živalske vrste. Hipotetični primer: njihovi vsakoletni selitveni poti.

površina države X = 100 km2 2 2 , Y = 10 km , Z = 10 km; pri tem

sta državi X in Y na osrednjem območju visoke gostote

zavarovane populacije (1000/km2 8 Koristi in škode pri elektroenergetskem vodu ) in država Z na obrobju

populacije z nizko gostoto (100/km2). Metoda za določanje vplivov elektroenergetskih vodov na

naravo je poponoma enaka kot metoda za določanje vplivov za



katerikoli proizvodni ali drugačen projekt. Vendar pa bi morali

elektroenergetske vode obravnavati celostno in drugače od

projektov, ki niso infrastrukturni. Pri odločanju o stroških

omilitvenih ukrepov za vod bi morali preveriti ali so ti stroški

upravičeni. Upoštevati moramo, da bodo stroški omilitvenih

ukrepov bremenili gospodarstvo, ki bo moralo ustvariti dodatno

dodano vrednost. Dodatna vrednost pa je na koncu uporabljena

za pokritje stroškov omilitvenih ukrepov elektroenergetskega

voda. Pri tem se vedno zanemari, da vsaka gospodarska

dejavnost povzroča negativne vplive na okolje. Zato je

pomembno, koliko gospodarstvo povprečno obremenjuje okolje.

Drugače povedano, »umazana industrija« bo ustvarila bistveno

večjo škodo v naravi za ustvarjanje dodatne dodane vrednosti,

kot pa bo korist za naravo zaradi izvedbe omilitvenih ukrepov za

elektroenergetski vod.



Slika 2: Lokacije treh podobnih vodov glede na območje 9 Strateška presoja

zavarovane živalske vrste V Sloveniji že desetletja izvajamo celovite presoje vplivov

planov na okolje. Pri tem se ne izvaja strateška okoljska presoja

Tabela 1 prikazuje delež prizadetih osebkov v odstotkih

plana ampak celovita presoja vplivov projektov, ki so vključeni

glede na populacijo v posamezni državi. Enak elektroenergetski

v plan, na okolje. Zato se običajno zahteva množica podrobnosti,

vod v osrednjem območju populacije prizadene enako število

ki bi jih morali preverjati pri presoji vplivov posameznega

osebkov, a je v državi X prizadet le 1 % populacije, ker je država

projekta na okolje (»Environmental Impact Assessment«,

10x večja od države Y. Posledično bo vpliv v državi X ocenjen

skrajšano EIA [11]). Šele zadnji predlog spremembe Zakona o

kot zanemarljiv, medtem ko bo vpliv v državi Y kljub enakemu

varstvu okolja uvaja neposredni prevod mednarodno

številu prizadetih osebkov ocenjen kot velik. V državi Z bo

uveljavljenega pojma »Strategic Environmental Assessment«

prizadetih 10x manjše število osebkov kot v državah X ali Y in

(skrajšano SEA [12]). Sprememba poimenovanja postopka v

bo vpliv ocenjen kot velik, ker bo prizadeto 10 % populacije v

»strateško okoljsko presojo« pa ne bo pospešila umeščanja

tej državi. Država Z bo tudi težje zagotavljala ugodne življenjske

elektroenergetskih vodov, če se bo izvajala kot do sedaj celovite

pogoje, ker so ti pogoji že sedaj očitno manj ugodni, ker je

presoje.

gostota populacije mnogo manjša. Ni izključeno, da bodo drugi

Naša posebnost je tudi, da izvajamo celovito presojo vplivov

negativni vplivi na populacijo v državi Z pripisani

elektroenergetskega voda na okolje, tudi če je pri projektu

elektroenergetskega vodu, pri katerem bo zato treba izvesti

obravnavanih več variant poteka. V številnih evropskih državah

obsežne omilitvene ukrepe, čeprav je v tej državi prizadeto

variante obravnavajo pri presoji vplivov voda na okolje (po EIA

najmanjše število osebkov zavarovane vrste.

postopku).

Značilen primer neustrezne strateške presoje pri nas je

Tabela 1: delež prizadetih osebkov v odstotkih presoja celotnega razvoja vodotokov. Ker ni izvedena strateška

okoljska presoja vseh vodotokov v Sloveniji, ki bi razvrstila vse



Država Površina vodotoke od naravovarstveno najvrednejših do najmanj Gostota Prizadeta Prizadet

(km2 naravovarstveno vrednih, tudi ni mogoče izvesti kakršnegakoli ) populacije populacija % v

(št./km 2 razvojnega projekta. Z razvrstitvijo po naravovarstveni vrednosti ) (št.) državi

X bi lahko zaščitili najvrednejše vodotoke in ostale namenili oskrbi 100 1000 1000 1

Y s pitno vodo, namakanju, proizvodnji električne energije, 10 1000 1000 10

Z turizmu, itd. Sedaj so zavarovana porečja Soče in Mure. Drava je 10 100 100 10

izkoriščena le energetsko, brez vizije za namakanje ali druge



vpliva elektroenergetskega voda na ptice selivke v obliki izgube Poseben primer ocenjevanja vpliva predstavlja ocenjevanje razvojne potrebe. Sava bo naravovarstveno preučevana le na odsekih, kjer so predlagane nove hidroelektrarne in tako dolgo, da bo tam najdenih dovolj zavarovanih živalskih vrst in vrednih procentov populacije, ker se uporabljajo podatki o populaciji na habitatov, da je ne bo mogoče uporabiti za katerokoli razvojno območju načrtovanega elektroenergetskega voda, pri čemer se funkcijo, razen, če bo uspešno izveden postopek prevlade druge

javne koristi nad koristjo narave. Primer neuspešno izvedenega



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia A. Kregar



postopka prevlade lahko opazujemo pri hidroelektrarni Mokrice projektov in nato vsakega posebej obravnavati s predhodnimi

in pripadajočih protipoplavnih in drugih razvojnih ureditvah. zahtevami po podrobnih podatkih kot se izvede presoja vplivov

Enako neustrezen pristop je bil v preteklosti izveden pri projektov na okolje (po EIA postopku).

presojah vplivov polj vetrnih elektrarn na ptice. Vedno se je Pri umeščanju in izvajanju okoljskih presoj

namreč preučevalo zgolj izbrano lokacijo in jo tudi po elektroenergetskih prenosnih vodov in druge infrastrukture

podrobnem preučevanju označilo kot naravovarstveno državnega pomena je nujno spremeniti tudi način usklajevanja

neprimerno, namesto, da bi se izvedla strateška okoljska presoja interesov v prostoru. Nov predpis Združenih držav Amerike [13]

za polja vetrnih elektrarn na območju celotne Slovenije. Da v dokazuje, da je nujno pospešiti postopke umeščanja tovrstne

Sloveniji ob enakih predpisih EU do sedaj nimamo niti enega infrastrukture v prostor in določa, kako to izvesti. Zakon o

delujočega polja vetrnih elektrarn, je dokaz, da ne izvajamo urejanju prostora omogoča učinkovitejše umeščanje z

pravilno strateških okoljskih presoj oziroma celovitih okoljskih vzpostavitvijo projektnih skupin, ki bi dejavno usklajevale

presoj in presoj vplivov projektov na okolje. interese v prostoru in sprejemale odločitve. Ker projektne

skupine niso prevzele odločevalske vloge oz. so to vlogo ohranili

formalni nosilci urejanja prostora, zakonska ureditev ni zaživela

10 Predlogi v praksi. Namesto ukinjanja projektnih skupin bi jih bilo

Najtežja bo sprememba gledanja na ohranjanje narave od ustrezneje nadomestiti s po eno projektno skupino za vsako vrsto

konzervatorskega pristopa nespremenljive narave k naravi kot infrastrukture, torej za elektroenergetske vode ena projektna

dinamičnem sistemu v katerem naj človek povzroča čim manjše skupina. Ta bi morala vključevati kompetentne člane, ki so

vplive. Spremembe gledanja na naravo pa ne bo mogoče doseči, sposobni sprejemati odločitve, ki bi nadomestile izdajo mnenj

če do teh spoznanj ne bo prišel izobraževalni kader, nato pa še z nosilcev urejanja prostora.

vzgojo nove generacije strokovnjakov za to področje. Največji

izziv bo razvoj metod za napovedovanje vplivov človeka na

naravo, kot najbolj zapletenega sistema soodvisnih spremenljivk. 11 Zaključek

Spremembo ocenjevanja vplivov elektroenergetskih vodov Postavitev množice obnovljivih virov v zadnjih letih, še posebej

na naravo bodo lahko izvedli le kadri, ki bodo sposobni videti v drugih državah članicah EU, ob enaki evropski zakonodaji,

tako pozitivne kot negativne učinke človekovih posegov v dokazuje, da je v Sloveniji nujno spremeniti način izvajanja

naravo. To pa bo mogoče šele takrat, ko bodo vplivi strateških okoljskih presoj planov in presoje vplivov projektov

kvantificirani na podlagi metod za napovedovanje vplivov, na okolje. Spremembe so nujne pri uresničevanju

potrjenih z monitoringi pri delujočih podobnih elektroenergetskih projektov, kot tudi pri vseh ostalih vrstah

elektroenergetskih vodih v podobnih habitatih. Čim prej je treba infrastrukture.

v presojo vplivov vključiti monitoringe ravnotežnega stanja Upravljavci infrastruktur podpiramo izogibanje ali vsaj

narave vodov izvedenih pred več desetletji in na njih utemeljiti omilitev vplivov načrtovanih projektov na naravo, če presoje

predvidene učinke omilitvenih ukrepov. Pri tem pa je še posebej temeljijo na preverljivih, količinsko določenih podatkih in so

treba upoštevati hitro prilagodljivost živali pri napovedovanju izvedene v razumnem času. Takšen pristop pomeni hiter razvoj

vplivov elektroenergetskih vodov na naravo. projekta in njegovo uresničitev vključno z izvedbo omilitvenih

Nujen bo tudi razvoj metod napovedovanja vplivov človeka ukrepov, kar pomeni tudi prispevek k vzpostavljanju

na naravo do te mere, da bo mogoče napovedati ločeno vplive kakovostnih habitatov.

posameznega elektroenergetskega voda in vplive zaradi V Sloveniji moramo kljub visokemu zavedanju o kakovosti

podnebnih in drugih sprememb. narave doseči nacionalni konsenz, da je nujno hitro sprejeti

Metode za kvantificirane napovedi vplivov odločitve in izvesti mnoge razvojni projekte, če želimo

elektroenergetskih vodov na naravo bo mnogo lažje razviti, če se zagotoviti vire za preživetje in ohraniti kakovost življenja

bo odstopilo od pristopa obravnave vplivov vsakega prebivalstva.

elektroenergetskega voda posebej in se bo vzpostavila »banka«

podatkov napovedanih vplivov in monitoringov dejanskih Literatura

vplivov podobnih vodov v podobnih habitatih na nivoju EU ali [1] ELES. 2023. Razvojni načrt prenosnega sistema Republike Slovenije za



celo globalno. obdobje 2023-2032, Ljubljana, (april, 2023), 55. DOI: https://www.eles.si/Portals/0/Documents/ELES_razvojni_nacrt_2023- Spremeniti bo treba tudi sistem vzpostavitve Natura 2000 2032.pdf .

območij v EU tako, da se pomembnost območij posameznih [2] ENTSO-E. 2022. System Needs Study, Opportunities for a more efficient

habitatnih tipov in populacij zavarovanih vrst določi glede na European power system in 2030 and 2040 – version for public

consultation, (julij 2022), 29. DOI:

njihovo redkost na nivoju celotne EU neglede na državne meje. https://eepublicdownloads.blob.core.windows.net/public-cdn-

Temu ustrezno bo treba vplive elektroenergetskih vodov na container/tyndp-documents/TYNDP2022/public/system-needs-report.pdf

naravo ocenjevati glede na celotno EU. [3] Mario Draghi. 2024. The future of European competitiveness Part A | A

competitiveness strategy for Europe, (september 2024), 41.

Pri ugotavljanju koristi vloženih sredstev za omilitvene DOI:https://commission.europa.eu/document/download/97e481fd-2dc3-

ukrepe infrastrukturnih projektov bi bilo treba preveriti 412d-be4c-

f152a8232961_en?filename=The%20future%20of%20European%20com

povprečno škodo na naravi zaradi delovanja gospodarstva na petitiveness%20_%20A%20competitiveness%20strategy%20for%20Eur

območju, ki je obremenjeno s stroški te infrastrukture. ope.pdf.

[4] Gene Wolf. 2024. Is It Innovation or Common Sense or Both?, T&D

V Sloveniji pri enakih direktivah EU ne uspemo umestiti World, (september 2024). 12-13. DOI:



elektroenergetskih vodov in mnogih drugih razvojnih projektov. https://issuu.com/docs/edd1f6b1151099893d523f9cc4efd281?fr=xKAE9 _zU1NQ&oly_enc_id=3489J1011234J1H . Zato je treba nujno spremeniti način izvajanja strateških [5] IUCN. 2024. Biodiversity, About biodiversity, (september 2024). DOI:

okoljskih presoj planov in opustiti sestavljanje planov iz https://iucn.org/our-work/biodiversity.



Natura 2000 in ohranjanje narave, spodbuda ali ovira pri razvoju

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

elektroenergetskega omrežja



[6] Rudi Vončina, et al. 2017. Okoljsko poročilo za DPN za prehod 220kV lex.europa.eu/LexUriServ/LexUriServ.do?uri=CELEX:31979L0409:SL:

omrežja na 400 kV Beričevo HTML. – Divača, (julij 2017), 27.

[11] Direktiva 2011/92/EU Evropskega parlamenta in Sveta z dne 13.

[7] [NB01] Velikost populacij izbranih vrst ptic, Komentar. (september

decembra 2011 o presoji vplivov nekaterih javnih in zasebnih projektov

2024). DOI: https://kazalci.arso.gov.si/sl/content/velikost-populacij-

na okolje (september 2024). DOI: https://eur-lex.europa.eu/legal-

izbranih-vrst-ptic-0.

content/SL/TXT/PDF/?uri=CELEX:32011L0092.

[8] Environmental Impact Assessment of Projects Guidance on Screening,

[12] Direktiva 2001/42/ES Evropskega parlamenta in Sveta z dne 27. junija

(Directive 2011/92/EU as amended by 2014/52/EU), (2017), 30. DOI:

2001 o presoji vplivov nekaterih načrtov in programov na okolje,

https://kliimaministeerium.ee/sites/default/files/documents/2021-

(september 2024). DOI: https://eur-lex.europa.eu/legal-

07/KMH%20eelhinnangu%20andmine.pdf.

content/SL/TXT/PDF/?uri=CELEX:32001L0042.

[9] Direktiva Sveta 92/43/EGS z dne 21. maja 1992 o ohranjanju naravnih

[13] United States of America, Department of Energy, Federal Energy

habitatov ter prosto živečih živalskih in rastlinskih vrst (EGT L 206,

Regulatory Commission. Building for the Future Through Electric

22.7.1992, s. 7). (september 2024). DOI: https://eur-

Regional Transmission Planning and Cost Allocation., Federal Registry,

lex.europa.eu/LexUriServ/LexUriServ.do?uri=CONSLEG:1992L0043:2

(maj 2024). DOI:

[10] https://www.federalregister.gov/documents/2024/06/11/2024-0070101:SL:PDF.

Direktiva Sveta z dne 2. aprila 1979 o ohranjanju prosto živečih ptic,

10872/building-for-the-future-through-electric-regional-transmission-

(79/409/EGS). (september 2024) DOI: https://eur-

planning-and-cost-allocation.



Sprejemljive in „nesprejemljive“ metode žlahtnjenja rastlin kot del boja za racionalno

odločanje v okoljevarstvu



Borut Bohanec, zaslužni profesor

Oddelek za agronomijo,

Biotehniška fakulteta,

Univerza v Ljubljani.

Jamnikarjeva 101,

1000 Ljubljana, Slovenija

borut.bohanec@bf.uni-lj.si



Izvleček



križanj in sicer in vitro reševanje nezrelih embrijev in

Žlahtnjenje rastlin se je na znanstvenih osnovah razvijalo

od pričetka 20. stoletja in močno vplivalo na prehrano vse fuzije protoplastov. Številne sorte so poliploidne. Nekatere

so take že po naravni poti, druge pa smo ustvarili ljudje.

večjega števila ljudi. Kratko opisujemo obstoječe metode

Poliploidi (pšenica, krompir, hmelj, jagode itd.) so

in njihovo prepoznavnost med prebivalstvom. Žal se

običajno robustnejši lahko pa tudi brezsemenski (banane,

najsodobnejši del teh metod – genski inženiring torej



genskih preurejanj vse prevečkrat na naši celini ocenjujejo Novejša uveljavljena metoda je formiranje F1 hibridnega semena. Gre za izkoriščanje heteroze (hibridne metode genskih transformacij in v zadnjem času tudi lubenice).

z negativnega vidika. Težko je razumeti tako odklonilen

odnos, če vemo kaj vse drugim delom sveta prinaša njihov bujnosti), ki se pojavi pri nekaterih križancih čistih linij v

njihovi prvi filialni generaciji. V naše kraje se je zlasti pri

sprejem. koruzi razširila po drugi svetovni vojni. Del prebivalstva

Abstract se je novosti bal, danes pa je hibridnost številnih vrst pogoj

za donosno pridelavo. V pomoč pri žlahtnjenju hibridnih

Plant breeding has been developed on a scientific basis sort je več biotehnoloških postopkov. S haploidi skrajšamo

since the early 20th century and has a major impact on the tvorbo čistih linij za mnogo let, z genskimi markerji pa si

diets of a growing number of people. We briefly describe pomagamo pri odbiri. Eno od takih metod smo pri zelju

the existing methods and their visibility among the razvili tudi pri nas (Rudolf Pilih in sod. 2019).

population. Unfortunately, the most modern part of the Mutageneza je tehnika 60ih let in je rešila mnoge

methods - genetic engineering, i.e. methods of genetic težave z nedostopnostjo drugih virov genske raznolikosti.

transformation and, more recently, genetic editing - are all Njen ekvivalent je najsodobnejša oblika biotehnološkega

too often viewed negatively on our continent. It is difficult žlahtnjenja – gensko preurejanje, ločita pa se po nekaj

to understand such a dismissive attitude, knowing all that pomembnih značilnostih. Izzvana mutageneza se izvede z

their adoption brings to other parts of the world. naključnim “obstreljevanjem” delov rastlin ali semen z

ionizirajočim žarčenjem ali mutagenimi kemikalijami.

Ključne besede: Geni, ki mutirajo, so naključno mutirani in zvečine

recesivni. Le velikemu obsegu dela lahko pripišemo vsaj

Križanja, Genske transformacije, In vitro embriji oz. delen uspeh, denimo odstranitev določene nezaželene

protoplasti, Izzvane mutacije, Gensko preurejanje lastnosti recimo erucične kisline v oljni ogrščici. Sodobna

metoda genskega preurejanja pa cilja točno določen gen,

lahko tudi istočasno v večih različicah (alelnih oblikah), ne



Metode žlahtnjenja pa tudi preostalih genov. Mutageneza se smatra kot

neproblematična metoda, o primernosti genomskega

preurejanja in njegovi regulativi, pa se v EU sučejo

Med starejše metode žlahtnjenja sodi uporaba znotraj pogajanja že 15 let. V tem času so bili drugod nekateri

vrstnega in medvrstnega križanja. Križanje omogoči rezultati že sproščeni (denimo GABA paradižnik na

premeščanje genov med sortami ali vrstami rastlin in s tem Japonskem), nekateri pa so najavljeni v kratkem (denimo

kopičenje za pridelavo ugodnih lastnosti med potomci. brezglutenska pšenica). Evropski parlament prejšnje

Medvrstna križanja so zahtevnejša in lahko rezultirajo z sestave je gensko preurejanje sprejel, naloga sedanjega pa

novo vrsto (tritikala) ali z novimi lastnostmi izhodiščne je izvedba zakonodaje.

vrste (paradižnik, itd.). Pogost vir genov so “divji



sorodniki”, če so dostopni. Denimo že več let je znana Spornost genskih transformacij

huda odvisnost ozkega nabora genov pri bananah, ki jih



zato težko branimo pred boleznimi. V Sloveniji je tak Gibanja aktivistov že tretje desetletje ovirajo aplikacijo

primer hmelj, katerega genska baza je ozka, uspešno pa je



bil rešen krompir, odpornost na virus je prišla iz sorodne uspešni primeri uporabe genskih transformacij ter sodobnih genskih tehnologij. V predavanju bodo zajeti

vrste. Dve biotehnološki metodi pripomoreta k uspehu avtorjeve osebne izkušnje pri sproščanju. Poudarjeno bo

tudi, da kljub pomanjkanju argumentov številne

aktivistične organizacije sprožajo kampanje tudi proti

najbolj nujno potrebnim sortam, denimo pravkar ponovno

na Filipinih proti zlatemu rižu (viru vitamina A) in to ne

glede na strahotne posledice, ki se merijo v sto tisočih

oslepelih otrocih v azijskih državah.



Vir



Rudolf Pilih K. et al (2019) Proposal of a new hybrid

breeding method based on genotyping, inter-pollination,

phenotyping and paternity testing of selected elite F1

hybrids. Frontiers in Plant Science. 2019, art. 1111, vol.

10



Ni fosilnih goriv – ni hrane



No fossil fuels - no food



Andrej Senegačnik†

Laboratorij za termoenergetiko

Fakulteta za strojništvo

Univerza v Ljubljani

Slovenija

andrej.senegacnik@fs.uni-lj.si



POVZETEK na to temo, večina držav ima tudi že strateške načrte kako to

izvesti, npr. Slovenija v Nacionalnem energetsko podnebnem

V prispevku je obravnavana poraba energije v sektorju načrtu (NEPN) [1]. Realni rezultati 20 letnega razogličenja so

kmetijstva in gozdarstva. Trenutna sektorska poraba končne ravno nasprotni. Emisije toplogrednih plinov se globalno

energije je na nivoju 70.000 ton dieselskega goriva letno. Glede povečujejo. V nekaterih državah, ki brezglavo politično forsirajo

na evropski "Green Deal", ki predvideva razogličenje in spremenljive obnovljive vire (VRE ang. Variable Renewable

opustitev fosilnih goriv do leta 2050, so raziskane možnosti Energy), kot so npr. fotonapetostne in vetrne elektrarne za

nadomestitve z obnovljivimi viri. Modelska predpostavka je, da proizvodnjo električne energije, se je cena le te v zadnjih dveh

se samooskrba s hrano poveča iz sedanjih 40 % na 80 %, letih izjemno zvišala, posledično se je pa konkurenčnost teh

hektarski donos pa se zmanjša zaradi zmanjšanja uporabe gospodarstev zmanjšala. Eklatanten primer je Nemčija in njene

fitofarmacevtskih sredstev. Obravnavanih je več alternativ gospodarske težave zaradi nekonkurenčnosti. Realno gledano, je

nadomestitve, s fizičnim delom ljudi in konj, biodieslom, pa vključevanje VRE virov v neko stabilno elektroenergetsko



izvedbo katerekoli alternative so realno zelo majhne. Ker je morejo obratovati in potrebujejo stabilnega gostitelja (klasične hrana nujno potrebna za naš obstoj bomo do odkritja kakih novih termoelektrarne na premog, plinske, jedrske, hidro). VRE viri električno energijo in sintetičnimi gorivi. Možnosti za dejansko oskrbo popolno tehnološko parazitiranje, saj VRE viri sami ne

tehnologij ostali na uporabi fosilnih goriv. torej lahko obratujejo izključno na račun drugega.



KEYWORDS / KLJUČNE BESEDE pomeni tudi povečevanje porabe energije je v nasprotju z drugim Zmanjševanje GHG ob sočasni gospodarski rasti, ki istočasno

razogličenje, fosilna goriva, pridelovanje hrane, glavnim zakonom termodinamike (II. GZT). Trenutno žal ne

znamo (in tudi nikoli ne bomo), kako zaobiti II. GZT. Politične

ABSTRACT direktive in zeleni aktivizem pa na fizikalne zakone ne deluje –

The paper deals with energy consumption in the agricultural and še sreča! Navidezno si politika pomaga z zelo ozko izbranimi

forestry sector. The current sectoral consumption of final energy mejami sistema. Danska npr. vse energijsko intenzivne

is 70,000 tons of diesel fuel per year. In line with the European polizdelke uvaža iz Azije, lesne pelete za ogrevanje Köbenhavna

"Green Deal", which envisages the abandonment of fossil fuels pa iz Brazilije in Kanade, itd., itd,… Npr. na Danskem je poraba

by 2050, the possibilities of replacing them with renewable jekla na prebivalca večja kot v Sloveniji, toda ker nimajo nobene

energy sources are examined. The model assumes that the degree jeklarne in druge energijsko intenzivne industrije… je njihova

of self-sufficiency in food production will increase from the energijska učinkovitost merjena kot razmerje med porabo

current 40% to 80% and that the yield per hectare will decrease energije in dodano vrednostjo bistveno manjša kot v Sloveniji in

due to the reduced use of pesticides. Several alternatives are celo najnižja v Evropi?? Tako navidezno zmanjševanje GHG na

being discussed, including manual labor by humans and horses, račun nekoga drugega je zelo prikladen politični pamflet, ko

biodiesel, electricity and synthetic fuels. Realistically, the prepričuješ ljudske množice o uspešnosti izvedenih "zelenih"

chances of actually implementing all alternatives are very slim. ukrepov in kako s tem rešuješ svet. Zelo podobno je tudi z

Since food is absolutely necessary for our existence, we will have električnimi avtomobili. Njihova emisija GHG je 0 gramov CO2

to rely on the use of fossil fuels until new technologies are na prevoženi kilometer? Emisija električne energija za polnjenje

discovered. baterije pa najbrž ne?

Ambiciozni evropski podnebni načrt "Green Deal" [2] je

KEYWORDS opustitev fosilnih goriv in prehod na obnovljive vire energije

decarbonization, fossil fuels, food production (OVE) do 2050. Glavna ideja razogljičenja je, da se bo vsa

fosilna in jedrska energija nadomestila z VRE električno energijo,

tam kjer pa prehod na električno energijo ni možen ali smiseln,

1 UVOD bo izvedena zamenjava z vodikom in sintetičnimi gorivi, ki se

Zadnje čase se globalno ukvarjamo kako zmanjšati bodo pridobivali iz VRE električne energije. Ideja je navidezno

kakršnekoli ogljične odtise v obliki antropogenih izpustov dobra in izvedljiva, dokler govorimo o lekarniškem in

toplogrednih plinov (GHG ang. Greenhouse Gases). V laboratorijskem nivoju izvedbe prehoda in zelo ozko postavljeno

preteklosti je bilo že veliko konferenc in strateških posvetovanj bilančno mejo sistema. Ko pa se bilančne meje razširijo in



preidemo na nivo držav ali globalni nivo, pa ugotovimo, da II. bistveno več mesa, ki porablja za pridelavo bistveno več energije

GZT še vedno velja. Realno se pojavi zelo veliko nepremostljivih (npr. najprej pridelavo koruze za živinsko krmo). Učinkovitost

fizikalnih preprek. Trenutni zeleni prehod lahko označimo kot se je povečala predvsem zaradi umetnih gnojil in

obupne poizkuse, da bi iz starih tehnologij dobili povsem nove fitofarmacevtskih sredstev.

rezultate (Einstein), po možnosti seveda takšne, ki so v nasprotju Ob predpostavki, da je izkoristek pretvorbe dieselskega

z II: GZT? Naj omenimo, da je bila vodikova gorivna celica goriva v mehansko delo 25 %, iz 200.000 t dobimo 600.000

odkrita že leta 1838, pa se do danes še ni uveljavila? Kje je MWh mehanskega dela (iz 1 kg goriva ~3 kWh mehanskega

vzrok? dela). Če to preračunamo v delovne dni, kjer se upošteva, da

Prispevek obravnava možnosti pridelave hrane v Sloveniji na človek lahko dnevno odda do 0,6 kWh [4] dobimo kar eno

tradicionalni in povsem OVE način s človeško in živalsko 9 milijardo (10) delovnih dni. Nadalje lahko predpostavimo, da se

delovno silo. Težko si namreč zamislimo, da bi polja obdelovali je v preteklosti s pridelavo hrane ukvarjalo več kot 90 %

z električnimi obdelovalnimi stroji, zaradi pogostih prekinitev prebivalstva, kar bi za trenutne Slovenske razmere pomenilo 1,8

dela zaradi več urnega polnjenja baterij. Delovna obremenitev milijona ljudi. Za 600 GWh mehanskega dela, bi moralo vseh 1,8

kmetijskih strojev je namreč trajno bistveno večja (večino časa milijona ljudi težko fizično delati kar 555 dni, torej več kot eno

vsaj na 80 % nazivne moči), kot pa električnih osebnih vozil. leto. Dobljena ocena potrebnih delovnih dni, se načeloma dobro

Tudi potrebne moči so bistveno večje npr. sekalnik lesnih ujema z dejstvom, da so nekoč ljudje večinoma delali samo za

sekancev, manjši kombajn; potrebuje moč ~500 kW. Ker bi hrano za preživetje. Pri tem so si pomagali tudi vprežnimi

uporaba električnih kmetijskih strojev kmetovalcem najbrž živalmi. Zelo groba ocena med oddanim delom konja in človeka

bistveno otežila obdelovanje polj, je zato najbrž bolj smiselna je ~10:1, pri čemer je potrebno upoštevati, da konj v naših

uporaba sintetičnih goriv. Pri teh pa je vložek energije za njihovo klimatskih pogojih poje četrtino "svojega" pridelka..

pridobivanje kar 3 do 4 krat večji kot njihova kurilnost. Zato v

prispevku hipotetično obravnavamo kakšne so možnosti

pridelave hrane na tradicionalen povsem OVE način, brez Tabela 1: Primerjava moči delovnih živali in človeka [5]

sodobnih tehnologij.

vlečna sila, hitrost, moč, W moč v

N m/s HP

2 METODOLOGIJA konj 600 1,2 650 0,86

Iz podatkov o porabi dieselskega goriva v kmetijstvu, ki so vol 600 0,8 430 0,57

zajeti iz Statističnega urada Slovenije (SURS) [3] je izračunana mula 300 1,2 321 0,43

nadomestna količina mehanskega dela. To mehansko opravljeno človek 90 1,2 67 0,09

delo se nato porazdeli na obnovljiva vira, človeško delo in delo

živali. Pri tem je dodatno upoštevano povečanje samooskrbe s Prikazana groba ocena zamenjave fosilnih goriv z delom

hrano iz današnjih ~40 % na ~80 % in zmanjšana uporaba živali in ljudi v kmetijstvu pokaže, da je to možno (neizpodbiten

fitofarmacevtskih sredstev in zato bolj ekološka pridelava dokaz je tudi v preteklosti, saj so naši predniki preživeli brez

poljščin, zaradi česar se pridelek glede na današnji nivo dodatno fosilnih goriv), toda cena zamenjave nazaj na OVE je visoka.

zmanjša za 25 %. Očitno je, da bi v tem primeru morali skoraj vsi obdelovati polja

in se ukvarjati s kmetijstvom. Pri današnji stopnji urbanizacije,

ko večina ljudi živi v velikih mestih (ali več milijonskih

3 PORABA ENERGIJE V KMETIJSTVU V velemestih) je tak prehod nazaj popolnoma nemogoč zaradi

SLOVENIJI transporta delovne sile na polja. Logična posledica je torej, da se

človeštvo ne more odreči trenutni kmetijski mehanizaciji na



energije v kmetijstvu ~75.000 t dieselskega goriva, s katerim se Po podatkih SURS je letna povprečna poraba končne fosilna goriva. Prehod na strojno obdelavo polj je bil temeljni pogoj za industrializacijo in splošni napredek [4] in današnji visok standard – izobilje dobrin. Delež delovnih mest v državah 2 obdeluje 1750 km njiv, skupne kmetijske površine v uporabi so z visoko razvitim in mehaniziranim kmetijstvom je samo še 2 % 2 4750 km [3] Njivskih površin na prebivalca imamo najmanj v do 3 %. Z večanjem produktivnosti se obseg produkcije linearno 2 EU, samo 875 m , kar bi moral biti izredno močen alarm za povečuje, poraba energije pa eksponentno. Sila trenja se politiko, ker še vedno dovoljuje gradnjo, večinoma ne življenjsko povečuje navadno s kvadratom hitrosti (hitrejši proces pomeni pomembnih objektov (nakupovalnih središč, nasedlih tovarn – večjo produktivnost, ki pa je vedno povezana z večjimi npr. štajerska Magna) na prvovrstnih kmetijskih zemljiščih! Ob izgubami), moč za premagovanje trenja pa torej s tretjo potenco. zelo grobi predpostavki, da se pridelki iz kmetijstva podvojijo do Delež porabe energije v kmetijstvu in živilski industriji je leta 2050 ob 25 % manjšem izplenu, bi se poraba goriv povečala glede na celotno porabo kljub vsemu relativno majhen, toda na ~200.000 t dieselskega goriva. Ob tem bi se morale bistveno izjemno pomemben " conditio sine qua non " (lat.: brez česar ne povečati tudi obdelovalne površine. Gozdnatost Slovenije je bila gre). Kljub relativno majhni porabi, je prehod na OVE izjemno v 19 stoletju samo ~30 %, samooskrba s hrano je bila ~80 % za težaven ampak dokazljivo mogoč (seveda ob bistveno nižjem ~1 milijon prebivalcev. Če torej primerjamo takratno in sedanjo življenjskem standardu in manj prebivalcih). Če bi hoteli fosilni učinkovitost kmetijstva, je razvidno, da je sedanja pridelava diesel nadomestit z biodieslom, ki ima EROI 2, pri povprečnem bistveno bolj učinkovita (na manjši obdelovalni površini 40 % izplenu 3 t/ha [6] in [7], bi za 200.000 t biodiesla potrebovali še hrane za 2 milijona = 80 % za 1 milijon), ob tem, da je v prehrani dodatnih ~70.000 ha (700 km 2 ) obdelovalnih polj in ker je EROI

2, tudi še vložek energije za predelavo oljne ogrščice v biodiesel, sredstev, izdelave obdelovalnih strojev,… Zelo lep pregled agro-

ki je pa 50 % energije biodiesla. Zato bi se poraba biodiesla ekonomskih faktorjev je prikazan v [8]. V tabeli 2 je navedenih

morala še dodatno povečati (npr. za oranje in obdelavo dodatnih samo nekaj najbolj zanimivih vrednosti.

700 km2…). Kot vidimo, je nadomeščanje z biogorivi bolj ali

manj nesmiselno. Še enkrat boljši za nadomeščanje je konj, ki

ima EROI vsaj ~4-5. Tabela 2: Nekateri agro-ekonomski faktorji povzeti iz [8]

Poraba končne energije v sektorju kmetijstva in gozdarstva

predstavlja le ~1,5 % porabe končne energije v Sloveniji, slika 1 skupna poraba za kWh/(ha leto) 315-2060

in 2. Relativno zelo malo, toda zelo težko jo nadomestimo z obdelovanje zemlje

obnovljivimi viri, tudi z OVE električno energijo, ker gre produkcija kWh/kg 75-117

prvenstveno za mobilne delovne stroje velikih moči. pesticida/fungicida

produkcija pest./fung. kWh/(ha leto) 140-550

za škropljenje 1 ha



produkcija dušikovih kWh/kg 12-18

umetnih gnojil

produkcija fosfornih kWh/kg 1,4-9

umetnih gnojil

produkcija kalijevih kWh/kg 1,5-14

umetnih gnojil

količina dušikovih kg/ha 30-210

gnojil za 1 ha

Slika 1: Delež porabe končne energije za kmetijstvo in



gozdarstvo [3] 4 ZAKLJUČEK





V prispevku je raziskana možnost prehoda iz uporabe fosilnih

goriv na druge obnovljive vire energije za sektor kmetijstva in

gozdarstva. Osnovno izhodišče modela so statistični podatki o

deležu samooskrbe s hrano in poraba končne energije sektorja.

Rezultati zelo enostavnega modela pokažejo, da bi bilo

mehansko delo iz fosilnih goriv hipotetično možno nadomestiti

z delom ljudi in konjev, ob robnem pogoju, da bi morali skoraj

vsi prebivalci, preko 80 %, trdo delati na poljih samo za

preživetje. Predviden je tudi dvig samooskrbe s hrano iz

Slika 2: Končna poraba energije za kmetijstvo in današnjih 40 % na nekdanjih 80 %. Tovrstna zamenjava s

gozdarstvo [3] povečanjem fizičnega dela ljudi je za današnji čas popolnoma

nesprejemljiva, čeprav je v preteklosti, ko še niso imeli toplotnih

Poraba končne energije industrijsko predelovalnega sektorja strojev in kapljevitih fosilnih goriv, to bila edina in zelo

proizvodnje hrane in pijač v Sloveniji je skoraj enaka porabi v samoumevna realnost. Druga alternativa je uporaba baterijskih

kmetijstvu, to je ~70.000 toe, pri čemer pa je poraba razdeljena kmetijskih strojev. Žal tudi ta alternativa tehnološko še ni zrela.

na dva glavna energenta, to je električno energijo in energijo Ti stroji so relativno bistveno bolj obremenjeni (glede na nazivno

goriv, slika 3, predvsem zemeljskega plina in nekaj malega moč) in to v daljših časovnih intervalih (dnevno več ur), torej

ostalih goriv. bistveno bolj kot npr. baterijska osebna vozila. Sredi polj bi

morali postaviti zelo zelo hitre polnilnice? Torej je tudi ta

alternativa zelo malo verjetna. Tretja alternativa je "klasični"



konjski pogon. Ta je najbolj ekološki, deluje na lokalno pridelano

hrano, proizvaja naravno gnojilo, konji se razmnožujejo sami,…

Ta pogon ima torej vrsto prednosti. Edina slabost zakaj se ga ne

uporablja več je pa specifična moč, ki je ~1 W/kg mase konja.

Sodobni kmetijski stroji imajo nekaj stokrat večjo specifično

moč. Četrta alternativa je pridelava biodiesla in njegova uporaba

za obdelavo polj. Glede na to, da je EROI 2, to pomeni da je

potrebni vložek energije za pridelavo biodiesla enak njegovi

razpoložljivi energiji. Zato bi se poraba energije povečala ob



Slika 3: Končna poraba energije za živilsko predelovalno 2 sočasnem povečanju obdelovalnih površin, vsaj za 700 km, v industrijo v Sloveniji [3] realnosti pa najbrž še bistveno več. Kot peta alternativa ostanejo



"posredne" energije v obliki umetnih gnojil, fitofarmacevtskih Poleg te neposredne porabe je v kmetijstvu še veliko še vodik in sintetična goriva, npr. sintetični diesel, ki ima EROI 0,23 [9], kar pomeni, da je neto porabnik energije in je potrebni zunanji vložek energije za sintezo kar ~350 %? Kot je razvidno



je vseh pet "zelenih" alternativ v skladu z "Green Deal-om" [5] San José State University, Department of Economics, Comparison of

realno (ne)izvedljivih iz česar sledi zelo neprijeten zaključek: če Power Produced by Various Draft Animals and Humans

https://www.sjsu.edu/faculty/watkins/animalpower.htm

ne bomo uporabljali fosilnih goriv tudi hrane ne bo. [6] I. van Duren, et.al. Where to produce rapeseed biodiesel and why?

Mapping European rapeseed energy efficiency, Renewable Energy, 74,

49-59, 2015, http://dx.doi.org/10.1016/j.renene.2014.07.016

[7] M. T. Firrisa, et.al. Energy efficiency for rapeseed biodiesel production in



VIRI different farming systems, Energy Efficiency 7:79–95, 2014, DOI 10.1007/s12053-013-9201-2

[1] Vlada Republike Slovenije, Nacionalni energetsko podnebni načrt – [8] O. T. Arodudu et. al. Integrating agronomic factors into energy efficiency

NEPN, 2023, dostopno na https://www.energetika- assessment of agro-bioenergy production – A case study of ethanol and

portal.si/fileadmin/dokumenti/publikacije/nepn/dokumenti/nepn_posod_ biogas production from maize feedstock, Applied Energy 198, 426–439,

dec2023.pdf 2017, http://dx.doi.org/10.1016/j.apenergy.2017.02.017

[2] European Commission, European Green Deal, 2019, dostopno na: [9] S. Hänggi, et. al. A review of synthetic fuels for passenger vehicles.

https://ec.europa.eu/commission/presscorner/detail/en/ip_19_6691 Energy Reports, 5, p. 555–569, 2019

[3] Statistični urad Republike Slovenije, SURS, podatkovne baze SiStat, https://doi.org/10.1016/j.egyr.2019.04.007

dostopno na: https://www.stat.si/statweb

[4] V. Smil, Energy and Civilization, MIT Press, Cambridge, 2017,

http://dx.doi.org/10.7551/mitpress/10752.001.0001



Nemški energetski prehod ter njegove politične in gospodarske



posledice



The German energy transition and its political and economic



consequences



prof.dr. Dušan Povh

Svetovalec

dusan.povh@t-online.de



POVZETEK V poznih osemdesetih letih prejšnjega stoletja so izpusti



V CO2 postali glavna tarča okoljevarstvenikov, ki so jih videli kot

prispevku na kratko opisujemo razvoj in položaj nemške glavni vzrok podnebnih sprememb. Leta 1988 je bil pod okriljem

Energiewende. Politiki so Energiewende načrtovali predvsem ZN ustanovljen IPCC Committee, ki od takrat igra vodilno vlogo

ideološko, ne da bi upoštevali možne posledice za življenjski v »boju« proti podnebnim spremembam. IPCC zbira raziskave in

standard ljudi, obstoj in razvoj gospodarstva ter socialno publikacije o podnebnih spremembah ter nato objavlja lastne

stabilnost države. Podane so posledice omenjenega energetskega zaključke.

prehoda, saj so te pomembne tudi za Evropo in države v njej, ki Podnebni problem je postal eno najpomembnejših vprašanj

imajo tesne gospodarske odnose z Nemčijo in lahko pričakujejo svetovne politike. Od leta 1992 je bilo organizirano veliko število

podobne probleme. letnih podnebnih konferenc. Najpomembnejša je bil leta 2015 v



ABSTRACT Parizu, kjer so bile sprejete splošne (vendar nezavezujoče) zaveze za zmanjšanje CO2. Nemčija je bila ves čas gonilna sila T he development and status of the German energy te politike, ne glede na to, da predstavlja le okoli 2 % skupnih

transition are briefly described. Politicians planned the energy emisij CO2. To politiko je nato v celoti sprejela tudi EU [1].

transition primarily ideologically, without considering the Drugi v svetu gledajo na ta problem bolj realno. Bolj kot

possible consequences for people's living standards, the podnebne spremembe jim je pomembna blaginja lastne države.

existence and development of the economy, and the social Nemčija se je zavezala k zmanjšanju izpustov CO2 v

stability of the country. The consequences resulting from the primerjavi z letom 1990 za 40 % do 2020, za 65 % do 2030 (tukaj

energy transition are described. These are particularly important je Nemčija zaostrila svoj cilj EU na 55 %), za 80 % do 2040 in

for Europe and the countries that have close economic relations do 2045 za dosego podnebne nevtralnosti (EU do leta 2050).

with Germany and can expect similar problems. Poleg nujne strukturne spremembe v proizvodnji električne

. energije in prestrukturiranja omrežja, to pomeni tudi dolgoročno

odločilne spremembe v industriji, pri toplotni izolaciji hiš,

KLJUČNE BESEDE kmetijstvu in prometu.

Energetska politika, energetski prehod, Energiewende, Nemčija Glavna usmeritev energetskega prehoda je bila gradnja

vetrnih turbin in fotovoltaike. Želeli so zapreti vse elektrarne na

KEYWORDS premog in pokriti vse potrebe z električno energijo z

Energy policy, Energy transition, Energiewende, Germany obnovljivimi viri. Ker pa to tehnično ni izvedljivo, bo treba

zgraditi veliko (40 do 50) dodatnih plinskih elektrarn. Nemčija

se je kot edina država EU tudi odločila zapreti vse delujoče

1 jedrske elektrarne, ki so desetletja zagotavljale poceni in varno ENERGIEWENDE



Zgodovina energetskega prehoda (Energiewende) v Od začetka energetskega prehoda do leta 2022 je bilo za električno energijo.

Nemčiji se je začela že veliko pred nastajajočim problemom subvencije porabljenih ogromno denarja. Toda največja,

podnebnih sprememb, domnevno zaradi izpustov CO2, ki jih ideološko vodena napaka je bila zaprtje vseh jedrskih elektrarn.

povzroča človek. Razlog je bilo gibanje proti jedrski energiji kot Študije [2] kažejo, da je bilo s tem izgubljenih 600 milijard EUR

posledica černobilske nesreče. brez večjega zmanjšanja CO2. Seveda pa ta znesek ne upošteva

morebitnih drugih dodatnih stroškov, povezanih z energetskim

prehodom. Seveda pa ti ogromni zneski manjkajo za druge

potrebne investicije in posebno na področjih razvoja, ki bi

Permission to make digital or hard copies of part or all of this work for personal or zagotovil konkurenčnost gospodarstva.

classroom use is granted without fee provided that copies are not made or distributed Izraz Energiewende se je razširil iz oskrbe z električno

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must energijo na vsa področja vsakdanjega življenja, kjer prihaja do

be honored. For all other uses, contact the owner/author(s). emisij CO2: industrija, promet, gradbeništvo in kmetijstvo.

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.env.20



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia D. Povh



Dodatni stroški Energiewende do leta 2035 naj bi znašali Posledice Energiewende za industrijo so alarmantne.

1,2 bilijona EUR. Primerljiv znesek se bo nato še potreben do Glavni razlog so visoki stroški energije (2 do 3 krat višji kot v

leta 2045, če politiki medtem ne bodo prisiljeni spremeniti USA ali na Kitajskem), ki nastanejo zaradi prisilnega opuščanja

časovnice in vsebine energetskega prehoda. premoga, plina in nafte. K temu je gotovo prispevala tudi ruska

Energiewende se nahaja že nekaj let v slepi ulici, čeprav vojna proti Ukrajini z izgubo poceni ruskega plina in trendi v

tega politiki (še) nočejo priznati. Večina projektov za zmanjšanje svetovnem gospodarstvu. Toda politiki bi se morali pravočasno

CO2 je modificiranih, ustavljenih ali preloženih. Jasno je tudi, da odzvati na te težave. Energetsko intenzivna industrija (jeklarska,

cilji za leto 2030 ne bodo doseženi. In kaj se bo zgodilo pozneje, kemična) v Nemčiji ni več konkurenčna. Tudi pomanjkljivosti

je politično vprašljivo. lokacij (okoljski ukrepi, preregulacija, birokracija) prispevajo k

temu, da se vse več podjetij izseljuje v države, ki ponujajo

cenejšo energijo in boljše lokacijske pogoje. Energetski prehod

2 POSLEDICE na vodik kot vir energije, ki naj bi omogočil zeleni prehod

Čeprav je še vedno vprašljivo, v kolikšni meri prispevek industrije, je zelo zelo drag in negospodaren. Trajalo bo še

človekovih emisij CO2 prispeva k podnebnim spremembam, mnogo let, preden bo sploh uporabljen v večjem obsegu.

politiki zmanjševanje CO2 razglašajo za brezalternativno in kot Denar, porabljen, za energetski prehod torej ni bil voljo za

vprašanje preživetja. In to brez predhodnega upoštevanja druge naloge, predvsem za potreben razvoj novih tehnologij.

tehničnih, ekonomskih in socioloških posledic. Energetski Glavni tekmeci v svetu (ZDA, Kitajska, Indija), pa tudi nekatere

prehod se je začel z obljubo, da ne bo stal skoraj nič (morda ceno druge države, so Nemčijo prehiteli na večini za bodočnost

ene kepice sladoleda na mesec), ponudil pa bo veliko novih ključnih področij. To velja seveda tudi za EU. V avstralski študiji

delovnih mest. Resničnost je popolnoma drugačna. [3] je med 44 razvojnimi področji, ki so ključna za prihodnost,

Cilj politike je bil neusmiljeno uveljavljanje ukrepov brez Nemčija navedena le na enem področju na 3 mestu, na nekaj

upoštevanja kakršnega koli nasprotnega mnenja in problemov, ki področjih na 4 ali 5 mestu. Pri veliki večini pa je ni med prvimi

so se pojavili. Primeri tega so: petimi. Tudi to so alarmantni znaki negotove gospodarske

prihodnosti Nemčije in EU.

 Stroški energije, elektrike in plina so se v nasprotju z

napovedmi politikov prekomerno podražili in močno  Nemčija je z vrha svetovne lestvice glede na

obremenili standard ljudi. Trenutno ponovno prihaja sicer BDP/prebivalca zdrsnila na 21. mesto. Prehitele so jo tudi

do znižanj cen plina, a cena električne energije (kar je bil manj razvite, a z viri bogate države.



cilj energetskega prehoda) dolgoročno ostaja visoka.  Nemčija je padla iz prve skupine na 24. mesto na svetovni

 Zakon o ogrevanju, ki naj bi dosegel prisilno zamenjavo lestvici po konkurenčnosti (kako privlačna je država za

ogrevalnih sistemov, ki bi povzročil enormne stroške za industrijo).



lastnike hiš in najemnike, je zaradi masovnih protestov  Nemško gospodarstvo že dve leti stagnira in je blizu propadel,

recesije in je pri tem na repu držav v EU. Izboljšanja v

 Uporaba toplotnih črpalk zastaja, potem ko se je izkazalo, naslednjih letih ni za pričakovati.



da povzročajo visoke stroške in ne prinašajo obljubljenih  Glede na študijo Zveze nemških industrij (BDI) je 20 %

ali pa le z zelo visokimi dodatnimi stroški. Zaradi nemške industrije v kritičnem stanju. koristi. Za večstanovanjske hiše rešitve sploh ne obstajajo

preobremenitev se pojavljajo težave v  Tuje investicije v Nemčiji so padle na minimum. Nemške

elektrodistribucijskih omrežjih. naložbe grejo v druge države, ki ponujajo boljše pogoje.



 Električni avtomobili niso splošno sprejeti. Stroški so  Številna, nekdanja vodilna industrijska podjetja odpuščajo

visoki (v primerjavi z avtomobili z motorji z notranjim delavce. Pričakujejo dvig brezposelnosti izgorevanjem), težave z njihovim dosegom in dostopom do

polnilnic ostajajo. Sledil je padec (40 %) prodaje teh  Izvoz (nekoč glavni steber nemške blaginje) je letos upadel

avtomobilov in posledično ekonomskie težave v za 15 %, tudi kot posledice nekonkurenčnosti.

avtomobilski industriji (npr. v VW nameravajo zapreti

30.000 delovnih mest). Ekonomske težave povzročajo tudi težave v družbi



 (socialne pomoči, pokojnine, migracije). Nezadovoljstvo ljudi,

Avtomobili so bistveni del življenjskega standarda in ljudje ne le zaradi energetske tranzicije, ampak tudi na drugih področjih CO2 emisije v prometnem sektorju še vedno naraščajo.

se jim nočejo odreči. V večini primerov tudi ni vsakdanjega življenja, vodi v nestabilne politične razmere.

enakovredne alternative s javnim prevozom za ljudi izven Razlike postajajo tako pereče, da ogrožajo demokracijo.

mest.



 Stroškovni problemi za toplotno izolacijo hiš so visoki in 3 SKLEP

bi bistveno povečali najemnine, predvsem za revnejše

sloje. Energiewende je v krizi in vidne so posledice za življenjski

standard ljudi, gospodarsko moč države in politične razmere. To

 Če povzamemo; ni dvoma, da je oz. bo zaradi energetskega velja tudi za EU in še posebej za države, ki so gospodarsko tesno



izrazite (negativne?) posledice tudi na političnem področju bi podnebno strategijo vsaj bistveno upočasnili in politične (vzpon skrajnežev). prioritete preusmerili v okrevanje gospodarstva in ohranjanje prehoda življenjski standard ljudi ogrožen. To pa ima povezane z Nemčijo. Konsolidacija razmer bi bila mogoča le, če

življenjskega standarda.





REFERENCE


[1] MIHALIČ, Rafael, POVH, dušan: German Energiewende Between

Ideology and Reality, Information Society, Oktober 2023, Ljubljana,

Slovenia .

[2] Jan EMBLEMSVÅG: What if Germany had invested in nuclear power? A

comparison between the German energy policy the last 20 years and an

alternative policy of investing in nuclear power, International Journal of

Sustainable Energy, Volume 43, 2024 - Issue 1

[3] Jamie GAIDO at al.: ASPI’ s Critical Technology Tracker (The global race

für future Power) Report No. 69/2023E



MITI IN RESNICE O ELEKTROMAGNETNEM SEVANJU /EMS/



Karel Lipič, univ.dipl.ing.

Zveza ekoloških gibanj Slovenije

– ZEG

Krško, Slovenija

zegslo20@gmail.com



Povzetek Zagovorniki varnosti brezžične tehnologije se sklicujejo na

zakonsko dovoljene meje (strokovno zastarelo in proti ustavno

V Sloveniji »lobi« mobilnih operaterjev in elektro podjetji v Uredbo o elektromagnetnem sevanju iz leta 1996 ) , ki naj bi

zadnjih 30-ih letih ni želel nobenih pozitivnih sprememb na zagotavljale varnost pred negativnimi učinki sevanj, ne omenijo

področju sevalne varnosti. Vsi uradni dokumenti in praksa na pa, da so te meje neustrezne (10 µT, napram 0.4 µT - mikro Tesla,

terenu se v zvezi z omenjenim varstvom sklicujejo na 26 let stare ki ga priporoča WHO, NIJZ) ) in kot take predmet resnih

smernice Mednarodne komisije za zaščito pred neionizirnim strokovnih pomislekov. Do danes se je namreč nabralo že

sevanjem - ICNIRP in 28 let staro Uredbo o elektromagnetnem ogromno dokazov, ki kažejo na škodljive učinke neionizirajočih

sevanju v naravnem in življenjskem okolju. Priporočila izpred sevanj daleč pod dovoljenimi mejami, zato bi morali po mnenju

več kot petindvajset let v teh smernicah določajo mejne vrednosti mnogih neodvisnih strokovnjakov nemudoma ukrepati in

glede na dokazane zaznavne učinke, ki temeljijo na termičnem regulativo temu ustrezno prilagoditi, obenem pa bi morala

segrevanju živih organizmov oziroma njihovih tkiv pod vplivi Mednarodna agencija za raziskave raka (IARC) radiofrekvenčno



umetnih izmeničnih elektromagnetnih polj. V času vse do danes , sevanje označiti kot “potrjeno rakotvorno” (skupina 1). so bile v svetu objavljene številne študije, ki kažejo škodljive



učinke na biološke oziroma celične procese znotraj živih Rakotvornost radiofrekvenčnih sevanj so v zadnjih letih potrdile

organizmov. Rezultati teh študij dokazujejo, da se škodljivi še dodatne epidemiološke raziskave in več raziskav na živalih

biološki učinki umetnih izmeničnih magnetnih polj pojavljajo pri (tudi govedi, ovc ..) , vključno z raziskavo National Toxicology

bistveno nižjih vrednostih v primerjavi z mejnimi vrednostmi, ki Program (NTP), eno največjih in najbolje zasnovanih tovrstnih

jih določajo smernice ICNIRP in omenjena uredba. Prav zaradi raziskav. A rak je le “končna postaja”, prej pa lahko pride tudi

vseh novejših dognanj, ter rezultatov mnogih epidemioloških do številnih drugih težav. Nekatere od bolje dokumentiranih so

študij o vplivih elektromagnetnih sevanj na zdravje ljudi, npr. poškodbe semenčic oziroma neplodnost pri moških in

predvsem otrok, je potrebno varovanju pred škodljivimi vplivi, živalih , oksidativni stres ter poškodbe DNK, raziskave pa

ob rekonstrukciji oziroma gradnji novih virov sevanja kot so poročajo tudi glavobolih, vedenjskih motnjah, anksioznosti in

BAT, GSM, Wi-Fi v šolah in visokonapetostni daljnovodi težavah s spanjem. Vse kaže, da so najbolj ranljivi še nerojeni,

visokih prenosnih moči, posvetiti bistveno večjo pozornost. dojenčki in otroci, škodljivi učinki pa so lahko kumulativni skozi

več generacij.

Ključne besede



neionizirna sevanja, uredba o EMS, previdnostno načelo, 1.cona 2 Raziskave in »neodvisne« raziskave EMS varovanja,5G,WI-FI,BAT, DV, GSM



Žal so organizacije, ki bi skrbele za naše zdravje, pod močnim



1 vplivom industrije; industrija pa je financirala tudi večji del Uvod

raziskav, ki niso pokazale vpliva na zdravje. Mednarodna

V zadnjem času je zaradi uvedbe omrežja 5G prišlo do polemike komisija za zaščito pred neionizirnimi sevanji (ICNIRP), ki

o škodljivosti t.i. neionizirajočih sevanj, med katere spada tudi določa mednarodna priporočila, na katerih temeljijo državne

radiofrekvenčno sevanje mobilnih telefonov in druge brezžične regulative, je bila pred časom s strani raziskovalnih novinarjev

tehnologije. Kot protiutež lobiju mobilne telefonije , ki javnost Investigate Europe označena kot “kartel” 14 znanstvenikov,

mirijo z razlago, da tveganja ni ; oziroma je kvečjemu katerih delo je izjemno pristransko in ne upošteva na tisoče

zanemarljivo. V ZEG bi radi podali svoja mnenje , ko nas recenziranih raziskav, ki so pokazale škodljive učinke.

nekontrolirano širjenje novih tehnologij skrbi (tudi) z vidika Ugotovitve komisije ICNIRP so v diametralnem nasprotju z

škodljivosti EM sevanja. izsledki več sto znanstvenikov, ki preučujejo vplive

neionizirajočih sevanj na zdravje in so resno zaskrbljeni nad

Permission to make digital or hard copies of part or all of this work for personal or razvojem dogodkov, na kar tudi neprestano opozarjajo. Leta

classroom use is granted without fee provided that copies are not made or distributed 2015 so npr. poslali poziv Združenim narodom, v katere

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must zahtevajo nujno ponovno ovrednotenje mednarodnih priporočil

be honored. For all other uses, contact the owner/author(s). in upoštevanje previdnostnega načela (pri nas 8. člen Zakona o

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia varstvu okolja), ki pravi, da je uvajanje nove tehnologije

© 2024 Copyright held by the owner/author(s).

http://doi.org/10.70314/is.2024.env.22 dopustno le, če ni pričakovati nepredvidljivih škodljivih učinkov

na okolje ali zdravje ljudi.



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Karel Lipič, univ.dipl.ing.



Tudi Resolucija 1815 Sveta Evrope iz leta 2011 pravi, da Pri ZEG podpiramo izključno uvajanje naprednih tehnologij, ne

elektromagnetna sevanja predstavljajo tveganje za ljudi, živali in pa spornih novotarij, ki ogrožajo telesno in duševno zdravje

rastline, in da je potrebno sevalne obremenitve po načelu njenih uporabnikov. Tehnologija omrežja 5G ne prinaša nobenih

ALARA (ang. as low as reasonably achievable) zmanjšati na omembe vrednih izboljšav komuniciranja na daljavo, povzroča

najnižjo možno raven. V resoluciji je zapisano, da imajo sedanja pa nepopravljivo duševno in telesno škodo, sploh ko gre za

priporočila komisije ICNIRP, na katera se zagovorniki varnosti bližnje občane.

brezžične tehnologije sklicujejo, resne pomanjkljivosti, saj ne

upoštevajo netermičnih učinkov, in bi jih bilo potrebno ponovno 3. Moderna in zastarela zakonodaja na področju

ovrednotiti. Resolucija priporoča zmanjšanje izpostavljenosti v EMS

notranjih prostorih na 0,2 V/m (gre za električno poljsko jakost),

kar je pri nas pogosto preseženo v urbanih okoljih, številnih

V ZEG si prizadevamo, da bi Vlada RS sprejela novelo Uredbe

vrtcih, šolah, bolnišnicah, domovih za upokojence itd.

o elektromagnetnem sevanju (ta je že šesto leto leži v predalih

MOP), s katero bi določila strožje mejne vrednosti od trenutno

V resoluciji obžalujejo, da se kljub njihovim svarilom ukrepi za

veljavnih. Poleg tega si prizadevamo, da bi Državni Zbor RS

zmanjševanje izpostavljenosti odvijajo prepočasi ali pa se sploh



zanesljivosti lahko pripelje do visokih zdravstvenih in odmika baznih antenskih postaj (BAT, 5G). Tako v mnogih gospodarskih stroškov, podobno kot se je zgodilo v primeru članicah EU se mora mobilna telefonija, Wi-Fi umakniti iz azbesta, osvinčenega bencina in tobaka. Po letu 2011 so se učilnic. Trenutno veljavna uredba, ki je bila sprejeta leta 1996, ne. Opozarjajo, da čakanje na visoko stopnjo znanstvene varovalne pasove visokonapetostnih daljnovodov (DV) in čimprej sprejel tako zakonodajo, ki bi določila širše preventivne



sevalne obremenitve zaradi novih baznih postaj, omrežja 4G, določa mejno vrednost elektromagnetnega sevanja daljnovodov uporabe pametnih telefonov in drugih brezžičnih naprav močno pri 10 µT. Številne tuje znanstvene raziskave kažejo, da je povečale, z uvedbo omrežja 5G pa se nam obeta še dodatno zdravju škodljiva že 50 krat nižja vrednost elektromagnetnega povečanje, tudi z novimi, še neraziskanimi frekvencami. sevanja (WHO in SCENIHR, telo Evropske komisije). V tem

času je znanost napredovala in vedno bolj se pritrjuje dejstvu, da

Opažamo, da so se nekateri slovenski predstavniki elektro stroke

je bivanje ali dolgotrajno zadrževanje blizu daljnovodov, BQAT,

v zadnjem času oddaljili od EU koncepta previdnostnega načela

WI-FI, 5G zelo škodljivo.

in na brezžično tehnologijo ter prihajajoče omrežje 5G gledajo

skozi rožnata očala industrije, drugih obolenj. V Sloveniji

V ZEG predlagamo, da se mejna vrednost elektromagnetnega

namreč ta čas , žal nimamo neodvisne institucije za zaščito



javnega zdravja ( zadnja leta tudi ne v NiJZ) , ki bi opozarjala na držav. Predlagamo, da nova uredba mejno vrednost sevanja zniža na vrednost primerljivo z zakonodajami razvitih

potencialne nevarnosti elektromagnetnih sevanj. elektromagnetnega sevanja iz maksimalno dovoljenih 10 µT



Zakon iz leta 1996 , ki pri nas določa zgornjo dopustno mejo aparatov, kot virov EMS lahko sami odločamo v kakšni meri se zniža na maksimalno 0,4 µT. Medtem, ko v primeru hišnih

EMS, ni predvidel ne bliskovite ekspanzije brezžične tehnologije,



niti njenega prodiranja v višja frekvenčna območja in porasta objektov kot so 5G, daljnovodi in bazno-antenske postaje temu bomo izpostavljali škodljivim učinkom, se v primeru energetskih

negativnih zdravstvenih posledic EMS. Do zdaj nam je na voljo

težko izognemo. Opozarjamo, da o škodljivih učinkih EMS

na stotine neodvisnih znanstvenih raziskav, ki ugotavljajo

govorimo predvsem v primeru dolgotrajne izpostavljenosti in da

korelacijo med bližino BAT in povečanim številom rakavih,

so negativni vplivi na zdravje posledica permanetnega delovanja

hormonalnih in nevroloških obolenj, zato je pri uvajanju nove

BAT tudi v povprečnem oziroma komulativnem smislu. Zaradi

generacije brezžične tehnologije (5G) nujno treba slediti logiki

fizikalnega pojava širjenja EMS in moči virov EMS je

načela previdnosti. Eden izmed dodatnih razlogov, zakaj je

pomembno, da v smislu upoštevanja načela previdnosti pri

načelo previdnosti tako pomembno, so tudi znanstvena dognanja



področje stranskih učinkov EMS, ki opozarja, da: »Varnostne predvsem otrok, kot najobčutljivejše skupine prebivalcev. Da smernice ICNIRP, FCC, ZDA in EU za področje EMS, temeljijo elektromagnetno sevanje vpliva na zdravje ljudi ni več sporno. na prepričanju, da je za napovedovanje bioloških učinkov in s Prav je torej, da bi se pri umeščanju novih virov takšnega sevanja dr. Martina L. Palla, vodilnega ameriškega strokovnjaka za nobenem primeru ne predstavljajo tveganja za zdravje ljudi, načrtovanju sledimo odmiku od virov sevanja tako, da v



tem varnosti mogoče uporabiti povprečne jakosti EMS in v prostor ravnalo skladno z načeli previdnosti in da bi se v čim povprečno SAR (specifična stopnja absorpcije). Izsledki analize večji meri izogibalo možnosti nastankom škodljivih posledic na osmih različnih vrst kvantitativnih ali kvalitativnih podatkov za zdravju ljudi. Tem ugotovitvam pa mora slediti tudi sodna praksa preverjanje, če to drži ali ne, so pokazali, da so varnostne in sicer tako pri obstoječih virih, kakor tudi pri umeščanju novih smernice v vsakem od teh primerov pomanjkljive, v večini virov v prostor, saj bo le tako lahko sčasoma prišlo do tega, da primerov pa popolnoma zgrešene. Učinki se pojavijo že pri bo državi in investitorjem načelo previdnosti postalo edino in približno 100.000 krat nižjih ravneh, zato je predvidena struktura osnovno vodilo pri sprejemanju odločitev v zvezi s tem. varnostnih smernic povsem neustrezna.«



Zakaj se pri umeščanju virov EMS ne upošteva načelo

A tudi če so INIS-ove meritve sprejmemo kot merodajne, ni



nobenega razloga, zakaj bi se morala BAT nahajati tako blizu optimalne rešitve umeščanja v prostor zaradi dokazanih previdnosti ter poveča oddaljenost od virov oziroma išče

naseljene hiše ,šole ,vrtca ,igrišča in vikenda… bioloških učinkov na žive organizme, ampak se sklicujemo na



Miti in resnice o elektromagnetnem sevanju (EMS) Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



zastarelo zakonodajo, ki bazira na mejnih vrednostih na osnovi Najpogostejši, tako rekoč začetni simptomi EMS so: 

segrevanja tkiv? anksioznost in nespečnost, ki ju pogosto spremlja glavobol 

zvonjenje v ušesih - tinitus  utrujenost, stres  izguba spomina

V postopkih načrtovanja in umeščanja visokoenergetskih in osredotočenega mišljenja (možganska megla)  veliko ljudi

daljnovodov BAT, RTV oddajnikov, Wi-FI-ja v šolah, 5G … kot razvije pretirano občutljivost na EM sevanja. Glede resnejših

virov EMS je potrebno v interesu varovanja zdravja ljudi obolenj bom navedel le tiste bolezni, ki so najpogosteje

upoštevati načelo previdnosti. omenjene v literaturi:  Porast raka glave pri ljudeh, ki že leta

intenzivno uporabljajo mobilne naprave  Porast raka na modih

V primeru načrtovanja in umeščanja npr. BAT in pri moških, ki imajo mobilne naprave v sebi. žepi in  Povečanje

visokoenergetskih daljnovodov, katerih nazivna napetost je večja raka dojke pri ženskah, ki imajo mobilne naprave v nedrčku 

kot 100 kV, načelo previdnosti zahteva, da je potrebno poiskati Poškodba enojne in dvojne vijačnice DNK, nenormalno

prostorske in tehnične rešitve (kabliranje) , ki se ognejo oziroma prepisovanje genov  Zmanjšana sposobnost popravljanja DNK

v najmanjši možni meri obremenjujejo občutljiva območja z matičnimi celicami  Zmanjšanje števila radikalov v telesu 

stanovanjskih stavb . Izbrati je treba rešitev, ki predstavlja čim Sperma morfološke spremembe in zmanjšana njihova gibljivost

večjo možno oddaljenost od naštetih občutljivih območij in ki - kar ima za posledico drastično zmanjšano plodnost mlajših

upoštevajoč načelo previdnosti zagotavlja varovanje zdravja, generacij itd. Vse te bolezni imajo dolgo latenco, kar pomeni, da

varnost in kvaliteto bivanja čim večjemu številu ljudi. Kadar se se statistično značilen porast teh bolezni pojavi šele po 10-15

v prejšnjem odstavku navedenim občutljivim območjem ni letih intenzivne uporabe mobilnih naprave. Zaradi tega dejstva

mogoče v celiti izogniti, je med več možnostmi potrebno izbrati mnogi uporabniki zanemarjajo izsledke znanstvenih raziskav,

tisto, ki v najmanjši možni meri obremenjuje občutljiva območja računajoč, da se to ne bo zgodilo »samo meni«, kot smo že rekli,

oziroma se izogne čim večjemu številu ljudi. imamo podoben pristop pri vožnji z avtomobilom, letalom ali pri

katerikoli potencialno nevarni dejavnosti. Mnogi znanstveniki, ki

4. Zdravstvene posledice neionizirnih sevanj že desetletja raziskujejo to problematiko, se zavedajo problema,

zato imajo svoje spletne strani in YouTube kanale z namenom

Znanstveniki že več kot 20 let pridno zbirajo podatke, delajo izobraževanja in ozaveščanja ljudi o potencialni nevarnosti EM

statistiko in povezujejo število bolnikov s kumulativnim sevanja. Za tiste, ki želijo vedeti več in znajo angleško, dajem

številom ur dela z mobilnimi napravami ali računalniki. Očiten nekaj dobrih povezav:  Portal EMF Univerze v Aachnu, kjer

je porast obolenj, a kljub na tisoče znanstvenih člankov, ki najdete več kot 30 tisoč znanstvenih člankov in več kot 6 tisoč

dokazujejo škodljivost mikrovalovnega sevanja za zdravje ljudi, povzetkov teh člankov, ki govorijo o vpliv EM sevanja na ljudi

nekateri "znanstveniki" in tudi vlade mnogih držav še vedno in živali. Povezava: https://www.emf-portal.org/en 

"tiščijo glavo v pesek" in nočejo priznati očitna dejstva. Še naprej BIOINITIATIVE 2012 – Neodvisna skupina znanstvenikov je na

podajajo polresnice, favorizirajo sponzorje in »močnejše« več kot 1500 straneh povzela rezultate 1800 neodvisnih študij, ki

stranke ter nočejo pogledati resnici v oči. dr. Henry Lai z potrjujejo pomemben vpliv EM sevanja na številne zdravstvene

Univerze v Washingtonu je nedavno naredil zbirko vseh vidike. Novejše rezultate lahko najdete na povezavi [3].

objavljenih člankov o škodljivosti EM sevanja in ugotovil, da so

rezultati raziskav zelo odvisni od tega, kdo je raziskavo plačal: Trenutno veljavni predpisi v Republiki Sloveniji niso sledili

če gre za neodvisno, pošteno raziskavo, potem več kot 70 % izsledkom medicinske stroke, ki je v zvezi z vplivom EMS na

raziskav potrjuje škodljivost EM sevanja, če pa je raziskava zdravje ljudi, predvsem otrok, v zadnjih desetletjih pripisala

sponzorirana s strani industrije, je situacija obratna, in sicer pomemben vpliv. Prepoved mobitelov v šolah, predvsem zaradi

skoraj 70 % raziskav trdi, da EM sevanje ni škodljivo (slika). Več EMS. Kot kažejo ugotovitve medicinske stroke, je vpliv EMS

o tej študiji in njegovih raziskovalnih rezultatih na splošno na na zdravje ljudi zelo velik, saj lahko že izpostavljenost

povezavah [1, 2]. vrednostim, ki so kar 50x nižje od tistih, ki so trenutno

predpisane v RS, bistveno poveča možnosti za nastanek malignih

Se sprašujete, kako je to mogoče, kako lahko znanstveniki obolenj.

"ponarejajo" rezultate? Tukaj je primer: med raziskovanjem

literature na to temo smo naleteli na dela , ki so delali na primer Ob dejstvu, da EMS od razdalj, (ki so pokazane v elaboratih kot

poskuse z mišmi. Skupino miši so vsak dan za uro ali dve vplivno območje) naprej pada počasi, so ob najslabših razmerah

izpostavljali EM sevanju določene frekvence (običajno v potrebni večji odmiki, če želimo zadostiti varovanju otrok kot

območju med 0,5 in 2,5 GHz) dva ali tri tedne. Nato so miši najobčutljivejše skupine ljudi na vplive EMS. Nacionalni inštitut

secirali in iskali spremembe v možganskem tkivu, preverjali za javno zdravje v svojem mnenju z dne 22. 6. 2016 ugotavlja,

kakovost ali mobilnost semenčic itd. ter rezultate primerjali z da je pri izpostavljanju prebivalstva, predvsem otrok EMS zelo

enakimi analizami kontrolne skupine miši, ki niso bile obsevane. nizkih frekvenc, treba nujno upoštevati načelo previdnosti do

In seveda, rezultati so bili pričakovani: med rezultati obsevane in dokončnega odgovora o vplivu tega dejavnika na zdravje – še

kontrolne skupine miši ni bilo razlike!! Kot smo že povedali, le posebej so izpostavili problematiko umeščanja BAT in

dolgotrajna izpostavljenost EM sevanju potencialno povzroča daljnovodov v bližino naselij.

zdravstvene težave... le nekajurna izpostavljenost seveda ne bo

pokazala statistično pomembne škode. Komentar je nepotreben! Zaradi fizikalnega pojava padanja jakosti EMS z razdaljo od vira

Na žalost je takšnih »znanstvenih«, torej naročenih prispevkov, je pomembno, da v smislu izpolnjevanja načela previdnosti pri

veliko in proizvajajo »hrup«, s katerim tisti, ki zastopajo interese umeščanju virov EMS v čim večji možni meri zmanjšamo

industrije, zasenčijo resne in neodvisne znanstvene prispevke. zdravstveno tveganje tako, da zagotavljamo čim večji odmik



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Karel Lipič, univ.dipl.ing.



virov, čim večjemu številu ljudi, predvsem otrok. Zaradi različne omejeval prenosnih moči zaradi negativnih vplivov EMS.

občutljivosti ljudi na vplive EMS je to še posebej pomembno na Izračuni negativnih vplivov na osnovi povprečnih vrednosti

območjih povišane koncentracije tam, kjer se dlje časa napovedanih obremenitev imajo preveč neznank, zato so možne

zadržujejo ljudje. Zdravstveno tveganje skratka pomembno številne manipulacije in napačno prikazovanje rezultatov.

zmanjšujemo, če se prednostno umikamo območjem, kjer se Morebitne vplive na okolje bi morali v fazi načrtovanja vedno

zadržuje večje število ljudi. Konkretizacija načela previdnosti v prikazovati na osnovi najneugodnejših pogojev obratovanja

bistvu zagotavlja, da se v primeru umeščanja nevarnih in zdravju oziroma maksimalni možni vrednosti gostote magnetnega

škodljivih objektov v prostor, kot so BAT, poiščejo in upoštevajo pretoka. Področje EMS, kot okoljski problem je zapisan v

variante, ki imajo najmanjši vpliv na ljudi. dokumentih : NPVO, NEPN



Znanstvene raziskave dokazujejo škodljivost elektromagnetnega

sevanja električnih daljnovodov in BAT na zdravje ljudi. Te

škodljive vplive priznavata tudi Svetovne Zdravstvena

organizacija in Evropska komisija. Le-ta je ustanovila SCENHIR Viri, Literatura

(the Scientific Committee on Emerging and Newly Identified [1] https://bioinitiative.org/research-summaries/



Health Risks), posvetovalno telo Evropske Komisije, ki proučuje [2] https://www.radiationresearch.org/articles/biological-effects-of-rf-prof-na novo odkrita oz. na novo identificirana tveganja za zdravje in henry-lai/

okolje. V najnovejšem poročilu SCENIHR ugotavlja, da

znanstvene [3] https://bioinitiative.org/whats-new/ raziskave potrjujejo škodljive vplive

elektromagnetnih sevanj na zdravje ljudi. [4] https://www.emf-portal.org/en  BIOINITIATIVE 2012



Trenutno v Sloveniji velja, da v bivalnih objektih sevanje ne sme [5] https://bioinitiative.org/whats-new/

preseči 10 µT. Raziskave o otroški levkemiji pa ugotavljajo, da

je povečana verjetnost levkemije že pri 50 krat nižji vrednosti [6] https://bioinitiative.org/research-summaries/



sevanja. Prav tako to velja za otroško astmo in otroško debelost. [7] https://www.radiationresearch.org/articles/biological-effects-of-rf-prof-Mimogrede, po naši zastareli zakonodaji oz. uredbi lahko henry

stanovanjski objekti, šole, vrtci stojijo pri 400 kV DV že na 40

metrih oz. na 16 metrih pri BAT?! Po podatkih ZEG je v [8] https://www.europarl.europa.eu/stoa/en/document/EPRS_STU(2021)690

Sloveniji več kot sto BAT v neposredni bližini šol in vrtcev. Pod 012

DV /110, 220, 400 kV/ pa živi na razdalji 100 m levo/desno [9] https://en.wikipedia.org/wiki/Martin_Pall približno 20.000 ljudi.

[10] https://www.researchgate.net/profile/Martin-Pall

Napovedi obratovanja za desetletja vnaprej so nezanesljive. Z

naraščanjem potreb po 5G in prenosa električne energije preko [11] CENHIR, ICNIRP, MOPE, NIJZ

mednarodnih tranzitnih daljnovodov, ki bi bili postavljeni na [12] Inštitut za neionizirna sevanja -INIS občutljivih poseljenih območjih, ki imajo visoke prenosne

kapacitete, jim zato zagotovo nihče nikoli v prihodnosti ne bo [13] Arhivi ZEG na www.zeg.si



Zbornik 27. mednarodne multikonference



INFORMACIJSKA DRUŽBA – IS 2024

Zvezek E



Proceedings of the 27th International Multiconference



INFORMATION SOCIETY – IS 2024

Volume E



17. Mednarodna konferenca o prenosu tehnologij



17th International Technology Transfer Conference



Uredniki / Editors



Urška Florjančič, Robert Blatnik, Špela Stres



http://is.ijs.si



9. oktober 2024 / 9 October 2024

Ljubljana, Slovenia



FOREWORD / PREDGOVOR



Dear guests, experts, panelists, and participants,

Welcome to the 17th International Technology Transfer Conference (17ITTC). Since its

inception, the Jožef Stefan Institute has proudly served as the initiator and main organizer of

this esteemed event, advancing innovation and knowledge transfer in Slovenia. This year, we

are honored to host the conference in collaboration with 13 public research organizations,

representing two national consortia of knowledge transfer offices (KTOs). The event is co-

financed and supported by the Ministry of Higher Education, Science, and Innovation, as part

of the "Mesec znanosti" campaign.



The ITTC has established itself as a crucial platform for exchanging ideas and fostering

collaboration between domestic and international stakeholders, significantly contributing to

the development of Slovenia’s national innovation ecosystem. The conference has been

instrumental in helping Slovenian public research organizations address challenges such as

securing funding for spin-outs, updating national legislation on research and innovation, and

building robust consortia for KTOs.



Collaboration among KTOs, both within and across the two consortia continues through joint

activities aimed at promoting KTO initiatives, raising awareness, and encouraging networking

and the exchange of best practices. These efforts focus on enhancing the skills and capabilities

of all stakeholders—from KTO employees and researchers to students—while improving the

implementation of intellectual property (IP) marketing and protection. Additionally,

coordinated efforts will establish common metrics and indicators, enabling effective

monitoring and evaluation of knowledge valorization processes at public research

organizations, ensuring long-term success.



This year’s conference theme, “Self-Evaluation of Research Organizations to Support the

Development and Strengthening of Knowledge Transfer,” aligns with our goal of bolstering

the role of KTOs and improving the commercialization of intellectual property, as well as to

promote the wider social relevance of knowledge transfer and the outputs and impacts of

KTO work on the well-being of society as a whole. The theme is being explored in a keynote

address focuses on the role of institutional self-evaluation within the Framework of proposed

amendments to the General Acts of the Slovenian Research and Innovation Agency (ARIS),

followed by a round table discussion. The panel will feature representatives from the Ministry

of Higher Education, Science, and Innovation, ARIS, the Slovenian Rectors' Conference,

KOsRIS, Leiden University, the Institute for Economic Research, and the University of

Colorado Boulder.



We present several prestigious awards during the conference, including the Conference Prize

for the Best Innovation in 2024, which aims to promote the commercialization of innovative

technologies developed at public research organizations. The WIPO National Award for

Enterprises is awarded to a Slovenian enterprise that has successfully developed a strategy for

commercializing university-based innovations. In addition, the WIPO National Award for

Inventors honors an individual researcher or a team of researchers from a Slovenian public

research institute whose patented invention has significantly contributed to Slovenia’s

economic and technological development.



The conference also features sessions on Opportunities Arising from Publicly Funded

Research Projects, where researchers and KTO experts showcase successful scientific projects



funded by the Slovenian Research Agency, highlighting their potential for innovation and

commercialization. In the session on Connecting the Educational System with the Academic

Sphere, presentations of selected research topics and collaboration proposals emphasize the

importance of bridging the gap between academia and education, fostering greater

cooperation and engagement.



We are especially excited about the ongoing growth of the conference, which, for the fifth

consecutive year, includes peer-reviewed contributions from researchers specializing in

knowledge and technology transfer. Since 2009, the entrepreneurial pitch competition for

research teams and their inventions, evaluated by international teams of commercialization

and investment experts, has remained a key feature, supporting over 100 research teams in

developing business models, with more than 30 winners recognized to date.

Together, we look forward to exploring new opportunities, including collaborations with the

Vesna DeepTech Fund, which plays a vital role in providing early-stage funding to spin-out

companies emerging from public research organizations. Established by the EIF in

partnership with Slovenian and Croatian development banks, the fund bridges the gap

between research and commercialization, offering financial backing to help transform cutting-

edge innovations into successful ventures. This collaboration fosters stronger partnerships

between research institutions and industry, further boosting the commercialization of

scientific discoveries.

Thank you for being part of this journey, and we look forward to an inspiring exchange of

ideas at the 17ITTC.



Programme Committee of the 17ITTC



ACKNOWLEDGEMENTS



We would like to acknowledge the valuable contributions of the scientific programme

committee to the scientific programme, review of the scientific papers on technology transfer

and intellectual property, and selection of publications in the conference proceedings, and the

efforts of the conference programme and organising committees for successful

implementation of the 17th International Technology Transfer Conference.





SCIENTIFIC PROGRAMME COMMITTEE




Niko Schlamberger, past President of Slovenian Society INFORMATIKA



Tamara Besednjak Valič, Faculty of Information Studies in Novo Mesto



Dolores Modic, Nord University Business School



Jana Hojnik, University of Primorska



Urška Fric, Faculty of Information Studies in Novo mesto



CONFERENCE PROGRAMME COMMITTEE



Robert Blatnik, Jožef Stefan Institute



Špela Stres



Peter Alešnik, University of Ljubljana



Simona Kustec, University of Primorska



Jana Hojnik, University of Primorska



ORGANISING COMMITTEE



Robert Blatnik, Jožef Stefan Institute



Marjeta Trobec, Jožef Stefan Institute



Doroteja Novak, University of Maribor



Maša Stošič, University of Maribor



Matej Draksler, Geological Survey of Slovenia



Urška Fric, Faculty of Information Studies in Novo mesto



Ana Hafner, Rudolfovo – Science and Technology Centre Novo mesto



Mateja Košir, Slovenian National Building and Civil Engineering Institute



Intellectual Property as a Success Factor for Startups:



Systematic Literature Review



Maja Fortun Novak

Faculty of information studies

Novo mesto, Slovenia

majafortunnovak@gmail.com



ABSTRACT Globally, the number of new companies, known as startups, is This paper presents a systematic literature review on the impact rapidly increasing daily [18]. Such companies, especially of intellectual property on startup success. By reviewing 21 innovative startups, often lack historical financial data or a track relevant articles published in the last five years, sourced from record, making it more difficult to establish a market reputation. Google Scholar, it analyses the influence of intellectual property Their innovative products or business processes also lack prior on key business factors such as financing, growth, experience or comparative standards [2]. This presents competitiveness, and innovation. The findings show that challenges that can lead to the failure of companies due to intellectual property plays a significant role in startup success, inadequate or non-existent business models, and insufficient though results vary regarding formal (patents) and informal business growth [4]. (market advantage, trade secrets) protection methods. A balanced approach to intellectual property management, tailored Díaz-Santamaría and Bulchand-Gidumal identify several factors to startups' needs and developmental stages, is recommended. that can influence the success of startups [6]. The results of their The article emphasizes the need for further research on different research indicate that the success of a startup can be measured in forms of intellectual property, considering regional contexts and two ways: the startup achieves significant revenue, and the long-term effects. Its value lies in offering both theoretical startup receives funding. In the following sections, we also insights and practical recommendations, particularly for highlight other indicators for measuring the business success of policymakers, investors, and startup owners seeking to promote a startup . innovation and growth through effective intellectual property

management. Despite numerous studies in recent years on the impact of



KEYWORDS innovative practices on the business development and success of

startup companies, no comprehensive and systematic analysis of

scientific literature has yet been conducted that specifically

Intellectual property, startup, trademark, patent, innovation, focuses on the influence of IP rights on the business success of

growth, business success startups. This gap in scientific research indicates the need for an

in-depth review of existing scientific sources that would enable



1 INTRODUCTION a holistic understanding of the impact of IP on the business

success of startups.



The purpose of our research is to systematically review the The research question is, how are a startup's IP and business

literature on the impact of intellectual property (hereinafter success connected, i.e., does IP affect business success, and how?

referred to as IP) on the business success of startup companies. In this way, we can better understand how IP contributes to the

Through an analysis of existing research results, we aim to competitiveness and long-term success of startups.

explore how IP contributes to achieving these success criteria.

Our objective is to determine whether, and how, IP influences 2 METHOD the business success of startups, which is crucial for

understanding their growth in a dynamic business environment. For this research, we used the systematic literature review

method, conducted between January and April 2024. During this

In the modern economy, there is a notable impact that new period, we reviewed foreign literature, focusing on the impact of

companies have on innovation [9, 15], and the economy as a IP on the success of startup companies. We examined

whole [4, 10]. In particular, startups drive innovation, create new professional, scientific, and research publications published in

jobs and introduce competitiveness into the business world [17]. the last five years to ensure the most up-to-date data and

The influence of IP is particularly interesting, as it can be crucial discussions in this field. We used the international bibliographic

for their success. database Google Scholar to collect information, which allows for

Permission to make digital or hard copies of part or all of this work for personal or the search of scientific literature and the ranking of documents in

classroom use is granted without fee provided that copies are not made or a manner used by researchers. To identify relevant sources, we

distributed for profit or commercial advantage and that copies bear this notice and used the following set of keywords: "Intellectual Property,"

the full citation on the first page. Copyrights for third-party components of this work "Startups," "Start-ups," "Patents," and "Trademark." In total, we

must be honored. For all other uses, contact the owner/author(s). obtained 30 relevant articles. After excluding 9 duplicates, we

© 2024 Copyright held by the owner/author(s). retained 21 suitable articles for further analysis. The four data Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

collection strategies were used to ensure a thorough and

comprehensive review of the relevant literature, tailored to the

research question. Each strategy contributed to refining the Schaberg explores various forms of IP protection and finds

search and eliminating irrelevant sources. that startups with a diverse portfolio of protected rights grow and

innovate more easily [19]. Various forms of protection, such as



patents, trademarks and copyrights, provide comprehensive

protection, increase investor confidence, and facilitate access to

capital, which encourages further innovation. Ljungqvist, Hegde,

and Raj add that the rapid granting of patents stimulates

innovation and facilitates the acquisition of capital, while delays

in patent granting negatively impact startups' growth [13].

Despite the numerous positive impacts of IP on startup

companies, some research shows mixed results. Power and Reid

caution that patents can negatively affect the success of startups,

while trademarks and licensing have a positive impact [16].

Teixeira and Ferreira, as well as YunQi and Lin, find that formal

methods such as patents often reduce companies'

competitiveness, while informal mechanisms such as market

advantage and trade secrets can improve competitiveness [22,

24].

Some studies recommend a balanced approach to IP

management. Silva Júnior, Siluk, Neuenfeldt Júnior, Rosa, and

Michelin believe that a combination of formal and informal

protection mechanisms, such as patents, trademarks, and

copyrights, is crucial for protecting innovations and enhancing

Figure 1: Diagram of the Results of the Systematic competitive advantage [21]. Audretsch, Colombelli, Grilli,

Literature Review Minola, and Rasmussen emphasise that policies for innovation

and IP protection must be tailored to the specific needs and stages



3 RESULTS AND DISCUSSION of startups' development [1]. Chou adds that patents help in

securing funding, but startups often face patent litigation, which

reduces their productivity [5]. The solution lies in

Most research confirms that IP significantly influences the commercialisation patents, which would reduce the negative

success of startup companies. Researchers have examined how impacts of disputes and enable better protection and marketing

IP affects various aspects of startups' operations, such as of innovations.

financing, growth, competitiveness, and innovation Research by EPO-EUIPO, Krauss, Breitenbach-Koller, and

performance. Below, we summarise the key findings of these Kuttenkeuler, Brandt, Laibach, Kamrath, and Bröring, Schaberg

studies. and Ljungqvist, Hegde and Raj emphasise the importance of

Hellström, Nilsson, Andersson, and Hakanson found that a effective IP management for startup success [7, 11, 3, 19, 13].

combination of patents and secrecy positively influences the However, these studies primarily focus on specific sectors or

protection of innovations and market opportunities, thereby particular cases and only certain forms of IP, which can lead to a

increasing companies' competitiveness [8]. Similarly, research limited understanding of the overall picture and strategies that

by EPO-EUIPO reports that 29% of European startups invest in would be beneficial for startups across different fields.

IP, which increases their chances of securing funding [7]. Krauss, Some studies focus on specific countries or geographical

Breitenbach-Koller, and Kuttenkeuler emphasise that IP is areas, such as EPO-EUIPO on European startups or Li, Gan, and

crucial for the success of biotech startups, as it enhances their Zhang on Chinese startups [7, 12]. These studies highlight

value and attractiveness to investors [11]. regional particularities in the use of IP, which can affect the

The role of IP is particularly pronounced in international innovation and competitiveness of companies in specific

markets, where the protection of innovations is essential for the geographical environments.

survival and growth of companies. Tula, Ofodile, Okoye, Nifise, University startups, as discussed by Shahidan, Latiff, and

and Odeyemi highlight that IP plays a key role in ensuring the Wahab, represent a special category where innovation

business success of startups in a global context, as it prevents the intertwines with academic knowledge [20]. These startups often

copying and exploitation of innovations by competitors, with face specific challenges in commercialising technologies, which

registered patents and trademarks serving as signals to investors can hinder the value creation process. Successful university

of startups' innovation and credibility [23]. startups must identify market opportunities, ensure

Brandt, Laibach, Kamrath, and Bröring also point out that IP entrepreneurial commitment, and continuously develop their

increases the value of companies and attracts investors, which is technologies to meet market demands.

crucial for success in corporate investments. Startups with well- The inclusion of perspectives such as that proposed by

protected IP find it easier to attract investors, as they perceive Panagopoulos and Park, where patents serve as negotiation tools,

protected technology as a lower-risk investment with greater highlights the potential for strategic use of IP in corporate

potential for profitability [3]. Additionally, IP enables startups to negotiations, not just as a defensive mechanism [14].

collaborate more easily with larger corporations, further Effective IP management generally has a positive impact on

strengthening their financial and strategic positions. the success of startups, but some results indicate the need for a



balanced approach that also includes informal protection REFERENCES methods. Therefore, it is essential that future research expands

existing methodological frameworks and includes analyses that [1] Audretsch, D., Colombelli, A., Grilli, L., Minola, T., & Rasmussen, E. (2020).

capture the complexity and interdependencies of different IP Innovative start-ups and policy initiatives. Research Policy, 49(10), 104027.

rights, including on an international level. [2] Backes-Gellner, U., & Werner, A. (2007). Entrepreneurial signalling via

education: A success factor in innovative start-ups. Small Business

Economics, 29, 173-190.



4 [3] Brandt, L., Laibach, N., Kamrath, C., & Bröring, S. (2023). Start-up selection CONCLUSION

criteria for corporate venturing: what matters for incumbents?. International

Journal of Entrepreneurial Venturing, 15(4), 381-408.



Based on the review of existing research, we can confirm that [4] Cantamessa, M., Gatteschi, V., Perboli, G., & Rosano, M. (2018). Startups’

roads to failure. Sustainability, 10(7), 2346.



competitiveness of startups. IP protects ideas and innovations commercialization patents can benefit startup innovation. Nw. J. Tech. & Intell. Prop. 17, 349. and contributes to securing capital and sustainable growth of IP is an important factor in supporting innovation and [5] Chou, R. (2019). Startups and investors and trolls, oh my!: How

[6] Díaz-Santamaría, C., & Bulchand-Gidumal, J. (2021). Econometric estimation

companies. of the factors that influence startup success. Sustainability, 13(4), 2242.

Although most research confirms the positive impact of IP, [7] EPO-EUIPO. (2023). Patents, trademarks and startup finance. Funding and

there are also studies that suggest a negative impact of certain exit performance of European startups. Retrieved from

forms. Further research is needed to clarify the impact of https://www.euipo.europa.eu/en/publications/2023-startup-finance (Accessed

on 4 Jan 2024).

different types of IP on the success of startups, including an [8] Hellström, A., Nilsson, S., Andersson, M., & Håkanson, U. (2019). Intellectual



analysis of specific geographical contexts and the long-term property for generating value for start-up companies in key enabling technologies. Biotechnology Research and Innovation, 3(1), 80-90. effects on company survival. [9] Khuan, H., Andriani, E., & Rukmana, A. Y. (2023). The Role of Technology With the doctoral dissertation currently in preparation, we in Fostering Innovation and Growth in Start-up Businesses. West Science

will explore the impact of IP on the success of innovative startups Journal Economic and Entrepreneurship, 1(08), 348-357.

that have received funding from the Slovenian Enterprise Fund [10] Kowalski, A. M. (2023). Entrepreneurship as the Factor of Competitiveness,

in the P2 tender between 2008 and 2023, analysing their and the Role of Start-Ups. Focus on Entrepreneurship and Competitive

Advantages, 29.



approach, combining quantitative and qualitative strategies. property rights and their role in the start-up bioeconomy–a success story? EFB Bioeconomy Journal, 1, 100002. Primary and secondary data will be collected from recipients of operations and registered IP. We will adopt a mixed-methods [11] Krauss, J., Breitenbach-Koller, L., & Kuttenkeuler, D. (2021). Intellectual

[12] Li, H., Gan, M., & Zhang, Y. (2023). The Impact of Initial Intellectual Property

the P2 grant (2008–2023), using IP databases (Espacenet, Global Decisions of Start-Ups on Innovation Performance. Entrepreneurship

Brand Database, DesignView), and conducting in-depth Research Journal, (0).

interviews with selected companies. Quantitatively, we will [13] Ljungqvist, A., Hegde, D., & Raj, M. (2021). Quick or Broad Patents?

analyse the IP portfolios of these startups and assess their Evidence from US Startups (No. 16320). CEPR Discussion Papers.

[14] Panagopoulos, A., & Park, I. U. (2018). Patents as negotiating assets: patenting

business performance through statistical analysis in JASP and versus secrecy for startups. The Economic Journal, 128(615), 2876-2894.



Excel, using univariate, bivariate, and multivariate methods. [15] Peniaz, L. (2024). THE ROLE OF STARTUPS IN CREATING INNOVATIVE ECOSYSTEMS. Věda a perspektivy, (1(32)). Qualitatively, we will conduct 20 in-depth interviews—10 with [16] Power, B., & Reid, G. C. (2021). The impact of intellectual property types on startups that have registered IP and 10 with those that have not— the performance of business start-ups in the United States. International Small

to uncover insights that quantitative methods cannot fully Business Journal, 39(4), 372-400.

address. This mixed-methods approach will allow us to [17] Rebernik, M., & Jaklič, M. (2014). Start: up Manifest: Slovenija, pripravljena

comprehensively explore the impact of IP on startup success, na prihodnost 2014–2020+. Start: up Manifesto. Slovenia, Ready for the

Future, 2020.



broader role of IP in innovation and business growth. Canciglieri Junior, O. (2020, July). Startup definition proposal using product lifecycle management. In IFIP international conference on product lifecycle Future research will thus contribute to a better theoretical offering both statistical analysis and qualitative insights into the [18] Reisdorfer-Leite, B., Marcos de Oliveira, M., Rudek, M., Szejka, A. L., &

management (pp. 426-435). Cham: Springer International Publishing.

understanding and offer practical recommendations for [19] Schaberg, U. G. (2023). "IP Is Paramount:" The Significance of IP in Early-

policymakers, investors, and startup owners regarding the Stage Start-Up Investment Decisions. In Intellectual Property Management for

optimal use of IP to promote innovation and long-term growth. Start-ups: Enhancing Value and Leveraging the Potential (pp. 91-116). Cham:

Springer International Publishing.

[20] Shahidan, N. H., Latiff, A. S. A., & Wahab, S. A. (2023). Sustainable



ACKNOWLEDGMENTS technology development during intellectual property rights commercialisation

by university startups. Asia Pacific Journal of Innovation and

Entrepreneurship, 17(3/4), 176-194.

[21] Silva Júnior, C. R., Siluk, J. C. M., Neuenfeldt Júnior, A., Rosa, C. B., &



possible without the support and guidance of my supervisor doc. competitiveness of startups: a systematized literature review. Gestão & Produção, 29, e13921. dr. Ana Hafner, from the Faculty of information studies in Novo The completion of this research paper would not have been Michelin, C. D. F. (2022). Overview of the factors that influence the

[22] Teixeira, A. A., & Ferreira, C. (2019). Intellectual property rights and the

mesto. I am deeply grateful for her dedication, insightful competitiveness of academic spin-offs. Journal of Innovation & Knowledge,

feedback, and encouragement throughout this process. I would 4(3), 154-161.

also like to extend my heartfelt thanks to my family for their [23] Tula, S. T., Ofodile, O. C., Okoye, C. C., Nifise, A. O. A., & Odeyemi, O.

support. (2024). Entrepreneurial ecosystems in the USA: A comparative review with

European models. International Journal of Management & Entrepreneurship

Research, 6(2), 451-466.

[24] YunQi, Y., & Lin, G. T. (2023). Bridging the Gap: Intellectual Property Rights

and Sustainable Development Goals in Innovation Ecosystems.



The Reversed European Paradox: do European Patents



have a High Market Value but Low Impact?



Ana Hafner

Centre for Technology Transfer and Intellectual Property Rights

Rudolfovo - Science and Technology Centre Novo mesto

Novo mesto, Slovenia

ana.hafner@rudolfovo.eu



ABSTRACT perceived to lag behind the U.S. in converting its academic results into economic outcomes” [4]. This lag may affect the economic growth of European countries and also their global The U.S. research institutions often serve as role models for competitiveness in industries that rely on technological European research institutions in knowledge and technology innovation. transfer. This paper investigates the widely held belief that



of patent family members, forward citations, backward and discovered that there is a difference between PRO-owned and PRO-invented patents (inventions). They discovered that EU citations, and claims. Results indicate that the dominance of PROs lag behind the U.S. because 80% of patents with academic U.S. organizations is not as clear as commonly perceived. The inventors are in the EU owned by private firms rather than PROs, study adds value by providing an additional understanding of and they are statistically not recognized as PRO patents. the technology transfer landscape, challenging the assumption of U.S. superiority. On the contrary, this study is not focused on the quantity of the patents, such as Crespi’s et al. [1], but on their quality. The top KEYWORDS European and U.S. PROs will be compared according to the Patents, patent quality, patent valuation, public research from leading U.S. and European universities and research deal with different aspects of KTT in Europe, especially in comparison to the U.S. For example, Crespi et al. [1] focused on institutes was done. The methodological approach involves a a comparison of European and U.S. academic patenting systems comparative analysis of key patent quality indicators: number the U.S. To explore this, an analysis of the quality of patents The aim of this study is to contribute to existing studies which European technology transfer performance is inferior to that of

value of their patents by indicators of patent value.

organizations, research institutes, universities, Europe-U.S.



comparison, number of patent family members, forward The research question is: If we compare the patents of the top citations. European and U.S. PROs by indicators of patent value, such as

the number of patent family members and forward citations, are



1 there any differences between Europe and the U.S.? INTRODUCTION

The European paradox is a term coined to describe that Europe Understanding this research problem is important because the

is strong in basic science but lags behind some other developed effective commercialization of scientific knowledge directly

countries in technological applications in world markets [1], impacts economic growth and innovation. If European PROs can

specifically in the commercialisation of scientific findings or enhance their KTT performance, it could lead to increased

what we call knowledge and technology transfer (KTT). competitiveness in global markets. By focusing on patent quality

rather than quantity, this study aims to provide some insights into

Many scholars have studied why some public research how Europe might overcome the perceived lag behind the U.S.

organisations (PROs) – which include universities and research

institutes – are more successful in commercializing knowledge.



has been conducted in the U.S., often identified as pioneers in Methods for patent valuation can be qualitative or quantitative this area [2]. [5]. We will focus only on quantitative and non-monetary Most of the research on university knowledge commercialization 2 INDICATORS OF PATENT VALUE



are still young, with half of them being established after 2000 [3]. citations and the existence of opposition and litigation [5]. Such However, this is probably not the only reason why “Europe is valuation has many advantages: the method is fast, objective and inexpensive and can be fully automated once the valuation In Europe, most university or PROs’ technology transfer offices status, international and technological scope, number of forward methods, i.e., patent indicators [5]. Typical indicators are legal

∗Article Title Footnote needs to be captured as Title Note

†Author Footnote to be captured as Author Note system is set up [5]. International scope (size of patent family)

Permission to make digital or hard copies of part or all of this work for personal or and forward citations (citations received from patents applied

classroom use is granted without fee provided that copies are not made or distributed later) are probably the most frequent measures for assessing

for profit or commercial advantage and that copies bear this notice and the full patent value. Patent valuation using forward citations has been



citation on the first page. Copyrights for third-party components of this work must increasingly used by practitioners when a patent’s value has not be honored. For all other uses, contact the owner/author(s). been otherwise established [6]. Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).



Og et al. [7] divide patent value indicators into ex-ante indicators 3 METHOD (family size, backward citations, backward references to non-For this study, the first methodological question was, how to patent literature, number of claims, and number of inventors) and determine the most important or innovative European and U.S. ex-post indicators (forward citations). PROs.



• [12]. From this report, five top PROs were chosen: Number of patent family members • Carnegie Mellon University • Number of backward citations • • For the U.S., the Heartland Forward’s report (2022) was used Number of claims We will consider the following indicators:

• University of Florida Number of forward citations • Columbia University



According to Squicciarini et al. [8], claims define the extent of • Stanford University • Harvard University the exclusive rights granted to a patent holder, as only the technologies or elements specified within these claims receive For Europe, the European Research Ranking list (2020) was used legal protection and can be enforced. Consequently, the scope of [13]. From this list, five top PROs were chosen: a patent's protection is determined by the number and specifics • Centre National de la Recherche Scientifique of its claims. Additionally, since patent fees typically depend on • Fraunhofer Gesellschaft zur Foerderung der Angewandten the number of claims included, having numerous claims can Forschung E V result in higher costs. Therefore, the number of claims in a patent • Commissariat a L'energie Atomique can indicate not just its technological scope but also its anticipated market value: more claims often suggest a higher • Eidgenoessische Technische Hochschule Zuerich (ETH) University Of Copenhagen expected value for the patent [8]. •



Patent family size – the number of countries in which the same Additionally, two not listed here PROs from Reuters' Top 100

invention is patented – is a very important indicator of patent report (2019) were selected [14].

quality [9]. Due to the expenses associated with obtaining patents

in various regions, patent holders typically choose to protect their For the U.S.:

most valuable inventions internationally. Besides considering • Massachusetts Institute of Technology (MIT) which was

raw family size, such as in this case, one variation of this method ranked at the world’s second place in this report.

is to look at triadic patents, which cover an invention in the three For Europe:

principal markets: the U.S., Japan, and the European Patent • KU Leuven (which was the top rank in Europe and took

Office (EPO). Alternatively, transnational patents, defined as seventh place on the Reuters' Top 100 report). patent families with at least one filing with the EPO or under the

Patent Cooperation Treaty (PCT), can be considered [10]. To access indicators of patent value for these selected PROs the

Orbis Intellectual Property database (Orbis IP) was used. Orbis

Backward citations reveal the prior art or existing knowledge that IP contains over 145 million patents linked to detailed company

a new patent builds upon. They are added by patent applicants, information and ownership structures [15]. examiners, and also by third parties (e.g. during opposition

proceedings), and are often used as measures of knowledge Excel was used to sort the data and draw the chart, and the open-

transfer [11]. A patent with numerous and relevant backward source program JASP was used for statistical analysis. We used

citations indicates that the patent applicants or inventors or the Student’s t-test (also called T-test) to compare the means

attorneys and examiners conducted a comprehensive search of between two groups [16], in the presented case, Europe and the

prior art. Such patents may also be less vulnerable to legal U.S.

challenges and can be protected from being invalidated due to

overlooked prior art. Additionally, if a patent references



patented invention is building on well-established and important From the selected institutions, we can first notice that in Europe, technology, potentially indicating a higher-quality patent. there are three research institutes listed and three universities, foundational and high-impact prior patents, it suggests that the 4 RESULTS



Forward citations are commonly used to measure the while in the U.S., there are five universities and only one research institute. technological impact of innovation [11]. We can say that this



indicator is the most understandable to us, as we are already Figure 1 below shows that selected European PROs outnumber familiar with it from scientific articles: when later patents quote the U.S. PROs in patents in the last at least 65 years. However, an earlier one, it suggests that the earlier patent has contributed since there are no reliable and comparable data about these to new developments in the field. The more forward citations a organisations' date of establishment, size and income (which can patent receives, the more significant its impact on subsequent all affect the presented number of patents), it is not possible to technological improvements. make any comparisons or conclusions.

Among these four indicators, the two most important can be

considered: 1) patent family size for reflecting the potential

commercial success of an invention and 2) forward citations,

which indicate the technological/scientific impact of the

invention.



16000 Number of patents from 1940 Group N Mean SD SE Coefficient of variation Number of Europe 112918 10.099 18.835 0.056 1.865 on family members U.S. 49752 8.610 10.977 0.049 1.275



12000 backward citations 14000 Number of Europe 112918 3.027 9.733 0.029 3.215 10000 U.S. 49752 6.789 27.966 0.125 4.120



8000 Number of Europe 112918 0.806 3.649 0.011 4.526 6000 forward 4000 citations U.S. 49752 2.960 11.729 0.053 3.962 2000

0

A closer look at individual PROs' patents reveals considerable



1940 differences between them. In the number of claims, MIT is the 1946 1952 1958 1964 1970 1976 1982 1988 1994 2000 2006 2012 2018 leading PRO with an average of 28 claims. In the number of

EU US family members (Table 3), Fraunhofer is the leader (with a mean

of more than 17 family members), followed by KU Leuven (with

more than 11 family members). PRO with the highest number of

Figure 1: Comparison of no. of patents of current top PROs backward citations is MIT again, but the leading PRO in the

from 1940 on number of forward citations (Table 4) is Carnegie Mellon

University, with a mean of 4,18. The best European PRO in the

Moreover, for the answer to the presented research question the number of forward citations is ETH, with a mean of 2,42.

past is not so important as in the current situation. Therefore,

patents from these organisations only from the last ten years were Table 3: Descriptive statistics - Number of family

selected, i.e., from 2014 on. members



In Table 1 below, we can see the results of the T-test. All the Atomiqu Columbi Copenha Carnegie Center ETH Florida Fraunhof Harvard Leuven MIT Stanford e National a gen er

differences in means are statistically significant (p < 0,05).

Descriptive statistics in Table 2 show us that U.S. PROs are Valid 35282 2860 34168 8196 1427 26 10191 37257 236 4758 15640 12629

better than European in the number of claims and backward and

forward citations. However, European PROs are better than the Mean 5,035 4,144 7,328 8,590 8,473 9,846 6,076 17,335 11,169 11,387 9,539 10,481

U.S. regarding the number of family members.

deviation 3,680 4,733 7,298 10,217 9,001 7,412 6,777 29,953 11,778 15,114 13,095 11,641 Std.



Table 1: Comparison of European and U.S. PROs (patents Minimum 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000



from 2014-2024) Maximu 48,000 29,000 147,000 130,000 44,000 22,000 52,000 300,000 93,000 74,000 106,000 78,000 m



Independent Samples T-test



t df p Table 4: Descriptive statistics - Number of forward

Number of claims -91,101 162668 < ,001 citations



Number of family members 16,447 162668 < ,001 Atomiqu Center Columbi Copenha Fraunhof Carnegie ETH Florida Harvard Leuven MIT Stanford e Number of backward citations -40,025 162668 < ,001 National a gen er

Number of forward citations -55,886 162668 < ,001 Valid 35282 2860 34168 8196 1427 26 10191 37257 236 4758 15640 12629



Mean 0,849 4,180 0,629 2,319 0,800 2,423 1,947 0,871 0,915 1,251 4,102 2,543



Std.

11,998 2,583 9,115 3,238 8,339 6,242 3,593 8,873 9,209 16,149 9,815

deviation 3,246



Minimum 0,000 0,000 0,000 0,000 0,000 0,000 0,000 0,000 0,000 0,000 0,000 0,000



Maximu

162,000 260,000 117,000 232,000 53,000 41,000 181,000 126,000 130,000 396,000 410,000 259,000

m



Table 2: Group descriptives (patents from 2014-2024)

5 DISCUSSION AND CONCLUSION



Number of Group Results of this study show something quite the opposite of the N Mean SD SE Coefficient of variation European paradox, which suggests that while European scientific



claims Europe lags behind in converting research results into U.S. 49752 23.823 29.331 0.131 1.231 innovations and gaining a competitive advantage [17]. European Europe 112918 14.082 13.774 0.041 0.978 performance is on par with its main international competitors,



paradox is a term that describes Europe's strength in basic science gap with Europe by filing and enforcing its patents in more

but its perceived lag in technological applications in the global countries.

market (for example, compared to the U.S.).



In this study, the top European and U.S. PROs were compared.

Results show that the U.S. top PROs are much stronger in the REFERENCES scientific performance of their patents: in the last ten years, an

average patent from top U.S. organisations received 3 forward



citations while a patent from top European organisations [1] Crespi, G. A., Geuna, A., & Verspagen, B. (2006). University IPRs and received only 0,8. Therefore, the scientific or technological knowledge transfer. Is the IPR ownership model more efficient. 6th influence of U.S. patents is more than three times higher than that Annual Roundtable of Engineering Research, Georgia Tech College of Management, 1-3. of Europe. [2] Vinig, G. T., & van Rijsbergen, P. (2009). Determinants of university

technology transfer-Comparative study of U.S., Europe and Australian

On the other hand, European PROs demonstrate larger patent universities. Europe and Australian Universities (January 8, 2009).

families than those of the U.S., which indicates a stronger [3] Bolzani, D., Munari, F., Rasmussen, E., & Toschi, L. (2021). Technology

emphasis on protecting intellectual property across multiple transfer offices as providers of science and technology entrepreneurship

jurisdictions and, thus, also a broader market potential for education. The Journal of Technology Transfer, 46, 335-365.

[4] Conti, A., & Gaule, P. (2011). Is the U.S. outperforming Europe in

patented inventions. That said, European inventions are much university technology licensing? A new perspective on the European

more focused on commercialisation or “competitive advantage”. Paradox. Research Policy, 40(1), 123-135. [5] Munari, F., & Oriani, R. (Eds.). (2011). The economic valuation of



U.S. PROs are also better than Europe’s in the number of claims patents: methods and applications. Edward Elgar Publishing. [6] Werner, D., & Dang, H. (2021). Patent Valuation Using Citations: A



and backward citations, but these indicators may not be so Review and Sensitivity Analysis. Journal of Business Valuation and Economic Loss Analysis, 16(1), 41-59. important for commercial and scientific/technological success. [7] Og, J. Y., Pawelec, K., Kim, B. K., Paprocki, R., & Jeong, E. (2020).

Measuring patent value indicators with patent renewal information.



patent claims, as well as backward and forward citations, and 16. [8] Squicciarini, M., Dernis, H., & Criscuolo, C. (2013). Measuring patent reduce the gap with the U.S., drafting patents more carefully with To help European PROs improve in terms of the number of Journal of Open Innovation: Technology, Market, and Complexity, 6(1),

more detected prior art can be suggested. This will result in more quality: Indicators of technological and economic value.

https://www.oecd-ilibrary.org/science-and-technology/measuring-patent-



backward citations of a particular patent and also in forward quality_5k4522wkw1r8-en [9] Nagaoka, S., Motohashi, K., & Goto, A. (2010). Patent statistics as an citations of quoted patents. It is also important to encourage innovation indicator. In Handbook of the Economics of Innovation (Vol. collaboration between different PROs and between PROs and 2, pp. 1083–1127). Elsevier. industry. Partnerships can create more comprehensive and https://www.sciencedirect.com/science/article/pii/S0169721810020095 [10] Kabore, F. P., & Park, W. G. (2019). Can patent family size and impactful patents that include more claims and are more composition signal patent value? Applied Economics, 51(60), 6476–6496. frequently cited. https://doi.org/10.1080/00036846.2019.1624914

[11] Aristodemou, L., & Tietze, F. (2018). Citations as a measure of

In conclusion, while the study highlights significant differences technological impact: A review of forward citation-based measures.

between European and U.S. PROs in terms of patent World patent information, 53, 39-44.

performance, it also points to areas where European PROs can [12] Heartland Forward (2022). Research to Renewal: Advancing University

Tech Transfer. Available at: https://heartlandforward.org/wp-



investigating the underlying factors contributing to these 17.08.2024). [13] European Research Ranking (2020). Institution Ranking 2020. Available disparities, particularly by examining how patent drafting enhance their impact. Future research should focus on content/uploads/2022/05/ResearchToRenewal.pdf (assessed:

practices, collaboration networks, and industry linkages affect at:

http://www.researchranking.org/index.php?orgtype=ALL&c=5&country



patent quality and citation rates. It should also be noted that this =&year=2020&action=ranking (assessed: 17.08.2024). [14] Reuters Top 100 (2019). The World’s Most Innovative Universities 2019. study referred to the top six PROs from each continent, and Available at: https://www.reuters.com/graphics/AMERS-different results might have been obtained if all PROs were REUTERS%20RANKING-INNOVATIVE-considered. But in any case, a methodological approach which UNIVERSITIES/0100B2JP1W1/ (assessed: 17.08.2024). [15] GOV.UK (2024). Orbis Intellectual Property. Available at: can combine quantitative analysis of patent metrics with case https://www.applytosupply.digitalmarketplace.service.gov.uk/g-studies of successful collaborations could provide deeper cloud/services/762715914988187 (assessed: 17.08.2024). insights into the mechanisms that drive patent performance. [16] Mishra, P., Singh, U., Pandey, C. M., Mishra, P., & Pandey, G. (2019).



Additionally, exploring policy interventions and strategies to Application of student's t-test, analysis of variance, and covariance. Annals of cardiac anaesthesia, 22(4), 407-411. strengthen technology transfer offices and foster innovation [17] Nagar, J. P., Breschi, S., & Fosfuri, A. (2024). ERC science and invention: ecosystems in Europe could offer actionable solutions to close Does ERC break free from the EU Paradox?. Research Policy, 53(8), the gap with the U.S. The U.S., on the other hand, may close the 105038.





The Importance of Technology Transfer Offices in




University Industry Collaboration: KTÜ TTM Example



Dilek İSKENDER BALABAN* Beril DEĞERMENCİ Harun YÜKSEL

Technology Transfer ARC Technology Transfer ARC Finance Department

Karadeniz Technical University Karadeniz Technical University Avrasya University

Trabzon, Türkiye Trabzon, Türkiye Trabzon, Türkiye

dilekiskender@ktu.edu.tr berildegermenci@ktu.edu.tr harun.yuksel@avrasya.edu.tr

Eren YILMAZ Sedanur KALYONCU Emrah AYVAZ

Technology Transfer ARC Technology Transfer ARC Technology Transfer ARC

Karadeniz Technical University Karadeniz Technical University Karadeniz Technical University

Trabzon / TÜRKİYE Trabzon / TÜRKİYE Trabzon / TÜRKİYE

erenyilmaz@ktu.edu.tr sedanursaglam@ktu.edu.tr emrahayvaz@ktu.edu.tr



Oktay YILDIZ Yalçın AYKUT İslam YILDIZ

Technology Transfer ARC Technology Transfer ARC Technology Transfer ARC

Karadeniz Technical University Karadeniz Technical University Karadeniz Technical University

Trabzon / TÜRKİYE Trabzon / TÜRKİYE Trabzon / TÜRKİYE

oktayyildiz@ktu.edu.tr yalcin.aykut@ktu.edu.tr islamyildiz@ktu.edu.tr



Gözde SAĞLAM Müslüm Serhat ÜNVER Hülya SABIR

Technology Transfer ARC Technology Transfer ARC Technology Transfer ARC

Karadeniz Technical University Karadeniz Technical University Karadeniz Technical University

Trabzon / TÜRKİYE Trabzon / TÜRKİYE Trabzon / TÜRKİYE

gozdesaglam@ktu.edu.tr serhatunver@ktu.edu.tr hulyahacisalihoglu@ktu.edu.tr



Kerim SÖNMEZ Güler Tuğba GÜLTEKİN Aleyna AYDIN

Technology Transfer ARC Technology Transfer ARC Technology Transfer ARC

Karadeniz Technical University Karadeniz Technical University Karadeniz Technical University

Trabzon / TÜRKİYE Trabzon / TÜRKİYE Trabzon / TÜRKİYE

kerimsonmez@ktu.edu.tr gulertugbagultekin@ktu.edu.tr aleynaaydin@ktu.edu.tr



ABSTRACT 1.INTRODUCTION



evaluated the place of Karadeniz Technical University universities, research institutes, students, investors, and Technology Transfer Application and Research Center (KTÜ companies. They engage in activities such as establishing TTM) within the framework of university-industry collaboration. connections and making matchings according to the necessary In the study, qualitative research technique was used and needs. “phenomenology” was used as the research design. The data in In Türkiye, TTOs provide consultancy and support to all this research was examined with the descriptive content analysis stakeholders throughout the entire process, from transforming method. Analysis was conducted in RStudio in order to analyze knowledge into products, selecting industrial partners, the collected qualitative data and determine emotional tendencies. identifying appropriate funding sources, project development According to the analysis results; It was seen that KTÜ TTM activities, intellectual property and industrial property rights the participants on the subject in order to answer the question national and international research projects, facilitating its “What is the place of technology transfer offices within the transfer to industry, and commercializing it. In general, TTOs framework of university-industry collaboration?” This study located within universities act as intermediaries between Within the scope of the research, interviews were conducted with structures that play a role in directing academic research towards Technology Transfer Offices (TTOs) are organizational



made significant contributions to raising the bar of success by applications, to commercialization and/or the establishment of using the potential of the university and had a positive effect in academic-based firms [10]. general. When similar studies in the literature are examined, it is As intermediary organizations, TTOs operate according to seen that TTOs play an important role in university-industry various strategies based on the past experiences of academic and collaboration. This study supports the theoretical discussions in industrial actors and the quality of the information conveyed in the literature with a practical example. Since the study provides the university-industry collaboration process. TTOs particularly an evaluation specific to KTÜ TTM, it makes a local and specific focus on enhancing cognitive and organizational domains. They contribution to the literature by examining the effects of TTOs in play a crucial role in bringing together actors with different a different university and geographical region. This could fill the visions, ways of interpreting life, and perspectives on the world gap in the literature on the functioning of TTOs across different [11]. institutions and regions. The collaboration between two distinct entities, universities and



KEYWORDS is achievable only if the process is managed with sound and industry, can contribute to national development; however, this

Technology, technology transfer office, university industry appropriate strategies. Nowadays, while collaborations between

cooperation, KTÜ TTM universities and industry can occur through various



*Dilek İSKENDER BALABAN is corresponding author of the ARC for

Technology Transfer Karadeniz Technical University in Trabzon, Türkiye. communication channels apart from the support of technology university-industry collaboration. Firms are prioritized and

transfer offices, the outcomes of these collaborations are often analyzed, and guidance is provided based on current calls for

quite weak [7]. proposals. Information on academicians' research that can be

A review of the literature reveals that technology transfer offices applicable in the industry or discussions on potential

(TTOs) play a significant role in enhancing and sustaining collaborations with industrial organizations are conducted.

university-industry collaboration. During the application phase of university-industry collaboration

projects, the entire process of project review, preparation, and

1.1 University Industry Collaboration submission of application documents is carried out. Legal



gained from research activities conducted at universities is not legal counsel. Once a project is approved for funding, support such as accounting transactions, completion of documentation, University industry collaboration ensures that the knowledge matters and contracts are prepared jointly with the university's

only published but also transformed into practical applications.

It is a collaborative method aimed at and implementing the and signature processes are provided to firms and academicians

transfer of technological developments and knowledge to by the administrative and financial affairs unit established within

production stages according to industrial needs [13]. The cultural KTÜ TTM.

differences between universities and industry contribute to the

diversification of research approaches [9]. Effective 2. METHODOLOGY

collaboration between industry and academia requires a special 2.1 Research Methodology and Research Design

alignment; understanding mutual interests, setting common In this study, the role of KTÜ TTM within the framework of

goals, and focusing on complementary skills form the basis for university-industry collaboration was evaluated. A qualitative

achieving successful collaboration [6]. research method was employed, and the research design was

As a result of collaboration, industry, whose goal is to increase based on "phenomenology." The phenomenological design

profits and expand its volume, has seen developments that typically focuses on phenomena that are recognized but not

positively impact production through the adaptation of deeply or thoroughly understood. Phenomenology is a method

technologically evolving and renewing processes to existing that concentrates on understanding and evaluating lived

systems. The aim of the university in collaborating with industry experiences [8]. This methodology aims to deeply examine and

is to develop a qualified human resource and support research comprehend individuals' experiences.

with a strong knowledge base, leading to the transformation of Although phenomenological data are obtained from the

theoretical work into practical applications and resulting in some experiences of a few individuals, the information gathered from

modifications [5]. Science plays an extremely important role in these individuals provides detailed insights into the phenomenon.

facilitating university-industry collaborations. It is a process born The fact that the phenomenon is experienced by different

from the mutual supply and demand between the university, individuals contributes to the provision of information from

which produces science, and industry, which converts science various perspectives by the research participants, thereby aiding

into economic benefits [2]. in understanding the phenomenon from a broad viewpoint. In this

University-industry collaboration highlights a partnership that way, the data obtained from the experiences of different

offers significant benefits for both parties. Through these participants support a comprehensive understanding of the

collaborations, universities strive to address global problems phenomenon [3]. For such research, the number of individuals to

using academic knowledge. Industry, on the other hand, benefits be included in the sample should generally not exceed ten. It is

from universities' research, expertise, and laboratories, leading to normal to limit the sample size in this type of research since the

the development of innovative products and improvements to interviews often require long and sometimes multiple meetings.

existing products. As a result of this partnership, mutual gains The limited number of individuals who have experienced the

are achieved in areas such as employment, education, innovation, phenomenon under investigation may also sometimes result in a

and economic growth, which significantly impact life. Thus, restricted number of people who can be included in the sample

these partnerships provide mutual benefits to the parties involved [12].

and contribute to society and the economy. In addition to their In addition, sentiment analysis was performed using RStudio to

research mission, universities also have educational and societal automatically detect and classify emotional expressions present

missions. While the educational mission is clear, societal in the texts. This analysis employs Natural Language Processing

missions have gained increasing importance in recent years. This (NLP) techniques to determine whether the sentiments in the

is reflected in factors such as the role of universities in university- texts are positive, negative, or neutral. During this process, words

industry technology transfer [4]. and expressions within the texts are analyzed to identify the

emotional content.

1.2 University Industry Collaboration Activities of To visually understand the key themes, topics, and word

KTÜ TTM distribution in the texts, a word cloud was generated in RStudio.

In order to increase R&D and innovation capacity and strengthen Word clouds, commonly used as part of text mining and data

university-industry cooperation activities, KTÜ TTM establishes visualization techniques, provide a quick representation of the

contacts with many new companies every year and develops frequency of words in a text or text corpus, indicating which

bilateral cooperation. These efforts are not limited to the region words are used more frequently.

but extend to firms across the country through online and face-

to-face meetings, integrating new companies into the 2.2 Universe and Sample of the Research

collaboration ecosystem. During these meetings, R&D topics The universe of this research consists of approximately 50

and requests are gathered, the needs of the firms are identified, faculty members working at KTÜ who have been involved in

and numerous firms are matched with academicians from KTÜ university industry collaboration processes. To align with the

for collaborative projects, involving online meetings and research objectives, the sample group was composed of 8 faculty

discussions. members who have both participated in university industry

Meetings are also organized with the boards of Organized collaboration processes and are knowledgeable about the KTÜ

Industrial Zones to discuss activities within the framework of TTM. The academic titles, faculty and department affiliations,



and the total number of projects funded by public and/or private 2.4 Research Findings

sector capital for the faculty members included in the sample The findings of the study are summarized as follows:

group are provided in Table 1. KTÜ TTM has made significant contributions to raising the bar



2.3 Data Collection Processes and Interview Questions of success by utilizing the potential of the university. Active

TTOs are essential units that every university must have. They



relevant participants to answer the question, "What is the role of institutional in industrial collaborations and play a critical role in KTÜ TTM in the context of university-industry collaboration?" Within the scope of the research, interviews were conducted with are crucial in presenting the university as professional and



recorded by the researcher, but their personal data would not be KTÜ TTM has been highly effective in conducting one-on-one shared with third parties. It was explained that the audio meetings with firms, matching academics, providing project recordings would be used for the purpose of data collection and writing support, analyzing industrial problems, and guiding both analysis. Initially, a pool of questions presumed to be relevant to Participants were informed that the interviews would be audio development process. reducing the risks researchers may encounter during the project



this pool were evaluated with experts deemed relevant to the the university’s recent collaborations. However, it is observed research content, and the most appropriate eight questions for the that KTÜ TTM faces a disadvantage due to its location being far study were finalized. The interview questions prepared for the from major industrial areas. It is suggested that the center could the study topic was created. Subsequently, the questions within process. It has also played a significant role in realizing many of parties on the appropriate course of action throughout the



participants are presented in Table 2. Before the interview become more effective by organizing events where industry questions were posed, a conversation with the participants was professionals and academics can come together and by placing initiated to foster a mutual trust relationship. After the audio greater emphasis on institutionalization. recordings were transcribed, with the permission of the It is noted that KTÜ TTM is preferred as an intermediary because participants, the next step was the analysis of the data. The it instills confidence in the industrial sector during company collected data was analyzed using descriptive content analysis, visits and ensures that academics feel secure. Its professional and and additionally, sentiment analysis was conducted, and a word corporate identity during industry visits, which represents the cloud was generated using the RStudio program. university, leads to a more positive and moderate view of the



Table1: Characteristics of the Study Sample Group project development processes among industrialists. Additionally, KTÜ TTM is favored for its objective approach to Number of Public Academic both academics and industry parties, its facilitation of smooth Faculty/Departments /Private Sector Title process progress, its role as a mediator, and its handling of Supported Projects accounting tasks. Forestry Faculty Professor 10 KTÜ TTM is believed to be doing its best to achieve its goals. Forestry Industrial Engineering Additionally, there are expectations for bringing academics Faculty of Engineering Professor 6 along on company visits, conducting matching processes more Mechanical Engineering meticulously, collecting project topic requests from academics Faculty of Engineering Professor 2 Mechanical Engineering based on company activity areas, and matching academics with

Associate Faculty of Engineering large-scale companies in the Technopolis where they have their

Professor Industrial Engineering firms. 4

Assistant Faculty of Science It is generally believed that KTÜ TTM does not have significant

Professor Computer Science shortcomings. However, suggestions have been made, including 6

Assistant Faculty of Science collecting R&D topic proposals from academics and forwarding

Professor Computer Science them to companies, grouping companies sectorally to hold 6

Vocational School of Health meetings with academics on specific days, providing support

Assistant Services Medical Services and 3 with sample project forms, and facilitating discussions and

Professor Techniques integration between academic entrepreneurs and companies.



Research Forestry Faculty To increase its activity, it is suggested that KTÜ TTM could 2 assistant Forestry Industrial Engineering increase its participation in fairs, fix the names of the companies

it works with on its website, enhance materials for promoting

Table 2: İnterview Questions TTO's module functions and staff, utilize international resources,

No İnterview Questions and raise awareness among companies about TTO activities.

1 It is believed that TTO plays a facilitative role in reviewing and

What are your opinions about KTÜ TTM? preparing contracts between academics and industrialists,

2 How would you define university industry collaboration? managing financial obligations, handling bureaucratic processes,

3 Do you think KTÜ TTM is effective in the processes of and establishing balances between the company and the

university industry collaboration? academic. Additionally, it is noted that TTO helps eliminate

4 Why would you prefer KTÜ TTM to be an intermediary problems by coordinating the project development processes for

in university industry collaboration processes? companies that are located far away.

5 What are your expectations regarding KTÜ TTM's The data obtained from the interviews was converted into a text

university industry collaboration module? file and sentiment analysis was performed using the RStudio

6 Is there any aspect of KTÜ TTM's university industry program. The graph showing the sentiment scores obtained from

collaboration processes that you find lacking? the analysis is presented in Figure 1.

7 In your opinion, how could KTÜ TTM become more Subsequently, a word cloud was created using the Rstudio

active in the context of university industry collaboration? program to analyze frequently used words within the text. The

8 resulting word cloud is presented in Figure 2. Has the solution/process of the problems you experienced

in university-industry collaborations at KTÜ TTM

become easier?

Figure 1: Distribution of Sentiment Scores In the word cloud presented in Figure 2, the words 'industry,'

'university,' and 'TTO' are prominently featured. The frequent



occurrence of the word 'industry' indicates a strong focus on how

KTÜ TTM interacts with and supports industrial partners. The

frequent mention of the word 'university' underscores the

importance of the academic side of the collaboration, suggesting

that researchers view the university's role as critical in partnering

with industry. The prominent presence of the abbreviation 'TTO'

highlights the central role of KTÜ TTM in facilitating these

collaborations. Overall, the word cloud demonstrates the

significant role that KTÜ TTM plays in supporting and

facilitating these interactions.

The current situation of KTÜ TTM has been evaluated, and the

following recommendations have been proposed:

Increasing awareness of the services provided by TTO and

conveying this awareness to the business ecosystem will enhance

Figure 2: Word Cloud the sustainability of new collaborations. Organizing events that

bring together universities and industry can foster more



communication between them. The strong relationships

established will increase the sense of trust, thereby creating

opportunities for further collaboration. Additionally, such efforts

will create internship and job opportunities for students trained

at the university for the business world.

In future studies, the place of TTOs can be examined within the

framework of commercialization of inventions within the

university and/or increasing academic entrepreneurship. By

increasing the number of study samples, the subject can be

analyzed in depth with different analysis methods and theoretical

frameworks can be tested. In addition, the role and impact of

TTOs in different universities, regions and different countries in

university-industry collaboration can be examined by conducting

multiple case studies. It will be useful to compare different



3. CONCLUSION AND RECOMMENDATIONS structures and operations of TTOs in terms of examining the impact of different regional and sectoral dynamics on In conclusion, it has been observed that the activities carried out collaboration processes. by the KTÜ TTM within the framework of university industry



collaboration have yielded positive results, as noted by REFERENCES participants who are familiar with the structure and functioning



of the TTO and have been involved in the university industry [1] Abdulai, A. F., Murphy, L., Thomas, A., & Thomas, B. (2022), Technology Transfer Offices and Their Role with Information Mechanisms for Innovation collaboration processes. Furthermore, it has been concluded that Performance in Firms: The Case of Ghana. Knowledge, 2(4), 719-734. the TTO has positively influenced its corporate image. The study [2] Akdoğan, A. (2007), The Functionality of Research and Application Centers

[1] supports this finding, as it concluded that TTOs positively in Universities (s. 85-104). Ankara: Detay Publishing

impact the innovation capacity within firms, with 47% of firms' [3] Baker, C., Wuest, J. & Stern, P.N. (1992), “ Method Slurring: the Grounded

innovation capacity being provided by TTOs. Theory/ Phenomenology Example”. Journal of Advanced Nursing, 17, 1355-

1360.

Upon evaluating the sentiment analysis scores presented in [4] Blankesteijn, M., Bossink, B., & van der Sijde, P. (2021), Science-based



Figure 1, it is observed that a high score corresponds to a positive entrepreneurship education as a means for university-industry technology transfer. International Entrepreneurship and Management Journal, 17(2), 779- sentiment. This result indicates that the university-industry 808.

collaborations carried out with KTÜ TTM generally have a [5] Eryanık, A. (2018), The Impact of University-Industry Collaboration on

positive impact. It demonstrates that stakeholders being satisfied Employment: The Case of Uşak. Published Master's Thesis. Afyonkarahisar:

with the collaboration and experiencing positive outcomes. Afyon Kocatepe University.

The high level of trust indicates that KTÜ TTM has established [6] Fasi, M. A. (2022), An Overview on patenting trends and technology

commercialization practices in the university Technology Transfer Offices in

a strong trust relationship between university and industry USA and China. World Patent Information, 68, 102097.

stakeholders and is recognized as a reliable partner. The high [7] Güler, M., & Kırbaşlar, İ. (2020), The Role of Technology Transfer Offices in

level of anticipation reflects the high expectations for future University-Industry Collaboration, Iktisad Publishing

[8] Jasper, M. A. (1994), “Issues in Phenomenology for Researchers of Nursing”.

projects and potential opportunities in collaborations with KTÜ Journal of Advanced Nursing, 19, 309-314.

TTM. [9] O’Dwyer, M., Filieri, R., & O’Malley, L. (2023), Establishing Successful

Based on the sentiment analysis results, we can conclude that University İndustry Collaborations: Barriers And Enablers Deconstructed. The

KTÜ TTM's role and significance in university industry Journal of Technology Transfer, 48(3), 900-931.

[10] Temel, S., Dabić, M., Ar, I. M., Howells, J., Mert, A., & Yesilay, R. B. (2021),

collaboration are highly positive. The high levels of positive Exploring The Relationship Between University İnnovation İntermediaries



emotions and trust demonstrate that the collaborations are being And Patenting Performance. Technology in Society, 66, 101665. [11] Villani, E., Rasmussen, E., & Grimaldi, R. (2017), How İntermediary conducted successfully and that stakeholders are satisfied, while Organizations Facilitate University–İndustry Technology Transfer: A the low levels of negative emotions suggest that the processes are Proximity Approach. Technological Forecasting And Social Change, 114, 86-

running smoothly and are being managed effectively. This 102.

proves that KTÜ TTM is a reliable and effective interface that [12] Yıldırım, A. & Şimşek, H. (2013), Qualitative Research Methods in Social

strengthens the collaboration between the university and [13] Yücel, İ. H. (1997), Science and Technology Policies and the 21st Century Sciences (9th Edition). Ankara: Seçkin Publishing. industry. Society. In Science and Technology Policies and the 21st Century Society (p.

76). Ankara: State Planning Organization.



The Impact of International Networks on Grants, R&D,



Knowledge and Technology Transfer - Case of COST



Network and KTU



Sedanur KALYONCU İslam YILDIZ Emrah AYVAZ

Technology Transfer ARC Technology Transfer ARC Technology Transfer ARC

Karadeniz Technical University Karadeniz Technical University Karadeniz Technical University

Trabzon / TÜRKİYE Trabzon / TÜRKİYE Trabzon / TÜRKİYE

sedanursaglam@ktu.edu.tr islamyildiz@ktu.edu.tr emrahayvaz@ktu.edu.tr



Gözde SAĞLAM Müslüm Serhat ÜNVER Oktay YILDIZ

Technology Transfer ARC Technology Transfer ARC Technology Transfer ARC

Karadeniz Technical University Karadeniz Technical University Karadeniz Technical University

Trabzon / TÜRKİYE Trabzon / TÜRKİYE Trabzon / TÜRKİYE

gozdesaglam@ktu.edu.tr serhatunver@ktu.edu.tr oktayyildiz@ktu.edu.tr



Güler Tuğba GÜLTEKİN Eren YILMAZ Kerim SÖNMEZ

Technology Transfer ARC Technology Transfer ARC Technology Transfer ARC

Karadeniz Technical University Karadeniz Technical University Karadeniz Technical University

Trabzon / TÜRKİYE Trabzon / TÜRKİYE Trabzon / TÜRKİYE

gulertugbagultekin@ktu.edu.tr erenyilmaz@ktu.edu.tr kerimsonmez@ktu.edu.tr



Hülya SABIR Beril DEĞERMENCİ Yalçın AYKUT

Technology Transfer ARC Technology Transfer ARC Technology Transfer ARC

Karadeniz Technical University Karadeniz Technical University Karadeniz Technical University

Trabzon / TÜRKİYE Trabzon / TÜRKİYE Trabzon / TÜRKİYE

hulyahacisalihoglu@ktu.edu.tr berildegermenci@ktu.edu.tr yalcin.aykut@ktu.edu.tr



Ayhan KOÇ Aleyna AYDIN Dilek İSKENDER BALABAN

Technology Transfer ARC

Technology Transfer ARC Technology Transfer ARC

Karadeniz Technical University

Karadeniz Technical University Karadeniz Technical University Trabzon / TÜRKİYE Trabzon / TÜRKİYE Trabzon / TÜRKİYE

aleynaaydin@ktu.edu.tr

ayhankoc@ktu.edu.tr dilekiskender@ktu.edu.tr



Seda BAŞ

Technology Transfer ARC

Karadeniz Technical University

Trabzon / TÜRKİYE

sedabas@ktu.edu.tr



ABSTRACT involved in the COST program from 2019 to 2023 was examined.

Additionally, Türkiye’s performance during this period was

analyzed to assess the effectiveness of the developed strategy. As

It is essential that researchers collaborate with international

a result of these analyses, in COST Actions in which a limited

colleagues and adhere to international standards to facilitate the

number of KTU researchers were involved at the beginning of

transfer of technology resulting from their research. In order to

2019, 252 researchers were involved in 528 actions by the end of

participate in these projects, it is essential for researchers to have 2023, becoming the first university in Türkiye in this context.

a broad international network and reliable collaborators to form

Thanks to the researchers and activation included in the COST

international consortia. This study examines the strategy adopted

by Karadeniz Technical University (KTU) in its pursuit of the Programme, a 92.3% increase in the number of international

project applications and a 366% increase in the number of project

COST (European Cooperation in Science and Technology)

acceptances were observed between 2019 and 2023. The

Programme and the subsequent results that enabled its

findings indicated that these exemplary practices could serve as

researchers to engage in prestigious international consortia and

an effective approach for the internationalisation of higher

access international funding sources. To increase the number of

education institutions.

Karadeniz Technical University (KTU) members in the

Management Committee (MC) and Working Group (WG) of

COST Actions, which have limited slots, each Action was KEYWORDS

carefully reviewed, and individual meetings were held with Technology Transfer, R&D, Internationalisation, COST researchers. Technical and administrative support was provided

to facilitate researchers' participation in the Actions. A statistical

analysis was conducted for researchers who participated in

COST Actions over the past five years. The status of countries



1 COST provide invaluable opportunities for researchers to INTRODUCTION

disseminate their knowledge and experience [5].

The involvement of researchers in internationally funded

projects presents a duality of opportunities and challenges. One

of the most significant challenges encountered during this 2 METHODOLGY

process is the formation of international consortia. Effective In accordance with the internationalisation strategy for KTU to

communication between researchers from different countries and become more effective in the international arena, activities are

disciplines is of great importance, as consortia require such a being undertaken with the objective of enhancing international

diverse group to come together [1]. Furthermore, operational collaboration and increasing the quantity and calibre of project

challenges, including the sourcing of funding, the optimal proposals submitted to international funding programmes. Given

utilisation of resources and the effective management of projects, the crucial role that a broad international network plays in

represent significant obstacles in the context of such projects [2]. securing participation in internationally funded projects,

Nevertheless, the intricacy of these procedures and the presence concerted efforts have been made to direct students towards the

of bureaucratic impediments can act as a deterrent for numerous COST Programme since late 2019, with ongoing initiatives still

researchers [3]. The existing literature frequently emphasises that in place.

participation in international projects has positive effects on In this context, awareness-raising activities, which commenced

researchers' career development, knowledge sharing and with information events held at various locations across the

innovation [4]. In particular, the formation of international university and its constituent departments, have yielded tangible

consortia hinges on the existence of reliable networks and outcomes through one-to-one interviews with academics. In this

cooperative networks, which are pivotal for the success of the regard, the strategy pursued by KTU Technology Transfer

projects. In this context, the effective utilisation of international Application and Research Center experts to enhance the

cooperation networks facilitates the efficiency and sustainability involvement of researchers in COST Actions is outlined below.

of projects. Due to the inability of KTU researchers to engage in - Organisation of information events:

sufficient international cooperation, it was observed that the - Analysing current COST actions: Existing COST actions are

university was quantitatively and qualitatively insufficient in analysed and listed with details covering objectives, research

international projects, resulting in limited scientific output and areas, participation requirements, duration, etc.

weakened competitiveness at the global level. This deficiency - Identification of researchers: Researchers with fields of study

prevented the university from fully exploiting its potential, compatible with COST actions are identified by analysing their

especially in areas such as access to international funding, academic profiles (research areas, project experiences, etc.).

exchange of knowledge and experience, and development of -Establishing contact with researchers: The identified

innovative solutions. This study presents the internationalisation researchers are contacted and one-to-one interviews (telephone,

strategy of Karadeniz Technical University, which enables e-mail or desk interviews) are conducted about the opportunities

researchers to participate in qualified international consortia and and benefits of participation in COST actions. It is assessed

access international funding sources, as exemplified by the whether the researchers are suitable for the identified actions.

COST programme. -Providing technical support services: Technical support is

The COST (European Cooperation in Science and Technology) provided to researchers during the application process for COST

programme was established in 1971 with the objective of actions (filling out the application form, preparation of necessary

promoting scientific and technological research in Europe. The documents, follow-up of the application process, etc.) and the

objective of COST is to facilitate the exchange of knowledge and application evaluation process is followed.

encourage innovation among researchers by fostering -Guidance to other COST-related support: Researchers are

interdisciplinary networks. The principal objective of the provided with the opportunity to benefit not only from MC/WG

programme is to facilitate international collaboration and enable assignments in actions, but also from other COST-related

researchers to collectively address global challenges. While support such as Short-Term Scientific Visit (STSM) and ITC

COST does not provide direct support for research and Conference Support. In addition, referrals were made to the

development, it plays a significant role in facilitating the COST 2515 Programme supported by TUBITAK, the COST

formation of international consortia, which allow researchers to researchers. National Coordinator in Türkiye, to provide R&D support to

collaborate on the advancement of their projects. Researchers The country-based data used in the study were obtained from the

may participate in the Actions in either the capacity of a member COST Annual Reports and the researcher-based data were

of the Management Committee (MC) or a member of a Working obtained from the COST website.

Group (WG). In these roles, they are afforded the opportunity to In order to assess the status of the developed strategy, a mini

engage in in-depth strategic planning, project management, and survey with open-ended questions was conducted with KTU

the exploration of specific research topics. Those engaged in researchers participating in COST Actions. In the survey, the

COST Actions enjoy significant advantages, including access to benefits of the participated actions for the academics were

information and resources, opportunities for career development evaluated with questions such as to what extent they benefited

and the strengthening of leadership skills as part of an from the action, what kind of effects it had at the academic level,

what kind of activities they participated in.

internationally recognised network. Such opportunities permit



researchers to make significant progress in their careers and to 3 RESULTS become more prominent figures within the international The COST programme facilitates extensive involvement from a scientific community. COST Actions facilitate the formation of diverse array of countries, encompassing 41 member countries new collaborative relationships and the expansion of existing and cooperation countries. In addition to Europe's leading networks, thereby enhancing the probability of securing countries in science and research, such as Germany, France, additional funding and support for research projects. Italy, Spain and the United Kingdom, Türkiye, Israel and some Furthermore, meetings, workshops and conferences organised by Western Balkan countries are also actively involved in this

programme. Table 1 presents detailed data illustrating the status institutions. In 2021, participation declined as a consequence of

of active participation in the COST programme between 2019 the impact of the pandemic. However, in 2022, there was a

and 2023 [6-10]. revival in participation and a success was achieved as in 2020.

This success continued to increase in the following years. As

Table 1: evidenced by the participation statistics provided by the COST Status of Countries in COST Programme

Organisation, Türkiye achieved notable success in 2022 and

Indicator 2019 2020 2021 2022 2023 2023. In 2022, Türkiye achieved the distinction of becoming the

Runnig third most participating country, with a participation rate of 99% COST

Action 294 291 289 302 269 in all active actions. Additionally, the country reached a notable

New COST action number of members, with 3,084 individuals participating in

launched 80 45 40 70 70 Working Groups. In the same year, Türkiye was the fifth most

Average number of successful country in terms of individual participation in COST

COST network activities, with 1,113 participants, and the fourth Members 30 30.8 31 33 33

per Action country with the highest budget allocation of approximately EUR

Average number of 1.5 million. Türkiye was the leading country in terms of

non-COST participation by young researchers, with a rate of 52.8%. In 2023,



countries per actions, thereby attaining the distinction of being the country - 4.3 6 5 6 Türkiye sustained its efficacy by participating in 99% of all Action

Articles with the highest number of working group members, with 7,096 479 921 1501 1253 -

Percentage of spin- working group members. Türkiye was the third most successful

off H2020* country in terms of individual participation in COST network 37% 39% 32% - -

proposals approved activities, with 1,849 participants, and the third most budgeted

Average value of country, with approximately EUR 2.27 million.

spin-off As a consequence of the activities conducted throughout this

projects per Action researchers, as well as an increase in the number of researchers 6M 5.8M 3.9M 9.5M 5.2M process, there has been an enhancement in the awareness of

(€) who have submitted applications on an individual basis.

*Horizon 2020 was the EU's research and innovation funding Moreover, the support provided by COST was not confined to

programme from 2014-2020 KTU but was also extended to TTO units and researchers at other

universities, thereby contributing to an increase in Türkiye's

Türkiye plays a significant role as an active participant in the participation in the COST Programme.

COST programme. Türkiye's involvement in COST Actions As the number of researchers engaged in the COST Programme

between 2019 and 2023, along with a comprehensive account of and active in its actions has grown, the graph below illustrates

its contributions and accomplishments, is presented in Table 2 the change in the number of international project applications and

[11-15]. acceptances submitted by KTU between 2019 and 2023 (Fig. 1).



Table 2: Türkiye's Position in COST Programme



Indicator 2019 2020 2021 2022 2023

Individual

participation in all 1075 - 20 1113 1849

action activities

Training

1 12 0 4 10

school/hosted

Short-term

scientific 14 103 8 12 23

missions/hosted

Short-term

Figure 1. Statistics of KTU projects with international funding

scientific 72 380 43 90 158

(2019-2023)

missions/participant



Trainess/participant 197 1001 33 219 310 4 DISCUSSION AND CONCLUSION

Trainers/participant 7 58 1 23 48 The formation of international networks plays a pivotal role in

Budget received (€) the establishment of dependable and collaborative international 0.9M 4.9M 0.3M 1.5M 2.3M

project consortia. This is achieved by the creation of scientific

networks comprising researchers and institutions, which

Türkiye demonstrates a notable level of involvement in the subsequently leads to an increase in the number of international

COST programme, exhibiting a discernible increase in project applications and acceptances.

participation on an annual basis. The number of individual This study examines the strategy employed in the process by

participants increased from 1,649 in 2019 to 103 short-term which KTU researchers were directed to the COST Programme,

scientific missions in 2020, with 380 participants being sent to an international organisation with the objective of uniting

these missions. Furthermore, 12 training schools were conducted scientists who are experts in their respective fields throughout

in 2020, with 1,001 individuals undergoing training at these Europe in scientific networks. This strategy facilitates the



integration of scientists engaged in national research projects into This, in turn, has resulted in KTU researchers establishing more

the international scientific community. robust international networks, which has directly influenced the

The strategy pursued has yielded notable results. At the inception number of international project applications and acceptances. In

of 2019, a modest number of KTU researchers were engaged in conclusion, the findings indicate that these exemplary practices

COST actions. By the conclusion of 2023, this number had may serve as an effective approach for the internationalisation of

grown to 252 researchers participating in 528 actions. With higher education institutions.

regard to the ongoing COST actions, KTU has been the most

successful university in Türkiye, with 29 active members of the REFERENCES

Management Committee. In the context of the ongoing actions, [1] Katz, J. S., & Martin, B. R. (1997). What is research collaboration?. Türkiye is playing a pioneering role, with 252 researchers Research policy, 26(1), 1-18.

engaged in diverse academic pursuits. KTU has become a [2] Wagner, C. S., & Leydesdorff, L. (2005). Network structure, self-

prominent hub for interdisciplinary studies, having participated Research policy, 34(10), 1608-1618. organization, and the growth of international collaboration in science. in approximately 61% of the 305 actions initiated during its five- [3] Olson, G. M., & Olson, J. S. (2000). Distance matters. Human–computer

year internationalisation strategy. interaction, 15(2-3), 139-178.

The COST programme has been instrumental in facilitating a [4] Glänzel, W., & Schubert, A. (2004). Analysing scientific networks through co-authorship. In Handbook of quantitative science and

significant increase in the number of international project technology research: The use of publication and patent statistics in studies

applications, with a 92.3% rise observed between 2019 and 2023. of S&T systems (pp. 257-276). Dordrecht: Springer Netherlands.

Additionally, there has been a notable surge in project [5] COST (European Cooperation in Science and Technology). “What are

acceptances, with a 366% increase during the same period. actions/ COST Actions?”. https://www.cost.eu/cost-actions/what-are-cost-

. 09 August 2024.

Participation in COST Actions was not only associated with an [6] COST 2023, COST Annual Report, COST (European Cooperation in

increase in the number of project applications and acceptances, Science and Technology), 2023.

but also with the administration of surveys to KTU researchers [7] COST 2022, COST Annual Report, COST (European Cooperation in

who took part in the actions in 2021 and 2023. The objective was [8] Science and Technology), 2022. COST 2021, COST Annual Report, COST (European Cooperation in

to ascertain the additional benefits that researchers derive from Science and Technology), 2021.

the COST Programme. The results of the surveys indicated that [9] COST 2020, COST Annual Report, COST (European Cooperation in

the researchers had participated in numerous training Science and Technology), 2020.

[10] COST 2019, COST Annual Report, COST (European Cooperation in

programmes, workshops and conferences, and had developed a Science and Technology), 2019.

network of contacts. They had also published more international [11] COST Türkiye 2023, Türkiye Factsheet, COST (European Cooperation in

collaborative papers, worked on multidisciplinary projects with Science and Technology), 2023.

[12] COST Türkiye 2022, Türkiye Factsheet, COST (European Cooperation in

researchers from other countries, and had access to research and Science and Technology), 2022. laboratory facilities that would not otherwise have been available [13] COST Türkiye 2021, Türkiye Factsheet, COST (European Cooperation in

to them in their home countries. Furthermore, they had Science and Technology), 2021.

[14] COST Türkiye 2020, Türkiye Horizon Factsheet, COST (European

disseminated their work more widely. Cooperation in Science and Technology), 2020. The findings of this study demonstrate that the strategy employed [15] COST Türkiye 2019, Türkiye Factsheet, COST (European Cooperation in

by KTU has led to an increase in participation in COST actions. Science and Technology), 2019.





The Effect of Evaluating Graduate Thesis Topics as




Invention Notification Form on Industrial and Intellectual



Property Applications: The Case of Karadeniz Technical



University



Kerim SÖNMEZ* İslam YILDIZ Yalçın AYKUT

Technology Transfer Application and Technology Transfer Application and Technology Transfer Application and

Research Center, Research Center, Research Center,

Karadeniz Technical University Karadeniz Technical University Karadeniz Technical University

Trabzon / TÜRKİYE Trabzon / TÜRKİYE Trabzon / TÜRKİYE

kerimsonmez@ktu.edu.tr islamyildiz@ktu.edu.tr yalcin.aykut@ktu.edu.tr



Hülya SABIR Gözde SAĞLAM Aleyna AYDIN

Technology Transfer Application and Technology Transfer Application and Technology Transfer Application and

Research Center, Research Center, Research Center,

Karadeniz Technical University Karadeniz Technical University Karadeniz Technical University

Trabzon / TÜRKİYE Trabzon / TÜRKİYE Trabzon / TÜRKİYE

hulyahacisalihoglu@ktu.edu.tr gozdesaglam@ktu.edu.tr aleynaaydin@ktu.edu.tr



Sedanur KALYONCU Müslüm Serhat ÜNVER Ayhan KOÇ

Technology Transfer Application and Technology Transfer Application and Technology Transfer Application and

Research Center, Research Center, Research Center,

Karadeniz Technical University Karadeniz Technical University Karadeniz Technical University

Trabzon / TÜRKİYE Trabzon / TÜRKİYE Trabzon / TÜRKİYE

berildegermenci@ktu.edu.tr serhatunver@ktu.edu.tr ayhankoc@ktu.edu.tr



Emrah AYVAZ Eren YILMAZ Dilek İSKENDER BALABAN

Technology Transfer Application and Technology Transfer Application and Technology Transfer Application and

Research Center, Research Center, Research Center,

Karadeniz Technical University Karadeniz Technical University Karadeniz Technical University

Trabzon / TÜRKİYE Trabzon / TÜRKİYE Trabzon / TÜRKİYE

emrahayvaz@ktu.edu.tr erenyilmaz@ktu.edu.tr dilekiskender@ktu.edu.tr



Beril DEĞERMENCİ Güler Tuğba GÜLTEKİN Oktay YILDIZ

Technology Transfer Application and Technology Transfer Application and Okta Arge Mühendislik Hiz. San ve Tic. Ltd.

Research Center, Research Center, Şti,

Karadeniz Technical University Karadeniz Technical University Üniversite Mh. Hastane Cd.Trabzon

Trabzon / TÜRKİYE Trabzon / TÜRKİYE Teknokent Trabzon / TÜRKİYE

sedanursaglam@ktu.edu.tr gulertugbagultekin@ktu.edu.tr oktayyildiz@hotmail.com



ABSTRACT necessary for KTU that is the application authority, to develop new strategies to increase industrial property assets. This study Industrial and intellectual property is an important structure aims to reveal the effect on the number of patent applications by that is popular all over the world. Each country has legal Karadeniz Technical University (KTU) as a result of the regulations in the field of intellectual and industrial property in evaluation of graduate thesis topics without request. Within the order to protect one's invention. The 6769 Industrial and scope of the new strategy, a methodology was applied for the Intellectual Property Law, which entered into force in 2017, evaluation of patent and utility model application data in the paved the way for universities in Türkiye to have rights in KTÜ patent portfolio, the distribution of data by year, and patent applications for inventions such as patents, utility models and registration documents. In this study, direct patent and utility designs. Thesis studies that young researchers start during their model application data were evaluated. When the application postgraduate period are focused solely on publication. The data was examined, it was seen that the new strategy commercialization and patenting potential of theses determined implemented increased the industrial property assets. without analyzing the needs of the industrial sector is low, and



this makes the thesis work of many engineers inefficient. It is KEYWORDS

*Lect. Kerim SÖNMEZ is Coordinator of the ARC for Technology Transfer Industrial rights, Intellectual rights, patent, utility model, 6769

Karadeniz Technical University in Trabzon, Türkiye. law, KTU



Permission to make digital or hard copies of part or all of this work for personal or

classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full 1 INTRODUCTION

citation on the first page. Copyrights for third-party components of this work must The concept of intellectual property refers to all rights that

be honored. For all other uses, contact the owner/author(s)

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia are the product of the human mind and have economic value even

© 2024 Copyright held by the owner/author(s). if they do not have a tangible equivalent. In other words, it PhD) were filtered from patent applications for each year and the

includes ideas that arise as a result of the creative efforts of a effect of the developed strategy on the applications was revealed.

person or organization, inventions, literary and artistic works, When the application data were examined, it was seen that the

symbols, names, shapes and designs used for commercial new strategy applied increased the industrial property assets.

purposes. With the application of an intellectual product, the

absolute right provided to the inventor in material and spiritual

terms is recognized for a certain period of time [1]. If these rights 3 HISTORICAL DEVELOPMENT OF

are defined in a different term, intellectual property rights (IPRs) PATENT RIGHTS IN TÜRKİYE

can be defined as the rights that enable sanctions to be imposed The first important legal arrangement regarding patent rights

on the products created by the human mind [2]. It is necessary to in Türkiye was made in 1879 during the Ottoman Empire. The

analyze the concept of intellectual property in two separate French Patent Law of that time was amended and translated into

sections. The first of these concepts includes industrial property the Ottoman Patent Law enacted in 1879. In parallel with the

rights including inventions (patents and utility models), developments in the world, valid patent laws could not be

trademarks, industrial designs, integrated circuit topographies enacted in Türkiye until 1995, except for the international

and geographical indications. The second concept includes all agreements signed, and the patent laws of the Ottoman Empire

intellectual and artistic works, including works of art, works of continued to be applied with some changes until 1995.

science, works of literature, music and musical works, fine art Subsequently, Türkiye became a party to the Paris Convention in

and cinematographic works, depending on copyright [3]. 1925 and signed the WIPO founding treaty in 1976. Furthermore,

Intellectual and industrial property rights give the inventor the Türkiye acceded to the London amendment in 1956, Articles 13

ability to manage all commercial activities thanks to the absolute through 30 of the Stockholm amendment in 1976 and Articles 1

rights it gives to the inventor. Both the desire of the inventor to through 12 of the Stockholm amendment in 1995 [6]. Türkiye

protect his/her invention and the desire to prevent imitation in signed the Customs Union Agreement in 1994. With this

commercial activities increase the number of applications of agreement, the "TRIPS" agreement, the "Strasbourg Agreement

Intellectual and Industrial Rights in legal protection processes. on the classification of patents" (IPC) and the "Patent

Patent, utility model, trademark, etc. industrial assets and Cooperation Agreement" (PCT) and the "Agreement

copyrights are subject to very serious court-based sanctions in Establishing the World Trade Organization" entered into force.

case of infringement of the intellectual assets in question. [4]. Later on, the "Budapest Agreement" on the international

Violations or infringements of rights by third parties have legal protection of microorganisms entered into force, as well as the

and criminal sanctions to protect the rights of right holders [5]. "Patent Law Treaty" (PLT) and the "European Patent

The purpose of patents is to provide protection that facilitates Convention" (Munich Convention) [6].

technological development. A patent not only gives the inventor Table 1 presents data on Türkiye's status as a party to the

exclusive rights to create an invention, but also provides conventions on intellectual property rights to which Türkiye is a

incentives for the technological development and party [7].

commercialization of that invention. Instead of obtaining a

patent, the inventor publishes the technical specifications of the Tablo 1: International Agreements to which Türkiye is a Party

invention, enabling others to make different new inventions [7]

based on the invention. An increase in the number of patents in a First Türkiye’s country indicates a high level of technological development in Participatio

Agreements Signatur Membershi

the country. The transformation of industrial and intellectual n Date

e Date p

assets into the economy through the sale of inventions, the World Intellectual

production of inventions and the sale of products positively Property Organization affects the welfare of the country. 1967 YES 12.05.1976

WIPO Articles of

This research aimed to determine if recognizing master's and Association

doctoral theses as invention disclosure forms, without requiring Treaty Establishing

additional notifications, would lead to an increase in the number the World Intellectual of granted patents. Since Karadeniz Technical University is a 1995 YES 26.03.1995



research university, the number of industrial property assets is (WIPO) Property Organization significant. It is necessary to develop new strategies to increase

European Patent

industrial property assets. When the application data was 1973 YES 01.11.2000



examined, it was seen that the new strategy implemented Stockholm Convention (EPC)

increased the industrial property assets.

(Articles 1-

Paris Convention for YES 12)



2 the Protection of 1883 (10.10.1925 01.02.1995 METHODOLOGY Industrial Property ) (Articles The application change status was revealed by using the 13-30 patent and utility model application data of Karadeniz Technical 16.05.1976) University. In addition, the registration numbers were evaluated Signed Patent Law (PLT) 2000 with a similar method. Among the patent applications in the KTU 02.06.200 patent portfolio, applications between the years 2017-2023 were Trademark Law (TLT) 1994 YES 01.01.2005 evaluated. In the relevant years, student applications (Master and



Singapore Agreement Signed The education institutions defined in subparagraph (c) of the first

on Trademark Law 28.03.2006 agreement paragraph of Article 3 of Law No. 2547 and higher education

2006 has not yet institutions affiliated to the Ministry of National Defense and the

entered into Ministry of Interior.

force (2) When an invention is realized as a result of scientific

BUDAPESTE studies or research conducted in higher education institutions, the



Agreement on the inventor is obliged to notify the higher education institution in 1977 YES 30.11.1998 writing and without delay. If a patent application has been made,

International Storage

of Microorganisms the higher education institution shall be notified of the

LAHEY Agreement application.

on the International (3) The higher education institution is obliged to file a patent

Registration of 1999 YES 1.01.2005 application if it claims right ownership over the invention.

Designs (Geneva Otherwise, the invention becomes a free invention

Text! On this occasion, studies for the protection of the knowledge

Protocol to the accumulation in universities with intellectual and industrial

MADRID Agreement 1989 YES 01.01.1999 rights have been carried out as of 2017.

Patent Cooperation

1970 YES 01.01.1996

Treaty (PCT)

LOCARNO 4 KARADENİZ TECHNICAL UNIVERSITY



Agreement on the INDUSTRIAL PROPERTY NUMBERS 1968 YES 30.11.1998

Restriction of Designs Founded in 1954, Karadeniz Technical University is the first

NIS Agreement on the technical university established in Anatolia in Türkiye. In

International addition, as of 2021, it continues to produce science in the

Classification of Research University category. New generation universities are

1957 YES 01.01.1996

Goods and Services in universities that transform the knowledge they produce into

Trademark added value while producing knowledge, coordinate these

Registration processes, and involve every individual from students to faculty

STRASBORG members in innovation-based commercialization activities.

Agreement on the Karadeniz Technical University has assumed an important role

International 1971 YES 01.10.1996 in serving this basic mission with the Technology Transfer

Classification of Application and Research Center (TTC).

Patents (IPC) With the Industrial Property Law No. 6769, which entered

VIENNA Agreement into force in 2017, Karadeniz Technical University has made a



on the Classification total of 262 industrial property applications with 166 national 1973 YES 01.01.1996 patents, 28 national utility models, 16 national designs and 52

of Figurative

Elements of Marks international patent applications with access to more than 300

inventors and more than 3000 students. With these applications,

39 national patents, 20 national utility models, 16 national

designs and 5 international patents were registered. In addition,



2.1 INDUSTRIAL PROPERTY LAW NO. 6769 6 copyright registrations and 24 trademark applications were also

In Turkish law, industrial property rights were first regulated carried out by Karadeniz Technical University [8].

by the Decree Law No. 551 on the Protection of Patent Rights,

No. 554 on the Protection of Industrial Designs, No. 555 on the

5 A NEW STRATEGY for INVENTION



Protection of Trademarks, which entered into force in 1995. DISCLUSURE Protection of Geographical Indications and No. 556 on the

These Decree Laws were repealed by the Industrial Property Law As stated in subparagraph c of Article 121 of the Industrial

No. 6769 ("IPL"), which entered into force in 2017. Property Law No. 6769, there is an obligation to notify the

According to the IPL No. 6769, Articles 113-122 include the university of an invention made at the university before applying

provisions on "Employee Inventions". Especially with Article for industrial property. This notification is referred to as the

121, universities and public institutions are entitled to have rights invention disclosure form in the literature [9]. As of 2017,

in patent applications [10]. universities collect invention disclosure forms and make patent

applications by evaluating these forms according to criteria such

ARTICLE 121 as patentability, commercialization, etc.

(1) Without prejudice to the provisions of special laws and As of 2023, a new strategy was developed by Karadeniz

regulations under this article, the provisions regarding the Technical University Technology Transfer Application and

inventions of employees shall apply to the inventions made as a Research Center and it was decided to consider the thesis

result of scientific studies or research conducted in higher subjects of graduate and doctoral students as invention disclosure



forms without notification. With this decision, the technology Karadeniz Technical University continues its support in the

and/or information of the relevant invention was protected at an field of Industrial Property with the 6769 IP Law published in

early stage. The number of invention notifications received by 2017. Before 2017, the number of patents belonging to

Karadeniz Technical University since 2017 is shown in academicians at the university was 7, while 262 industrial

following figure 1. property applications were realized under the university's

ownership as of 2024.

In 2020, there was a global COVID-19 pandemic, the effects

Number of Invention Notifications of which continued in 2021, and a national stagnation in



n 200 industrial property applications in 2022-2021. However, even 156 o ti during these periods, the know-how at the university was n s 119 n 107 111 108 transformed into industrial property assets. Figure 2 shows an ve 81 In 100 increase in industrial assets with the normalization process after f 64 o ficatio the pandemic. ti er o 22 The 25 national patent applications until 2022 increased to N mb u 31 national patent applications with the evaluation of the thesis 0 N subjects of master's and doctoral students in postgraduate



Year which was put into effect in 2023. In the first 8 months of 2024, education as invention disclosure forms without notification,



Figure 1 : the number of national patent applications reached 45 Number of Invention Notifications Received at

Karadeniz Technical University applications, and 12 applications are based on given information

from theses.

While international applications are examined, 18

The number of national patents, national utility model,

international patent applications were submitted as of 2024.

national design and international patent applications applied as In 2023, a new strategy was put into effect as a new strategy

Karadeniz Technical University since 2017 is given in Figure 2. in which the thesis subjects of master's and doctoral students in

Number of National Patent Applications postgraduate education were evaluated as invention disclosure

Number of International Patent forms without notification, and the knowledge accumulation at

Industrial Property Numbers Applications Karadeniz Technical University was protected at an early stage. Number of National Utility Model

Applications It is thought that the use of this practice in all universities will

Number of National Design Applications

60 produce positive results and increase the number of national and

international patents.

40 The new strategy provides early awareness to young

20 researchers and supports patent applications that adopt industrial

needs and have high commercialization potential.

0 The strategy of evaluating theses as invention notifications

was introduced to international partners (8 European Countries)

as an example of good practice in the projects of strengthening

technology transfer with innovative approaches, in which KTU

Figure 2: Karadeniz Technical University Industrial is a partner, within the scope of the ERASMUS+ and

Property Numbers INTERREG NEXT Programs.

REFERENCES

Table 2 shows the comparative number of patent applications [1] Geven, Zeki. 2011, "Fikri Mülkiyet Hukukunda yenilik ve

of Karadeniz Technical University compared to Trabzon orijinallik." İnönü Üniversitesi Hukuk Fakültesi Dergisi 2.2, 327-401.

[2] Dıjk, T., 1994, The Economic Theory of Patente: A Survey”, Merit

province. Research Memorandum, 2/94-017.

[3] WEB, (Access Date: 12.08.2024), Telif Hakları Nelerdir?



Table 2: Number of patents in Trabzon province - Karadeniz [4] https://www.tahanci.av.tr/telif-hakki/

Başlar, Y., 2019, Fikir ve Sanat Eserleri Kanunu’nun Ek Madde 4

Technical University and the contribution of the university to the Hükmünün İhlali Suçları. Ankara Barosu Dergisi, 77(4), 37-74.

number of patents [5] Turan, H. S., 2012, Fikir ve Sanat Eserlerinin Cezai Himayesi, Seçkin

Yayıncılık, Ankara, s. 119, Türkiye

Number of 2017 2018 2019 2020 2021 2022 2023 [6] Gökovalı, U., & Bozkurt, K., 2006, Fikri Ve Sinaî Mülkiyet Hakki (Fsmh) Trabz

Olarak Patentler: Dünya Ve Türkiye Açisindan Tarihsel Bir Bakiş. Muğla

National on 48 Üniversitesi Sosyal Bilimler Enstitüsü Dergisi, (17), 135-146. 28 29 24 28 51 51

Patent [7] Toker Köse, M., 2018, Fikri mülkiyet hakları çerçevesinde patentin (City)

iktisadi boyutu ve Türkiye örneği, Master's Thesis, Hitit Üniversitesi



Applicatio Sosyal Bilimler Enstitüsü, Türkiye 11 14 7 9 12 23 41 [8] WEB, (Access Date: 13.08.2024), KTÜ Bülten, KTÜ TTM 100.Sayı ns KTU DEERGİ https:/ /ktuttmebulten.ktu.edu.tr/bulten/subat2024 23% University 50% 24,1 37,5 42,8 45% 80,3 [9] WEB, (Access Date: 13.08.2024), Buluş Bildirim Formu,

Contribution % % 5% 9% https://www.ktu.edu.tr/dosyalar/tto_BiaUH.docx

[10] WEB, (Access Date: 12.08.2024), 6769 SMK,

https://www.mevzuat.gov.tr/MevzuatMetin/1.5.6769.pdf



6 CONCLUSION





Using Open-Access Resources and Platforms to Create a




Technology Transfer Ecosystem *



Viatcheslav Britchkovski†

Informatiom Center/National Library of Belarus

v_britch@nlb.by



ABSTRACT offices. There are several main approaches to technology

transfer.

Technology transfer is a complex process that requires up-to-date Business assistance: companies providing services on specific

and reliable information on various aspects of a technological issues related to technology can be considered as specialized

solution. Approaches to improving the efficiency of technology organizations working in the scientific and technical sectors. In

transfer systems through the use of open access resources and order to correctly navigate among such companies, many TTOs

platforms are considered. have extensive databases.

Technology dissemination means the transfer of specific



KEYWORDS knowledge from research institutes to a group of small and

medium-sized enterprises with common technology needs.

open access resources, open access services, digital ecosysytem, Technology search consists of analyzing the national and

technology transfer office international market in order to acquire promising technologies

and commercial opportunities that can be used by companies in

a certain region.

1. INTRODUCTION This task is often carried out independently of specific industry

needs.

Improving the national innovation system is a key factor in

In addition to these direct approaches to technology transfer,

increasing the country's competitiveness in the modern

TTOs are increasingly focusing on the use of various indirect

environment, often defined as a "knowledge economy" and

technology transfer mechanisms, such as technology exchange

focused on the commercialization of scientific results. Of

through networks of companies, technology and innovation

particular importance in this context are studies aimed at

support centers, product development centers, outsourcing, etc.

improving technology transfer and organizing effective

This means that attention is paid not only to technology transfer

interaction between all participants.

from research institutes to industry, but also to stimulating

Technology transfer is considered as one of the most important

technology exchange directly within industry.

instruments for national and regional economic growth.

Participation in network organizations allows TTOs:

Much attention in Belarus is paid to improving the functioning

of technology transfer offices (TTOs). • develop and maintain high standards for their services;

The activities of TTOs are aimed at commercializing the results • significantly increase opportunities for finding

of R&D, ensuring the acceleration of solutions for technical and partners for technology commercialization projects.

technological problems of enterprises, improving the quality of • implement innovation policy at the interregional and

their products, and mastering the production of new types of international levels.

products. TTOs, participating in the work of technology transfer

The main activities of TTOs aimed at the implementation of a set

networks, can more effectively provide their clients with the

of measures related to transferring innovations from the sphere

following services:

of their development to the sphere of practical application. They

include: • search for partners for the joint implementation of

• technology commercialization projects for R&D, entry conducting market research to identify opportunities

for implementing innovations by universities, into new markets, etc; scientific and other organizations; • dissemination of technological information is a

• relevant service for scientific organizations that are performing work to ensure legal protection and

interested in widely informing industry and companies

introduction of innovations into civil circulation;

• about their research capabilities and competencies;

providing engineering and consulting services. • promotion of technological projects using various

The implementation of new technologies and research results

from the scientific and technical sectors in industry is a networking tools; traditional task, and often the main activity of technology transfer • a primary analysis of supply and demand in certain

subject areas of research.

The main role in the Belarusian technology transfer network

Permission to make digital or hard copies of part or all of this work for personal or infrastructure is played by the Republican Technology Transfer

classroom use is granted without fee provided that copies are not made or distributed Center (RCTT).

for profit or commercial advantage and that copies bear this notice and the full The organizational structure of the RCTT network includes

citation on the first page. Copyrights for third-party components of this work must members, clients, partners and a coordinating organization.

be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Members of the network are research organizations, higher

© 2024 Copyright held by the owner/author(s). education institutions, enterprises and organizations of all forms

of ownership that have TTOs or divisions responsible for



technology transfer in their structure. Within the framework of Taking into account such limitations, in 2009 the World

the methodology adopted in the RCTT network, network Intellectual Property Organization (WIPO) launched an

members help their clients prepare proposals for cooperation, international project to create a network of Technology and

requests for the implementation of R&D. [1]. There are 3 options Innovation Support Centers (TISCs), the purpose of which is to

for disseminating them:

simplify access to technical knowledge and improve the

• they are posted by network members on the RCTT efficiency of using patent information.

Internet portal; The National Intellectual Property Center (NIPC) is creating a

• they are posted at the request of network members by network of TISCs in the Republic of Belarus in accordance with

the coordinating organization in foreign technology the Agreement between WIPO and NIPC dated October 10,

transfer networks; 2016. As part of its implementation, NIPC performs the

• they are posted on the websites of foreign partners of functions of a coordinating body.

the coordinating organization. Currently, there are 29 TISCs operating in the Republic of

Network members help their clients prepare information on the

Belarus.

products and services of the organization for posting them at a

The creation of the TISCs has improved information and

virtual exhibition on the Internet portal of the Russian Center for

Technology Transfer. Network members also monitor external scientific-methodological support for information and patent

and internal markets to find the target consumers for the activities, increasing the efficiency of using IP objects.

organizations. Clients of the network are suppliers and High-quality scientific information is also one of the most

consumers of technologies (research and design organizations, important factors facilitating technology transfer.

educational institutions, enterprises and organizations of all Underdeveloped information infrastructure and the lack of

forms of ownership). objective data on advanced scientific knowledge and

The RCTT is a consortium for coordinating activities in the field developments create serious barriers to the further development

of technology transfer, which includes of science and its commercialization, significantly reducing the

• efficiency of TTOs. The so-called "serial publication crisis" [2] head office in Minsk;

• has a negative impact on the quality of information support for 5 branches in the regions of the Republic of Belarus

and 30 branches at research organizations, higher TTOs, caused by the fact that the traditional commercial

education institutions and enterprises of the Republic economic model of scientific communication leads to a rapid

of Belarus; increase in subscription prices with relatively unchanged budgets

• 97 foreign organizations in 23 countries. for organizations. The problem is that both TTOs specialists and

The main objectives of the activities of the RCTT branches at researchers working in various subject areas face significant

manufacturing and industrial enterprises are:

difficulties in the process of searching, obtaining and using

improving the quality and reducing the cost of manufactured

information. In the context of the constant growth of scientific

products;

output and the simultaneous increase in the cost of access to

assisting in the expansion of sales markets.

information resources due to the fact that publishers seek to

maximize their profits through the sale of subscriptions to

OPEN ACCESS SERVICES FOR scientific journals, scientists and other consumers of scientific

INFORMATION SUPPORT OF TTOs information experience serious difficulties when it is necessary

to find a potentially useful scientific result and get acquainted

with it [3]. The deficit of high-quality scientific information

TTOs use various information system and resources to

resources deprives specialists of the opportunity to analyze and

effectively manage and transfer technologies. Some of them are:

objectively evaluate the quality of research and development

• Patent and invention databases: these resources help results.

track and manage intellectual property. The important place in the activities of the technology transfer

• Scientific publications databases: these resources offices is occupied by legal problems and issues of protecting

allow TTOs to stay abreast of the latest advances and intellectual property, including problems of legislative and

innovations in various fields. judicial protection of copyright. The lack of relevant and up-to-

• Collaboration and knowledge sharing platforms: such date information in this area significantly reduces the

platforms help researchers and developers exchange

effectiveness of the commercialization of scientific research [4].

ideas, find partners, and collaborate on projects.

• To overcome these challenges, we suggest to use open access

Project management information systems: these

resources and platforms for facilitating information support of

systems help coordinate project work.

• business processes during transfer knowledge and technology.

Marketing and analytics tools: these tools are used to

analyze the market and identify needs and Open Access (OA) as a movement has been steadily gaining

opportunities strength for roughly the last two decades. This is due to the for the commercialization of

technologies. following factors:

To operate effectively TTOs need high-quality and timely The number of publications in open access reaches 47% [5].

information. Research funding programs and foundations require that research

Although patent information has become more accessible in results must be published in the OA repositories or OA journals.

recent years, including through services provided via the Internet Many organizations support the requirements for the openness of

primary data and research results.

on a paid or free basis, the coverage and availability of patent

The citation rate of OA scientific publications is higher than that

data in some countries, including Belarus, remain limited.

of those distributed by subscription [6].



Open access resources are increasingly considered like an option AMiner integrates academic data from multiple sources by data

to replace expensive commercial databases necessary for the mining and social network analysis and mining technology to

information services of TTOs [7]. catch paper indexes.

Using OA resources can significantly improve the efficiency of AMiner cooperates with scholars and academic institutions to

Technology Transfer Offices (TTOs). Below are some of the share papers and scholar data and purchase copyright.

ways in which they can be used. CORE provides access to the world's largest collection of open

1. Access to scientific publications. Open access provides free access scholarly papers by collecting and indexing research from

and unrestricted access to scientific articles and research. This repositories and journals. It is a non-profit service dedicated to

allows TTOs to stay up-to-date with the latest advances and the open access mission and a signatory to the Principles for

innovations in various fields of science and technology.

Open Scholarly Infrastructures (POSI) [8]. CORE serves a global

2. Analysis of patent information. Many patent databases also

network of repositories and journals by improving

provide open access to information. This helps TTOs track new

discoverability and preventing misuse of their content; ensuring

patents, analyze trends, and find potential partners for licensing

and commercializing technologies. For example, Espacenet that metadata records are uniquely identified; supporting data

provides free access to millions of patents worldwide providers in applying best practices by providing tools for

3. Collaboration and knowledge sharing. OA platforms facilitate metadata validation, content management, enrichment, and OA

collaboration between researchers and developers. This allows compliance; and facilitating machine access to open research.

TTOs to find partners for joint projects and share ideas. CORE's mission is to index all open access research worldwide

Examples of such platforms include ResearchGate and and make it accessible to all. In doing so, CORE: Academia.edu.

4. Training and professional development. OA resources can also • enriches scientific data using modern text and data

be used for the training and professional development of TTOs. mining technologies to make it easier to find;

Online courses and webinars available on platforms such as • enables others to develop new tools and use cases

Coursera and edX help TTOs staff stay up-to-date with new based on the CORE platform;

techniques and technologies. • supports the network of OA repositories and journals

5. Market Research. Using open data and analytics tools helps with innovative technical solutions;

TTOs conduct market research. This allows for a better • promotes the creation of a scalable and cost-effective

understanding of the needs and opportunities for way to provide open scientific information.

commercializing technologies. Examples of such tools include CORE aggregates research papers from data providers around

Google Scholar and Microsoft Academia. the world, including institutional and subject repositories and

There are many integrated platforms and services for facilitating journal publishers. This process, also called data harvesting,

OA resources usage.

enables CORE to offer search, text mining, and analysis

OpenAIRE (the Open Access Infrastructure for Research in

capabilities not only on metadata but also on the full text of

Europe) enables the search, discovery and monitoring of the

publications and datasets from 100,000+ research projects. research articles, making CORE a unique service in the research

OpenAIRE actively supports the Open Science initiative. On the community.

one hand, OpenAIRE is the network of dedicated Open Science BASE is one of the largest search engines in the world,

experts promoting and providing training for Open Science. particularly for academic web resources. BASE provides over

On the other hand, OpenAIRE is a technical infrastructure 300 million documents from over 10,000 content providers. Full

harvesting research output from connected data providers. texts of about 60% of indexed documents are available free of

OpenAIRE aims to establish an open and sustainable scholarly charge [9]. BASE indexes the metadata of all types of

communication infrastructure responsible for the overall academically significant resources (journals, institutional

management, analysis, manipulation, provision, monitoring and

repositories, digital collections, etc.) that provide an OAI

cross-linking of all research outcomes.

interface and use OAI-PMH to provide their content. The index

This combination of knowledge and a pan-European Research

Information platform enables OpenAIRE to provide services for is constantly being expanded by integrating new sources/content

researchers, research support organizations, funders, content providers. Database managers can integrate the BASE index into

providers and TTOs such as: their local infrastructure (e.g., metasearch engines, library



• catalogs).

Integrate scientific information.

• International research collaborations can bring TTOs new

Monitor and report on research outcomes for funders

opportunities for collaborative research, increase the impact of

and partners.

• their research, and boost the commercialization of scientific

Train and support on all subjects related to OA.

• results. For instance, knowing which institutions globally work

Discovery of OA output per project, funder, and data

on similar research can help identify new partnership

provider.

AMiner is a new generation of scientific and technological opportunities. Identifying existing connections among

intelligence analysis and mining platform with completely researchers between those institutions can help drive

independent intellectual property rights. It was established by a development opportunities. These data have come through costly

team led by Professor Tang Jie from the Department of Computer subscriptions to restricted databases. OpenAlex now provides the

Science and Technology of Tsinghua University. data required for international intelligence freely to all users

AMiner's scientific research data includes 331 million papers, across the globe [10]. 135 million scholars, 1.122 billion paper citation relationships

and 8.79 million knowledge entities (this data is in dynamic

change).





CONCLUSION REFERENCES


[1] Подготовка и управление профилями в Сети Республиканского

центра трансфера технологий. Методическое руководство / Под ред.

Open access platforms and services provide technology transfer А.А. Успенского. Изд. 2–е переработ. и доп. – Мн.: Центр системного

анализа и стратегических исследований НАН Беларуси, 2023. – 65 с

offices with effective tools for searching, disseminating and [2] T Jurchen S. Open access and the serials crisis: The role of academic using scientific publications for the purposes of commercializing libraries //Technical Services Quarterly. – 2020. – v. 37. – №. 2. – p. 160-

research. Their use contributes to the acceleration of technology 170.

[3] Бричковский, В.И. Инициатива открытого доступа в

transfer and the increase in the efficiency of innovation activities.

информационном обеспечении инновационной деятельности / В.И.

OA resources may be integrated using API into the structures of Бричковский // Наука и инновации. – 2019. – № 12. – С. 76–79.

the digital ecosystem of technology transfer, which includes [4] Young P. The serials crisis and open access //A White Paper for the

agents (scientific organizations and teams), objects (information Virginia Tech Commission on Research, University Libraries, Virginia

and knowledge) and infrastructure (services and information Tech. – 2009.

[5] The State of Open Data Report 2019. Digital Science. Report.

systems). https://doi.org/10.6084/m9.figshare.9980783.v2

[6] The effect of data sources on the measurement of open access: A



ACKNOWLEDGMENTS comparison of Dimensions and the Web of Science / Basson I. [et al.] //

PLoS ONE. 2022. № 17 (3).e0265545.

Author would like to thank people from the State Committee on https://doi.org/10.1371/journal.pone.02655459999

[7] Colavizza G., Hrynaszkiewicz I., Staden I., Whitaker K., McGillivray B.

Science and Technology of the Republic of Belarus for their The citation advantage of linking publications to research data // PLoS

constant support of OA activities and express gratitude to all my ONE. 2020. Vol. 15. № 4. e0230416.

colleagues from the National Library of Belarus, Belarusian https://doi.org/10.1371/journal.pone.0230416

[8] CORE Services // CORE – Aggregating the world’s open access research

State University for their help. papers - https://core.ac.uk/services.

[9] Indexed content providers by date // BASE - Bielefeld Academic Search

Engine –: https://www.base-search.net/about/en/about_sources_date.php

[10] Priem J., Piwowar H., R. Orr R. Openalex: A fully-open index of scholarly

works, authors, venues, institutions, and concepts, 2022.

arXiv:2205.01833.

Fostering Open Innovation and Technology Transfer:



Insights from the Euro-Mediterranean Innovation Camp



(EMIC)



Prof. Dr. Abdelhamid El-Zoheiry Karen Gladović, FirstName Surname

President Lecturer of Law Department Name

Forum for Euro-Mediterranean Euro-Mediterranean University Institution/University Name

Innovation in Action, France (EMUNI) City State Country

zoheiry@gmail.com Piran, Slovenia email@email.com

karen.gladovic@emuni.si



ABSTRACT / POVZETEK transfer within the Euro-Mediterranean region, using the Euro- Mediterranean Innovation Camp (EMIC) as a case study. The

The global marketplace is rapidly evolving, demanding paper explores how EMIC serves as a successful model for

innovative approaches to technology transfer that can bridge the bridging the gap between academic research and commercial

gap between research and commercial application. The Euro- applications in a region that presents both challenges and

Mediterranean region, with its diverse socio-economic landscape, opportunities due to its socio-economic diversity. Overall, the

presents both challenges and opportunities for such endeavors. paper contributes to the literature on technology transfer by



(EMIC) as a successful model for implementing open innovation programs like EMIC can drive economic growth, address and technology transfer, particularly within the strategic This paper presents the Euro-Mediterranean Innovation Camp providing a detailed exploration of how structured innovation

pressing global challenges, and create marketable solutions,



institutions. The analysis not only draws on the outcomes of the particularly in the complex and diverse Euro-Mediterranean framework of the Jožef Stefan Institute (JSI) and its partner

recent EMIC initiatives but also aligns these practical insights context.

with the theoretical foundations of open innovation as discussed

in the doctoral disposition on technology transfer. The EMIC initiative has attracted applicants from over 17

countries, with a significant portion of the applications coming

KEYWORDS / KLJUČNE BESEDE from Egypt and Jordan. The charts below illustrate the diversity

and distribution of applicants by country of residence across the



Region, EMIC, Jožef Stefan Institute two seasons of the program. Open Innovation, Technology Transfer, Euro-Mediterranean

(1)



1 INTRODUCTION



The global marketplace is rapidly evolving, demanding

innovative approaches to technology transfer that can bridge the

gap between research and commercial application. The Euro-

Mediterranean region, with its diverse socio-economic

landscape, presents both challenges and opportunities for such

endeavors. This paper presents the Euro-Mediterranean

Innovation Camp (EMIC) as a successful model for

implementing open innovation and technology transfer,

particularly within the strategic framework of the Jožef Stefan

Institute (JSI) and its partner institutions. The analysis not only

draws on the outcomes of the recent EMIC initiatives but also

aligns these practical insights with the theoretical foundations

of open innovation, as discussed in the doctoral disposition on

technology transfer. The main purpose of the paper is to

examine the application of open innovation and technology



Permission to make digital or hard copies of part or all of this work for personal or

classroom use is granted without fee provided that copies are not made or distributed Figure 1: Number of applicants by country of residence –

for profit or commercial advantage and that copies bear this notice and the full Season 1 citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).





The Euro-Mediterranean Innovation



Camp (EMIC): A Model of Open



Innovation



The Euro-Mediterranean Innovation Camp (EMIC) is a flagship

initiative that embodies the principles of open innovation within

the Euro-Mediterranean region. Launched by the Euro-

Mediterranean University (EMUNI) in collaboration with the JSI

and the EuroMed University of Fes (UEMF), EMIC provides a

platform for young innovators to develop and showcase their

ideas in response to pressing global challenges. The camp

focuses on three critical areas: health, renewable energy, and

climate change—fields that are not only relevant to the region

but also globally significant.



The EMIC initiative is structured to promote iterative learning

and development. Participants, who are selected through a

rigorous process, receive mentorship and technical support from

Figure 2: Number of applicants by country of residence – experts in their respective fields. This support is crucial in

Season 2 helping them refine their ideas and develop viable prototypes.

The average age of participants was 24 years, with the youngest

being 18 and the oldest 35, showcasing the youthful energy

Theoretical Foundations: Open driving innovation in the Euro-Mediterranean region.



Innovation and Technology Transfer

Selection Process of EMIC

Open innovation, a concept popularized by Henry Chesbrough,

has significantly reshaped our understanding of how innovation The Euro-Mediterranean Innovation Camp (EMIC) follows a

occurs in the modern business environment. Unlike traditional rigorous and multi-stage selection process to identify and support

closed innovation models where companies rely solely on the most promising young innovators from across the Euro-

internal resources for R&D, open innovation promotes the use of Mediterranean region. Here’s an overview of the selection

both internal and external knowledge sources. This approach process:

accelerates the innovation process and expands the potential

market for new technologies (Chesbrough, 2005a). In the context

1. Application Submission:

of technology transfer, open innovation facilitates the o Eligibility: Applicants must be between 18 commercialization of intellectual property (IP) through various

and 35 years old, reside in one of the Euro-

channels, including licensing, joint ventures, and spin-offs, thus

Mediterranean countries, and possess at least

driving economic growth and enhancing competitiveness

a high school diploma. The innovation they

(Chesbrough, 2003b).

propose must address one of the three

thematic areas: Health, Renewable Energy,

The Euro-Mediterranean region is a fertile ground for applying or Environment (including Climate Change)

open innovation principles. However, the integration of public and must be capable of being converted into

research outputs with industry needs has been a persistent a prototype within three months.

challenge. Slovenia, for instance, excels in scientific output, o Application Process: Interested candidates

ranking high in terms of research publications. Yet, as the submit their applications through an online

OECD’s 2012 report highlights, the country struggles with the form available on the EMUNI website. The

commercialization of research findings, particularly in application requires a detailed description of

converting scientific discoveries into marketable products and the innovative idea or invention,

services. This gap underscores the critical role of initiatives like highlighting its novelty, feasibility, and

EMIC, which aim to bridge the divide between academic potential for commercialization.

research and industrial application through structured innovation 2. Initial Screening:

programs. o A panel of experts reviews all submitted

applications. The review process evaluates

The thematic focus of the EMIC projects aligns well with global the novelty of the idea, its practical

challenges, as evidenced by the distribution of project types: 45% applicability, and the feasibility of

focused on health, 35% on renewable energy, and 20% on implementation within the specified

environmental issues, reflecting the alignment of participant timeframe. interests with critical global needs. o Shortlisting: Based on the initial screening,

a subset of applicants is shortlisted to

advance to the next phase. For instance, in

Season 2, out of 124 applications,

approximately 40 candidates were



shortlisted for the online pitching phase The Euro-Mediterranean Innovation Camp (EMIC) Season 2

(EMUNI). brought together some of the brightest minds across the region to

3. Online Pitching: develop innovative solutions addressing critical challenges. This

o Pitch Preparation: Shortlisted candidates section highlights the top three projects that stood out for their

prepare a pitch presentation, which they creativity, technical expertise, and potential for real-world

deliver via an online platform. During this impact.

phase, they present their ideas to a jury

comprising experts from relevant fields.

o 1. Muhammad Mounir (Egypt) – "SugarHeal"

Jury Evaluation: The jury evaluates the o Project Overview: Muhammad Mounir, a pitches based on several criteria, including

Molecular Biotechnology student from

innovation, feasibility, potential impact, and Galala University, developed "SugarHeal,"

the candidate’s ability to articulate and an innovative wound dressing material

defend their idea. The most innovative and designed to accelerate the healing process of

viable projects are selected to move forward.

chronic and acute wounds. During his time



4. Innovation Bootcamp: at the Jožef Stefan Institute (JSI), o Workshops and Mentorship: The selected Muhammad explored two main fabrication

finalists, known as the "Innovation Squad,"

techniques:



bootcamp held at facilities like the Jožef cellulose-based are invited to participate in a 10-week ▪ Electrospinning: He created a solution with

Stefan Institute in Slovenia or the EuroMed

antibacterial properties for

University of Fes in Morocco. During this

electrospinning, resulting in a

period, they receive technical assistance,

biodegradable wound dressing

mentorship from subject matter experts, and

that promotes faster healing.

support in refining their prototypes. ▪ 3D Bioprinting: Muhammad also

o Elimination Stages: Throughout the

developed a cellulose-based ink

bootcamp, participants go through multiple

enriched with natural antibacterial

elimination stages. These stages are

extracts, which was used in 3D

designed to progressively challenge the

bioprinting to produce

innovators, focusing on proof of concept,

customizable wound patches.

engineering, prototyping/testing, and o Current Stage: Muhammad has

customer validation. The best performers in

successfully developed prototypes of the

each stage advance to the next round

wound dressing through 3D bioprinting,

(EMUNI).

which have shown promising results in terms

5. The Finale:

o of mechanical stability and biological

Final Presentation: The competition

response. His next steps include further

culminates in a live finale event where the

optimization and exploring commercial

remaining candidates present their fully

applications (EMUNI).

developed prototypes. This event is attended 2. Rahma M. Tolba (Egypt) – "Interactive

by a live audience, including mentors,

Augmented Reality for Lisp Correction"

representatives from partner institutions, and o Project Overview: Rahma Tolba, a PhD other stakeholders.

o researcher from Ain Shams University,

Scoring: Final scores are determined by

developed an interactive Augmented Reality

both the jury and audience voting, with each

(AR) application designed to assist in speech

accounting for 50% of the final score. The

therapy for individuals with a lisp. Her

winners are announced based on the

project focuses on improving phonetic

combined results of these evaluations

learning through the use of 3D animated

(EMUNI).

models that demonstrate correct tongue

movements. The app guides users through

This selection process is designed to ensure that the most pronunciation exercises, providing real-time

innovative and feasible ideas are given the support they need to feedback through an AI-based Automatic

develop into successful market-ready products. It also Speech Recognition (ASR) system.

emphasizes the importance of mentorship and iterative o Current Stage: Rahma has developed a

development, helping young innovators turn their ideas into fully functional prototype for Android

impactful solutions. devices, which has been tested on a small

group of individuals. The next steps involve



Case Studies: Innovations from EMIC gathering user feedback from speech

therapists and phoniatricians to refine the

Season 2 design and functionality (EMUNI).



The impact of EMIC is best illustrated through the success stories 3. Med Aziz Mhalla (Tunisia) – "Drowsy Driver

of its participants. The most recent season of EMIC, concluded Detection System" in June 2024, showcased a range of groundbreaking innovations o Project Overview: Med Aziz Mhalla, an

that have the potential to address significant challenges in health electronics engineering student from the

and sustainability. National Engineering School of Sousse,



created the "Drowsy Driver Detection initial stages of development, these institutions can help ensure

System" (DDDS). This system leverages that the innovations produced at EMIC reach their full potential.

machine learning, computer vision, and



embedded hardware to monitor drivers in Opportunities for Enhancing Open

real-time, detecting signs of drowsiness,

distraction, and sleep onset. The system uses Innovation

a Convolutional Neural Network (CNN)

model to classify eye states and detect blinks

and yawns, which are key indicators of Despite these challenges, the EMIC model also presents

drowsiness. significant opportunities for enhancing open innovation in the

o Euro-Mediterranean region. One of the key opportunities lies in Current Stage: Med Aziz has successfully

developed a proof of concept and prototype the potential for cross-border collaboration. By bringing together

that participants from different countries and backgrounds, EMIC has been tested in controlled

environments and on a laptop. He is fosters a rich exchange of ideas and approaches. This diversity is

currently optimizing the system for real-time a strength as it allows for the development of solutions that are

performance using NVIDIA Jetson Nano informed by a wide range of perspectives and experiences.

and preparing for on-road testing (Med Aziz

- Drowsy Driver…). The full cycle of open innovation and technology transfer has yet

to be fully achieved, as both processes require more than just

These projects not only exemplify the innovative spirit fostered innovative ideas and technological breakthroughs. For these

by EMIC but also demonstrate the potential for significant cycles to reach their full potential, business entities must engage

contributions to healthcare, road safety, and speech therapy. early and consistently, starting from the initial stages of research

Each of these top three finalists utilized the mentorship and and development. Their investment and involvement are crucial

resources provided during the EMIC bootcamp to bring their in ensuring that ideas and technologies not only progress beyond

ideas closer to real-world application. the conceptual phase but also successfully transition from labs to

the market. Without the proactive participation of businesses, the

promise of open innovation and effective technology transfer

The diversity of innovations emerging from EMIC highlights the

program’s success in fostering creativity across different fields. may remain unfulfilled, with many promising projects never

realizing their full impact.

These projects are not just theoretical exercises; they represent

tangible solutions that can have a real-world impact, addressing

some of the most pressing challenges in the Euro-Mediterranean The collaborative model of EMIC, supported by the JSI’s

region and beyond. extensive research infrastructure, offers valuable insights into

how open innovation can be effectively implemented in a

complex and diverse region. The partnerships between academic

Challenges in Implementing Open institutions, industry players, and government bodies are crucial



Innovation in providing the necessary resources for young innovators to

translate their ideas into impactful technologies. These

partnerships also help to ensure that the innovations produced at

Despite its successes, the implementation of open innovation EMIC are aligned with market needs, increasing their chances of

within the EMIC framework has not been without challenges. success.

One of the primary challenges is the alignment of the diverse

objectives of the program’s international partners. The Euro- Another opportunity for enhancing open innovation in the Euro-

Mediterranean region encompasses a wide range of economic,

Mediterranean region is through the development of stronger

social, and political contexts, each with its unique set of networks and ecosystems. By fostering closer ties between

challenges. Coordinating efforts across such a diverse region research institutions, industry, and government, it is possible to

requires careful planning and robust frameworks for create a more supportive environment for innovation. This

collaboration.

includes not only providing funding and resources but also

creating opportunities for mentorship, networking, and

Intellectual property (IP) management is another critical collaboration. Such ecosystems can help to sustain the

challenge in open innovation environments. While open momentum generated by initiatives like EMIC, ensuring that the

innovation encourages the sharing of ideas and resources, it also innovations produced continue to evolve and have a lasting

raises questions about how IP is managed and protected. In the impact.

context of EMIC, ensuring that participants retain control over



their innovations while still benefiting from the collaborative Conclusion environment is crucial. This requires clear guidelines and

agreements on IP management, which can be complex to

negotiate across different legal and regulatory frameworks. The Euro-Mediterranean Innovation Camp (EMIC) serves as a

compelling example of how open innovation can be successfully

Another challenge is the scalability of the solutions developed implemented within a structured technology transfer framework.

through EMIC. While the innovations produced during the camp By leveraging the strengths of regional partnerships and focusing

are often groundbreaking, bringing these solutions to market on on critical areas such as renewable energy, health, and climate

a larger scale requires resources that may not be immediately change, EMIC has successfully fostered a culture of innovation

available to the participants. This is where the support of across the Euro-Mediterranean region. The initiative has not only

institutions like JSI and the involvement of industry partners provided a platform for young innovators to develop their ideas

become critical. By providing continued support beyond the but has also facilitated the transfer of these ideas from research



to market, demonstrating the potential of open innovation to initiatives like EMIC can help to unlock the full potential of the

drive economic growth and address global challenges. region’s young innovators, driving economic growth and

addressing some of the most pressing challenges of our time. As

we look to the future, it is clear that open innovation will continue

Moving forward, it will be crucial to address the challenges of IP

management and resource allocation to sustain the momentum to play a critical role in shaping the global innovation landscape,

and initiatives like EMIC will be at the forefront of this exciting

generated by these initiatives. The ongoing collaboration

between academic institutions like JSI, industry partners, and journey.

government bodies will be key to enhancing the commercial

viability of the innovations emerging from EMIC. As Slovenia REFERENCES and the broader Euro-Mediterranean region continue to refine [1] Chesbrough, H. W. (2003a). Open Innovation: The New Imperative for their approaches to technology transfer, the lessons learned from Creating and Profiting from Technology. Harvard Business School Press.

EMIC will serve as a valuable guide for future innovation [2] Chesbrough, H. W. (2003b). The era of open innovation. MIT Sloan

policies and practices. Management Review, 44(3), 35-41.

[3] Chesbrough, H. W. (2005a). Open Business Models: How to Thrive in the

New Innovation Landscape. Harvard Business School Press.

The success of EMIC highlights the importance of fostering [4] EMIC Status Report 2022-23.

[5] EMIC Status Report 2023-24.

innovation among young people across the Mediterranean [6] OECD (2012). Science, Technology and Industry Outlook. OECD

region. By providing the necessary support and resources, Publishing.





Research Organisation-Industry Cooperation and State




Aid Rules in Slovenia and Europe



Tomaž Lutman Urška Florjančič Urška Fric

Office for Substantive Project Office for Substantive Project Knowledge and Technology

Support, Technology Transfer and Support, Technology Transfer and Transfer Office

Innovation Innovation Faculty of Information Studies in

Jožef Stefan Institute Jožef Stefan Institute Novo mesto

Ljubljana, Slovenia Ljubljana, Slovenia Novo mesto, Slovenia

tomaz.lutman@ijs.si urska.florjancic@ijs.si urska.fric@fis.unm.si



ABSTRACT good practices have been collected from different types of

groups, i.e. researchers, industry, technology transfer managers,

This study provides an international comparative view on state contract research managers and accounting officers. We have

aid regulation in infrastructure use and intellectual property focused on Slovenian research organisations, in addition to

rights transfer in cooperative research and development projects which we included two European research organisations in order

within the European Research and Innovation Ecosystem. to make international comparison. In spring 2024 we concluded

Technology transfer officers or similar profiles at research 8 in-person or online semi-structured interviews with R&D

organisations were interviewed. Additionally, a desk research managers from 7 research organisations. was performed. Annual reports were studied in order to identify

the differentiation of economic and non-economic activity as

well as good practices. Desk research included also rulebooks 2 DESK RESEARCH RESULTS and related Slovenian & EU legislature in the field of contract

and collaborative research. 2.1 Share of economic activity, rulebooks and



KEYWORDS Our study comprised 13 Slovenian, 1 Czech and 1 Italian public pricelists

research organisation – industry cooperation, research services, research organisation (Table 1). As foreseen in articles 16 (ff)

intellectual property rights transfer, state aid rules, Slovenian and 19 of EC Communication (2022/C 414/01), the research

and European research organisations, research and innovation organisation has to account for the costs and the revenues of the

ecosystem

economic activities separately. Different practices on how to do

INTRODUCTION this exist among European research organisations. It was

The European Commission has set specific rules in the field of observed that most research organisations generate up to 20% of

research, development, and innovation (R&D&I) to prevent their revenues from economic activities. This correlates well

market distortion. These rules are described in the European with the maximum 20% capacity limit as foreseen in 2022/C

Framework for State aid for R&D&I (2022/C 414/01 [1]) and 414/01 (it should however be noted that % of income may differ

relate to the Article 107 (1) Treaty on the Functioning of the from % of capacity, which is the actual threshold value).

European Union. Some organisations in the study have around 50% of their

We believe that knowledge and implementation of state aid rules activities classified as economic in nature. They most likely

regarding research services (economic activity), collaboration surpass 20% of economic activities’ capacity limit. For this

projects (non-economic activity) and intellectual property rights reason, they as whole cannot be considered as research

(IPR) are insufficient and could be improved, which was organisations according to 2022/C 414/01. Only departments,

identified also by other authors [2], [3]. laboratories or similar subunits of such organisations which do

not surpass 20% of economic activities’ capacity limit can be

considered ‘research organisations’.

1 METHODOLOGY In Slovenia, the new Law on Scientific Research and Innovation

In order to understand how state aid rules in academia-industry Activities (ZZrID) entered into force on 1 January 2022 [15]. In

cooperation work in practice, we have performed a detailed the same period, a Rulebook on procedures for implementing the

analysis with an international comparative view. Experience and budget of the Republic of Slovenia was updated. Article 119(b)

requires each public research organisation to have an internal

rulebook and pricelist regarding sale of products and services,

Permission to make digital or hard copies of part or all of this work for personal or

classroom use is granted without fee provided that copies are not made or distributed i.e. economic activity [16]. Up to date, several Slovenian

for profit or commercial advantage and that copies bear this notice and the full research organisations have prepared their rulebooks and

citation on the first page. Copyrights for third-party components of this work must pricelists, while many have not yet (at least they are not publicly

be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia available).

© 2024 Copyright held by the owner/author(s).



The sizes of analysed organisations vary a lot, i.e. from 58 to

9560 employees. This strongly affects the organisational

structure and extent of experience in a specific organisation.



Table 1: Selected Slovenian and European research organisations and their info on economic activities.



Organisation Country Share of Rulebook for sale of Pricelist of products Number of Source

economic activity products and services and services employees in

in 2023 2023

Slovenian National Building and Civil

SI 53.6% 16.08.2022 11.01.2022 254 [4], [5]

Engineering Institute

University of Ljubljana, Faculty of unpublished / under unpublished / under

SI 48.2% 829 [4]

Medicine preparation preparation

unpublished / under

Agricultural Institute of Slovenia SI 23.7% 28.11.2023 258 [4], [6]

preparation

Geological Survey of Slovenia SI 15.0% 11.08.2022 23.02.2023 124 [4], [7]

unpublished / under unpublished / under

Institute of Metals and Technology SI 14.9% 58 [4]

preparation preparation

University of Ljubljana, Faculty of unpublished / under unpublished / under

SI 12.0% 362 [4]

Electrical Engineering preparation preparation



Jožef Stefan Institute unpublished, to be

SI 10.2% under preparation 1206 [4]

updated

University of Ljubljana, Faculty of unpublished / under unpublished / under

SI 10.0% 433 [4]

Mechanical Engineering preparation preparation

National Institute of Biology SI 8.7% 3.11.2023 9.02.2023 194 [4], [8]

different guidelines, not identified,

Czech Academy of Sciences CZ 7.4% 9560 [9], [10]

decentralised decentralised

University of Ljubljana, Faculty of unpublished / under

SI 6.2% unpublished 199 [4]

Pharmacy preparation

National Institute of Chemistry SI 5.0% 24.08.2022 24.08.2022 437 [4], [11]

unpublished / under 6.11.2023 (UM-FVV),

University of Maribor SI 4.5% 2121 [4], [12]

preparation decentralised

not identified, 8457 (year

Consiglio Nazionale delle Ricerche IT 0.47% not identified [13], [14]

decentralised 2022)

Faculty of Information Studies in Novo

SI 0.03% under preparation under preparation 82 [4]

mesto



2.2 outside Slovenian Research and Innovation Agency – ARIS; or Good and bad examples of transparent

bookkeeping and economic activity everything related to for-profit organisations).

management

Several Slovenian institutes, such as National Institute of

The transparency of studied research organisations is good. Biology, Institute of Metals and Technology, Slovenian National

Yearly reports support this observation. However, due to Building and Civil Engineering Institute, National Institute of

differentiation in the reports’ structure, the comparison is Chemistry, and Agricultural Institute of Slovenia present their

sometimes difficult. An additional challenge is the lack of contract and collaborative research activities well. Some

standardisation in terminology. unclarity persists, which is also highlighted below in the

translated sections of the annual reports. We assume that this is

One of the important messages of EC Communication 2022/C caused due to the use of vague terminology, as explained above,

414/01 is the requirement to differentiate economic (such as and the lack of differentiation between economic and non-

research service) and non-economic activities (such as economic activities. collaborative research and knowledge transfer activities). It

should be noted that in the Slovenian legislature and The Consiglio Nazionale delle Ricerche defines its income and

consequently other documents, terminology ‘market activity’ is outcome well, but unlike other annual reports, its annual report

used, which is not well defined. Sometimes it is used as is not supplemented with qualitative description. The annual

‘economic activity’ and sometimes as activity on the ‘market’, report of the Czech Academy of Sciences (CAS) is very

with again different interpretations (‘market’ as everything informative. Subunits (i.e, institutes of CAS) have their own

annual reports, which present their technology transfer activities as a market activity in the annual report 2023. However, it is not

well, while contract and collaborative research are inadequately clear from the annual report that this is a non-economic activity.

described. Additionally, the financial part contains non-machine

readable text, which cannot be easily translated. In Table 2, we can see a very good delimitation between contract

research (‘laboratory services’) and collaborative research

Examples (‘research projects with industry’). However, these two

At the Jožef Stefan Institute, Horizon 2020 and Horizon Europe categories are later wrongly joined into one category, ‘income

projects were classified as market projects, which were changed from goods and services on the market’ (Table 3). We believe

in 2022. From the total number of market activities' income, it

this is not an isolated case among research organisations.

was thus unclear which activities were economic and which non-

economic. Table 2. Income from market activity of the Institute of

Metals and Technology (IMT). Annual report of IMT for

At the National Institute of Chemistry, royalties and other 2023, p. 99 [1].

revenues from patents are classified as a group of market



revenues instead of a group of non-economic activities. In the

case that market revenues are considered economic activities,

this classification is false.



At the National Institute of Biology as well as some other

Slovenian research organisations, collaboration projects with

industry such as ARIS applied projects (TRL1-4) are classified



Table 3. Statement of income and expenses of the Institute of Metals and Technology (IMT). Annual report of IMT for 2023, p.

80 [1].





allowed to be performed, and accountants that believe there

should be no margin included in prices of public research

3 QUALITATIVE ANALYSIS organisations’ services.



Awareness and knowledge about state aid legislation varied 3.1 Contract research

among the interviewees. Most of them are well acquainted with Different approaches to establish a pricelist of services and goods

EC Communication 2022/C 414/01, especially those whose main exist. They can be structured using either a cost based approach

profession is technology transfer or accounting. In accounting, or market based approach. In a market based approach,

managers get familiar with state aid rules when there are organisations observe the prices of other service providers, while

investments in bigger infrastructure and financer monitors the in a cost based approach, the costs are summed up and a margin

economic/non-economic activities of the unit using the is added.

infrastructure.

In a cost based approach, direct costs are sometimes joined in

During the semi-structured interviews, the organisational cost blocks composed of work costs, depreciation of the cost of

structure of academia – industry cooperation management was the infrastructure, costs for maintenance and running of the

discussed. Intellectual property management (patenting, infrastructure (electricity, water, heating, ventilation ...) and

licensing etc.) is often centralized, even at large research materials. Such cost blocks are then multiplied based on the

organisations such as Consiglio Nazionale delle Ricerche, number of samples or complexity of the task. Some organisations

University of Ljubljana and University of Maribor. On the add direct costs in % of work, while others in % of all direct costs.

contrary, contract research is decentralised and managed in Both options can be found in public funded calls.

smaller units. Comprehensive and standardised management,

established rulebooks and pricelists often lack at the institutions Prices are often established and then regulated by the inflation

that have a very low percentage of contract research. The rate or other changes in cost structure. Sometimes this does not

management is often left to departments which leads to different take place, especially when the activities are less important for

approaches in price setting etc. Interestingly, there are still some the department or institute.

researchers that are surprised to hear that economic activity is



3.2 Intellectual property in collaboration from the Slovenian Ministry of Higher Education, Science and

projects and start-up companies Innovation, which will consider this recommendation.



Management of intellectual property rights (IPR) in ACKNOWLEDGMENTS collaboration projects was discussed with technology transfer

managers. In most cases it is advised to discuss and agree to IPR This project has received funding from the European Union’s

Horizon 2020 research and innovation programme under grant

in advance. Sometimes even the IPR’ price is evaluated in

agreement No. 101004462, project ATTRACT, ExSACT. We

advance. Internal policy of one interviewed research organisation

would also like to thank all interviewees for taking time to share

regarding the 2022/C 414/01 article 29(c) is that a company can

be co-owner of invention or other IPR only when they provide their valuable experiences.



intellectual contribution, not financial or other in-kind REFERENCES contribution such as equipment usage. Another research institute

has interpreted this article in a way that a financial contribution [1] Communication from the Commission. (2022). Framework for State aid

for research and development and innovation (2022/C 414/01).

of the company to the project, for example 25% in cash, can https://eur-lex.europa.eu/legal-

result in an automated 25% co-ownership of IPR, generated in content/EN/TXT/?uri=CELEX%3A52022XC1028%2803%29.

the project. It is important to note that this institute has a policy 15th [2] Belusky M. (2022). Subsidizing Knowledge Transfer with Public Funds. International Technology Transfer Conference.

that in case of IPR exploitation, the co-owner has to (financially) http://library.ijs.si/Stacks/Proceedings/InformationSociety/2022/IS2022_

compensate the other co-owner(s). Volume-E%20-%20ITTC.pdf.

[3] Schwendinger, G. (2013). State Aid and Intellectual Property in Contract

Research and R&D&I Collaboration. European State Aid Law Quarterly.

The most common procedure to conclude a licence or sale Vol. 12, no. 4, p. 685–698. https://doi.org/10.21552/ESTAL/2013/4/385.

[4] Agency of the Republic of Slovenia for Public Legal Records and Related

agreement is thus to use arm’s length negotiations – article 30(c). Services. Slovenian Business Register (2024). https://www.ajpes.si/prs/.

The licence agreement commonly involves lump sum and [5] Zavod za gradbeništvo Slovenije. (2022). Pravilnik ZAG o prodaji blaga

royalties. One research organisation mentioned that they in storitev na trgu. Cenik storitev ZAG, november 2023.

https://www.zag.si/o-nas/katalog-informacij-javnega-znacaja/.

negotiate with their spin-out companies in the same manner as [6] Kmetijski inštitut Slovenije. (2024) Cenik storitev KIS.

with other companies, which is fair. However, the specific https://www.kis.si/Storitve/.

[7] Geološki zavod Slovenije. (2023). Pravilnik GeoZS o prodaji blaga in

characteristics of start-up companies should be taken into storitev na trgu. Cenik GeoZS za opravljanje prodaje blaga in storitev na

consideration. trgu. https://www.geo-zs.si/?option=com_content&view=article&id=286.

[8] Nacionalni inštitut za biologijo. Storitve in produkti. (2023). Ceniki NIB.

Pravilnik o prodaji blaga in storitev na trgu.

With the implementation of the new Law on Scientific Research https://www.nib.si/images/Prodaja-blaga-storitev_2023.pdf.

and Innovation Activities, research organisations in Slovenia are [9] Czech Ministry of Education and Culture. (2024). Research

Organisations. https://www.msmt.cz/vyzkum-a-vyvoj-2/vyzkumne-

now permitted to establish spin-off companies, which they enter organizace.

into ownership with equity. To our knowledge, no such [10] Czech Academy of Sciences. (2023). Annual Report.

https://www.avcr.cz/cs/o-nas/vyrocni-zprava/.

companies have yet been established in Slovenia. In the last 2 [11] Kemijski inštitut. (2022). Cenik in pravilnik o tržni dejavnosti.

years, CAS has established 5 such companies encountering many https://www.ki.si/za-gospodarstvo/cenik-in-pravilnik-o-trzni-dejavnosti/.

difficulties throughout the process. One of the main concerns is [12] Univerza v Mariboru. Fakulteta za varnostne vede. (2024). Dokumentno

središče. Cenik FVV UM za prodajo blaga in storitev na trgu.

the accuracy and changeability of IPR price. https://www.fvv.um.si/vstopna-stran/o-fakulteti/dokumentno-sredisce/.

[13] Consiglio Nazionale delle Ricerche. (2024) Documenti di bilancio.

https://www.cnr.it/it/documenti-bilancio.

[14] Consiglio Nazionale delle Ricerche. https://www.cnr.it/.

4 CONCLUSIONS [15] Pravno-informacijski system Republike Slovenije. (2022) Slovenian Law

on Scientific Research and Innovation Activities.

Differentiation between contract research and collaborative https://pisrs.si/pregledPredpisa?id=ZAKO7733.

research in Slovenian research organisations is not well known [16] Pravno-informacijski system Republike Slovenije. (2023). Rulebook on

and could be improved. There are many projects between https://pisrs.si/pregledPredpisa?id=PRAV7654 procedures for implementing the budget of the Republic of Slovenia.

.

research organisations and companies that fall somewhere [17] Kaiser, M. Neu, F. Teernstra. (2021) State Aid on R&D&I – The Right

between contract research and collaborative research. For Way. EARTO Report. https://www.earto.eu/wp-content/uploads/EARTO-

Report-on-State-Aid-on-RDI-The-Right-Way-Final.pdf.

example, different methodologies in literature were tested ([17], [18] Kebapci, H., Von Wendland, B. & Kaymaktchiyski, S. (2020). State Aid

[18]), which produced two different results for the same project. Rules in Research, Development & Innovation. Addressing Knowledge

and Awareness Gaps among Research and Knowledge Dissemination

Nevertheless, due to the obligation to account for the economic Organisations. Decision Tree, Kaiser, L. (Ed.), Neu, M. (Ed.), Teernstra,

activities separately, each such project should be labelled as F. (Ed.), Nicolaides, P. (Ed.), EUR 30436 EN, Publications Office of the

either (i) a contract research or (ii) a collaborative research. European Union, Luxembourg. https://op.europa.eu/en/publication-

detail/-/publication/10740aa6-223a-11eb-b57e-01aa75ed71a1/

According to our discussion, this happens only rarely. As

discussed above, the activities in Slovenia are divided into public

service and market activities which makes it more complicated

to introduce another set of classification. Label ‘economic

activity’ or ‘non-economic activity’ should be assigned during

the process of bookkeeping, i.e. when the invoice is issued or

contract concluded. A more significant, but important change of

replacing wording ‘public service/market activity’ with

‘economic/non-economic activity’ in Slovenian legislation

should be made. We communicated this with representatives





Feasibility Analysis for the New Mechanism of Knowledge




Transfer within the INDUSAC Project



Duško Odić Urška Mrgole Marjeta Trobec

Office for Project Informatics, Office for Project Informatics, Office for Project Informatics,

Organization of Thematic Events Organization of Thematic Events Organization of Thematic Events

and Conferences (SPIK) and Conferences (SPIK) and Conferences (SPIK)

Jožef Stefan Institute Jožef Stefan Institute Jožef Stefan Institute

Ljubljana, Slovenia Ljubljana, Slovenia Ljubljana, Slovenia

dusko.odic@ijs.si urska.mrgole@ijs.si marjeta.trobec@ijs.si



ABSTRACT

1 INTRODUCTION

In September 2022, the Horizon Europe INDUSAC project

introduced a novel mechanism for knowledge transfer, extending

the usual company-researcher partnerships to include students as In September 2022, the Horizon Europe INDUSAC project

well. Between March and May 2024, thirteen co-creation (www.indusac.eu; EU project number 101070297) introduced a

projects involving international teams of students and novel mechanism for knowledge transfer, extending the usual

researchers solving companies’ challenges, were carried out. company-researcher partnerships to include students as well. It

This study describes results of surveys given to companies, comprises a methodology that would allow for a streamlined



and the level of usefulness of solutions made possible by the an online platform to support that methodology [1]. In November collaboration. We analyzed data collected from 10 companies, students and researchers about their experience in the projects, facilitation of collaboration between industry and academia, and

2023, the INDUSAC project, coordinated by the Jožef Stefan



satisfaction of companies with technical aspects of the Institute, commenced its piloting phase wherein universities, 57 students and 4 researchers. Measured on the Likert scale,

methodology ranged from average to good (average values public research organisations, and companies were invited by the

between 3.1 and 4.2), whereas their satisfaction with the solution international project consortium to join the project. The idea

to their challenge, and with the work done by the team, had a behind the methodology is to bring together a company and an

narrower range between 3.2 and 3.8. Financial support to student international team of 3-6 students and/or researchers to solve a

members of co-creation teams, in the amount of up to 1,000 EUR company challenge within 4-8 weeks, with the company

gross per student, was perceived as sufficient by 67% of students. providing assistance during regular meetings with the team. The



relatively well accepted among companies, with room for deliverables specific for the type of challenge, and the improvement in certain aspects such as the user-friendliness of deliverables are evaluated by companies. Being the main target Initial results indicate that the INDUSAC mechanism is team delivers results in the form of pre-defined types of



around 30 % of co-creation projects have demonstrated true audience, during the project, special attention was given to the platform and the time allowed to solve a challenge. Overall,

value to the company involved, and there is potential in the students / researchers from EU widening countries, and

further 50 %. Selected testimonials from companies, geographical and gender balance was ensured by the criteria that

complimenting the work of students and expressing their own team members must be from at least three different countries, and

belief that the students are richer for the experience as well, must include representatives of at least two gender groups;

demonstrate that the INDUSAC mechanism shows promise in student members of the co-creation teams were financially

knowledge transfer. rewarded for successfully completing the project. First such



KEYWORDS 2024. This study describes results of surveys given to companies, collaborations started in March 2024 and wrapped up in May

INDUSAC project, international cooperation, student-industry students and researchers about their experience in the project,

cooperation, knowledge transfer and the level of usefulness of solutions made possible by the

collaboration. Implications for the feasibility of this concept of

knowledge transfer are discussed.



2 METHODS

Permission to make digital or hard copies of part or all of this work for personal or

classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full As per the methodology of the project, students and researchers

citation on the first page. Copyrights for third-party components of this work must were surveyed before they started working on the solution to the

be honored. For all other uses, contact the owner/author(s).

Information Society 2023, 9–13 October 2023, Ljubljana, Slovenia company's challenge, and after they finished. Topics in the

© 2020 Copyright held by the owner/author(s). survey, relevant to the scope of this study, included the students'

feedback on how the collaboration affects social impact, and how

appropriate the funding is. In addition, companies were surveyed

after the project, mainly about the quality of work performed by

the students / researchers, and the impact that their work has. All

surveyed individuals were asked to provide short testimonials

about their impressions and satisfaction. Students / researchers

were asked to fill in separate surveys for separate co-creation

projects (maximum three) and companies were likewise asked to

fill in separate surveys for each team they worked with. Further

details are indicated in the Results section. In this study, we

analyzed data collected from 10 companies, 57 students and 4

researchers.



3 RESULTS



Figure 2: Satisfaction of companies with work of the co-

In the first round of the INDUSAC co-creation projects, creation team and the solution delivered. Average values ±

taking place between March and May 2024, thirteen co-creation sd are shown (n = 10). Of the categories surveyed, Satisfaction,

projects took place that resulted in proposed solutions, two of Quality of Work, and Soundness refer to the work done by the

which were rejected and eleven approved by companies. co-creation team, whereas Relevance, Market Potential,

Companies' overall satisfaction with the INDUSAC process after Improvement over existing solutions, Creativity, and

the projects, expressed as various aspects of the methodology, is Innovativeness refer to the solution delivered. Satisfaction was

shown in Figure 1. Satisfaction was evaluated on a Likert scale measured on a Likert scale: 1 – very poor, 2- poor, 3 – average,

from 1 to 5. On average, the processes of registering on the 4 – good, 5- very good.

platform, publishing Challenges, and reviewing Motivation

Letters (ie., students' applications) ranked highest at 4.0, 4.0, and

4.2, respectively, while the user-friendliness of the INDUSAC In terms of delivery of results, the companies have reported

platform and the time allowed to solve a challenge ranked lowest, that all requested deliverables had been delivered by the co-

each at 3.1. creation teams in all cases except one (representing one of the

Companies' overall satisfaction with the solution to their projects where the solution was rejected). In terms of follow-up

challenge, and with the work done by the team, expressed as on the solution within the company, indicating its usefulness, two

various attributes, is shown in Figure 2. Satisfaction was companies have already started, a third company has confirmed

evaluated on a Likert scale from 1 to 5. On average, relevance of that they will follow up on the solution, while 5 have not yet

the solution, quality of work of the team, and satisfaction with decided and in two cases it will probably not happen.

the work of the team ranked highest at 3.8 each, while the market Since the INDUSAC project put a fair amount of emphasis on

potential of the solution ranked lowest, at 3.2. social aspects such as geographically and gender-balanced

collaboration, the survey for students and researchers included



questions on agreement with (i) incorporation by the co-creation



process of customer research and insights to understand the end-

users' needs and preferences, (ii) solutions that specifically

addressed gender-related issues or considerations, and (iii)

successful prioritisation of the human aspect (Inclusivity, Gender

dimension, Interdisciplinarity, User Perspective, Collaboration,

Iterative Feedback, Ethical Considerations) and creation of a

meaningful and inclusive environment. Results are shown in

Figure 3. Agreement was evaluated on a Likert scale from 1 to 5.

On average, all three categories ranked fairly high, between 4.0

and 4.5.

Lastly, questions about the adequateness of financial support

to students, were also included in the survey. As per the

INDUSAC methodology, each student received up to 1.000 EUR

gross for a successfully finished project, and this amount was

reduced as the number of students per team increased, as each

team received up to 3.000 EUR gross. Results, demonstrated as

Figure 1: Satisfaction of companies with technical aspects of distribution of opinions among different geographical groups

the methodology. Average values ± sd are shown (n = 8 for (ie., EU member states, widening countries, and EU associated

assistance to the team, n = 9 for support material, submission of countries), are given in Figure 4, and indicate that overall,

deliverables and punctuality of submission, and n = 10 for the between 58% and 70% of students agree that funding was

other nine categories). Satisfaction was measured on a Likert sufficient.

scale: 1 – very poor, 2- poor, 3 – average, 4 – good, 5- very

good.





and taking two out of thirteen solutions rejected into account,

companies have expressed a fair level of general satisfaction with

the solutions and the work done by the teams (Figure 2). It is

likely that this was aided by the methodology sections which

defined interim reviews and evaluation steps (eg., reviews of

challenges before publishing, reviews of Motivation Letters

before starting, etc.), and regular communication between

companies and co-creation teams during the project. In all except

Figure 3: Agreement of students / researchers with in one case, all deliverables were satisfactorily produced by the

incorporation of customer-oriented and human-focused teams, indicating that the supporting documents that comprised

elements in the projects. Average values ± sd are shown (n = the deliverables, and which were developed within the

61). Agreement was measured on a Likert scale: 1 – very poor, INDUSAC consortium, served as useful guidelines for particular

2- poor, 3 – average, 4 – good, 5- very good. type of challenge.

Having the project open to a wide range of challenge types

also proved beneficial as among the 13 projects for which

solutions were provided, seven out of nine possible challenge



types were represented, and distribution among different

challenge types was fairly even, with 'Marketing campaigns' and

'Service and product ideas' being most preferred.

An additional advantage was presented by the fact that the

efforts to facilitate knowledge transfer between industry and

academia are financially supported within the INDUSAC

scheme. This type of support is particularly welcome, as the lack

of funding is a frequent barrier for student-industry collaboration

[5,6]. Around two thirds of surveyed students found funding to

be adequate, and the largest percentage of this opinion was found

among students from widening countries (Figure 4) indicating

that the funding scheme shows promise for the major target

Figure 4: Perception of adequacy of funding within the group of the project.

scheme in the INUDSAC project, as surveyed among There is, however, room for improvement – not least based

students and researchers from different countries of on comments given by the companies themselves. Geographical

residency. Within EU member states, there were no opinions balance, for example, may in some cases be an obstacle, as, in

towards [moderate]. Total number of individuals responding one company's opinion, having a team with members from

was 5 in EU Member States, 37 in widening countries, and 12 different countries can make it difficult to work on projects that

in associated countries. require physical experience with a product. It is likewise

important to be able to streamline the process, which needs to be

backed by a reliably functioning platform, as well as to unify the

4 DISCUSSION working space, as it was, in one company's opinion, difficult to

keep track which information they received from which

The INDUSAC approach set out to bring several advantages platform. Lastly, as mentioned, companies have expressed

to the existing landscape of knowledge transfer practices, such as interest in a more flexible data management, as the project's

inclusiveness represented by gender balance, international timeline may prove too rigid. In terms of funding, one student

cooperation by mandatory geographical diversity, enhanced pointed out that it would have been preferable to receive funding

support to widening countries by mandatory representation in the during the project rather than after, to allow for traveling to

teams, and expansion to include students via mandatory companies and collecting data. The problem of limited mobility

participation of at least one student per team solving companies' was also perceived by companies, two of which stated that the

challenges; some of these have already shown to be biggest challenge in projects related to physical products was that

advantageous for companies [2-4]. Our results point to initial the participants cannot get to know and test the products live, and

indications that the INDUSAC mechanism, comprising the that creativity may be limited due to the lack of face-to-face

methodology and the platform, is relatively well accepted among interaction with products and colleagues.

companies (Figure 1), with room for improvement in certain

aspects such as the user-friendliness of the platform and the time

allowed to solve a challenge. The latter points to a general 5 FUTURE PERSPECTIVES

enthusiasm among companies to engage in finding solutions for

more serious challenges as well, which is encouraging – in two The INDUSAC project set out to show that companies

cases, work is already under way to continue with the projects, benefit from a particular type of knowledge transfer in the form

and overall, around 30 % of co-creation projects have of creative young minds, that this knowledge transfer brings

demonstrated true value to the company involved, and there is satisfactory results and useful solutions, and that the gender and

potential in the further 50 %. So even with the constraints given, geographical balance, as well as the inclusion of social elements



(Figure 3) have a positive effect on the overall process ACKNOWLEDGEMENTS (satisfaction by teams, satisfaction by companies). While we did

not perform any control studies (for example, with single-gender

Work described in this manuscript has received funding from

teams) to truly test the effect of gender balance, there was a slight

the European Union’s Horizon Europe Programme under grant

positive effect of (i) number of team members and (ii) ratio of

agreement No 101070297.

female-vs-male team members, on company's satisfaction with



results and quality of work (unpublished data). Other results and REFERENCES selected testimonials from companies, complimenting the work

of students and expressing their own belief that the students are [1] Odić D, U Mrgole, M Trobec. 2023. New initiatives for knowledge

transfer between industry and academia : the INDUSAC Project. In: 16th

richer for the experience as well, demonstrate that the INDUSAC International Technology Transfer Conference : 8 October 2023,

mechanism shows promise in knowledge transfer, and the Ljubljana, Slovenia : Information Society - IS 2023 : Proceedings of the

26th international multiconference E: 58-61

rejected solutions stand as reminders that even following the [2] Cavero-Rubio JA, A Collazo-Mazón, A Amorós-Martinez. 2019. Public

careful process of team assembly and selection, monitoring of recognition of gender equality in the workplace and its influence on firms'

performance. Women’s Studies International Forum 76: 102273

the work done needs to be vigilant for it to lead to satisfactory [3] Chang T-LS, C-M Chuang, W-S Jan. 1998. International Collaboration of

results. With this in mind, the INDUSAC methodology is Law Firms: Modes, Motives and Advantages. Journal of World Business

continuously improving and mechanisms are put in place to 33, 3: 241-262

[4] Ratten V. 2016. International Collaboration and Knowledge Transfer

minimize such occurrences. The challenges that remain also among Universities and Firms Affecting Regional Competitiveness.

include attracting larger numbers of companies and students / Thunderbird International Business Review 58, 1: 91-93

researchers to engage into cooperation, but the level of success [5] Ejubovic A, G Cerinsek, T Davey, A Meerman, V Galan-Muros, B

Orazbayeva. 2017. State of University-Business Cooperation: Slovenia -

described here represents a strong starting point. University Perspective. EC University-Business cooperation in Europe;

Country Reports - Slovenia

[6] Morisson A, M Pattinson. 2020. University-Industry Collaboration. Lille:

Interreg Europe Policy Learning Platform





Approaches to Monitoring and Impact Assessment in




Research Infrastructures



Jure Plaskan Barbara N. Brečko

Faculty of Social Sciences Faculty of Social Sciences

University of Ljubljana University of Ljubljana

Ljubljana, Slovenia Ljubljana, Slovenia

jure.plaskan@fdv.uni-lj.si barbara.brecko@fdv.uni-lj.si



ABSTRACT In recent decades the significance of research infrastructures has

become increasingly evident across all fields.

Impact assessment is a critical process in understanding the

broader effects of research infrastructures (RI) on various sectors Although RIs are primarily designed to meet research needs,

such as science, society, the economy, and policy-making. It their influence extends well beyond promoting scientific

helps RI identify their strengths, weaknesses, and areas for excellence. The advanced technological capabilities and

concentration of skilled expertise they provide can stimulate

improvement. The paper addresses the challenges of monitoring

innovation, create or expand markets, attract foreign investment,

and evaluating the impact of RI, focusing on the distinction

boost economic activity, and potentially enrich the social and

between performance monitoring and impact assessment. It cultural life of a region. [2]

emphasizes the importance of demonstrating the broader

societal, economic, and scientific impacts of RIs to inform public RIs necessitate relatively large and long-term financial

policy and secure funding. In the article we address different investments, making it crucial for investors, policymakers, and

other stakeholders to ensure that these infrastructures operate

methodological approaches to impact assessment and self-

successfully and effectively, contributing to scientific

evaluation of RIs as well as the possible challenges in these

advancement and addressing societal and economic challenges.

processes. The paper advances the integration of multiple

evaluation approaches to provide a robust and detailed Although reflections and publications on defining and measuring

assessment of the contributions RIs make to society, the impact have increased in recent years, there is still no unified

economy, and scientific development. framework or consensus on how to assess the impact of RIs.

Therefore, it is crucial to explore the potential for developing

such a framework and investigating its practical application.

KEYWORDS



Impact assessment, monitoring, research infrastructures 2 PERFORMANCE MONITORING AND

IMPACT ASSESSMENT

1 INTRODUCTION

Research Infrastructures (RIs) are essential facilities that offer To this end, various solutions have been developed to enable

resources and services to research communities, enabling them stakeholders to monitor performance and evaluate the impact of

to conduct research and drive innovation. Beyond their primary RIs. However, there is a distinction between these two activities,

role in research, these infrastructures can also support education, which this paper aims to clarify. The concepts of performance

public services, and other non-research activities. They may take monitoring and impact assessment represent two distinct yet

various forms, including single site, distributed, or virtual setups. related processes for evaluating the activities of institutions.

Although both processes involve data collection and analysis of

RIs encompass human resources, major equipment, and/or sets

RI performance outcomes, their focus, scope, and objectives

of instruments, as well as resources containing knowledge, such

differ.

as collections, archives, and databases. They are used by

scientists from various disciplines – e.g. astronomy, biology,

Performance monitoring, often simply referred to as

chemistry, physics, human and social sciences, etc. RIs can “monitoring”, involves the systematic and regular collection and

maintain their competitive advantage only if they keep pace with

analysis of data related to activities and outcomes. This process

the latest advancements in relevant scientific fields and the

is crucial for assessing progress toward predefined goals,

newest techniques and technologies. Therefore, it is crucial for

identifying areas where activities are achieving success, and

RIs to connect with the research community and industry to stay

pinpointing areas that require improvement. Typically,

aligned with developments in both science and technology. [1]

performance monitoring focuses on tracking key performance

indicators (KPIs), which serve as measurable values that reflect

the effectiveness and efficiency of the activities being evaluated

(e.g. Number of publications, Number of master and PhD

Permission to make digital or hard copies of part or all of this work for personal or students using the RI, Outreach through media, ...). classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full Impact assessment, in contrast, focuses on identifying and citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s). evaluating the changes within the broader ecosystem that result

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia from the activities and outcomes of RIs. This process aims to

© 2024 Copyright held by the owner/author(s).

determine which specific RI activities lead to impacts across



various domains. A well-established approach, developed 2.1 Defining Areas for Impact Assessment

through European initiatives (such as the RI-PATHS [3] project),

is the concept of “impact pathways”. This method enables Impact assessment becomes especially crucial in times of limited

evaluators to trace the different routes through which activities

public funding for science. By highlighting the effects of RIs on

translate into impacts at various socio-economic levels. science, society, the environment, the economy, and other

sectors, impact assessments can demonstrate the value of both

Impact assessments can be conducted either before or after the potential and actual investments in RIs. This analysis helps to

implementation of RI. When carried out during the planning

phase, this process is known as an “ex-ante” impact assessment. underscore the relevance of these investments in addressing

societal needs. Moreover, impact assessments provide

Its purpose is to forecast the potential impacts of the RI,

policymakers with a clear picture of the broader benefits that RI

anticipate its effects, and inform strategic planning to ensure activities offer, thereby supporting the development of informed

those outcomes are realized. This type of assessment is largely

public policies and decision-making.

conceptual and, to some degree, abstract. Once the RI is

established and fully operational, an “ex-post impact



assessment” is conducted to evaluate whether the RI has Impact assessment is closely tied to the goals of RIs and the 1 expectations they set. The ESFRI working group on RI successfully met its intended objectives.

performance monitoring has identified nine objectives which are

When determining criteria and indicators for monitoring and relevant to RIs [7], and largely correspond to the following five

evaluating (e.g., research infrastructures, measures, programs, impact areas:

policies), it is crucial to recognize the differing roles of these two Contribution to Scientific Excellence: At the heart of •

processes. Monitoring focuses on real-time oversight of every RI is the drive for scientific excellence. RIs

implementation: as a funder, one needs to know the current contribute in numerous ways, including data collection and

status, whether progress is on track, whether funds have been preservation, providing access to infrastructure and

appropriately allocated, whether a sufficient number of target databases, sample collection and dissemination,

audiences has been engaged, etc. While monitoring can alert us facilitating analytical experiments, offering software, and

that things are not proceeding as planned, it does not reveal the providing general support to researchers. These activities

causes of deviations nor provide adequate information for are fundamental to the research process, fostering scientific

making necessary corrective actions [4]. progress by advancing innovative research, expanding the

frontiers of knowledge, and generating new insights and

On the other hand, the role of evaluation is to explain how the discoveries.

institution/measure/program/policy functioned, how successful Addressing Societal Challenges •: In recent years,

it was in achieving its objectives, and what its impacts were. addressing societal challenges has become an increasingly

Evaluation allows us to determine success, identify what worked important focus for RIs. Their impact ranges from

and what did not, and, if not, what changes need to be made in contributing to the United Nations’ Sustainable

future planning. The focus of evaluation may be on assessing the Development Goals and the European Green Deal to

degree to which objectives are achieved, or it may focus on the enhancing public understanding of science.

process of implementing the instrument/program/policy itself. •

Contribution to Innovation and Economic

Development: Given the substantial financial investments

Impact assessment is beneficial for RIs when used to evaluate required by RIs, it is crucial to highlight their role in

and enhance their functioning. It plays a crucial role in the

driving innovation and economic growth. This can be

strategic planning of an RI by informing decisions on internal reflected in job creation, economic development, or

resource allocation and driving continuous improvement and

increased competitiveness, particularly at local, regional,

alignment of services with the needs of users and other and national levels. Large RIs, in particular, employ a

stakeholders. Additionally, impact assessment fosters

significant workforce and, in some cases, make substantial

accountability and transparency, thereby enhancing the

investments in constructing and offering high-value-added

legitimacy, visibility, and overall value of the RI. Furthermore, it

components.

serves as a platform for meaningful dialogue and exchange •

Contribution to Policy-Making: Research facilitated by

among relevant stakeholders regarding the objectives, direction,

RIs can significantly inform policy-making across various

and operations of RIs, which can be exceptionally valuable. [5]

thematic areas. This is especially important for



The OECD defines impact as “the extent to which an intervention organizations responsible for policy development at the

European or national level.



intended or unintended effects at a higher level.” [5] The Contribution to Human Resource Development: Many has produced, or is expected to produce, positive or negative, •

European Commission mandates the implementation of impact RIs also focus on education and training, often dedicating

assessments for every policy intervention or law (including significant resources to these efforts. As centers of

investments in research infrastructure and their activities) scientific excellence, they play a crucial role in developing

expected to cause significant effects or require substantial human resources and training the next generation of

financial resources. Impacts represent all “direct or indirect scientists. They impact their users and their careers through

changes” relative to the baseline scenario. Such impacts may enhanced scientific excellence, productivity, networking,

occur over different time periods, affect different stakeholders, and training opportunities.

and be relevant at different levels (local, regional, national, and

EU) [6].



1 European Strategy Forum on Research Infrastructures



Listed areas are not relevant only to RIs, but can be relevant also defining “impact pathways”. The impact pathway approach was

to other research organizations. further developed in the RI-PATHS project, which explores

more details than the logical framework and provides a



3 descriptive vision with more information on causes and effects. METHODS AND APPROACHES FOR

6. Case studies: This approach involves an in-depth analysis of

MEASURING IMPACT a specific case to understand the effectiveness of a policy or

In the RI-PATHS project [3] a comprehensive review of project. The analysis focuses on the specific context, identifying

literature was conducted on methodologies for evaluating and factors contributing to success or failure and deriving lessons that

measuring the socio-economic impacts of RIs. The project can be applied to future policies and projects. When used in

focused on ex-post impact evaluation methodologies, which are impact evaluations, case studies aim to better reflect the

employed during the operation of RIs when it is possible to uniqueness and complexity of RIs.

ascertain whether they are creating certain impacts and in what

manner. The effectiveness of the analysis is demonstrated It is evident that some approaches are more suitable for assessing

quantitatively (e.g., through indicators) or qualitatively (e.g., economic rather than social or scientific impact, and vice versa.

through case studies). [8] In general, these approaches can complement each other—some

are more quantitative, such as macroeconomic modelling or cost-

Six main approaches/methods for measuring impact based on the benefit analysis (CBA), while others are more qualitative, like

literature review were identified: case study descriptions.

1. Socio-economic assessment based on impact multipliers: The RI-PATHS project systematically evaluated each of the

This approach evaluates the socio-economic impact of a policy mentioned approaches using criteria such as reliability, validity,

or project by quantifying the effects on various economic sectors. precision, cost and time efficiency, and relevance to both

The assessment is based on impact multipliers that estimate the policymakers and research infrastructure managers. It is evident

indirect effects of the policy or project on the economy. This that no single methodological approach can comprehensively

approach expresses impacts on aggregated macroeconomic address all the questions intended for impact evaluation.

variables such as GDP, gross value added, or employment. The However, combining different approaches can offer greater value

main advantage of this methodology, which is grounded in a and effectiveness compared to relying on existing methods alone.

well-established theory and uses input/output analysis tools, is its

reliability in producing reproducible and comparable project 4 IMPACT PATHWAYS AND

results. However, its limitation is its restricted validity, as it often INDICATORS cannot reliably measure non-monetary effects (e.g., cultural,

social, and environmental). While there is not a universally accepted approach to impact

2. Methodologies utilizing the knowledge production assessments in RIs, the work of the RI PATHS project has, as

function: mentioned, become well established in Europe. Indeed, results This approach focuses on the impact of research and

development activities on the economy. The knowledge from the survey conducted by ESFRI among RIs [7] show that

production function method quantifies the relationship between impact pathways have become a common method for impact

research and development investments and economic growth. assessments among European RIs. Several RIs have conducted

The approach focuses on only a small portion of the expected their impact assessments with the help of impact pathways as part

socio-economic impacts of RIs. of the RI-PATHS pilot exercises (for example, ALBA, ELIXIR,

3. Cost-benefit analysis (CBA): EATRIS) [9]. Identifying impact pathways was also an integral This approach compares the

advantages and disadvantages of a policy or project and component of the impact assessment of ICOS [10].

determines whether the benefits outweigh the costs. The analysis

considers both quantitative and qualitative factors to enable well- The mechanism of impact pathways is recommended as a way to

informed decision-making. All costs or benefits are monetized, demonstrate causal links between inputs, various activities and

even if the effects are not solely financial. Governments and outputs of RIs, and their identifiable impacts [3] [11]. These can

economists frequently use this approach to assess the impact of be both intended or unintended – while impact pathways always

various investment projects. It is reliable for comparing positive have a clear origin in one or few related activities, which are

and negative effects and can capture numerous RI impacts. under control of RIs, these activities branch out into different

However, it can be expensive and time-consuming and has directions and trigger effects in different areas, which can be

limited causal explanatory power. Additionally, it may not outside the sphere of influence of RIs. An example of exploring

always capture all drawbacks. impact pathways according to spheres of control, influence, and

4. Multi-methods multiple partial indicators: interest can be seen in AnaEE’s position paper [12], which This approach

combines multiple methods and indicators to evaluate the impact sought to build a framework that would specify AnaEE’s position

of policies or projects. Methods can include surveys, focus in the chain of actors generating impact in its scientific field.

groups, and statistical analysis, while indicators encompass

economic, social, and environmental factors. An example of this In order to map the path from activities of RIs to outcomes and

approach is the OECD framework for socio-economic impacts, impacts, it is crucial to systematically collect data. This is

which includes a list of 25 essential impact indicators and 58 recommended for both performance monitoring and impact

additional standard indicators. assessment. Several lists of indicators have been proposed in

5. Theory-based approaches: recent years (OECD, RI PATHS, ESFRI WG). The indicators These approaches rely on

established economic or social theories to evaluate the impact of can vary – from those that primarily measure performance, also

a policy or project. They depend on theoretical models and known as key performance indicators (KPIs) [5], and those

empirical evidence to predict impact. A typical example is the which are focused on impact (e.g. OECD prepared a list of

“logical framework/model”, which is based on a logical impact indicators) [13]. The purpose of impact indicators is to

sequence of steps from inputs to impacts. Theory-based create a link to strategic objectives of RI, as well as to different

approaches share common features such as considering the areas of impact that RIs create. In addition to the connection of

broader context and external factors that can affect success and indicators with strategic goals, the OECD recommends that the



indicators provide information related to operational issues and It is important to note that some challenges may be specific (or

that the data is measured in a specific time frame. more common) to a certain type of RI or to certain thematic areas

they cover. To address this challenge, the recommendation is to

Impact indicators can be quantitative or qualitative, e.g. in form avoid directly comparing impacts of RIs, and to consider the

of “narratives”. This information is usually collected via tailored diversity of RIs. When deciding on a methodology, it is advisable

methods, such as interviews, surveys, or case studies. These to tailor the selected methodology to each RI, and first establish

indicators are more difficult to be standardised and must be a consensus between RIs, funders, governments and other

tailored for specific RIs and depend on the context. These relevant stakeholders. This agreement should establish clear

methods can help RIs to report on intangible impacts. expectations regarding the objectives of the RI and the

assessment itself. However, all RIs should strive to demonstrate

impact in the field of scientific progress, while considering

5 various other socio-economic impacts. INTERNAL AND EXTERNAL IMPACT



Both external and internal evaluations are relevant for assessing EVALUATION Providing adequate resources for the implementation of an impact assessment is indeed challenging, in particular as it is necessary to adopt a long-term plan for evaluation in order to the impact of RIs, each with its advantages and disadvantages. capture impacts that take years to reveal. At the same time the External evaluations are conducted by independent evaluators data collection needs to be done systematically and begin early who assess the institution's impact. This approach ensures an enough, which can be more resource intensive although also objective and impartial assessment, as external evaluators are not helps to lower the amount of “ad-hoc” data collection when affiliated with the evaluated institution and are, therefore, less conducting impact assessments. likely to be influenced by internal biases or personal interests.



Additionally, external evaluations can provide new perspectives In spite of these challenges, impact assessments provide

and insights that are not available to internal evaluators. important information for all RI stakeholders, as well as the

However, external evaluations are often costly and time- general public, as they allow RIs to demonstrate their

consuming and may sometimes fail to account for the contextual contributions to science, society and the economy, and help

nuances and priorities of the evaluated institution. improve their performance. As such, they can be used as means

to communicate about RI activities. Promoting and

In contrast, internal impact evaluations rely on an evaluation disseminating the results of these evaluations can subsequently

process conducted by the institution's staff or stakeholders. This help promote positive RI development and funding. approach is more cost-effective and efficient, as internal



evaluators are already familiar with the institution and its 7 CONCLUSIONS operations. Internal evaluations may also better consider the

institution's contexts and priorities and be more adaptable to Despite the growing focus on this area, there remain significant

changes in RI goals. However, internal evaluations may be challenges in developing a unified and comprehensive

biased due to internal motivations and conflicts of interest and framework for evaluating such impacts, particularly when

may lack the objectivity and independence of external accounting for both economic and non-economic factors. There

evaluations. Moreover, internal evaluators may be limited by is a reason for that – the unified methodology cannot adequately

their knowledge and expertise, reducing their ability to bring new address all aspects of variety of RIs and the diversity of fields

insights and perspectives. where they operate. There is a number of methods which can be

applied, and future work could explore how combinations of

The choice between external and internal evaluation often different methods (e.g. quantitative, such as macroeconomic

depends on internal capabilities, available resources, evaluation modeling and cost-benefit analyses and qualitative such as case

objectives, and so on. To ensure a comprehensive and balanced studies or theory-based assessments) can be effectively balanced.

assessment, it is beneficial to combine both approaches. It is also This could provide more holistic view on RI impacts, especially

increasingly common for institutions (including RIs) to in understanding intangible impacts like societal and

periodically self-evaluate, thereby preparing for external environmental changes.

evaluation.

There are already lists of indicators suggested to be used for



6 impact assessment, nevertheless the selection of indicators CHALLENGES AND OBSTACLES should be done with a great deal of prudence and not to be used There are several challenges that RIs encounter while conducting to compare RIs, given the diversity in their structure and impact assessments. Some of them were outlined by respondents objectives. to an ESFRI survey among RIs (2023). According to the survey,

a recurring challenge was to identify an appropriate method or

framework or finding appropriate indicators. Other respondents ACKNOWLEDGEMENT

mentioned the amount of resources required and the time frame

needed to properly evaluate the impacts of their infrastructure. In The research was supported by Slovenian research agency, grant

general, some RIs are concerned that impacts may not be number V5-2283. properly detected. This is a similar issue to what was described



in the ERIC forum’s “Report on Socio-economic impact REFERENCES framework” [14] as a “traceability” problem – there is [1] ESFRI Scripta Vol. 2 (2017). Long-Term Sustainability of Research uncertainty about how to link RI activities or data generated Infrastructures

within an RI to their subsequent use. One of the RIs responded https://www.esfri.eu/sites/default/files/ESFRI_SCRIPTA_SINGLE_PAGE_19102

that measuring innovation or social impacts could take several 017_0.pdf.

decades. [2] Griniece, E., Reid, A., Angelis J. (2015). Evaluating and Monitoring the Socio-

Economic Impact of Investment in Research Infrastructures. Technopolis Group.

Available at:



https://www.researchgate.net/publication/275037404_Evaluating_and_Monitoring

_the_Socio-Economic_Impact_of_Investment_in_Research_Infrastructures

[3] RI-PATHS. Research Infrastructures’ Impact Assessment Toolkit. https://ri-

paths-tool.eu/en

[4] BUČAR, M. (ur.), ČRNIGOJ, M. (ur.), LIPNIK, A. (ur.). (2020). Vrednotenje

sodelovanja med znanostjo in gospodarstvom. Ljubljana: Fakulteta za družbene

vede, Založba FDV

[5] OECD (2021). Applying Evaluation Criteria Thoughtfully, OECD Publishing,

Paris, https://doi.org/10.1787/543e84ed-en.

[6] EC (2023). Better Regulation Toolbox

[7] ESFRI (2019). Working Group Report. Monitoring of Research Infrastructures

Performance. Available at:

https://www.esfri.eu/sites/default/files/ESFRI_WG_Monitoring_Report.pdf

[8] Giffoni F, Schubert T, Kroll H, Zenker A, Griniece E, Gulyas O, Angelis J, et

al (2018) State of play - literature review of methods to assess socio-economic

impact of research infrastructures. RI-Paths project. H2020 grant 777563. Available

at https://ri-paths.eu/wp-content/uploads/2018/03/D3.2_Report-on-stocktaking-

results-and-initial-IA-model.pdf (last accessed 1 August 2024)

[9] Griniece,E., Angelis,J., Reid,A., et al. (2020). Guidebook for Socio-Economic

Impact Assessment of Research Infrastructures.

Doi: https://doi.org/10.5281/zenodo.3950043

[10] Van Belle,J., Van Barneveld,J., Bastiaanssen,V., et al. (2018). ICOS Impact

Assessment Report. Technopolis Group. Available at https://www.icos-

cp.eu/sites/default/files/2018-09/ICOS_Impact_Assessment_Report_2018.pdf

(last accessed 18th September 2024)

[11] Lutz, G. Kolar, J., Brečko, B. N., Plaskan, J., et al. (2023). Assessment of

impact of RIs : policy brief. [Brussels]: European Strategy Forum on Research

Infrastructures (ESFRI), 2023. DOI: 10.5281/zenodo.8091633

[12] Boer,M., Chavel,S., Mahé,S., and Plaude,S. (2021). Setting a relevant

framework to assess AnaEE’s impact. Available at

https://www.anaee.eu/sites/default/files/Mediatheque/Resources/reportsdocuments

/setting_a_relevant_framework_to_assess_anaees_impact.pdf (Last accessed on

18th September 2024)

[13] OECD (2019). Reference framework for assessing the scientific and socio-

economic impact of research infrastructures. OECD Science, Technology and

Industry Policy Papers, No. 65, OECD Publishing, Paris

https://doi.org/10.1787/3ffee43b-en

[14] ERIC Forum Implementation Project (2022). Report on SEI ERIC Framework.

Retrieved 16 August 2024, from https://www.eric-forum.eu/wp-

content/uploads/D4.3-EF-Report-on-SEI-ERIC-Framework.pdf





Intellectual Property Valuation




in the Cyber Security Sector



Marta E. Wachowicz

Technology Transfer Unit

NASK - National Research Institute

Warsaw, Poland

marta.wachowicz@nask.pl



ABSTRACT / POVZETEK to understand the economic value of cyber IP assets by carrying

out an IP valuation. The article addresses a topic specific and

relevant to the digital economy, the literature abounds with

This paper explores the applicability of intellectual property

methodologies for different approaches to valuing IPR [9], [1],

rights (IPR) valuation methods in cyber security by using the

but there is no guidance on how to consider the importance of

criteria of the Artificial Intelligence development phase model.

data, learning models, and all aspects of AI in new inventions.

After analysis of the interconnections and interdependency in

The challenges faced by the cyber security sector are defensive

cyber security products, an approach to data quality is proposed.

AI and machine learning technology, sophisticated cyber attacks,

It is worth emphasizing that the process of valuating IPRs is

reinforcement learning-based cyber attacks, AI-enabled

highly contextual and requires professional judgment based on

malware, the vulnerability of IoT technology, cloud security

the experience of the appraiser, now also in terms of data

issues, and the involvement of cryptography. However, future

management. This issue has not been discussed in the literature,

directions, in cyber security, such as quantum-secure encryption,

an article is a contribution to the discussion on the importance of

biometric authentication, advanced artificial intelligence, and

valuation in the cyber sector, given the specific characteristics of

machine learning, may be able to address these issues.

cyber start-ups using AI and machine learning solutions. Despite

all these difficulties, IPR valuation will become increasingly

necessary and induce further questions regarding the valuation of 1.2 Cyber security products

a given intellectual property (IP). Firstly, how to value a patent According to the American Authorities, precisely Cyber Security

with Artificial Intelligence (AI), secondly how to assess the level and Infrastructure Security Agency, cyber security is “the art of

of sophistication of model training, and thirdly how to rate and protecting networks, devices, and data from unauthorized access

value data quality, or more broadly data sets. The findings can or criminal use and the practice of ensuring confidentiality,

help practitioners, especially from Technology Transfer Offices, integrity, and availability of information” [2]. The current cyber

to develop roadmaps for IP valuation in the cyber security security situation is characterized by the regular emergence of

industry. new cyber threats. The most common types of cyber threats

include malware, phishing attacks, ransomware, threats against

KEYWORDS / KLJUČNE BESEDE data or availability, disinformation, supply chain targeting, and

IPR valuation, IP in cyber security, data quality in AI model distributed denial of service (DDoS) attacks. The level of digital

resilience varies from different industries and countries,

however, effective cyber security remedies use security



1. technologies and techniques such as intrusion detection and CYBER SECURITY SPECIFICITY



VERSUS IPR VALUATION [8]. Cyber attackers are constantly evolving their approach to prevention systems, firewalls, antivirus software, and encryption

penetrate the computer systems of enterprises which means that

1.1 Introduction organizations must continuously monitor their networks against

The growth in importance of IPR is unquestionable, in every potential attack vectors, using a broad array of cybersecurity

sector of the economy, and in those key to the digital solutions to protect the entire ecosystem, including clouds and

transformation an undisputed. The IPR valuation is gaining in several applications [8]. importance and reliable valuation is relevant in the cyber security Typical products are software for stopping the biggest,

sector. The valuation approach dedicated to the cyber market is bandwidth-busting DDoS attacks, software that proactively

not described in the literature and represents an unexplored reduces attack surfaces, Edge DNS, authentication services,

research question. In IPR valuation, whatever the method, the clouds for protecting customers and providing data security,

essential characteristics of an intellectual asset should be taken which reduce friction during registration, authentication, and

into account. There is a fundamental complication arising from sign-ins while making it easy for customers to control their

the difficulty of determining the essential characteristics of IP, accounts from any device. Products are, on the one hand, closely

the scope of protection, and the need to consider source data related to IT or ICT. On the other hand, they use the latest

related to potential cyber exploitation on an unprecedented scale. developments in biometrics, behaviorism, psychology, and the

IP assets can be independently identified, are transferrable, sociology of human behavior. They use, as in criminology,

protected and that protection can be enforced. In the case of knowledge about human behavior, but the implementation of

cyber, they have an economic lifespan, defined by their knowledge is strictly technical, in a digital world.

characteristics. Depending on the nature of intangible assets,

there are different legal instruments by which protection is

possible and ultimately benefit from using them. It is important



2. MULTIDIMENSIONAL IP appear numerous, including AI inventorship, patent eligibility,

PROTECTION and AI-related copyright issues, particularly important are data

2.1 Impact of cyber product features on IP issues.

protection 2.2. FLDX system – an example of IPR

The development of new solutions to combat or prevent protection cybercrime requires the proactive action and the creation of new

inventions combating the criminal incidents. Since AI is widely An example of a cyber security product is the FLDX system,

used in cyberspace, AI-based products are also tools for patent protected by NASK, a Polish National Research Institute,

mitigating attacks. Ransomware and phishing attacks encrypt whose mission is to develop and implement solutions that

critical data, demand high ransoms, and disrupt a wide range of facilitate the development of information and communication

operations. The growing use of Internet of Things devices is networks in Poland, in addition to improving their effectiveness

introducing new security vulnerabilities, while cyber attacks and security. Patent – PL241005-Method and system for adaptive

targeting the software supply chain are exploiting third-party creation of network traffic filtering rules on a network device

vulnerabilities to gain access to sensitive information. Artificial spontaneously detecting anomalies and automatically

intelligence technologies enable cybercriminals to launch suppressing volumetric attacks (DDoS) protects digital services

sophisticated attacks. These AI-based threats are often not and network devices from DDoS attacks and a sudden and

subject to traditional security measures, making them difficult to unpredictable increase in user activity. Sudden and unexpected

detect bursts of Internet traffic can saturate network links or and mitigate. The World Intellectual Property

Organization (WIPO), the United Nations agency that serves the overloading application servers. Therefore, protecting networks

world’s innovators, is following the trend of consumer interest in and digital services from intentional attacks must go along with

AI in various economic, social, and cultural sectors, having fair distribution of network resources. The FLDX system is a fast

published some very interesting and important reports on AI over and extremely effective way to protect the availability of services

the past few years. [6], [10]. WIPO Technology Trends 2019 – on the network - whether the source of the threat is a volumetric

Artificial Intelligence reveals trends in patenting of artificial DDoS attack or a sudden increase in user activity. Maintaining a

intelligence innovations [10]. AI-related patents disclose AI fair distribution of network bandwidth is the primary goal of the

techniques and applications and refer to an application field or FLDX system, achieved in a time of up to 10 seconds. Unlike the

industry. WIPO analysis shows that many sectors and industries solutions currently offered in the anti-DDoS market, the FLDX

are system is not based on a database of signatures and static rules. exploring the commercial exploitation of AI,

telecommunications (mentioned in 15 % of all identified patent It dynamically self-adjusts filters to the current situation. This

documents), transportation (15 %), life and medical sciences (12 approach allows us to react extremely quickly to the observed

%), and personal devices, computing and human-computer changes in network load, as well as forecast them. The FLDX

interaction (11 %), the rest - other sectors including banking and system is therefore not only a protection tool - it is also a network

security. In the WIPO patent landscape report on Generative AI, knowledge discovery tool. The object of the invention is a

there are the latest patent trends for GenAI with a comprehensive method for adaptively creating network traffic filtering rules on

and up-to-date understanding of the GenAI patent landscape, a network device spontaneously detecting anomalies and

alongside insights into its future applications and potential automatically suppressing volumetric attacks (DDoS).

impact. The report explores patents relating to the different That FLDX example may illustrate the challenges of protecting

modes, models, and industrial application areas of GenAI. Deep IPR in this area. The speed and precision of the FLDX system

neural networks can be adapted to be either discriminative or are the result of years of scientific research in the fields of control

generative tasks, which has led to the development of various theory and adaptive signal processing, the IP behind the solution

types of GenAI models, which can support different types of is not only a patent, but also a copyright protecting the software

input and output data. This opens up a new perspective on the and the user's system, trade secret, the implicit knowledge of the

protection of inventions and products. implementation as well as the knowledge contained in the

There is a need to answer the threshold question of whether such technical documentation. Solutions are sporadically planned to

AI-related inventions qualify for patent protection. The United be patent-protected, due to non-compliance with requirements

States Patent and Trademark Office (USPTO) has issued for implementations of mathematical theorems or new

guidelines to clarify the requirements for patenting AI-assisted applications of functional analysis. However, even an obtained

inventions. For an invention to be patentable, there must be exclusive right is not sufficient protection in the market. It is

significant human input into its conception. Human inventors necessary, as with other software-based products, to supplement

must make a significant contribution to the invention that goes protection not only with copyright protection due to the nature of

beyond the mere use of AI tools. Otherwise, the invention is not the solution but also to keep in secret any know-how resulting

eligible for patent protection. In addition, the USPTO has created from the implementation and to circumvent technical problems

five principles for evaluating AI-assisted inventions, the fifth is arising from software development and installation in the cloud

worth mentioning here - namely, merely owning or supervising or at the customer's site.

an AI system does not qualify a person as an inventor without a 3. IPR VALUATION ISSUES

significant contribution to the concept of the invention [7]. This 3.1 Valuation approach selection

principle ensures that human ingenuity remains at the heart of Valuation of IPR regardless of the subject of valuation

patentable inventions while recognizing the supporting role of AI strictly depends on the potential area of application of the

in the inventive process. In the cyber industry, solutions can be protected technology. In the cyber sector, the issue of the

protected by patents and then there is a need to value IPR in the valuation of IP goods is becoming increasingly challenging, for

form of a patent on an AI-related property. The number of cyber several reasons. First, this is due to the obvious development of

security patent applications per year shows that the amount of the cyber market and the growing demand for all kinds of

investment going into finding new ways to help prevent cyber services and products protecting digital assets. Secondly, AI

attacks is huge. However, it is usually a bundle of different IPs technologies are finding applications in this sector, which makes

that is valued. Apart from the fact that AI-related IP problems the valuation problem more complicated, and thirdly, a complex



method of product IPR protection is common. The issue of IPR related know-how, there is an issue directly related to the

valuation in high-growth sectors, for new technologies, and understanding of the operation and use of AI models [4].

cutting-edge technologies, has been addressed in the literature

for years. Major researchers (such as Damodaran) describe the 3.2. Data in Artificial Intelligence model challenges of estimating value for technology [1], [9]. However,

During training, the artificial intelligence model is exposed to a

the growing cyber market introduces a significant level of

prepared dataset and tries to learn the patterns and relationships

complexity to the subject, due to the dynamics of development,

present in the data. This process involves adjusting the internal

key development trends, market estimation, and the scalability

parameters of the model based on the input data and the desired

and adaptability of solutions in this market.

outcome. In a situation where AI is used, another problem arises.

Depending on the nature of intangible assets, various

When is the product in question completed? AI models need to

legal instruments are offered to protect and ultimately profit from

be taught. What does AI model training include?

them. IP management is a key element of the business strategy

AI model training includes three main aspects:

of entities developing cyber services. The linkage of copyright

a) data collection

protection, patent, trade secret, and confidential know-how

There are ready-to-use open-source data sets. Data collection and

protection makes IP valuation difficult. Trade secrets may be

other resources are also collected and used. Internal data

preferable to patents in several circumstances, such as when the

collection provides access to proprietary information and control

patentability requirements may not be satisfied; the cost of

over data quality. Web scraping is the process of extracting data

pursuing patent protection outweighs the benefits; and/or the

from websites using various tools. Automation eliminates the

need for IPR protection extends beyond the available patent term

need for manual data collection, which in itself is impractical

[9].

when it comes to training AI models. Regardless of the data

Regardless of the method used, the valuation process

collection technique, the data should be relevant, accurate,

requires gathering a lot of information about intellectual property

consistent, presentable, and complete. Such data increases the

assets, as well as an in-depth understanding of the economy,

accuracy of the AI model, reduces bias, and increases user

industry, and specific businesses that directly affect their value.

confidence and trust in the AI model.

It is well known that there are three basic categories of valuation

b) data processing

methods for evaluating intellectual property and intellectual

Having a rich data set, it is necessary to validate the data. Data

property rights: income-based, market-based, and cost-based.

validation involves preparing the data to match the requirements

The choice of the appropriate method for valuing intellectual

of the specific learning mechanism used by the artificial

property depends on the type of intellectual property, the stage

intelligence model. Each learning technique requires the data to

of development, the purpose of the valuation, and the available

be presented in a specific way. An artificial intelligence model

data. The cost method establishes the value of an IP asset by

incorporating algorithms that learn through supervised learning

calculating the cost of a similar (or exact) IP asset. The cost

aims to predict or classify new data points. So, to select data for

method is particularly useful when the IP asset can be easily

an artificial intelligence model equipped with supervised

reproduced and when the economic benefits of the asset cannot

learning algorithms, label your data. Then divide the selected

be accurately quantified. This method does not account for

data into training, validation, and test sets. Using the training set

wasted costs, nor does it consider any unique or novel

is needed to teach the artificial intelligence model, the validation

characteristics of the asset. Although a cost-based method is used

set to evaluate performance, and the test set to evaluate the final

for software value estimation, the combination of various

model. For unsupervised learning, the artificial intelligence

elements of protection makes one think about the wisdom of

model aims to reveal underlying structures, group similar data,

choosing a revenue-based method [9]. The income method

and discover patterns without the help of labels. The model needs

values the IP asset based on the amount of economic income that

to understand the data by finding commonalities and

it is expected to generate, adjusted to its present-day value. This

understanding the features that define a particular dataset. In this

method is easiest to use for IP assets with positive cash flows, for

case, feature-based clustering of the data is required. This makes

those whose cash flows can be estimated with some degree of

it easier for the AI model to navigate and learn from unlabelled

reliability for future periods, and where a proxy for risk can be

data. The situation becomes a little more complicated taking into

used to obtain discount rates. The market method is based on a

account reinforcement learning (learning through interaction)

comparison with the actual price paid for the transfer of rights to

[5]. Artificial intelligence models involving reinforcement

a similar IP asset under comparable circumstances. This method

learning learn by exploring the specifics of a task in a particular

has the advantage of being simple and based on market

environment and performing functions by trial and error. In

information, so it is often used to establish approximate values

reinforcement learning, an environment must be simulated for

for use in determining royalty rates and inputs for the income

the AI model to interact with. However, another level of

method. For cyber industry this type of approach can be highly

complication relates to deep learning (neural networks and

problematic, since products in the cyber crime market are

beyond) , it is an advanced learning mechanism that drives the

evolving very quickly and there is considerable difficulty in

AI model and enables it to handle complex actions. AI models

comparing them. Often, it is only possible to make inferences on

with deep learning algorithms require large-scale data collection

the level of effects offered, i.e. expected rather than concrete

based on what the model is supposed to do. As deep learning

results, due to the widespread confidentiality of information.

algorithms use multiple layers of learning, the goal is to have

Companies do not necessarily boast about the ineffectiveness of

different versions of large data sets.

protecting their computers, resources, or access to the cloud.

c) providing selected data to the AI model and iterative

While one approach may seem particularly well-suited, the final

refinement

value estimation should merge the value indications obtained

Once the data has been structured based on the AI model's

under different approaches [1], [9]. Irrespective of the choice of

learning technique, the data is fed into the AI model. The model

valuation approach, in the situation of innovation, patent, or AI-

learns from the algorithms on which it is built. During the



learning stage, the capabilities of the model should be explored based on technology; personal, sensitive, and business data are

for refinement. Without iteration, the model cannot adapt to stored on computers, smartphones, and tablets, so an extensive

changing data and cannot improve its performance when exposed range of concepts are covered by cyber security - from

to other data sets. communication to transport, and shopping to healthcare. It is

This raises further questions regarding the valuation of crucial to consider the interrelationships and relationships

a given IP. Firstly, how to value a patent with AI, secondly how between the different types of IP. Depending on the business

to assess the level of sophistication of model training, and thirdly needs, an appropriate valuation method should be chosen, taking

how to assess and value the quality of training and validation into account whether the IP relates to AI. When analyzing an AI-

data, or more broadly data sets. In addition, in the cyber area, related patent, the relationship to the data, the individual datasets,

matters are further complicated by the use of sensitive or and the way the models are taught should be explored. Particular

confidential data, such as tools for detecting illegal, offensive or care should be taken to analyze the quality of the data and to

harmful content based on data from law enforcement agencies. understand the principles of data management (from collection,

An additional legal complication arises. description, sharing, archiving, etc., including the FAIR principle

– it is an acronym for Findable, Accessible, Interoperable, and

4. RECOMMENDATIONS AND Reusable). In addition, the origin of the data in the cyber sector

CHALLENGES OF VALUATION IN should be taken into consideration. Moreover, in the cyber area,



CYBER SECURITY SECTOR confidential data, such as tools for detecting illegal, offensive, or further complications are caused by using sensitive or

harmful content based on data from law enforcement agencies.

4.1 Exploring difficulties Without high data quality, even the most advanced artificial

Nowadays, cyber security plays a crucial role in the global intelligence models will fail. Data quality in the new era of AI

economy. The risk of cyber threats becomes more prevalent and highlights the key role of data quality in shaping effective data

cyber attacks can have devastating consequences leading to strategies. The task of the IPR evaluator in cyber products or

financial losses, reputational damage, and national security solutions is to evaluate the AI model, and assess how each dataset

breaches. Therefore, it is imperative that governments prioritize is used, how the evaluation process works, which IPRs use it, and

cyber security measures to safeguard their interests. In addition to what extent, and what parameters influence the business aspect

to economic implications, cyber attacks also pose significant of the entire evaluation process. The process of valuing IPRs is

risks to national security. Governments around the world are highly contextual and requires professional judgment based on

increasingly concerned about hacking activities that aim to steal the experience of the appraiser, now also in terms of data

sensitive information or disrupt critical infrastructure systems. management and understanding of AI development and

Cyber attacks usually modify, access, or destroy sensitive application phase . information, extort users' money, or disrupt normal business

processes. In 2024, the cyber security industry is expecting a

REFERENCES

paradigm shift in a more coherent and business-involved

[1] Damodaran, A., Investment Valuation, John Wiley &

approach that reflects a better understanding and management of

Sons, New York, 1994

cyber threats [8]. This shift concerns the latest technology

[2] https://www.cisa.gov

adoption and revolution, associated liability, maturity,

[3] Marius Schneider, Intellectual property rights, the new

integration, regulatory, quantification, communication, and

currency, Journal of Intellectual Property Law &

behavioral shifts. As the market grows, there will undoubtedly

Practice, Vol 14, Iss.11, 2019, p.825–

be an increased demand for intellectual solutions to support the

826, https://doi.org/10.1093/jiplp/jpz106

fight against cyber crime. Hence, the growth in importance of

[4] Kathi Vidal, The Applicability of Existing Regulations

IPR will be indisputable, which in turn will result in a significant

as to Party and Practitioner Misconduct Related to the

increase in the valuation of IP and its need in the market [3].

Use of Artificial Intelligence, 2024

Therefore, IP valuation is an important issue, and reliable

[5] Lee J. Tiedrich, Gregory S. Discher, Fredericka

valuation is important for multinational corporations involved in

Argent, and Daniel Rios 10 Best Practices for Artificial

IP transactions. IP valuation guidelines and regulations are also

Intelligence Related Intellectual Property Intellectual

changing around the world due to different statutory provisions

Property Technology Law Journal,.vol.32, nr.7, 2020.

in each country. Valuing intellectual property involves assigning

[6] Patent Landscape Report - Generative Artificial

a monetary value to the intangible assets of a business entity.

Intelligence (GenAI), World Intellectual Property

However, the intangible nature of intellectual property means

Organization Geneva, Switzerland, 2024,

that it is often difficult to value and define, making it challenging

https://doi.org/10.34667/tind.49740

to set a fair price.

[7] https://www.uspto.gov



4.2 Recommendations Abebe Abeshu Diro, Cyber security: State of the art, [8] Wasyihun Sema Admass, Yirga Yayeh Munaye,

The most challenging tasks are determining the scale of the challenges and future directions, Cyber Security and

valuation portfolio, determining the role of AI in an invention, Applications Vol.2, 2024, 100031.

patent, or confidential know-how, determining the strength of a https://doi.org/10.1016/j.csa.2023.100031

patent using AI and comparing it to other similar solutions, and [9] William J. Murphy, John L. Orcutt, Paul C. Remus,

determining the extent of model validation and database quality. Patent Valuation: Improving Decision Making through

Of course, issues related to the market, comparison of coverage, Analysis, Wiley, 2012

the scale of adaptation, etc. are also in force. However, [10] WIPO Technology Trends 2019 – Artificial

completely new problems are gaining importance, the valuation Intelligence, Geneva, Switzerland: World Intellectual

of IPR in the cyber sector will be a further stage of complication Property Organization, 2019,

and will require knowledge of a great level of AI invention https://doi.org/10.34667/tind.29084

protection and data management. A large part of everyday life is





The Challenge of Licensing Artificial Intelligence




Technology for Cybersecurity Applications



Michał Rotnicki†

Technology Transfer Department

NASK – national research

institute

Warsaw, Poland

michal.rotnicki@nask.pl



ABSTRACT It's difficult to quantify the impact of cybercrime on the banking sector, but public data from the US[17] and the EU [16] The central question of this article is whether the transfer of cyber suggest it is around €4 billion each. The criminals are highly security technology based on neural networks into a production effective in the search for the optimal strategy of action in order environment poses significant challenges due to the complexity to steal money from Internet users, while at the same time and time variation of the technical environment, constantly minimizing the legal risk and the resources (effort) involved evolving threats, and regulatory requirements. Error! Reference source not found.

The article uses observational research techniques for The criminal’ss resources involved are the use of a technical

cybercrime activities, and experimental research for product method, a socio-technical method, or both, leading to a

management since 2011. successful theft [5].



The article presents an application case study of behavioural

biometrics and artificial intelligence (AI) techniques to detect

remote desktop attacks, and technology transfer adaptations to 2 METHODOLOGY

changing conditions. Since 2007, cybercrime data has been based on natural

The added value of the paper is to draw conclusions from a real observations. NASK provides Computer Emergency Response

business case observed in internal business activities. Team (CERT) services at the national level and commercial

threat intelligence services to the main financial institutions in

KEYWORDS Poland.

Cybersecurity, Artificial Intelligence, software licensing, The case study is an original commercialisation case provided as

software development, low compliance, behavioral biometric, AI part of the BotSense product offered by NASK.

licensing.



3 THE BANKS, THE THIEVES AND EVEN



1 THE SCIENTISTS INTRODUCTION

NASK activities are focused on issues of security in Poland has a population of about 38 million and in the first

cyberspace. quarter of 2024, the Polish banking sector operates

One of the areas of influence on cyberspace [6] is the approximately 43,5 million accounts retail accounts with

provision of new technologies for counteract cybercrime and contracts allowing access to internet banking. About 23 million

transfer them to commercial IT products. The goal is to increase accounts are actively accessed via internet banking and about 22

resilience of the banking services and key services supplier [1]. million users access via mobile applications [15]. Since 2007,

The banking sector is particularly vulnerable to the activities NASK has been working with the Polish banking sector to

of commercially motivated criminals [7], who are believed to be identify and counteract theft from Internet banking users. Over

personally motivated in their criminal activities. These are the years, with the improvement of technical methods of

criminals who directly seek to make a profit by seizing the funds protecting electronic banking, both on the side of the banks and

of electronic banking users. on the side of the end user, attacks based on vulnerabilities of IT

systems, have been significantly reduced [2] [3]. They required

sophisticated technical knowledge, considerable technical

resources and centralised malware management, making such

Permission to make digital or hard copies of part or all of this work for personal or criminal infrastructure vulnerable to law enforcement. classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full Socio-technical attacks , on the other hand, have experienced

citation on the first page. Copyrights for third-party components of this work must a renaissance, using voice communication techniques to

be honored. For all other uses, contact the owner/author(s). persuade the victim to provide the criminal with login and

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s). authentication credentials for banking transactions and, crucially,

to give the criminal access to their device's desktop via a legal must operate, we can distinguish between technical layers:

remote access application. device category, hardware, operating system, components of

As a result, the attack scenario does not require any specialist operating system, web browser.

IT knowledge [9], which has made this method of criminal



activity accessible to a wider group of criminals, resulting in a

sharp increase in the number of remote desktop attacks.

At the same time, in a social engineering and remote desktop

attack scenario, there is virtually no event that can be classified

as technically incorrect. The user voluntarily provides his or her

credentials to the criminal, voluntarily agrees to open a remote

desktop connection, and is often persuaded by the criminal to

deliver the final blow by turning off the monitor. This means that

no cybersecurity incident occurs in the data transmission channel

between the endpoint and the bank's server.

Analyzing the above, it can be said that a dynamic market

model is emerging in which criminals are effectively and

efficiently adapting to the limitation of increasing the resistance

of information systems to cyber-attacks. Criminals are creatively

and rationally searching for new effective techniques and crime

scenarios to carry out successful theft. The specific type of Figure 1: Technical layers

attacks mentioned above are those carried out with the unwitting



legislation [11], are forced to search for ever new technical However, banking institutions in particular, burdened by the diversity of devices, operating systems and web browsers does not guarantee the achievement of a stable, transferable participation of the victim. However, overcoming the additional complications posed by



solutions to identify electronic banking sessions compromised by technology. The whole technical environment described above is evolving. For example, the major web browsers, Chrome [13] criminals. and Firefox [14], are released on a monthly basis. This means From a technology transfer perspective, this raises the non- that, the technical conditions under which cybersecurity obvious problem of how to organize the process of technology technology should operate are constantly changing. transfer to combat criminals. It should also be noted that changes affecting the security of

the operating system and web browser may be made between the

scheduled release dates of new versions and may involve

4 unpredictable technical changes. Case Study - Behavioral biometrics and

artificial intelligence techniques to detect a In addition, there are other elements in the formal-legal field

[11] that we should consider, such as: international and national

access via Remote Desktop

legislation, technical legislation, standards, norms,

NASK set up an internal research project to work on an AI recommendations and internal company regulations.

model capable of analysing how an end user uses a keyboard to

identify themselves. In a laboratory environment, this is a task



that requires a certain number of experiments, the construction

of relevant data sets and the application of technical expertise,

but in principle the level of scientific risk is limited. However,

when it comes to transferring the developed technology to a

production application that is expected to operate at a certain

minimum level of effectiveness for the entire population using e-

banking, the issue becomes much more complicated.

Even if the expected level of efficacy is auxiliary, e.g. 70%,

and unrepresentative individuals are discarded from the user

population.



4.1 The cybersecurity technology ecosystem

Cybersecurity technologies require deep and precise technical

integration with the environment to be protected. For example,

tracking the use of a keyboard via a web browser, as a function Figure 2: Formal layers

of the time can be disrupted by the security mechanism

We are also seeing dynamic changes in the way criminals

embedded in that web browser. One of the security mechanism

operate: variability of attack scenario and variability of tools.

implemented by vendors is randomization of selected user

behavior data and disrupt the time line data.

If we take the oversimplification of identifying the main

layers of the environment in which cybersecurity technology





4.2 The challenge of licensing

As a result, a solution developed in the laboratory will either

start to fail immediately when deployed on the entire population,

or it will start to fail over a finite period of time (as a function of

time). This phenomenon has no risk characteristics, but is an

inherent feature of the cybersecurity technology ecosystem.



The question is how to structure the process of technology

transfer and licensing in this dynamic ecosystem?

In the process of technology transfer, we can distinguish:



1. Stage I - licensing the results of the R&D team and

transferring them to the product development team (in

this case software),

Figure 3: Criminal activity 2. Stage II - advising the product development team on

how to incorporate the innovation into the

And, of course, criminals are constantly identifying banking manufacturing environment.

security techniques and bypassing or neutralising them [10].

We can think of cyber technology as a black box influenced Such an approach is not rational and will fail if we apply it to the

by the forces of many independent parameters. transfer of cybersecurity technology.

As in physics, the degrees of freedom (DOF) of a mechanical

system is the number of independent parameters that define its This is because there is a high probability that the transferred

configuration or state. technology will need to be modified before it is fully

implemented in the product.

This will make the whole process infeasible and banks will



start looking for non-IT methods to fight crime.



4.3 Practices applied

For technology transfer in the cyber domain, the NASK

has adopted its own specific operating procedures.

First, the cybersecurity technologies developed in the NASK

R&D teams are transferred to internal development teams.

By technology, we mean the form of a method, algorithm or

learned AI model. The development team then builds a finished

software component on top of it.

After that, the R&D team still, support and develop

technology. The R&D teams are prepared for long-term

development of the technology for detection of specific types of

Figure 4: Forces affecting cyber security technology attacks, including its modification in the event of a change in the

conditions of the technical environment in which the technology

A useful technology should be in balance between these is to operate (e.g. loss of access to data relevant for detection).

parameters. If one vector increases, it means there's a need for Such organisation of technology production and preparation

action. for transfer enables the temporary licensing of the finished

A multi-layered dynamic model of the variability of the software component, which allows the use of the cybersecurity

environment is thus created, which seeks an equilibrium that technology for implementation in software. The license contains

includes the success rate of attacks. The stimulating (agonistic) a number of specific conditions tailored to the cybersecurity

factor is the activity of criminals and the antagonistic (inhibitory) ecosystem, the main ones are:

factor is the development of security technologies. Another two



parameters, which can be both agonistic and inhibitory, are 1. an assurance that the licensee will adapt the technology

changes in the technical or formal domain. Both can improve or to changes in the technological environment,

reduce the effectiveness of cybersecurity technology. What is 2. an obligation on the licensee to improve the technology

certain, however, is that all of these parameters create a need for in the event of a decline in the effectiveness of attack

constant review and adaptation of the technology. detection,

3. an limitation of the licensor's liability for failure to

adapt the technology to changes in the technological

environment or to changes in the activity patterns of

the perpetrators.



These points are almost impossible to define precisely. They

are declarative in nature, with no strict guarantees from either For future work, it is possible to approach techniques to assist in

party. The technology provider cannot reliably guarantee the predicting the effects of the changing environment.

effectiveness, cost or time it will take to modify its technology,

and the recipient cannot guarantee the conditions under which it REFERENCES

expects the technology to be effective.



In other words, the factor that determines the balance between [1] Directive (EU) 2016/1148 of the European Parliament and of the Council of 6 July 2016 concerning measures for a high common level of security

the technology provider and the technology recipient is a of network and information systems across the Union

common rational economic interest. And the basis for deciding [2] Raport Roczny z Działalności Cert Polska 2023, NASK 2023

on such a cooperation model should not be so much an [3] Annual Report From The Actions Of Cert Polska 2022, NASK 2023

[4] Successful Technology Licensing, World Intellectual Property

assessment of the effectiveness of the technology at the time of Organization (2015), ISBN 978-92-805-2633-2,

its production, but rather the ability of the technology provider to [5] Lella I., Tsekmezoglou E., Theocharidou M., Magonara E., Malatras A.,

modify and develop the technology to keep pace with changes in Svetozarov Naydenov R., Ciobanu C. Enisa Threat Landscape 2023,

October 2023, ISBN: 978-92-9204-645-3

the application environment within a reasonable period of time. [6] Hogan M., Newton E, Supplemental Information for the Interagency

Report on Strategic U.S. Government Engagement in International

Standardization to Achieve U.S. Objectives for Cybersecurity, NISTIR

8074 Volume 2, 2015



5 Conclusion [7] Troy E. S., A Conceptual Review and Exploratory Evaluation of the Motivations for Cybercrime , August 2013

The transfer of ICT technologies for cybersecurity may force [8] Gargir Sarkar, Sandeep K. Shukla, Behavioral analysis of cybercrime:

a different way of thinking about building a collaborative model Paving the way for effective policing strategies, Journal of Economic

Criminology, Volume 2, December 2023

with business [12]. Thinking of collaboration with business as a [9] Phiri. J., Lavhengwa, T., Segooa M.A., Onlinebanking fraud detection:



one-off design phenomenon may prove to be a dead end. To Acomparative study of casesfrom South Africa and Spain, South African Journal of Information Management 26(1), a1763., 2024 ensure a steady flow of solutions for business in a rapidly [10] Tapiwa Mazikana A., Development of a Predictive Model for Online

changing environment, it is worth considering a process model fraud Detection in the Banking Sector. Case Study of First Capital Bank,

of collaboration [4]. Journal of Machine Learning Research, May 2024,

[11] Harnay S., Scialom L., The influence of the economic approaches to

The example case study analyses the behavioral biometrics regulation on banking regulations: A short history of banking

project and the AI technology used. However, the issue seems regulations, Cambridge Journal of Economics 40(2), April 2015

[12] Larsson B, Rolandsson B. , Ilsøe A., Masso J., Digital disruption

relevant to any application of technology or methodology in an diversified-FinTechs and the emergence of a coopetitive market

unstable cyberspace environment. ecosystem, Socio-Economic Review, July 2023

The fundamental value of collaboration is to ensure the ability [13] Chrome RoadMap, https://chromestatus.com/roadmap , 10.08.2024

[14] Firefox Release Calendar

to solve a class of research problems within a reasonable time https://wiki.mozilla.org/index.php?title=Release_Management/Calendar

and cost. The process for sharing further technologies developed &redirect=no, 10.08.2024

on the basis of the first project should be proposed in advance. [15] Barbrich P., Nocoń B., NetB@nk bankowość internetowa i mobilna,

płatności bezgotówkowe, Polish Bank Association Raport I Kwartał,



The initial project or technology licensing is therefore only a 2024 [16] 2024 Report On Payment Fraud, European Central Bank 2024 starting point for long-term collaboration. It's also necessary to [17] Global Banking Fraud Index 2023,

rethink the organisation of technology transfer agreements. https://www.ecb.europa.eu/press/pr/date/2024/html/ecb.pr240801~f21cc

Towards a collaborative framework and the definition of a 4a009.en.html , 18.09.2024

dynamic research process.



Technology Transfer: Revenues Estimation



in the Cyber Security Sector



Michal J. Falkowski† Jaroslaw Kaminski Marta Wachowicz

Technology Transfer Department Technology Transfer Department Technology Transfer Department

NASK National Research Institute NASK National Research Institute NASK National Research Institute

Warsaw, Poland Warsaw, Poland Warsaw, Poland

michal.falkowski@nask.pl jaroslaw.kaminski@nask.pl marta.wachowicz@nask.pl



ABSTRACT 1.1 Defining cyber security product This study investigates the complexities of technology transfer Given the multidisciplinary nature of cyber security and its within the cyber security sector, focusing on the financial and widespread impact on society, it is essential to establish, utilize, operational challenges posed by its dynamic nature. The primary and elaborate a standardized terminology and develop a research problem is understanding how to define final cyber comprehensive, shared understanding of what constitutes cyber product and estimate associated costs, particularly in the context security product and economic risks associated with it [7]. of both traditional and new economy revenue models. In defining a cyber security product, it is crucial to recognize Preliminary findings reveal significant discrepancies in cost the role of interdisciplinary contributions, ranging from estimation and revenue forecasting, particularly due to the non-computer science and engineering to law, economics, and human linear contributions of scientists, which complicate the creation factors. For instance, a cyber security product may include not of effective license agreements. The paper offers a framework to only technical components, such as encryption algorithms or better align technology transfer processes with the unique intrusion detection systems, but also legal frameworks and characteristics of cyber security innovations, thus improving the organizational practices that enhance security. The integration of accuracy of cost projections and licensing strategies. these diverse elements requires a standardized terminology that



KEYWORDS can be universally understood across disciplines, enabling



Technology transfer, cyber security sector, revenue estimation, Moreover, the definition of a cyber security product must effective communication and collaboration.



AI models, new economy, science contribution, license account for its intended purpose and scope. Products may vary agreements significantly in their focus - some are designed to prevent

unauthorized access, others to detect intrusions, and yet others to



1 respond to or recover from cyber incidents. This diversity UNCERTAINTIES IN CYBER SECURITY

TOOLS SPECIFICATION necessitates a clear classification system that categorizes

products based on their functionality, target environment, and the

Cyber security is a term with widely varying definitions that are specific threats they address. For example, network security

frequently subjective and, in some cases, lack precision. tools, endpoint protection software, and identity management

According to the America’s Cyber Defense Agency (CISA), it is systems each serve different purposes but collectively contribute

defined as the art of protecting networks, devices, and data from to a comprehensive cyber security strategy.

unauthorized access or criminal use and the practice of ensuring Economic considerations also play a critical role in defining

confidentiality, integrity, and availability of information [11]. cyber security products. The value of a cyber security product is

The absence of a clear, universally accepted definition that often measured by its effectiveness in mitigating risks, which are

encapsulates the multidimensional nature of cyber security themselves subject to economic assessment. The economic

hinders progress in technology and science [6]. This is because it impact of cyber threats, the cost of deploying and maintaining

reinforces a technical perspective on cyber security, while cyber security products, and the return on investment are all

simultaneously isolating disciplines that should be collaborating factors that influence how a cyber security product is defined and

to address complex cyber security challenges effectively. The evaluated. This underscores the importance of aligning technical

complexities involved significantly affect the determination of definitions with economic realities to ensure that cyber security

what constitutes a cyber security product, the criteria for deeming investments are both effective and sustainable.

it complete, and the estimation of production costs within defined Furthermore, the lifecycle of a cyber security product must be

timeframes and budgetary constraints. clearly delineated, from initial development through deployment,

operation, and eventual decommissioning. A comprehensive

†Author Footnote to be captured as Author Note understanding of this lifecycle is necessary to establish criteria

for when a product can be considered complete and to identify

Permission to make digital or hard copies of part or all of this work for personal or potential risks and vulnerabilities that may arise at various stages.

classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full This lifecycle approach also highlights the importance of

citation on the first page. Copyrights for third-party components of this work must adaptability in cyber security products, as they must evolve to

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia address emerging threats and changing environments. be honored. For all other uses, contact the owner/author(s). © 2024 Copyright held by the owner/author(s).



In summary, defining a cyber security product requires a vulnerabilities, which can involve both software

multidisciplinary approach that integrates technical, legal, patches and hardware upgrades. Additionally,

economic, and operational perspectives. Standardized operational costs include the resources required to

terminology and clear classification systems are essential to monitor the product's performance, respond to security

fostering a shared understanding across disciplines, while incidents, and conduct regular security assessments.

economic considerations and lifecycle management provide the The need for highly skilled personnel to manage these

framework for evaluating the effectiveness and sustainability of tasks further contributes to operational costs, as cyber

cyber security products. security expertise is often in high demand and short

supply.

1.2 Estimating cyber product costs 4. Decommissioning Costs: At the end of its lifecycle, a

A standardized method for measuring and managing the costs cyber security product must be decommissioned,

associated with implementing cyber security programs has yet to which involves safely removing it from the

be established. To advance research and practice in this field, environment and ensuring that no residual

various cost estimation frameworks related to the development vulnerabilities remain. Decommissioning costs may

and deployment of cyber security products have emerged in include data migration, system reconfiguration, and the

recent years [9]. Estimating the costs associated with cyber disposal of outdated hardware. Additionally,

security products is a critical aspect of cyber security planning organizations may need to invest in new cyber security

and management. However, this task is fraught with uncertainties products to replace those being decommissioned,

due to the dynamic and evolving nature of cyber threats, the adding to the overall cost.

complexity of cyber security products, and the diverse Estimating these costs is complicated by several factors,

environments in which they are deployed [8]. Unlike traditional including the unpredictability of cyber threats, the rapid pace of

products, cyber security products must continuously adapt to an technological change, and the variability in organizational needs

evolving threat landscape, where new vulnerabilities and attack and environments [10]. It means that a cyber security product

vectors emerge regularly. This requires ongoing updates, patches, may require extensive customization and integration efforts,

and upgrades, leading to unpredictable and often escalating which further complicates cost estimation. For example, the

operational costs over time. introduction of disruptive technologies, such as quantum

Cost estimation for cyber security products involves several computing, can render existing cyber security products obsolete,

key components: development costs, deployment costs, necessitating additional investments.

operational costs, and decommissioning costs. Each of these The need for specialized personnel to manage and maintain

components must be carefully assessed to provide an accurate cyber security products, combined with the scarcity of cyber

estimate of the total cost of ownership (TCO) for a cyber security security expertise, adds another layer of complexity to cost

product. forecasting. Furthermore, the consequences of underestimating

1. the costs must be carefully considered, as they are often Development Costs: These include the expenses

incurred during the design and creation of the cyber significant and far-reaching, potentially resulting in insufficient

security product. Development costs can vary widely protection and increased risk exposure. This contrasts with other

depending on the complexity of the product, the products, where cost overruns might primarily affect financial

technologies involved, and the level of expertise performance without posing immediate security risks. Therefore,

required. For example, developing an advanced threat the cost estimation of cyber security products must account for

detection system may involve significant investment in not only the tangible costs of development, deployment, and

research and development, including the use of maintenance but also the intangible costs associated with risk

machine learning algorithms, data analysis tools, and management and the potential impact of cyber incidents.

security protocols. Additionally, the need for To address these uncertainties, organizations must adopt a

compliance with industry standards and regulations flexible and adaptive approach to cost estimation. This may

can add to development costs, as products must be involve using scenario analysis, which considers different

designed to meet specific security requirements. potential future states and their impact on costs, as well as

2. incorporating risk assessments to identify and quantify potential Deployment Costs: Once a cyber security product is

developed, it must be deployed within the target cost drivers. Additionally, organizations should consider the total

environment. Deployment costs include the expenses cost of ownership over the entire lifecycle of the cyber security

related to integrating the product with existing systems, product, rather than focusing solely on upfront costs. This

configuring it to meet organizational needs, and approach ensures that all relevant costs are accounted for and

training personnel to use it effectively. In some cases, provides a more accurate estimate of the long-term financial

deployment may also involve significant infrastructure commitment required to maintain cyber security.

upgrades, such as installing new hardware or



enhancing network capabilities. These costs can be 2 METHODOLOGY substantial, particularly in large or complex

organizations with extensive IT environments. To address issues, this study employs a mixed-method approach.

3. Operational Costs: The ongoing operation of a cyber An extensive literature review is conducted. Relevant academic

security product generates costs related to maintenance, journals, industry reports, and government publications are

monitoring, and updates. Cyber security products must examined. Additionally, qualitative data is collected through

be continuously updated to address new threats and semi-structured interviews with key specialists and experts.



Cyber security is the practice of protecting individuals’ and COMPANIES VALUATION the belief that for a potential investor, a particular asset is worth as much as he can get income from that asset. The risk of the 3 REVENUE ESTIMATION AND The income method of technology valuation is grounded in



categorize the main components of cyber security into the indicated as the most appropriate for valuing technology for following areas: cyber security Governance, Policies, and which there is a high degree of confidence in the forecasts of Procedures, User Identity and Access Management, Network operating income. Security, Application Security, Data Protection, Business Market (comparative) methods of valuing intellectual Continuity and Disaster Recovery Plan, Education. The number property, on the other hand, involve estimating the value of of fields results in miscellaneous cyber security business models, technology based on a comparison to market transactions for reflecting various comprehensive solutions in the evolving similar assets. However, information on transactions for the landscape of cyber threats and swift pace of technological purchase or sale of intellectual property is rarely publicly advancement. The differences are both in revenue streams, cost available. Therefore, the method often uses an analogy with the devices, sensitive data, and financial assets against any digital determination of the period of economic usefulness of the valued attacks [3]. It refers to any technology, measure, or practice for technology. It is done based on projected cash flows discounted preventing cyberattacks or mitigating their impact. We could at an appropriate discount rate. The income method is most often organizations’ systems, networks, applications, computing Valuation of technology using the income approach requires business and the time value of money should be considered.

structures and scalability. valuation of technology companies, whose value depends largely



3.1 or entirely on the technology they own. The main shortcoming Cyber security business models

of this method is the inability to identify comparable technology.

We can distinguish three basic revenue streams: subscriptions, As a rule, each innovative technology is unique and has specific

professional services, and licensing [5]. In first case cyber parameters, which leads to limited possibilities of comparison to

security firms offer their services on a subscription basis, existing solutions known to date. providing continuous protection with regular updates and



steady and predictable revenue flow, development of customer sector relationships, mutually beneficial vendor relationships with The Market Multiples method is a key tool for valuing companies support in exchange for a recurring fee. This model ensures a 3.3 Companies’ valuation in cyber security



consulting, threat assessment, and response services. These firm by comparing it to similar private or recently acquired companies in the sector. Specifically, it focuses on two primary include penetration testing, incident response teams and security types of multiples: Revenue Multiple and EBITDA (Earnings audits. Finally, many companies operate under licensing model - focused on professional services as business model often offer in the cyber security industry. This approach involves valuing a major focus on customer procurement. Cyber security companies



selling licenses for proprietary security software or technology Before Interest, Taxes, Depreciation, and Amortization) Multiple. For startups (especially those that are pre-profit) the solution could be significant revenue stream, creates an easier Revenue Multiple is often more relevant. It compares the entry into foreign markets, does not require capital investment or company's value to its revenue, offering a perspective on how the presence of the licensor in new geographical regions. market values the revenue generated. For more mature



3.2 companies (with significant earnings), the EBITDA Multiple Classic technology valuation

provides a view of the company's value relative to its profitability

Tech spending as a percentage of revenue has increased from before accounting for financial and accounting factors.

3.28% in 2016 to 5.49% in 2023 [4]. With bigger budgets often Applying the Market Multiples method effectively requires a

comes increased oversight and expectations from the business- deep understanding of market trends and financial metrics

tech leaders are becoming thoughtful about allocating capital for specific to the cyber security sector. The rapidly evolving nature

tech investments. 2023 Deloitte research shows that 6 in 10 of cyber security, with frequent technological innovations and

executives struggle with measuring the value of these varying threat landscapes combined with investor confidence in

investments. The choice of an appropriate valuation method the sector's growth can significantly influence these multiples.

depends on the circumstances, scope, and purpose of the The most common purpose of technology valuation is the needs

valuation – the three main approaches concentrate on the cost, for commercialization of completed development work in R&D

market, and income. Units. It is determined as part of the commercialization of

Cost methods determine the value of intellectual property technology, the value of the sale to an external investor or in-kind

based on the historical cost of production or the estimated cost of contribution to a special purpose vehicle (SPV or Spin-off). Prior

replacement with assets of comparable utility. These methods to the commercialization of intellectual property, there is often a

involve considering any expenses that need to be incurred to need to determine the value of these intangible assets and whole

remanufacture the asset or replace it with an asset comparable to company. Another reason, also encountered, for the valuation of

the one being valued. Cost methods are applied mostly to technology is the need to recognize the fair value in the

unfinished or easily manufactured technologies. It is possible to accounting books. Less common are cases of estimating the

imagine situations in which a relatively considerable sum of value of technology for litigation, where it is required to

money has been spent on a technology that does not produce the determine the value of the subject matter of the dispute or under

anticipated benefits. In such a case, the valuation of technology collateral for financial instruments. In the case of cyber security

by the cost method may significantly overestimate its value, and technology and company valuations, it is useful to define the

income methods will come to the rescue. circumstances valuation determines purpose: accounting, market



(for the current owners or new investors) or liquidation. It would programmers and software developers, and there is no space for

be desirable to strike a balance between qualitative and discovering independent universal truths in the sense of

quantitative measures. breakthrough ideas or inventions. We observe the non-linear

contribution of the researcher to the development of the cyber



4 security product. For TTOs, this is an additional complication, IMPLICATIONS FOR TECHNOLOGY

the connection of the author to his work is strong, and the cyber

TRANSFER OFFICES security market forces not only close teamwork but also IT and

From the point of view of technology transfer and data professionals themselves are gaining in importance. Data

commercialization of scientific results, managing the process of stewards have a significant impact on the development of AI

new solution building using AI models is particularly difficult. models and thus cyber products. For TTO is a difficulty related

The problematic question of revenue estimation implies further to the progress and commercialization plans for a specific

issues related to the creation of licensing or distribution solution.



characteristics of AI models. First, there are several problems agreements; additional complications also arise from the very 4.3 Risks in license agreement

associated with the application and obtaining Intellectual Forming a license agreement for a product or solution using an

Property Rights (IPR) protection for such solutions. Secondly, AI model requires considering the strict characteristics of

cooperation with scientists is done in close cooperation with training AI models, the difficulty of determining milestones for

software developers, and scientific input is expected not in the model development, and the system of subscription or license

entire process. Third, the solutions for specific markets generate fees depending on the stage of learning or re-learning the model.

several difficulties in shaping models for licensing agreements The fundamental difficulty in estimating and establishing profit

for the cyber security industry. or revenue models depending on the development of machine

learning lies in the indefiniteness of the solution itself. Models

4.1 Intellectual Property Rights protection need successive iterations, the cost of software development

When considering patenting AI-related inventions, there is a changes, and the demand for certain solutions also changes,

need to answer the fundamental questions of whether inventions which makes it exceedingly difficult to forecast profits and build

qualify for patent protection. In European system, while a a model of fees and payments in a license agreement. The

computer program or software may not be patentable, artificial described problem of revenue estimation forces the adaptation of

intelligence and machine learning that serve or achieve a cyber security solutions using AI models of licensing agreements

technical purpose may be a desirable alternative. The newest and billing systems, a thorough reflection is needed in the society

EPO guidelines [2], require the mathematical methods and of technology transfer professionals on this subject.

training data used by an AI-related invention to be disclosed in



sufficient detail to reproduce the technical effect of the invention 5 CONCLUSION over the whole scope of the claims. To address these issues and

prepare a commercialization plan for the cyber security market, Developing a more precise and universally accepted definition of

Technology Transfer Offices should identify the territories for cyber security products is essential for standardizing cost and

patent protection for their AI inventions and assess whether such revenue estimation processes. Authors will focus on robust

inventions meet the relevant subject matter eligibility criteria. If methodologies to account for the non-linear contributions of

AI-related patent protection seems unfeasible and ineligible, R&D teams in cyber security, as current models are inadequate.

TTO should consider protection using trade secrets or other These areas will dictate the trajectory of future research, reducing

alternatives. Protecting rights to training data, AI output, and uncertainties in product finalization and financial forecasting.

other crucial training data requires attention, awareness, and

careful action. REFERENCES

[1] Science Council, August 2024, <https://sciencecouncil.org/about-



4.2 Relations with scientists science/10-types-of-scientist/> [2] Guidelines for Examination in the European Patent Office (2024),



AI is forcing a change in the attitude of scientists, from that of a ISBN 978-3-89605-361-9 [3] CISCO, August 2024, < https://www.cisco.com/site/us/en/learn/topics strict researcher to one that is far more oriented toward creating /security/what-iscyber security.html > a working IT system. In terms of describing the types of scientists [4] Global Technology Leadership Study, Deloitte 2023 [5] Cyber security Startup Valuation Report, Finro 2024 according to the Science Council, one can explain the change in [6] Diakun-Thibault, Nadia. (2014). Defining Cybersecurity . Technology attitude of the Explorer Scientist to the Developer Scientist [1]. Innovation Management Review. 2014. DOI: 10.22215/timreview/835 [7] Cains M.G., Flora Liberty, Taber Danica, King Zoe, and Henshel Diane. This reflects a commitment to the area of creating AI solutions (2021). Defining Cyber Security and Cyber Security Risk within a for specific and demanding markets. “The Explorer Scientists Multidisciplinary Context using Expert Elicitation . Risk Analysis. 42.



rarely focus on a particular outcome or impact, rather they want DOI: 10.1111/risa.13687. [8] Leszczyna Rafal and Litwin Adrian. (2020). Estimating the Cost of to know the next piece of the jigsaw of scientific understanding Cyber security Activities with CAsPeA: A Case Study and Comparative and knowledge. […] The Investigator Scientist digs into the Analysis. DOI: 10.1007/978-3-030-65610-2_17. [9] Radziwill Nicole. (2017). Cyber security Cost of Quality: Managing the unknown observing, mapping, understanding, and piecing Costs of Cyber security Risk Management . Software Quality together in-depth knowledge and data, setting out the landscape Professional. 19. DOI: 10.48550/arXiv.1707.02653



for others to translate and develop” [10] Basholli Fatmir and Juraev Davron. (2024). [1]. The scientist is needed Framework, tools and challenges in cyber security . 1. 96-106. DOI: at specific moments, the innovation forces seasonal involvement, 10.13140/RG.2.2.21009.24161 the product is created more as a result of collaboration with [11] America’s Cyber Defense Agency, August 2024, < https://www.cisa.gov >



Prospects for the Use of AI Tools in the Republican Centre



for Technology Transfer Network



Alexander Uspenskiy Aliaksei Uspenski Maxim Prybylski

Republican Centre for Republican Centre for Republican Centre for

Technology Transfer Technology Transfer Technology Transfer

Center for System Analysis and Center for System Analysis and Center for System Analysis and

Strategic Research of the National Strategic Research of the National Strategic Research of the National

Academy of Sciences of Belarus Academy of Sciences of Belarus Academy of Sciences of Belarus

Minsk, Belarus Minsk, Belarus Minsk, Belarus

uspenskiy@mail.ru auspen79@gmail.com m.pribylsky@hotmail.com



ABSTRACT documentation associated with files that will remain accessible even several years after the file is closed (the data is the The paper informs about services and information resources property of the client); provided by the Republican Centre for Technology Transfer – automation of interaction between all participants in (RCTT) to innovation activity agents and prospects for the use international cooperation; of AI tools in the RCTT Network in order to improve the – direct access to file information from anywhere where quality and speed of preparing profiles (technology there is an Internet connection, even from mobile devices and offers/requests, business offers/requests and R&D requests), smartphones. creating promotion and marketing content to find potential United Nations System White Paper on AI Governance [2] partners, drafting contracts, etc. suggests an increasing recognition of AI's role in amplifying the



KEYWORDS work of governments and international bodies. Additionally,

Gartner predicts a staggering 80% of project management tasks

Technology transfer (TT), AI, generative AI, technology will use AI by 2030 [3], a testament to the growing reliability

transfer offices (TTOs) and trust in AI technologies within structured operational

frameworks.



1 Since 2023 the use of AI in the work of technology transfer INTRODUCTION

offices (TTOs) has been regularly discussed at webinars of the

"Will AI be replacing people in the near future?" "It looks to me Association of University Technology Managers (AUTM).

like, and for a while, AI is much better at doing tasks than On May 5, 2023, a webinar "Generative AI has Arrived:

doing jobs. It can do these little pieces super well, but Essential Knowledge for TTOs" was held, which explained:

sometimes it goes off the rails. It can't keep very long What is generative AI? Why should you care? The current state

coherence. So, people are instead just able to do their existing of AI and applications, such as ChatGPT, that are already at

jobs way more productively, but you really still need the human your disposal. How you can implement these tools in your

there today." Sam Altman, CEO of Open AI. office, some of the most pressing risks and concerns your office

As noted in the UNECE White Paper on the use of Artificial might face, and a look into what's coming next.

Intelligence in Trade Facilitation [1], artificial intelligence (AI) On March 3, 2024, a webinar "The AI Enabled TTO" was

is an enabling technology impacting the global economy and held, where the use of AI by TTOs to improve the efficiency of

international trade. Combined with business-process-oriented their work was discussed, in particular to: automate routine

automation and more efficient data flow exchanges, AI further tasks, analyze market trends, assess competitors, assess

promises to lift barriers to international trade, stimulate growth intangible assets, speed up decision-making and optimize

in global electronic commerce and allow for better predictions resources.

and associations to inform policy decisions. On May 2, 2024, a webinar "Tailoring Your AI Tools for

The benefits of AI-based systems include: Tech Transfer Transformation" was held. This webinar explores

– reducing the time spent on working with one document by customizing AI tools to better support unique tech transfer

more than 80%; processes and goals.

– reducing the number of errors in procedures; Video recordings and presentations of these webinars can be

– creating centralized repositories of information and found on the Internet portal of AUTM [4].

Participation in the above webinars, as well as the analysis

Permission to make digital or hard copies of part or all of this work for personal or of publications [1–6] shows that, AI can be used in the work of

classroom use is granted without fee provided that copies are not made or

distributed for profit or commercial advantage and that copies bear this notice and the TTOs to:



the full citation on the first page. Copyrights for third-party components of this – improve the quality of profiles (technology offers/requests, work must be honored. For all other uses, contact the owner/author(s). business offers/requests and R&D requests) published in the TT Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s). networks;



– analyze big data to identify potential technologies for TT, the NAS of Belarus"; "New partnership opportunities", to

as well as to predict market trends and demand; present in real-time offers and requests from RCTT, NATT,

– monitor patents and publications – AI can monitor AUTM and EEN networks; "Catalogs"; "Manuals"; "Media";

publications and patents related to a particular technology to "Commercialization", with subsections "IP auctions",

assess interest from the scientific community and industry; "Investment and venture funds", "Crowdfunding" and

– search for technologies – use machine learning algorithms "Technoparks of Belarus"; "IP insurance"; "Legislation",

to search for and compare technologies, patents and research covering the laws and regulations applicable to innovation

results that can be commercialized or licensed; activity in Belarus and foreign countries, and others.

– automate processes – AI can help automate routine tasks RCTT provide services to more than 250 Belarusian state

such as technology and intangible asset assessment, document organizations, private enterprises and individuals. The National

management and licensing processes; Academy of Sciences, Belarusian State University, Belarusian

– predict risks – analyze risks and possible obstacles in National Technical University are among its clients. In 2003–

technology transfer using machine learning methods to predict 2024 with the support from RCTT more than 8500+ persons

the likelihood of project success; improved their skills in the filed of technology transfer at 650+

– improving communication – using chatbots or neural national and international events (workshops, conferences,

networks to interact with potential partners and clients. exhibitions).

RCTT is the coordinator of the Republican Center for



2 Technology Transfer Network which contains more than 3000 SERVICES AND INFORMATION

technology offers, technology requests, business offers,

RESOURCES PROVIDED BY THE business requests, and offers for cross-border R&D

REPUBLICAN CENTER FOR collaboration of Belarusian enterprises and organizations.



Tasks set for RCTT: INNOVATION ACTIVITY AGENTS – 1120+ cooperation offers from NASB organizations in Russian language and 1010+ in English language, TECHNOLOGY TRANSFER TO As of August 2024 the Internet portal of RCTT contains:



serving clients in the technology transfer sector; in Russian, English and Chinese, – provide RCTT clients with access to foreign technology – information about 250+ exhibitions, 50+ brokerage events, transfer networks; – create and maintain information databases meant for organizations of the National Academy of Sciences of Belarus – 45+ catalogs, presenting services and products of



promotion of their innovation and investment projects; where organizations of the NAS of Belarus took part (will take – train specialists in research- and innovation-related part) in 2019–2024, entrepreneurship; – 50+ educational materials in the field of IP management, – assist innovation activity agents in development and management, transfer and commercialization of technologies 210+ webinars and events in the field of intellectual property



unified national network of technology transfer centers; – establish RCTT offices across the country, to create a technology transfer and commercialization.

– promote international technical and scientific cooperation



and exchange of experts. 3 PROSPECTS FOR THE USE OF AI TOOLS RCTT is a consortium with the headquarters in Minsk that IN THE RCTT NETWORK comprises [7, 8]: RCTT plans to use AI tools to solve the following problems: – 5 regional offices and 30 branch offices at research – automation and improvement of the quality of profile organizations, institutes, universities, enterprises in Brest, preparation (technology offers/requests, business Vitsebsk, Homel, Hrodna, Lida, Minsk, Mahileu, Novapolatsk offers/requests and R&D requests); and other cities and towns across Belarus; – creation of promotion and marketing content to find – 98 foreign partners in 23 countries: Armenia (3), partners; Azerbaijan (2), China (25), the Czech Republic (2), Denmark – automatic scanning and analysis of Internet resources, (1), Germany (4), Georgia (1), India (1), Iran (1), Italy (1), scientific publications, patents, catalogs and other data sources Kazakhstan (6), Lithuania (1), Moldova (1), Poland (3), Russia to identify competitors and potentially valuable technologies; (25), South Africa (1), South Korea (4), Sweden (1), UK (2), – identification of technologies that can be successfully the USA (3), Ukraine (7), Uzbekistan (1), Vietnam (2); commercialized by matching the proposed technologies and – 2 overseas field offices in China. services with market needs; RCTT has implemented over 400 projects, including over – determination of optimal product promotion channels and 100 international projects funded by UNDP, UNIDO, CEI, EU, optimization of marketing strategies; the Swedish Institute, etc. – support of the negotiation process by providing RCTT experts are certified members of 14 foreign information on market prices, transaction terms, etc.; technology transfer networks. – monitoring and management of the commercialization RCTT offers its services to innovation activity agents in process. After a contract is concluded, AI can be used to track Belarus as well as foreign companies and investors. progress in commercializing the technology and identify RCTT has a web-portal [9], with several subject sections possible problems or opportunities; and databases such as: "Virtual exhibition of the NAS of – improvement of communication – use of chatbots or Belarus"; "Catalogue of innovation offers by organizations of neural networks to interact with potential partners and clients. A



chatbot is created to build a dialogue with the user. It simulates – Dataminr is a platform for monitoring news and social

a conversation between real people and can respond briefly to a media using machine learning. It allows you to discover events,

simple request or construct a complex conversation with a high trends and competitors that may be important for a specific

level of personalization. Neural networks are a type of machine business or research;

learning in which a computer program works on the principle of – Crayon is an online competitor and technology monitoring

the human brain, using various neural connections. A neural platform. It uses machine learning to automatically scan

network can be either a learning or self-learning system. websites, social media, and other data sources to provide

As part of the modernization and development of the insights into the competitive landscape and emerging

automated system of information support for innovation technologies;

activities and technology transfer in the NAS of Belarus (ASIS – Cortico is a data analysis tool that uses artificial

IATT), commissioned in December 2021 [10], on the basis of intelligence to process and classify information from various

which the RCTT network operates, the following work is sources, such as the Internet, news articles, and social media. It

planned: can help identify trends, competitors, and new technologies.

1. Analysis, selection and adaptation of AI models for Integration of AI tools into the ASIS IATT subsystems will

carrying out work aimed at integrating the selected AI models reduce time, labor, and technological costs and improve the

into the ASIS IATT subsystems; quality and speed of preparing profiles, creating promotion and

2. Integration of AI tools into the subsystems of the ASIS marketing content to find potential partners, and preparing

IATT. contracts in the RCTT Network.

Here are some examples of generative AI tools that can be



marketing content, scanning and analyzing Internet resources, 4 used when preparing profiles, creating promotion and CONCLUSIONS

and solving other problems: The paper informs about prospects for the use of AI tools in the

a) AI tools for profile descriptions and other texts could be: RCTT Network for reduce time, labor, and technological costs,

– OpenAI GPT-3 or Generative Pre-trained Transformer 3 is improve the quality and speed of services provided.

a powerful neural network model capable of generating text Examples of generative AI tools, that planned to be used in

based on provided contextual data. It can be used to the RCTT Network for preparing profiles, creating promotion

automatically generate technology descriptions, technical and marketing content, scanning and analyzing Internet

concepts, and other text materials; resources, preparing contracts, and solving other problems are

– IBM Watson Natural Language Generator is a tool that given.

allows you to automatically generate text based on specified

templates and parameters. It can be used to create descriptions

of technology features, product specifications, and other



technical materials; ACKNOWLEDGMENTS – Copy.ai is a platform that provides a wide range of tools We would like to thank the National Academy of Sciences of for generating text content, including descriptions, headlines, Belarus and the State Committee on Science and Technology of articles, and more. It can be used to create excellent the Republic of Belarus for their constant support of RCTT descriptions of technologies and products; activities and express gratitude to all our colleagues who work – Jasper (Adobe's AI Copywriting Assistant) is a tool that in technology transfer for their help and advice. uses AI to generate text that can be used to create technology

descriptions, blogs, and advertising materials;

– ChatGPT by OpenAI is a generative neural network model



user input. It can be used to chat with the user, provide [1] that can hold a conversation and generate text content based on REFERENCES information about technology, and answer questions; Sray Agarwal (Ed.). 2023. UNECE White Paper on the use of Artificial

Intelligence in Trade Facilitation. UNECE – UN/CEFACT.

– Writesonic is another AI-powered writing assistant that [2] Inter-Agency Working Group on Artificial Intelligence (IAWG-AI).

enables users to generate a variety of content types quickly and 2024. United Nations System White Paper on AI Governance: An

analysis of the UN system's institutional models, functions, and existing



users generate long-form content efficiently. – Nations System. ShortlyAI is an AI writing assistant focused on helping [3] Katie Costello. 2019. Gartner Says 80 Percent of Today's Project efficiently; international normative frameworks applicable to AI governance. United

Management Tasks Will Be Eliminated by 2030 as Artificial Intelligence



available to automatically scan, analyze, and identify b) There are a number of AI and machine learning tools Takes Over (March 20, 2019). Retrieved September 11, 2024 from https://www.gartner.com/en/newsroom/press-releases/2019-03-20- gartner-says-80-percent-of-today-s-project-management. competitors and potentially valuable technologies. Here are [4] AUTM. Webinar Library. Retrieved from https://imis.autm.net/customer/customer/Ext/Webinar_Library.aspx some of them: [5] Berna Uygur, Steven M. Ferguson. 2024. Will Artificial Intelligence – Scite.ai is a platform for analyzing research articles and Shape The Future Of Technology Transfer? A Guide For Licensing



academic publications using AI. It allows you to identify Professionals. Retrieved September 11, 2024 from https://www.techtransfer.nih.gov/sites/default/files/documents/Ferguson connections between studies, evaluate their reliability, and find %20-%20les%20Nouvelles%20Vol%20LIX%20no%201%20pp%201- new technological directions; 11%20(March%202024)%5B2%5D.pdf. [6] Ming-Hui Huang, Roland T. Rust. 2021. A strategic framework for – PatSnap is an AI-powered patent and intellectual property artificial intelligence in marketing. Retrieved September 11, 2024 from scanning tool that helps you research competitors, identify new https://link.springer.com/article/10.1007/s11747-020-00749-9.

technologies, and assess their business potential;



[7] Успенский А.Ал. 2024. Республиканский центр трансфера

технологий: 20 лет в национальной инновационной системе

(история развития, структура, методология, деятельность,

перспективы). Центр системного анализа и стратегических

исследований НАН Беларуси, Минск. ISBN 978-985-6999-29-4.

[8] Alexander Uspenskiy, Aliaksei Uspenski and Maxim Prybylski. 2021.

Technology Transfer in Belarus. In Proceedings of the 24th International

Multiconference INFORMATION SOCIETY – IS 2021 Volume E – 7

October 2021, Ljubljana, Slovenia. 62–64. ISBN 978-961-264-224-2

(PDF).

[9] Republican Centre for Technology Transfer. Retrieved from

https://www.ictt.by.

[10] Григянец Р.Б., Успенский А.А., Венгеров В.Н. 2023. Формирование

и ведение единого информационного ресурса по обеспечению

инновационной деятельности и трансфера технологий в НАН

Беларуси. Материалы конференции "Развитие информатизации и

государственной системы научно-технической информации

(РИНТИ-2023)" 16 ноября 2023. Минск. Беларусь. 164–167.



Zbornik 27. mednarodne multikonference



INFORMACIJSKA DRUŽBA – IS 2024

Zvezek F



Proceedings of the 27th International Multiconference



INFORMATION SOCIETY – IS 2024

Volume F



Demografske in družinske analize



Demographic and Family Analyses



Uredniki / Editors



Janez Malačič, Mari Jože Osredkar, Matjaž Gams



http://is.ijs.si



7. oktober 2024 / 7 October 2024

Ljubljana, Slovenia





PREDGOVOR




Demografska vprašanja postajajo čedalje pomembnejša za Slovenijo, Evropo, svet. Poleg

prenaseljenosti se pogosto pojavljajo teze kot pretirana migracija in depopulacija, vse pa se

demonstrirajo kot resen politični in gospodarski izziv.



Slovenija, podobno kot mnoge evropske države, se sooča s posledicami nizke rodnosti in

staranja prebivalstva, kar močno vpliva na delovno silo, socialno varnost in produktivnost.

Poleg tega obsežne migracije ustvarjajo dodatne pritiske na politične in socialne strukture v

Sloveniji, Evropi in ZDA. Naraščanje svetovne populacije je bilo zadnjih sto let eksponentno,

a ta rast se upočasnjuje, zato depopulacija morda ni več tako nujna kot pred desetletji.



Kljub splošnemu prepričanju, da na demografske trende ni mogoče bistveno vplivati, izjemne

demografske spremembe niso prišle same od sebe. Z ustreznimi politikami in inovativnimi

pristopi lahko vplivamo na demografske razmere ter jih prilagodimo, da bodo najboljše za

ljudi in celotno človeško civilizacijo. Zavedanje o nujnosti iskanja primernih rešitev je

ključno, saj tako prenaseljenost kot depopulacija s sabo prinašata številne negativne posledice,

kot so problemi delovne sile, izzivi pri vzdrževanju socialnih sistemov in upad gospodarske

rasti.



Migracijski pritiski poudarjajo potrebo po uravnoteženih in trajnostnih politikah, ki hkrati

naslavljajo migracijske tokove ter ohranjajo socialno in kulturno kohezijo. Slovenija, ki je

obdana z dinamičnimi migracijskimi tokovi, se sooča z izzivom integracije priseljencev,

hkrati pa želi ohraniti jezik, nacionalno identiteto in družbeno stabilnost. To zahteva

učinkovite politike, ki bodo omogočale uspešno prilagoditev spremenjenim demografskim

razmeram.



V tem kontekstu je razumevanje in obvladovanje demografskih trendov nujno. Prilagodljive

politike, inovativne rešitve ter sodelovanje med strokovnjaki, politiki in družbo so ključni

elementi, ki nam lahko pomagajo soočiti se s spreminjajočim se demografskim okoljem. Po

mnenju nekaterih strokovnjakov bo že v nekaj desetletjih v številnih evropskih državah prišlo

do dramatičnega demografskega obrata, kjer bo avtohtono prebivalstvo postalo manjšina. V

nekaj stoletjih bi lahko prišlo celo do izumrtja malih evropskih narodov, kar predstavlja resen

izziv za ohranjanje kulturne in jezikovne raznolikosti ter identitete evropskih držav.



Prispevki na konferenci ponujajo osnovo za poglobljeno razpravo o teh ključnih vprašanjih s

poudarkom na Sloveniji, Evropi in globalnem kontekstu. Naš cilj je združiti strokovnjake in

odločevalce, da skupaj raziskujejo strategije za učinkovito soočanje s kompleksnimi

demografskimi izzivi in tako prispevajo k boljši prihodnosti za vse.



S skupnimi prizadevanji in izmenjavo znanja lahko oblikujemo trajnostne in učinkovite

rešitve, ki bodo prispevale k dolgoročnemu demografskemu razvoju.



Janez Malačič, Mari Osredkar, Jože Sambt in Matjaž Gams





FOREWORD




Demographic issues are becoming increasingly important for Slovenia, Europe, and the

world. In addition to overpopulation, topics such as excessive migration and depopulation

frequently arise, manifesting as serious political and economic challenges.



Like many European countries, Slovenia faces the consequences of low birth rates and an

aging population, which significantly impact the labor force, social security, and productivity.

Furthermore, large-scale migrations create additional pressures on political and social

structures in Slovenia, Europe, and the United States. While the global population has grown

exponentially over the last hundred years, this growth is slowing, suggesting that

depopulation may no longer be as urgent as it was decades ago.



Despite the common belief that demographic trends cannot be significantly influenced, major

demographic changes did not occur on their own. With appropriate policies and innovative

approaches, we can influence demographic conditions and adapt them to be the most

beneficial for people and human civilization. Recognizing the urgency of finding suitable

solutions is essential, as overpopulation and depopulation bring numerous negative

consequences, such as labor force issues, challenges in maintaining social systems, and

declining economic growth.



Migration pressures emphasize the need for balanced and sustainable policies that address

migration flows while maintaining social and cultural cohesion. Slovenia, surrounded by

dynamic migration flows, faces the challenge of integrating immigrants while striving to

preserve its language, national identity, and social stability. This requires effective policies

that will enable successful adaptation to changing demographic conditions.



In this context, understanding and managing demographic trends is crucial. Flexible policies,

innovative solutions, and cooperation between experts, policymakers, and society are key

elements that can help us face the changing demographic environment. According to some

experts, within just a few decades, many European countries will experience a dramatic

demographic shift where the indigenous population may become a minority. In a few

centuries, this could even lead to the extinction of small European nations, presenting a

serious challenge to preserving cultural and linguistic diversity and the identity of European

states.



The contributions at this conference provide a foundation for in-depth discussions on these

critical issues, with a focus on Slovenia, Europe, and the global context. Our goal is to bring

together experts and policymakers to jointly explore strategies for effectively addressing

complex demographic challenges and thus contribute to a better future for all.



Through collective efforts and the exchange of knowledge, we can shape sustainable and

effective solutions that will contribute to long-term demographic development.



Janez Malačič, Mari Osredkar, Jože Sambt, and Matjaž Gams



PROGRAMSKI ODBOR / PROGRAMME COMMITTEE



Janez Malačič, predsednik



Mari Osredkar, sopredsednik



Matjaž Gams, organizator



Jože Sambt



Drago Čepar



Christian Gostečnik



Majda Černič Istenič



Boštjan Kerbler



Karin Kasesnik



Dušan Kidrič



Marko Krevs



Tomaž Merše



Janja Pečar



Janja Povhe



Jože Ramovš



Milivoja Šircelj



Petronela Vertot



Božidar Voljč



Razvoj empatije, vloga spola in socialne vrednote pri



mladostnikih



The Development of Empathy, the Role of Gender, and



Social Values in Adolescents



Barbara Simonič

Katedra za zakonsko in družinsko terapijo ter psihologijo in sociologijo religije

Teološka fakulteta, Univerza v Ljubljani

Ljubljana, Slovenija

barbara.simonic@teof.uni-lj.si



Povzetek states of others, is of paramount importance during this

developmental period. The development of empathy in

V obdobju mladostništva se mladi srečujejo z vrsto izzivov in adolescents has a significant impact on not only their

sprememb, ki oblikujejo njihovo osebnost, vrednote in socialne interpersonal relationships but also the formation of their social

spretnosti. Empatija, kot sposobnost razumevanja in deljenja values, which are essential for the maintenance of a healthy and

čustev drugih, ima v tem obdobju ključno vlogo. Razvoj empatije cohesive society. Social values represent the fundamental

pri mladostnikih ne vpliva samo na njihove medosebne odnose, principles that inform the behaviour of individuals and groups

temveč tudi na oblikovanje njihovih socialnih vrednot, ki so and are instrumental in the formation of social order. The values

bistvene za zdravo in povezano družbo. Socialne vrednote so of solidarity, respect and justice are of significant importance in

temeljni principi, ki usmerjajo vedenje posameznikov in skupin the formulation of solutions to demographic challenges, life the



Vrednote, kot so solidarnost, spoštovanje in pravičnost, so comparative analysis of cognitive and emotional empathy among pomembne pri iskanju rešitev za demografske izzive, kot so npr. adolescents, based on the findings of a survey conducted among ter so ključnega pomena za oblikovanje družbenega reda. ageing of populations and migration. This paper presents a

staranje prebivalstva in migracije. V prispevku s pomočjo 231 adolescents in Greece, Slovenia and Portugal. It is evident

rezultatov raziskave, ki je vključevala 231 mladostnikov iz that adolescents achieve higher levels of cognitive empathy

Grčije, Slovenije in Portugalske, prikazujemo razlike v compared to emotional empathy. Girls score higher on both types

kognitivni in čustveni empatiji pri mladostnikih. Razvidno je, da of empathy. These findings highlight the necessity for tailored

mladostniki dosegajo višjo stopnjo kognitivne empatije v approaches to social values education and the promotion of

primerjavi z emocionalno. Dekleta dosegajo višje rezultate pri empathy according to gender, which is crucial to effectively

obeh vrstah empatije. Te ugotovitve poudarjajo potrebo po address demographic change and promote social inclusion and

prilagojenih pristopih pri vzgoji za socialne vrednote in understanding of others among adolescents. spodbujanju empatije glede na spol, kar je ključno za učinkovito



soočanje z demografskimi spremembami, za spodbujanje Keywords socialne vključenosti in za razumevanje drugih med mladostniki.

Adolescence, social change, cognitive empathy, emotional



Ključne besede empathy

Adolescenca, družbene spremembe, kognitivna empatija,

emocionalna empatija 1 Socialne vrednote in soočanje z



Abstract demografskimi spremembami



Adolescence is a period of significant transition and change, Socialne vrednote so temeljni principi in prepričanja, ki usmerjajo vedenje posameznikov in skupin v družbi. during which young people encounter a range of challenges and Opredeljene so kot standardi, ki jih posamezniki in družbene experiences that influence the development of their personality, skupine uporabljajo za opredelitev osebnih ciljev ter bistveno values and social skills. The capacity for empathy, defined as the oblikujejo naravo in obliko družbenega reda v kolektivu, tj. kaj ability to comprehend and vicariously experience the emotional je sprejemljivo in kaj ne, kaj mora biti in kaj ne, kaj je zaželeno

in kaj ne [1]. Te vrednote vključujejo spoštovanje, enakost,

Permission to make digital or hard copies of part or all of this work for personal or svobodo, pravičnost, solidarnost, odgovornost in empatijo [2].

classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full Vrednote so vir konkretnih smernic za ravnanje v praktičnih

citation on the first page. Copyrights for third-party components of this work must situacijah vsakdanjega življenja. Tesno so povezane s sorodnimi

be honored. For all other uses, contact the owner/author(s). pojmi, kot so stališča, prepričanja in norme, saj vplivajo nanje.

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s). Na vrednotah temeljijo temeljna prepričanja, na tem stališča in

https://doi.org/10.70314/is.2024.dem.1 norme, od tod izvirajo vedenjske namere in na koncu samo



vedenje [3]. Socialne vrednote so ključnega pomena za [12]. Koncept družbene empatije, ki združuje individualno

oblikovanje vedenja in odločitev, tudi ob soočanju z empatijo, razumevanje konteksta in družbeno odgovornost,

demografskimi spremembami, ki prinašajo veliko izzivov. zagotavlja okvir za obravnavanje strukturnih neenakosti in

Vrednote, kot so solidarnost, spoštovanje in pravičnost, spodbujanje družbene pravičnosti [13]. Te ugotovitve poudarjajo

usmerjajo pri iskanju rešitev za staranje prebivalstva, migracije pomen empatije in vrednot pri spodbujanju prosocialnega

in spreminjajočo se družbeno strukturo. Z močnimi socialnimi vedenja in trajnostnih odnosov med ljudmi in naravo, tudi v

vrednotami lahko spodbujamo medgeneracijsko sodelovanje, kontekstu soočanja z izzivi in različnostjo, ki jih prinašajo

vključevanje marginaliziranih skupin in pravično porazdelitev demografske spremembe.

virov. Te vrednote pomagajo graditi kohezivne skupnosti, kjer so

posamezniki pripravljeni prispevati k skupnemu dobremu, sicer

se krepi nelagodje in tesnoba ob soočanju z različnostjo [4]. Zato 3 Mladostnik, socialne vrednote in empatija

je krepitev socialnih vrednot nujna za trajnostno in pravično Mladostništvo je kritično obdobje za razvoj identitete in tudi

prilagajanje demografskim spremembam. oblikovanje in ponotranjenje vrednot, kjer sta značilni vse večja

notranja skladnost in stabilnost [14]. Na vrednote, ki usmerjajo



2 Empatija in njen družbeni pomen vplivajo kulturni, družbeni in verski dejavniki [15, 16]. Razvoj mladostnikovo vedenje in oblikujejo družbene interakcije,

Empatija, kot socialna vrednota, igra ključno vlogo v vrednot v mladostništvu vključuje neprekinjen cikel notranjega

medsebojnem razumevanju in podpori v sodobni družbi ter se vse potrjevanja z izbiro in udejanjanjem ciljev ter zunanjega

bolj uveljavlja kot temelj medosebnih odnosov in skupnostnega potrjevanja s socialno primerjavo med vrstniki [17].

sobivanja. Po definiciji gre pri empatiji za sposobnost, da Raziskave kažejo, da so vrednote mladostnikov tesno

razumemo in občutimo čustva ter izkušnje drugih ljudi [5]. Pri povezane z njihovo socialno kompetentnostjo in prilagoditvijo.

tem zasledimo dve različni, a med seboj povezani komponenti Socialne vrednote igrajo pomembno vlogo pri tem, saj vplivajo

empatije: kognitivno in emocionalno. Kognitivna empatija se na njihovo vedenje, odnose in odločanje. Vrednote, kot so

nanaša na sposobnost razumevanja perspektive, misli ali čustev spoštovanje, empatija, pravičnost in odgovornost, pomagajo

druge osebe na intelektualni ravni. Vključuje razumevanje mladostnikom razumeti njihov položaj v družbi in pomembnost

mentalnega ali čustvenega stanja druge osebe, ne da bi nujno prispevanja k skupnemu dobremu. Mladostniki, ki internalizirajo

delili ali doživljali to stanje. Emocionalna empatija je sposobnost pozitivne socialne vrednote, kažejo višjo stopnjo empatije in

deljenja in doživljanja čustev druge osebe na način, kot da bi bila sodelovanja [18]. Ti mladostniki so bolj verjetno vpleteni v

ta čustva lastna, pri tem pa sega onkraj zgolj golega razumevanja prostovoljne dejavnosti in imajo manj konfliktov z vrstniki, kar

[6]. kaže na pomembnost socialnih vrednot pri spodbujanju zdravega

Empatija, zlasti kognitivna, omogoča posameznikom, da socialnega vedenja [19]. Poleg tega socialne vrednote pomagajo

bolje razumejo perspektive drugih, kar je ključno za mladostnikom krmariti skozi zapletene družbene izzive, kot so

vzpostavljanje medsebojnega zaupanja in sodelovanja. Empatija pritiski vrstnikov in moralne dileme [20].

pomaga tudi pri zmanjševanju konfliktov in povečanju socialne Empatija ima v mladostništvu več pomembnih vidikov.

kohezije v družbi [7]. Največja tegoba sodobnega sveta je morda Pomembna je za povezovanje z drugimi, vzpostavljanje in

pomanjkanje empatije, kar vodi v individualizem in egoizem. krmarjenje v odnosih: empatija služi kot most, ki spodbuja

Ljudje pogosto pričakujejo, da bodo drugi čustvovali in razumevanje in povezovanje z vrstniškimi skupinami in

razmišljali tako kot oni, kar vodi v nestrpnost in razočaranje. družinskimi člani mladostnikov [21, 22]. Empatija je pomembna

Empatija nas uči, da je vsak posameznik drugačen in ga je treba za reševanje konfliktov, saj posameznikom omogoča

spoštovati ter jemati resno. Zahodna večkulturna družba pogosto razumevanje perspektiv in čustev drugih. To posledično olajša

pojmuje empatijo kot most za demokratični dialog, saj omogoča učinkovito komunikacijo in pogajalske veščine [23]. Empatija je

preseči egocentrični pogled na svet in zavzeti sociocentrični tesno povezana s prosocialnim vedenjem, vključno z dejanji

pogled, ki vključuje tudi druge [8]. prijaznosti in altruizma. Mladostniki z dobro razvitim občutkom

Empatija torej ni le individualna lastnost, ampak ključen empatije se pogosteje vključujejo v dejavnosti, ki koristijo

gradnik, ki oblikuje našo kolektivno identiteto in družbeno drugim, ter tako prispevajo k pozitivni in podporni skupnosti

strukturo. Dejavna empatija je ključna za socialni in moralni [24]. Za najstniška leta so značilne povečane čustvene izkušnje

razvoj posameznika. Kot socialna bitja potrebujemo socialne in razvijanje spretnosti čustvenega uravnavanja. Empatija

vezi za fizično in čustveno preživetje. Empatija pomaga mladostnikom pomaga prepoznati in razumeti lastna čustva in

usklajevati dejanja posameznikov, razumeti misli in namene čustva drugih, kar vodi k boljšemu čustvenemu uravnavanju in

drugih ter spodbuja solidarnost [9]. Pomaga pri oblikovanju mehanizmom spoprijemanja [25, 26]. Učinkovita komunikacija

trajnih socialnih vezi, spodbuja altruistično vedenje in preprečuje je temelj uspešnih odnosov. Empatija mladostnikom omogoča,

medosebno nasilje [8]. Pomembno je tudi zavedanje, da empatija da se uglasijo s čustvi, ki so v ozadju verbalnih in neverbalnih

omogoča prepoznati, kdaj so naša dejanja do drugih škodljiva, znakov, kar omogoča natančnejšo in sočutnejšo komunikacijo

kar nam omogoča, da ta dejanja popravimo [10]. [27].

Raziskave kažejo na močno povezavo med empatijo, Raziskave dosledno kažejo pozitivno povezavo med

socialnimi vrednotami in prosocialnim vedenjem. Empatija, empatijo, prosocialnim vedenjem in socialnim razumevanjem pri

zlasti čustvena empatija, je pozitivno povezana z altruističnimi mladostnikih. Empatija, zlasti empatična skrb, je močan

vrednotami, negativno pa z vrednotami, ki so usmerjene samo na napovedni dejavnik prosocialnega vedenja [28]. Razlike med

posameznika [11]. Ta povezava se ujema s hipotezo o empatiji spoloma so očitne, saj dekleta na splošno izkazujejo višjo stopnjo

in altruizmu. Empatija in relacijske vrednote so tesno prepletene empatije in prosocialnega vedenja kot fantje [28, 29]. Te



ugotovitve opozarjajo na zapleteno prepletanje med empatijo, in kognitivni empatiji pri mladostnikih in utemeljuje smiselnost

socialnim razumevanjem in prosocialnim vedenjem v razvoju upoštevanja različnih izhodišč pri vključevanju empatije v

mladostnikov. odzivanju mladostnika na zaznane družbene probleme in

potrebe.



4 Razvoj empatije v mladostništvu 5.1 Metoda

Razvoj empatije v mladostništvu je zapleten in večplasten 5.1.1 Udeleženci. V raziskavi je sodelovalo 231 mladostnikov,

proces, ki vključuje čustvene, kognitivne in vedenjske od tega 43,7% fantov in 56,3% deklet. Njihova povprečna starost

komponente. V obdobju mladostništva se razvoj empatije je bila 16,11 let (Min. 14; Max. 20; SD = 1,185).

bistveno spremeni in ima ključno vlogo pri oblikovanju socialnih 5.1.2 Pripomočki in postopek. Za merjenje stopnje kognitivne

interakcij in prosocialnega vedenja. Raziskave kažejo, da se in čustvene empatije je bil uporabljen Vprašalnik z za merjenje

empatija običajno povečuje od otroštva do mladostništva [27] in mladostnikove empatije in sočutja - AMES (ang. Adolescent

naprej v odraslost [30]. V zorenju nevronskih mrež in čustvenih Measure of Empathy and Sympathy) [34]. Instrument je

odzivov se že zgodaj pojavijo razlike med spoloma in vplivajo zasnovan za merjenje in razlikovanje kognitivne, emocionalne

na razvoj empatije [31]. empatije in sočutja/simpatije (te dimenzije v našo raziskavo

Pot razvoja empatije od zgodnjega mladostništva do nismo vključili). Udeleženci so posamezne trditve ocenjevali na

odraslosti ostaja nepopolno opisana, saj obstajajo notranji lestvici od 1 (nikoli) do 5 (vedno). Višji rezultati na posamezni

(biološki, kognitivni) in zunanji (socialni in okoljski) dejavniki, lestvici pomenijo višjo stopnjo kognitivne oz. čustvene empatije.

ki oblikujejo empatijo v mladostništvu. Kar raziskave Koeficient zanesljivosti (Cronbachov alfa) za dimenzijo

konsistentno omenjajo, je značilna razlika v stopnji empatije kognitivne empatije je znašal ɑ = 0,748, za dimenzijo

(tako čustvene kot kognitivne) glede na spol, kar je tudi emocionalne empatije pa ɑ = 0,767.

pomemben dejavnik pri na empatiji oblikovanem odnosu do Raziskava je bila izvedena v okviru ene izmed faz projekta

družbe in okolja. »Using virtual reality for inclusive and action-oriented empathy

Raziskave [32, 33] kažejo na razlike med spoloma pri razvoju in schools« (VR4Empathy) [35], katerega namen je oblikovati,

kognitivne (in čustvene) empatije. Čeprav tako fantje kot dekleta preizkusiti in razširiti ozaveščeno uporabo virtualne resničnosti

na splošno doživljajo rast empatičnih spretnosti, dekleta pogosto za razvoj kognitivne empatije v šolah ter tako povečal kakovost

kažejo nekoliko zgodnejši in naprednejši razvoj kognitivne in vključenost digitalnega izobraževanja na ravni EU. Projekt

empatije v mladostništvu. Pri dekletih se kognitivna empatija temelji na stičišču najsodobnejših raziskav VR, psihologije in

začne povečevati pri 13 letih. Čustvena empatija ostaja izobraževanja ter predlaga v prihodnost usmerjen pristop na

razmeroma visoka in stabilna v celotnem obdobju odraščanja. Pri ravni EU, ki bo zagotovil znanja in vire za uporabo VR kot

fantih se kognitivna empatija začne povečevati pri 15 letih. Med izobraževalne podpore pri poučevanju tem v več učnih načrtih,

13. in 16. letom pride do začasnega upada čustvene empatije, ki ki zajemajo teme, povezane z evropskimi zgodovinskimi dejstvi,

pa se v poznih najstniških letih ponovno vzpostavi. Upad skupnimi evropskimi socialnimi vrednotami ter evropskim

čustvene empatije pri mladostnikih je lahko delno povezan s državljanstvom in različnostjo, razvijajo kognitivno empatijo in

hkratnim povečanjem ravni testosterona. Povečanje ravni izboljšujejo učinkovitost učnih izkušenj. Udeleženci so v tej fazi

testosterona je povezano z željo po prevladi in moči nad drugimi poleg stopnje empatije ocenjevali še svoje digitalne kompetence

in ni težko razumeti, da so ti interesi pogosto v nasprotju s ter motivacijo za uporabo virtualne tehnologije v učnem procesu.

sposobnostjo sočutja in razumevanja drugih. Čeprav torej fantje Udeleženci iz treh sodelujočih držav (Grčija, Slovenija,

začenjajo razvijati kognitivne sposobnosti empatije in se Portugalska) so izpolnili vprašalnik v spletni obliki. Anketiranje

teoretično lahko postavijo v čevlje nekoga drugega, bodo morda je potekalo aprila 2024. Rezultate smo izračunali z uporabo

težje ravnali v skladu s tem, zlasti v skupinskih situacijah. Te programskega paketa SPSS 20.

ugotovitve poudarjajo dinamično in spolno pogojeno naravo

razvoja empatije v mladostništvu [28]. 5.2 Rezultati in razprava

Mladostniki so na vprašalniku AMES, ki razlikuje med



5 Raziskava: Razlike v čustveni in kognitivni emocionalno in kognitivno empatijo, dosegli višjo stopnjo



empatiji pri mladostnikih empatijo (M = 4,8). Na vsaki lestvici je bilo mogoče doseči kognitivne empatije (M = 10,2) v primerjavi z emocionalno

Razlike med spoloma v razvoju empatije v mladostništvu so največ 20 točk.

lahko pomembne za razvoj prosocialnega vedenja, saj empatija Razlike v stopnji kognitivne in emocionalne empatije so

igra ključno vlogo pri motivaciji posameznikov za pomoč vidne tudi glede na spol. Fantje (N = 101) so v povprečju pri

drugim in pri oblikovanju moralnih stališč. Razumevanje, kako kognitivni empatiji (M = 10,5) dosegli manj točk kot dekleta (N

se empatija razvija različno pri dečkih in deklicah, lahko osvetli = 130; M = 12,7). Prav tako so pri emocionalni empatiji fantje

mehanizme, ki spodbujajo ali zavirajo prosocialno in socialno (M = 3,3) dosegli manj točk kot dekleta (M = 9,0) (Slika 1).

odgovorno vedenje pri mladostnikih. Zato je smiselno zavedanje Preverili smo tudi, ali so te razlike statistično pomembne, za kar

teh razlik, da bi razvili učinkovite strategije za spodbujanje smo uporabili Mann-Whitneyev U-test, saj se podatki niso

prosocialnega vedenja pri obeh spolih in tako prispevali k bolj distribuirali normalno. Rezultati so pokazali, da so pri

sodelujoči in sočutni družbi. Za osvetlitev teh razlik mladostnikih glede na spol razlike v kognitivni (p =,000) in

predstavljamo rezultate raziskave, ki, tako kot številne druge emocionalni empatiji (p =,000) statistično pomembne. Dekleta

raziskave, prikazuje in potrjuje razlike med spoloma v čustveni dosegajo statistično pomembno višjo stopnjo kognitivne

empatije v primerjavi s fanti. Prav tako dekleta dosegajo Raziskave o empatiji in socialnih vrednotah pri mladostnikih

statistično pomembno višjo stopnjo emocionalne empatije v poudarjajo njen ključen pomen za razvoj socialnih kompetenc in

primerjavi s fanti. pozitivnih družbenih odzivov. Empatija se v mladostništvu

običajno povečuje, pri čemer je pri dekletih praviloma višja kot

pri fantih. Tako kognitivna kot čustvena empatija prispevata k

15 12,7 razvoju socialne kognicije in vplivata na interakcije

10,5 10,2

10 9,0 mladostnikov z različnimi perspektivami v družbi [36]. Razvoj



4,8 empatije v adolescenci napoveduje socialne kompetence v



5 3,3 odrasli dobi, vključno s komunikacijskimi spretnostmi in



zadovoljstvom v odnosih [27]. Spodbujanje in krepitev empatije



0 lahko predstavlja pomembno izhodišče za programe



KOGNITIVNA EMPATIJA EMOCIONALNA opolnomočenja mladostnikov, kar lahko prinese dolgoročne



EMPATIJA družbene koristi. Razumevanje razlik v empatiji med spoloma v

FANTJE obdobju najstništva omogoča bolj ciljno usmerjene pristope v DEKLETA VSI MLADOSTNIKI



vzgoji za socialne vrednote in učinkovitejše soočanje z



demografskimi spremembami, saj lahko prilagodimo strategije



Slika 1. Povprečne vrednosti kognitivne in emocionalne in programe, ki spodbujajo enakopravnost, povezanost in

empatije po spolu pri mladostnikih medsebojno razumevanje med mladostniki.



Rezultati kažejo, da so v splošnem sposobnosti kognitivne Zahvala

empatije pri mladostnikih boljše v primerjavi z emocionalno Financirano s strani Evropske unije. Izražena stališča in mnenja

empatijo, kjer so vrednosti precej nizke. To pomeni, da so so zgolj stališča in mnenja avtorja(-ev) in ni nujno, da odražajo

mladostniki razmeroma sposobni pri razumevanju perspektive stališča in mnenja Evropske unije ali Evropske izvajalske

ali čustvenega stanja druge osebe na intelektualni ravni, medtem agencije za izobraževanje in kulturo (EACEA). Zanje ne moreta

ko so njihove sposobnosti dejanskega deljenja in doživljanja biti odgovorna niti Evropska unija niti EACEA. čustev druge osebe razmeroma nizke. Pri kognitivni empatiji

dekleta dosegajo višje povprečne vrednosti v primerjavi s fanti, Literatura

vendar je stopnja kognitivne empatije še razmeroma enaka oz. [1] Tsirogianni, S., in Gaskell, G. (2011). The role of plurality and context in primerljiva. Večje razlike so pri emocionalni empatiji, kjer fantje social values. Journal for the Theory of Social Behaviour, 41(4), 441–465.

dosegajo bistveno nižje rezultate v primerjavi z dekleti. Ti vidiki DOI:https://doi.org/10.1111/j.1468-5914.2011.00470.x.

[2] Sekulic, D. (2019). Values: Global. The Blackwell Encyclopedia of

predstavljajo relevantna izhodišča pri ukrepih za spodbujanje Sociology, 1–5. https://doi.org/10.1002/9781405165518.wbeosv005.pub2



empatije pri mladostnikih ter pomenijo tudi različna izhodišča, (pridobljeno 20. julija 2024). DOI:https://doi.org/10.1002/9781405165518.wbeosv005.pub2. ki jih je potrebno upoštevati pri nagovarjanju empatije [3] Ives, C. D., in Kendal, D. (2014). The role of social values in the

mladostnika (npr. dekleta bodo lažje pri osvajanju neke vsebine management of ecological systems. Journal of Environmental

in situacije izhajala iz emocionalne empatije kot fantje, medtem Management, 144, 67–72.

DOI:https://doi.org/10.1016/j.jenvman.2014.05.013.

ko bo kapaciteta kognitivne empatije pri obeh spolih relativno [4] Brooks Masters, S. (2020). Change is Hard: Managing Fear and Anxiety

podobna). about Demographic Change and Immigration in Polarized Times. https://welcomingamerica.org/wp-

Soočanje z demografskimi spremembami, kot so večja content/uploads/2021/01/Changeishard_FINAL.pdf (pridobljeno 22.

kulturna raznolikost in različni družbeni izzivi, zahteva, da julija 2024).

vzgojitelji in oblikovalci politik upoštevajo te razlike v empatiji [5] Decety, J., in Jackson, P. L. (2004). The functional architecture of human

empathy. Behavioral and Cognitive Neuroscience Reviews, 3(2), 71–100.

med spoloma. Dekleta, ki so bolj nagnjena k emocionalni DOI:https://doi.org/10.1177/1534582304267187.

empatiji, bodo morda lažje sprejemala in se vključevala v [6] Preston, S. D., in de Waal, F. B. M. (2002). Empathy: Its ultimate and

proximate bases. Behavioral and Brain Sciences, 25(1), 1–20.

medkulturne in medosebne dinamike, ki zahtevajo čustveno DOI:https://doi.org/10.1017/s0140525x02000018.

povezanost in razumevanje. Pri fantih pa bo potrebno dati več [7] Baron-Cohen, S. (2011). The Science of Evil: On Empathy and the Origins

poudarka na spodbujanje razvoja emocionalne empatije, da bi of Cruelty. Basic Books.

[8] Tudor, K. (2011). Understanding empathy. Transactional Analysis

dosegli enako stopnjo socialne vključenosti in razumevanja. Journal, 41(1), 39–57.

Pri spodbujanju kognitivne empatije, kjer so razlike med DOI:http://dx.doi.org/10.1177/036215371104100107.

[9] Anderson, C., in Keltner, D. (2002). The role of empathy in the formation

spoloma manj izrazite, je smiselno uporabljati enake strategije za and maintenance of social bonds. Behavioral and Brain Sciences, 25(1),

oba spola, saj imajo mladostniki relativno podobne kapacitete za 21–22. DOI:http://dx.doi.org/10.1017/S0140525X02230010.

[10] Rifkin, J. (2009). The Empathic Civilization. Jeremy P. Tarcher/Penguin.

razumevanje perspektive drugih na intelektualni ravni. Prav tako [11] Persson, B. N., in Kajonius, P. J. (2016). Empathy and universal values

je kognitivna empatija pri mladostnikih višja v primerjavi z explicated by the empathy-altruism hypothesis. The Journal of Social

emocionalno, kar kaže na to, da je ta vidik empatije bolj smiselno Psychology, 156(6), 610–619.

DOI:https://doi.org/10.1080/00224545.2016.1152212.

nagovarjati in povezovati s socialnimi vrednotami in vključevati [12] Hagen, E. J., in Gould, R. K. (2022). Relational values and empathy are

v soočanje s socialnimi izzivi. To je mogoče doseči z closely connected: A study of residents of Vermont’s Winooski River

vključevanjem aktivnosti, ki spodbujajo razmišljanje o čustvenih watershed. Ecology and Society, 27(3), art19.

DOI:https://doi.org/10.5751/ES-13406-270319.

stanjih drugih in reševanje problemov skozi razumevanje [13] Segal, E. A. (2011). Social Empathy: A model built on empathy,

perspektiv drugih. contextual understanding, and social responsibility that promotes social

justice. Journal of Social Service Research, 37(3), 266–277.

DOI:http://dx.doi.org/10.1080/01488376.2011.564040.

[14] Daniel, E., in Benish‐Weisman, M. (2019). Value development during

6 adolescence: Dimensions of change and stability. Journal of Personality, Sklep

87(3), 620–632. DOI:https://doi.org/10.1111/jopy.12420.



[15] Wong, D. S. W., in Yeung, C. T. Y. A. (2020). Values in adolescence. V intelligence in adolescent cyberaggressors and cybervictims. International

J. D. Jewell in S. Hupp (ur.), The Encyclopedia of Child and Adolescent Journal of Environmental Research and Public Health, 17(13), 4681. Development (str. 1–11). Wiley. DOI:https://doi.org/10.3390%2Fijerph17134681.

[16] Stattin, H., in Kerr, M. (2001). Adolescents’ values matter. V J.-E. Nurmi [27] Allemand, M., Steiger, A. E., in Fend, H. A. (2015). Empathy

(ur.), Navigating Through Adolescence (str. 21–58). Routledge. development in adolescence predicts social competencies in adulthood.

[17] Kindschi, M., Cieciuch, J., Davidov, E., Ehlert, A., Rauhut, H., Tessone, Journal of Personality, 83(2), 229–241.

C. J., in Algesheimer, R. (2019). Values in adolescent friendship networks. DOI:https://doi.org/10.1111/jopy.12098. Network Science, 7(4), 498–522. DOI:https://doi.org/10.5167/uzh- [28] Van der Graaff, J., Carlo, G., Crocetti, E., Koot, H. M., in Branje, S. 175078. (2018). Prosocial behavior in adolescence: Gender differences in

[18] Wentzel, K. R. (2014). Prosocial behavior and peer relations in development and links with empathy. Journal of Youth and Adolescence,

adolescence. V L. M. Padilla-Walker in G. Carlo (ur.), Prosocial 47(5), 1086–1099. DOI:https://doi.org/10.1007/s10964-017-0786-1. Development: A Multidimensional Approach (str. 178–200). Oxford [29] Sánchez-Queija, I., Oliva, A., in Parra, Á. (2006). Empatía y conducta University Press. prosocial durante la adolescencia. Revista de Psicología Social, 21(3),

[19] Berk, L. E. (2018). Development Through the Lifespan. Pearson 259–271. DOI:https://doi.org/10.1174/021347406778538230.

Education, Inc. [30] Kim, E. J., Son, J. W., Park, S. K., Chung, S., Ghim, H. R., Lee, S., Lee,

[20] Steinberg, L. (2017). Adolescent brain science and juvenile justice S. I., Shin, C. J., Kim, S., Ju, G., Park, H., in Lee, J. (2020). Cognitive and

policymaking. Psychology, Public Policy, and Law, 23(4), 410–420. emotional empathy in young adolescents: An fMRI study. Journal of the DOI:https://psycnet.apa.org/doi/10.1037/law0000128. Korean Academy of Child and Adolescent Psychiatry, 31(3), 121–130.

[21] Rieffe, C., in Camodeca, M. (2016). Empathy in adolescence: Relations DOI:https://doi.org/10.5765%2Fjkacap.200020.

with emotion awareness and social roles. The British Journal of [31] Decety, J., in Svetlova, M. (2012). Putting together phylogenetic and Developmental Psychology, 34(3), 340–353. ontogenetic perspectives on empathy. Developmental Cognitive DOI:https://doi.org/10.1111/bjdp.12133. Neuroscience, 2(1), 1–24. DOI:https://doi.org/10.1016/j.dcn.2011.05.003.

[22] Miklikowska, M., Tilton-Weaver, L., in Burk, W. J. (2022). With a little [32] Gaspar, A., in Esteves, F. (2022). Empathy development from adolescence

help from my empathic friends: The role of peers in the development of to adulthood and its consistency across targets. Frontiers in Psychology, empathy in adolescence. Developmental Psychology, 58(6), 1156–1162. 13, 936053. DOI:https://doi.org/10.3389/fpsyg.2022.936053. DOI:https://psycnet.apa.org/doi/10.1037/dev0001347. [33] Van Hoorn, J., Van Dijk, E., Güroğlu, B., in Crone, E. A. (2016). Neural

[23] Van Lissa, C. J., Hawk, S. T., De Wied, M., Koot, H. M., Van Lier, P., in correlates of prosocial peer influence on public goods game donations

Meeus, W. (2014). The longitudinal interplay of affective and cognitive during adolescence. Social Cognitive and Affective Neuroscience, 11(6), empathy within and between adolescents and mothers. Developmental 923–933. DOI:https://doi.org/10.1093/scan/nsw013. Psychology, 50(4), 1219–1225. DOI:https://doi.org/10.1037/a0035050. [34] Vossen, H. G. M., Piotrowski, J. T., in Valkenburg, P. M. (2015).

[24] Weisz, E., Chen, P., Ong, D. C., Carlson, R. W., Clark, M. D., in Zaki, J. Development of the Adolescent Measure of Empathy and Sympathy

(2022). A brief intervention to motivate empathy among middle school (AMES). Personality and Individual Differences, 74, 66–71. students. Journal of Experimental Psychology: General, 151(12), 3144– DOI:https://doi.org/10.1093/scan/nsw013. 3153. DOI:https://psycnet.apa.org/doi/10.1037/xge0001249. [35] VR4 Empathy – Using virtual reality for inclusive and action-oriented

[25] MacDonald, H. Z., in Price, J. L. (2019). The role of emotion regulation empathy in schools. (6. februar 2024). https://virtualreality4empathy.eu/

in the relationship between empathy and internalizing symptoms in (pridobljeno 23. julija 2024). college students. Mental Health & Prevention, 13, 43–49. [36] Jacob, L., in Reddy, K. J. (2024). Role of cognitive and emotional empathy DOI:http://dx.doi.org/10.1016/j.mhp.2018.11.004. in the development of social cognition among adolescents. IP Indian

[26] Segura, L., Estévez, J. F., in Estévez, E. (2020). Empathy and emotional Journal of Anatomy and Surgery of Head, Neck and Brain, 10(1), 4–9.

DOI:http://dx.doi.org/10.18231/j.ijashnb.2024.002.



Prednosti IKT tehnologije pri pouku športne vzgoje



Advantages of ICT technology in physical education



lessons



Polona Rajher

III. Gimnazija

Maribor, Slovenija

polona.rajherl@gmail.com



Abstract / Povzetek are often confronted with the importance of healthy eating in our

learning system, and healthy eating is often a topic in the media.

V prispevku je predstavljeno, kako se lahko s pomočjo IKT Despite the frequent mention of the importance of a healthy diet,

tehnologije lotimo izziva prekomerne telesne teže dijakov in research has shown that adolescents often resort to unhealthy

dijakinj ter spregovorimo o pomenu zdrave prehrane in presnove, foods. More than half of the adolescents surveyed said they were

ki je v času odraščanja za mladostnike ključnega pomena. eating healthy, but the results showed that only 28% actually did.

Mladostniki se v našem učnem sistemu pogosto srečajo s Consumers often buy food at the store based on the appearance

pomenom zdrave prehrane, prav tako pa je zdrava prehrana of the packaging, and adolescents are even more receptive to the

pogosto tema v medijih. Kljub pogostemu omenjanju pomena visual appearance and advertising of a particular product. As a

zdrave prehrane, je raziskava pokazala, da mladostniki result, unhealthy foods are often mistaken for healthy foods and

velikokrat posežejo po nezdravi hrani. Več kot polovica eaten in good faith. In order to approach adolescents in a different

anketiranih mladostnikov je mnenja, da se prehranjujejo zdravo, way the importance of a healthy diet for their development and

vendar so rezultati pokazali, da je v resnici takšnih zgolj 28%. metabolism, a study was conducted in which students first tested

Potrošniki pogosto v trgovini kupujemo hrano glede na izgled their knowledge of healthy eating, completed a survey of their

embalaže, mladostniki pa so glede na vizualni izgled in eating habits, and then participated in workshops on the

oglaševanje določenega izdelka še bolj dovzetni. Posledično importance of healthy eating and metabolism. Through the

nezdravo hrano pogosto zamenjajo za zdravo in jo v prepričanju, Energy for life application, students learned the importance of a

da je zanje ustrezna, jedo. Da bi mladostnikom na drugačen način healthy diet and how nutrients are metabolised in our bodies, and

približali pomen zdrave prehrane, je bila opravljena raziskava, v then tested their knowledge with a questionnaire. At the end of

kateri so dijaki najprej preverili svoje znanje o zdravi prehrani, the workshop, students used the application Veš kaj ješ? and

izpolnili anketo o njihovih prehranjevalnih navadah in bili nato made examples of healthy menus. If writing in Slovene, add

vključeni v delavnice o pomenu zdrave prehrane in presnove. Na English abstract below the Slovene one. delavnicah so dijaki s pomočjo aplikacije Energy for life na



interaktivni način spoznali pomen zdrave prehrane in kako se Keywords hranila presnovijo v našem telesu. Svoje znanje so nato preverili

z vprašalnikom. Na koncu delavnice so dijaki s pomočjo ICT technology, healthy nutrition, overweight, workshop,

aplikacije Veš kaj ješ? naredili primere zdravih jedilnikov. metabolism, aplication.



Ključne besede 1 Uvod

IKT tehnologija, zdrava prehrana, prekomerna telesna teža,

Zdrava prehrana je nekaj o čemer se zadnje čase zelo pogosto

delavnice, presnova, aplikacija.

govori. Pomen zdrave prehrane je ključen za zdrav razvoj, saj

Abstract telo za izgradnjo notranjih ogranov, mišic, kostnine in živčevja

nujno potrebuje hranila, ki jih vsebuje hrana. V zgodnjem

The article presents how we can tackle the challenge of excessive

otroštvu se običajno prehranjujemo bolj zdravo in uživamo

body weight in male and female students and address the

raznoliko prehrano, ki nam omogoča zdrav razvoj. Otrokov

importance of healthy diet and metabolism using ICT technology,

jedilnik določajo starši oz. skrbniki in posledično ne moremo kot

which is crucial for adolescents during their growth. Adolescents

posamezniki vplivati na izbor hrane in uživati zgolj hrane, ki nam

je všeč. V času pubertete, pa mladostniki počasi začnejo sami

določati svoj jedilnik in pri tem večkrat ne upoštevajo smernic

Permission to make digital or hard copies of part or all of this work for personal or zdravega prehranjevanja. Nepravilen izbor živil in njihovo

classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full uživanje pa lahko pusti posledice, ki se pokažejo kot nepravilna,

citation on the first page. Copyrights for third-party components of this work must neenakomerna rast za vse življenje, prav tako pa lahko vpliva na

be honored. For all other uses, contact the owner/author(s). oblikovanje prehranjevalnih navad v kasnejšem življenju. [6]

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s). Zaradi upada upoštevanja smernic zdrave prehrane skozi

https://doi.org/10.70314/is.2024.dem.2 obdobje pubertete je bila posledično opravljena raziskava katere



namen je ugotoviti prehranske navade mladostnikov, v kateri so kot je telesna neaktivnost in kajenje. Bolezni katere lahko

sodelovali dijaki tretjih letnikov III. gimnazije Maribor. Dijaki neposredno povežemo s prehrano so:

so sprva izpolnili vprašalnik, ki je preveril njihove prehranske • beri-beri (posledica, dolgoročnega pomankanja vitamina

navade, nato pa so preko didaktične delavnice, pri katerih so B1),

uporabljali interaktivni aplikaciji, ki mladostnike spodbudita k • skorbut (nastane zaradi pomanjkanja vitamina C v

boljšem razumevanju pomena zdrave prehrane in njene daljšem časovnem obdobju),

presnove, spoznavali načela zdravega prehranjevanja. Z vidika • hipertenzija (povišan krvni tlak zaradi hrane, ki vsebuje

varovanja zdravja je namreč zdrava prehrana za mladostnike visoke vrednosti nitratov),

izredno pomembna, saj imajo ravno v tem obdobju veliko večje • zobna gniloba (posledica prekomernega uživanja sladke

potrebe po beljakovinah, vitaminih in mineralih ter energiji. [1] hrane in slabe ustne nege).



Obstaja še množica drugi bolezni, vendar njihovega nastanka

ne morejo povezati izključno z slabimi prehranskimi navadami

2 Pomen zdrave prehrane [8].

Mladostniki se že v času osnovne šole pri pouku Ker so bolezni, ki jih povzroča nezdrava prehrana velik

gospodinjstva srečajo s pomenom zdrave prehrane, kjer se učijo javnozdravstveni problem, se veliko pozornosti namenja k

o prehranski piramidi in različni pripravi hrane. Nato se o omilitvami posledic le teh.

prehrani in presnovi učijo pri pouku biologije, kjer največ

poudarka namenijo presnovi ter vgradnji hranil v človeško telo

in posledicah nepravilne prehrane in prehranskih navad. Njihove

prehranske navade so v času osnovne šole običajno dobre, vsaj v 3 Interaktivno poučevanje

prvi in drugi triadi. V tretji triadi pa mladostniki med glavnimi Ustaljena praksa poučevanja mladostnikov o pomenu zdrave

obroki pogosto uživajo prigrizke za katere je značilno, da prehrane, očitno ne prinaša dobrih rezultatov, saj se delež

vsebujejo veliko energije, zaradi prisotnosti sladkorjev in mladostnikov, ki imajo slabe prehranske navade in

maščob, ne vsebujejo pa hranilnih snovi, ki jih telo potrebuje za posledično zdravstvene težave, iz leta v leto povečuje. Razlog za

krepitev odpornosti. Ena izmed praks mladostnikov v tretji triadi slabe navade mladostnikov, bi lahko bili tudi zastareli načini

je opuščanje zajtrka in neustrezen ritem prehranjevanja ter poučevanja v šoli, saj so učenci zaradi uporabe interneta,

premajhne količine zaužite zelenjave. Vendar se prehranske mobitelov in televizije navajeni na interaktivno pridobivanje

navade osnovnošolskih otrok v tretji triadi v veliki meri informacij. Frontalen način poučevanja, katerega uporablja

razlikujejo od navad srednješolskih otrok, saj le ti velikokrat večina učiteljev, v večini primerov ne doprinese k trajnostnem

nimajo primerne šolske malice. [5] znanju, saj so učenci do takšnega pridobivanja informacij vedno

manj motivirani pri učnem procesu [4]. Učitelj mora dan danes

Zdravo prehrano bi lahko opredelili z naslednjimi učence spodbujati, da sami pridejo do določenih spoznanj, ne da

smernicami: jim zgolj posreduje znanje. Z uporabo in upoštevanjem

• redna prehrana (vsaj 5 obrokov dnevno), konstruktivističnega pristopa pri izkustvenem učenju lahko

• uživanje hrane glede na prehransko piramido, učitelj učencem pomaga povezati ugotovitve s praktičnimi

• raznolikost hrane, zakonitostmi. Uporaba IKT (informacijsko komunikacijske

• počasno uživanje hrane, ki jo dobro prežvečimo, tehnologije) lahko omogoči posredovanje učne snovi na bolj

• izogibanje sladkim in mastnim jedem, atraktiven način. Pri pouku lahko uporabimo tako računalnike,

• izogibanje prigrizkom [2]. kot tudi mobilne telefone, sploh v primeru, ko imamo v vzgojno

izobraževalnem zavodu težave zagotoviti primerno število

Iskanje krivca slabih prehranskih navad mladostnikov ni računalnikov za večjo skupino učencev. Uporaba mobilnih

enostavno, saj na njihove navade vpliva mnogo različnih telefonov pri pouku ima tudi svoje omejitve in predstavlja

dejavnikov. Živimo v okolju, ki spodbuja debelost, saj smo kot določene izzive, sploh od učiteljev. Prednost mobilnih telefonov

družba izredno potrošnikov naravnani in nam mediji dnevno je v tem, da jih ima večina učencev in da so običajno zmogljivejši

posredujejo vedno nove in nove proizvode. Poleg tega, pa nas od šolskih računalnikov [3].

okolje spodbuja k temu, da smo vedno manj aktivni (električna

kolesa, električni skiroji, uporaba osebnega vozila...). Podatki za

Slovenijo kažejo naraščajoče gibanje prekomerno hranjenih in

debelih mladostnikov in odraslih pri nas, kar kaže, da se navade 4 Aplikacije za poučevanje o zdravi prehrani in

pogosto prenašajo s starejše na mlajšo generacijo [7]. Debelost presnovi

pri mladostnikih je namreč eden izmed največjih problemov

današnjega časa, saj prekomerna telesna teža v mladostniškem Na spletu je množica različnih aplikacij in spletnih strani, ki

obdobju zelo pogosto nakaže na prekomerno telesno težo tudi v opisujejo zdravo prehranjevanje vendar je zgolj peščica

odraslem obdobju. Pri mladostniku neuravnotežena prehrana didaktično primernih za uporabo v učnem procesu. Večina

povzroča manjšo delovno storilnost, utrujenost, manjšo aplikacij je namenjena ali izobraževanju ali zabavi, vendar le

sposobnost pomnenja in kronične bolezni. Strokovnjaki so malo je takšnih, ki bi združevale oba spektra. Ena izmed

ugotovili, da so številne bolezni neposredno povezane s hrano, aplikacij, ki je tako poučna, da uporabnik pridobi in utrdi svoje

kar pomeni da je od prehrane odvisno naše zdravje. Nepravilna znanje, kot tudi zabavna, da se ne naveliča in je pripravljen dlje prehrana pa je večkrat povezana z različnimi slabimi navadami

časa ostati motiviran, se imenuje Energy for life. Aplikacija sicer kot učnega pripomočka med poukom oceniti smiselnost, pri

ne ponuja uporabe v slovenskem jeziku, temveč je zgolj v kateri učni snovi ga bo uporabil in katere učne cilje bo s pomočjo

angleškem in španskem jeziku. Posledično se lahko aplikacijo njih usvojil. Prav tako ni namen vsakodnevne uporabe mobilnih

uporabi kot primer medpredmetnega povezovanja biologije in telefonov pri učnem procesu, saj bi s tem njegova atraktivnost

angleščine. Prednost aplikacije je tudi v tem, da je popolnoma izzvenela, prav tako pa bi lahko postal moteči faktor v razredu

brezplačna in omogoča več različnih učnih modulov. Aplikacija zaradi morebitne nedovoljene uporabe iz strani učencev.

s pomočjo igrifikacije – uporaba dinozavra, ki pritegne pozornost

uporabnika, uči novih konceptov in ga motivira za interakcijo,

kar izboljša učni proces. Zaradi enostavnega umesnika je ta

aplikacija dostopna vsem učencem različnih starosti, tako v

osnovni, kot tudi v srednji šoli. Vizualen prikaz in strokovna

natančnost izrazov, se določi s pismenimi sposobnostmi

uporabnika. Velik poudarek je na znanstveni pismenosti,

povezan z vizualnimi dražljaji, ki uporabnika spodbujajo k

aktivni uporabi aplikacije. Aplikacija je zasnovana tako, da se jo

lahko uporablja ob spremstvu vodje učnega procesa ali pa se jo

uporablja samostojno. Slika 1 prikazuje dinozavra, ki

mladostnika nagovarja, kaj vse lahko s pomočjo aplikacije izve.

Uporabnik lahko izbira med različnimi področji, ki jih lahko s

pomočjo aplikacije osvoji in utrdi svoje znanje. Slika 2 prikazuje

modul prebavil, katerega uporabnik lahko izbere v aplikaciji in s Slika 2: Prikaz modula hranila pomočjo njega pregleda kaj vse njegovo telo potrebuje za

delovanje. Nato uporabnik preveri kaj se zgodi, če tega hranila

ne dobi (nastanek različnih obolenj). Prav tako, ga aplikacija

spodbuja k nadaljnjem razmišljanju in preizkušanju naučenega

znanja.





Slika 3: Prikaz aplikacije Veš kaj ješ?





5 Raziskava



Da bi ugotovili odnos do zdrave prehrane mladostnikov in

preverili znanje o pomenu zdravega prehranjevanja in presnove,

je bila opravljena raziskava, v kateri so sodelovali dijaki tretjih

Slika 1: Prikaz različnih modulov apliakcije

letnikov III. gimnazije Maribor. Za namen raziskave je bilo

izbranih 42 dijakov, ki so sprva izpolnili vprašalnik, ki je bil

Aplikacija Veš kaj ješ? omogoča uporabniku preveriti energijsko

sestavljen iz 14 vprašanj, katerih namen je bil ovrednotiti znanje

vrednost hrane. Prav tako uporabniku omogoča, da s pomočjo

dijakov o zdravi prehrani in presnovi, prav tako pa je bil namen



opravljanjem nakupov. Aplikacija nam omogoča uporabo učni uri je bila vloga učitelja zgolj kot usmerjevalec učnega procesa in ne posredovalca znanja. Po uporabi aplikacije so bili kamere pri pametnem telefonu in s tem skeniranje črtne kode dijaki pozvani kakšne so njihove prehanske navade in opisu izdelka, katerega prehransko vrednost želi uporabnik preveriti. njihovega tedenskega jedilnika. Ker so imeli pri jedilniku večje Aplikacija nam nato s pomočjo barvnega semaforja prikaže težave, so nato s pomočjo aplikacije Veš kaj ješ? preverili kakšne katerih hranil je v izdelku v zadostnih, prekomernih ali jedi so primerne za določen del dneva in katerih jedi se morajo premajhnih količinah. Pri sliki 4 lahko vidimo, kako je razčlenjen izogibati pri svoji prehrani. Po zaključku dela z aplikacijo so prehranski semafor, ki uporabniku sporoča še dopustno vrednost dijaki s pomočjo razgovora posredovali določenih hranil v izdelku. Aplikacija Veš kaj ješ? ni namenjana kaj so se naučili in izročke oblikovane skupaj z učiteljem zapisali zabavi in je posledično manj atraktivna kot aplikacija Energy for aplikacije je pomoč pri sestavi jedilnika, katerega lahko vsak izvedene 3 učne ure, pri katerih so dijaki s pomočjo aplikacije Energy for life ugotovili pomen zdrave prehrane in presnove, posameznik sestavi glede na svoje prehranske navade. Slika 3 nato pa so s pomočjo aplikacije tudi preverili svoje znanje. Pri prikazuje primer uporabe aplikacije na pametnem telefonu med prekoračenih vrednosti določenih hranil izogiba. Namen vprašalnika preveriti njihove prehranske navade. Nato so bile prehranskega semaforja ugotovi, katerih živil naj se zaradi



life, zato je vsaj uvodna uporaba aplikacije priporočljiva pod v zvezke. Nato so ponovno izpolnjevali vprašalnik, s katerim je bilo preverjeno, kaj so se naučili. Določeni rezultati vprašalnika vodstvom učitelja. Učitelj mora pred uporabo mobilnih telefonov



so bili pričakovani, saj se je tako znanje kot tudi odnos do zdrave dostopne preko mobilnih telefonov, kar pomeni, da lahko tudi

prehrane izboljšal. Presenetljivo je bilo to, da je bilo 54% vzgojno izobraževalni zavodi z omejitvami pri omogočanju

odstotkov dijakov mnenja, da se prehranjuje zdravo, vendar je dovolj velikega števila računalnikov za vse učence, omogočijo

bilo nato v nadaljnjem delu rezultatov vprašalnika prikazano, da svojim učencem izvedbo pouka na takšen način. Ali bo uporaba

so zmotno prepričani v svoj zdrav način prehranjevanja. Kar 48% takšnih aplikacij drastično vplivala na prehranske navade

dijakov navaja premalo časa kot glaven krivec za njihove slabe učencev je neodgovorno trditi, vendar njihova uporaba lahko

prehranske navade, 35% dijakov pa navaja, da stres močno pripomore k boljšem ozaveščanju zdravega načina

pripomore k njihovim prehranskim navadam. Kot glaven razlog prehranjevanja in razumevanju presnove hranil ter zdravemu

zakaj kupijo določen proizvod so dijaki podali izgled in njegovo načinu življenja.

oglaševanje. Večina dijakov ne pregleda hranilnih vrednosti pred

nakupom produkta. Pri vprašanju katera pijača ima najvišjo

vrednost sladkorja, References / Literatura

je sprva od 42 dijakov pravilno odgovorilo zgolj 5 dijakov, kar

nakazuje na njihovo zmotno znanje o hranilih v sladkanih [1] [1] Gavin, L. M., Dowshen, S. A., Izenberg, N.(2007). Otrok v formi.

pijačah. Praktični vodnik za vzgojo zdravih otrok – od Pri zapoznelen vprašalniku pa je pravilno odgovorilo 18

dijakov. Večina dijakov, kar 34 od 42 je odgovorila pri [2] rojstva do najstniških let. Ljubljana: Mladinska knjiga.

[3] [2] Kodele, M., Suwa – Stanojević, M., Gliha, M. Prehrana. Ljubljana:

zapoznelen vprašalnikov, da bodo več pozornosti namenili DZS, 2000.

svojim prehanskim navadam [4] [3] Lobe, B. in Muha, S. Internet v vsakdanjem življenju slovenskih otrok .

in mladostnikov (online). 2012. (citirano 15.

[5] 5. 2023). Dostopno na naslovu:

http://www.safe.si/uploadi/editor/1297947005MLADINAN

ETU_porocilo.pdf.

[6] [4] Ormrod, J.E. (2014). How Motivation Affects Learning and Behavior.

6 Zaključek Education Psychology, Developing Learners’

[7] (online). 2014. (citirano 21. 5. 2023). Dostopno na naslovu:

https://www.education.com/reference/article/motivationaffects-



prej v začetku izobraževalnega procesa. Vendar zgolj frontalno Ključno je, da se o zdravi prehrani začnemo pogovarjati čim [8] learning-behavior/. [9] [5] Polajžar, A, Razgor, E, Regner, A. (2009). Zakon spreminja prehranske navade srednješolcev »bolje pozno kot

poučevanje ne prinaša željenih rezultatov, saj se delež [10] nikoli«. Celje.

prekomerno debelih otrok in mladostnikov v zadnjem času [11] [6] Sizer, F., Whitney, E. (2006). Nutrition. 10th ed. Belmont: Thomson

Wadsworth.

drastično povečuje. V raziskavi je predstavljen eden od načinov, [12] [7] Zupančič, A., Hoyer, S. Prehranjevalne navade študentov.

kako lahko izobrazimo mlade o pomenu zdrave prehrane. [13] [8] Žitnik, N. Trtžnokomunikacijski program zdrave prehrane med dijaki

in študenti. Ljubljana: 2009.

Aplikacije, ki jih dan danes učenci uporabljajo pri pouku, so



E-branje in e-knjige: posebnosti in izzivi v šolski praksi



E-reading and e-books: specifics and challenges in school



practice



Dr. Maja Šebjanič Oražem

Osnovna šola Puconci

Puconci 178

9201 Puconci

maja.sebjanic@gmail.com



pretvarjanja v glas, tj. izgovorjave. Gre za človeku specifično

POVZETEK dejavnost, ki nam ni dana, ampak se je (na)učimo. Da postane

V prispevku predstavljamo e-knjige in e-branje v šolski avtomatiziran proces, je potrebno (dovolj) vaje. Kot tako

praksi. Osredotočamo se na posebnosti in izzive, ki se porajajo predstavlja najučinkovitejše sredstvo usvajanja znanja. Učenje

pri njihovem vključevanju v pouk slovenščine. Povzemamo branja poteka stopenjsko (več o tem v [18]).

nekatere raziskave e-branja, njegove specifike in možnosti, ki jih Če se osredotočimo na namene prispevka, bomo v

nudi za bralno motivacijo pri današnjih šolajočih se generacijah, nadaljevanju s pojmom “branje” zajeli opredelitev, uveljavljeno

ki so v digitalni svet vpete praktično na vseh korakih. v slovenistični stroke. Ta branje1 razume kot temeljno dejavnost,

ki vodi k bralni pismenosti [5].



KLJUČNE BESEDE

1.2 Kratek pregled raziskav e-branja

E-knjiga, e-branje, pouk slovenščine, motivacija za branje,

digitalna generacija V slovenskem prostoru je relativno malo raziskav, ki bi se

osredotočale na raziskovanje razlik med “klasičnim” branjem, tj.



ABSTRACT branjem s papirja, in med “digitalnim” oz. e-branjem. Med raziskavami izstopa S. Pečjak [20], ki s sodelavko ugotavlja, da In this paper we present e-books and e-reading in school je generacija Y 2 pri e-branju nekoliko hitrejša od branja s papirja, practice. We focus on the specifics and challenges that arise, kar pripisuje verjetnemu pogostemu pregledovanju družabnih when we include them in Slovenian language lessons. We omrežij, brskanju po spletu in branju informativnih vsebin, pri summarize some research of e-reading, its specifics and the čemer zaradi same narave tovrstnih besedil redkeje uporabljajo possibilities, that e-books offer for reading motivation among metakognitivne strategije in torej manj preverjajo svoje today's school-age generations, who are embedded in the digital razumevanje med samim branjem. Posledično lahko z world at practically every step. računalnika berejo hitreje kot s papirja, vendar manj natančno in



KEYWORDS več časa za naloge bralnega razumevanja in odgovarjanjem na imajo zato več težav s ponovnim iskanjem informacij, porabijo

E-book, e-reading, Slovene language lesson, reading vprašanja. Glede razumevanja prebranega pa med tistimi, ki so

motivation, digital generation brali s papirja, in tistimi, ki so brali na zaslonu, ni pomembnih

razlik [20].

Za primerjavo z rezultati slovenske raziskave naj navedemo

1 BRANJE IN E-BRANJE: PODOBNOSTI IN še primer raziskave na Norveškem, ki jo je s sodelavci izvedla E.

RAZLIKE Mangen [4]. Ugotovili so, da so učenci, ki so brali besedila s

papirja, dosegli boljše rezultate pri nalogah bralnega

1.1 Opredelitev (e-)branja razumevanja od tistih, ki so brali besedila na zaslonih. V tem

Pri strokovnih (in znanstvenih) razpravah o (e-)branju je segmentu se raziskavi med seboj razlikujeta.

vedno potrebno vzeti v obzir čas in s tem družbene okoliščine, A. Kepic Mohar [1] pa ugotavlja, da je proces pridobivanja

znotraj katerih se pogovarjamo o branju. Prav tako je potrebno znanja in s tem celotnega razvoja posameznika, za katerega je v

jasno definirati, o katerem segmentu branja oz. stopnji razvoja preteklosti veljalo, da potekalo (s knjigami in) z branjem,

bralne zmožnosti razpravljamo. dandanes le še pot do pragmatičnega cilja, torej orodje za lažji

Če se dotaknemo opredelitve branja, gre za proces vstop na trg delovne sile. Novi digitalni mediji, ki praktično

dekodiranja/prepoznavanja zapisanih črk in njihovega vsakodnevno vstopajo v vse sfere našega življenja, vplivajo in



1 2 V prispevku se bomo osredotočili na branje digitalnega bralnega gradiva in Tudi t. i. digitalna generacija, ki je odraščala v drugačnih okoliščinah in je

ga mestoma primerjali s (klasičnim) tiskanim. Stroka pa poleg teh dveh med vrste bolj vajena branja z računalniškega zaslona, zato ima tudi najbrž temu primerno

bralnih gradiv uvršča še zvočne in interaktivne (za več glej razvojni projekt Bralna razvite bralne strategije [20]. pismenost in razvoj slovenščine – OBJEM (Ozaveščanje, Branje, Jezik, Evalvacije,

Modeli), ki je potekal pod okriljem Zavoda Republike Slovenije za šolstvo).



spreminjajo oblike sporazumevanja/komunikacije, jezik, ki ga knjige, ki v zadnjih letih (zlasti po epidemiji covida) vedno bolj

pri tem uporabljamo, značilnosti pisanja in podobno, nenazadnje prodirajo v vsakdanje življenje in šolsko prakso, zato je prav, da

pa tudi naše zmožnosti sprejemanja besedil različnih vrst, kamor nekaj mesta namenimo tudi razmejitvi teh dveh pojmovanj.

spada branje.

2.1 Opredelitev e-knjige

1.3 Posebnosti branja elektronskih besedil Slovar slovenskega knjižnega jezika (dostopen na fran.si v

Digitalizacija je spremenila naš vsakdan in močno posegla nadaljevanju SSKJ) opredeljuje e-knjigo oz. elektronsko knjigo

tudi v šolstvo. Ne le da je omogočila hiter dostop do velike kot knjigo v elektronski obliki, namenjeno zlasti za branje na e-

količine informacij, spremenila je tudi naš način razmišljanja ter bralniku ali drugih elektronskih napravah. Kot sopomenko e-

branja in posledično učenja, če ostajamo na predpostavki, da je knjigi najdemo tudi pojem digitalna knjiga (prim. [24]). Tukaj

branje (še vedno) najučinkovitejši način usvajanja znanja, čeprav najdemo še nekoliko obširnejšo opredelitev, saj gre za

se učenci več ne učijo samo s papirja, ampak tudi z zaslona.3 enciklopedičen članek, in sicer da je to knjižna publikacija,

Da bi lahko razumeli posebnosti branja elektronskih besedil izdana v digitalni obliki. Lahko je izdana samostojno ali kot

in jih upoštevali v svoji praksi, je potrebno poznati značilnosti različica tiskane knjige. Namesto v fizični obliki do nje

digitalnih virov. Ti imajo dve temeljni značilnosti: dostopamo s pomočjo računalniške naprave in se izpisuje na

zaslonu (za primerjavo glej [6]).

(1) nematerialnost oz. neoprijemljivost in E-knjiga je torej prvenstveno namenjena branju (in

(2) večzaznavno naravo interakcij.

posledično v šolski praksi bralnemu razumevanju).

Učenec torej pri uporabi digitalnih virov nima konkretnih oz.



oprijemljivih predmetov za manipulacijo (prijemanje, 2.2 Opredelitev zvočne knjige premikanje …), poleg tega pa je v digitalnem okolju prekinjena

prostorsko-časovna povezanost med aktivnostjo učenca in V SSKJ ne najdemo zadetka pod iztočnico “zvočna”, ko

njegovim želimo najti definicijo zvočne knjige, prav tako ne pod iztočnico zaznavanjem. Digitalna/elektronska besedila

zmanjšujejo eno ključnih navigacijskih sposobnosti bralca, tj. “knjiga”. Ponujena nam je razlaga pod iztočnico “avdioknjiga”

nadzor nad branjem, ki je ena od osrednjih regulacijskih in je v SSKJ opredeljena kot “obsežnejše literarno besedilo, ki je

sposobnosti [19]. posneto na nosilec zvoka, navadno na CD-ploščo”. Nadalje jo

A. Kepic Mohar [1] (po Kovač in van der Well 2018) najdemo še pod iztočnico “zvočnica”, ki jo SSKJ opredeljuje kot:

povzema tri načine branja, ki so posledica digitalizacije in ki so (1) “zvočni posnetek zlasti literarnega besedila, ki ga

povezani z dolžino besedil ter medijem, na katerem beremo: navadno berejo profesionalni bralci, igralci” oz. kot

(1) (2) “knjig[o] za otroke, ki omogoča zaznavanje branje s preletom (angl. skimming );

(2) predstavljene vsebine (tudi) s sluhom”. vživeto branje (angl. immersive/absorbed reading ), ki

poteka pri branju žanrskega leposlovja; Iz slovarskega sestavka izvemo še, da je pojem “zvočnica”

(3) nastal iz “zvočne knjige” po zgledu angleškega izraza “audio poglobljeno/globoko branje (angl. in-depth/deep

reading book” in da se zanj uporabljata še sinonimna pojma ), ki pa je kognitivno zahtevnejše oz. študijsko

branje in vključuje predznanje.4 “avdioknjiga” in “zvočna knjiga”.5

Da je e-branje drugačno od branja s papirja, je pokazala tudi Zvočna knjiga/zvočnica je torej knjiga, namenjena poslušanju

mednarodna raziskava (posledično v šolski praksi imenovanemu slušnemu Spretnosti odraslih (PIAAC), kakor

povzema njene rezultate P. Javrh [15]. Že sama narava razumevanju).

elektronskih besedil narekuje “drugačno” branje, je potrdila

raziskava. Poleg tega pa je za učinkovito obdelavo informacij v 2.3 Kaj pa (e-)bralnik?

digitalnem okolju predpogoj dobro usvojena bralna E-bralnik (pogosto tudi elektronski papir ali e-papir) je

tehnika/veščina. Ta je namreč potrebna, da lahko pri e-branju in naprava, zelo podobna tabličnemu računalniku, namenjena

obdelavi e-gradiv stečejo zahtevnejši kognitivni procesi. branju e-knjig. Zanj je značilen zaslon e-ink oz. zaslon s črnilom.

“Zaslon s črnilom je vrsta zaslonske tehnologije, ki natančno

posnema videz knjige. E-ink zaslon je tudi zelo varčen in manj

2 E-KNJIGA IN ZVOČNA KNJIGA: ISTA obremenjuje oči.” [10] (več o tem tudi [11])

STVAR, RAZLIČNO POIMENOVANJE?



različna izraza za isto stvar, kar vnaša marsikatero zmedo in 3 APLIKACIJE, NAMENJENE BRANJU E- V šolski praksi se pogosto dogaja, da se uporabljata dva

nejasnost. Tudi kar zadeva e-branje (ali elektronsko branje), ki se KNJIG

odvija v e-knjigah (ali elektronskih izdajah knjig), se dogajajo V nadaljevanju navajamo pregled aplikacij, namenjenih

(napačne) interference z drugim pojmom. Gre za t. i. zvočne branju e-knjig. Pri tem se osredotočamo na e-knjige v



3 Za lažje razumevanje posebnosti e-branja (tudi digitalnega branja, kakor se življenj, tj. povsod navzoča prisotnost tehnologije« (62) Nadalje avtorica povzema

pogosto imenuje v slovenistični stroki, primer [2]), na kratko po S. Pečjak [19] rezultate raziskav, ki so pokazale, da „digitalne” generacije nimajo odpora do branja

povzemamo nekaj značilnosti generacij, ki so v času nastajanja prispevka vključene z zaslona, a lažje razumejo slišano/prebrano in prikličejo bistvo prebranega, če

v sistem formalnega izobraževanja. Gre za dve generaciji, in sicer za (1) generacijo poslušajo oz. preberejo na papirju kot v e-obliki. Z (rojeni v letih 1995–2010), imenovano tudi digitalni pionirji, in (2) generacijo 4 Na podlagi nove informacije si širimo besedni zaklad in s kognitivnimi alfa (rojeni po letu 2010), ki jo imenujejo tudi digitalni domorodci. »Za obe procesi gradimo znanje [1] (po Kovač in van der Well 2018).

generaciji je značilno, da sta v svojem življenju doživeli največ družbenih 5 Za več glej [12]. sprememb od vseh generacij pred njima, najbolj pa ju je zaznamovala digitalizacija

slovenskem jeziku, saj želimo osvetliti priložnosti in ovire pri zmožnost literarnega branja kot specifično podvrsto

njihovi uporabi pri pouku slovenščine 6 (kot prvega oz. sporazumevalne zmožnosti ter se usposabljajo za branje in

maternega jezika): interpretacijo literarnih besedil [22] (7).

Eden od gradnikov bralne pismenosti je motiviranost za

(1) KOBI je aplikacija, namenjena začetnemu branju, tj.

branje, ki zajema interes za branje, pozitiven odnos do branja

učenju in treningu branja (bralne tehnike). Uporaba

različnih vrst besedil (za različne namene, in sicer branje za

aplikacije je brezplačna. Ker je namenjena učenju branja znanje in branje leposlovja oz. razvojanje bralne kulture) ter

in je torej primerna za prvo triletje osnovne šole, je

izbiranje tiskanih in drugih informacijskih virov glede na namen

podrobneje ne bomo obravnavali. branja v povezavi z uporabo različnih lokacij bralnih virov. Drug

(2) Biblos, spletna platforma, je prva slovenska e-knjižnica (za naš prispevek relevanten) gradnik je še tekoče branje, ki med

(in e-knjigarna), nastala leta 2013 pod okriljem drugim zajema tehniko branja (in njeno prilagajanje zahtevam

Beletrine, zavoda za za založniško dejavnost, in s

besedila).

sodelovanjem mreže slovenskih knjižnic ter inštitutom

IZUM, skrbnikom sistema COBISS [13]. Prvenstveno je 3.1.2 Zbirka e-knjig v slovenščini

bila namenjena izposoji e-knjig, kasneje tudi njihovemu Razpoložljiva baza e-knjig v slovenskem jeziku zajema zlasti

nakupu. Za uporabnike, člane splošnih knjižnic, je bila literarna/umetnostna besedila, kot je razvidno s Slike 18 in ki jih

izposoja brezplačna. S 1. 6. 2024 jo je za člane splošnih obravnavamo pri pouku književnosti. Ker gre za branje literarnih

knjižnic nadomestila platforma COBISS Ela (več o njej besedil, to vrsto branja v slovenistični stroke imenujemo

v nadaljevanju). Beletrina ima od takrat naprej novo literarno branje9 (prim. [5, 3]).

platformo, Beletrina Digital, ki ob plačilu naročnine



nudi izposojo in nakup e-knjig, zvočnih knjig, filmov,

podkastov in prispevkov.

(3) sBiblos je Beletrinina spletna platforma, namenjena

izposoji knjig za učence zadnjega triletja osnovne šole in

srednješolce. Uporaba je zanje brezplačna, prijava je

mogoča z AAI računom.

(4) Mladinska knjiga PLUS je aplikacija založbe Mladinska

knjiga, ki ob plačilu izbranega paketa nudi različne

digitalne vsebine (e-knjige, zvočnice, animirane

pravljice in podkaste ter drugo, glede na izbran paket).

(5) COBISS Ela je platforma, ki nudi članom splošnih

knjižnic ob namestitvi ustrezne programske opreme

izposojo e-knjig.



3.1 Za šolsko prakso uporabne aplikacije

Pregled aplikacij za e-branje, ki so uporabne v šolski praksi, Slika 1: Tipologija e-knjig v COBISS Eli (zajem zaslona)

bomo naredili na podlagi temeljnih ciljev pouka književnosti, ki

jih opredeljuje učni načrt za osnovno šolo v zadnjem triletju (7.– Tudi baza e-knjig, ki so brezplačno na voljo za učence in

9. razred) in učni načrt za gimnazijo.7 Kot podlago smo vzeli še dijake v platformi sBiblos zajema literarna besedila, kot je

gradnike bralne pismenosti projekta OBJEM [8], ki se nanašajo razvidno s Slike 2.

na branje in bralno motivacijo.





3.1.1 Cilji pouka slovenščine (kot prvega oz. maternega

jezika)

Naj na kratko zaradi lažjega razumevanja in za argumentacijo

izbora povzamemo temeljne cilje pouka slovenščine, ki se

nanašajo (tudi) na (motivacijo za) branje.

Učenci in dijaki pri pouku slovenščine razvijajo

sporazumevalno zmožnost v slovenskem (knjižnem) jeziku, torej

zmožnost kritičnega sprejemanja (in tvorjenja) besedil različnih

vrst. Branje (zlasti v okviru pouka književnosti) prepoznavajo

kot užitek, prijetno doživetje in intelektualni izziv. “Pri

razvijanju sporazumevalne zmožnosti ob dejavnem stiku z

neumetnostnimi in umetnostnimi besedili z uporabo digitalne

tehnologije varno, ustvarjalno in kritično pridobivajo ter Slika 2: Tipologija e-knjig v sBiblosu (zajem zaslona)

uporabljajo podatke/informacije.” [21] (7) V gimnaziji razvijajo



6 8 Osredotočili se bomo na pouk slovenščine od 6. do 9. razreda osnovne šole in Podatke sicer navajamo za platformo COBISS Ela, vendar podobno velja tudi

pouk slovenščine v srednješolskem izobraževanju. za Biblos v začetku leta 2017 [13]: razmerje gre v prid leposlovja, in sicer je

7 Na ostale programe srednješolskega izobraževanja se nismo posebej predstavljajo 71 % vseh razpoložljivih e-knjig, strokovna literatura pa 29 %.

osredotočali, saj gre v večini za nadaljevanje in manjšo nadgradnjo osnovnošolskih 9 V stroki pojmujemo literarno branje kot najbolj kompleksno branje, ki nudi ciljev. največ možnosti oblikovanja vsestranskega bralca [5, 3].



E-knjige v slovenskem jeziku najdemo tudi na slovenskem 3.2.3 Čemu v šolsko prakso vključiti e-knjige?

Wikiviru [23] (in drugih spletnih mestih), kjer so digitalizirana Nanizali bomo zgolj nekaj argumentov, zakaj v šolsko prakso

dela slovenskih avtorjev različnih obdobij slovenske književnosti (občasno) vključiti e-knjige.

(prim. [14, 25]). Ker gre za postavitev literranih besedil na splet, Prvič se nam zdi takšen način zelo pragmatičen, saj

do katerih ne dostopamo preko aplikacije, jih v našem prispevku potrebujemo malo opreme, zadostuje že kakšen (pametni) telefon

ne bomo podrobneje obravnavali. ali tablični računalnik in lahko manjša skupina (predlagamo do

4) učencev/dijakov bere skupaj eno e-knjigo. Na ta način

3.2 Možnost uporabe izbranih aplikacij pri pokrijemo tako sodelovalno učenje, razvijanje digitalnih

pouku kompetenc učencev in še hkrati jih motiviramo za branje ter

Pri predstavitvi možnosti uporabe aplikacij za branje e-knjig pogovarjanje o prebranem (izmenjavo mnenj, različnih

smo se odločili za vključitev tistih, ki nudijo brezplačno uporabo interpretacij prebranega, razjasnjevanje manj jasnih mest ipd.).

za učence in dijake, saj na ta način omogočimo vsem učencem Na ta način lahko pokrijemo tudi skupno daljše branje izbranih

enake možnosti dostopa. To sta dve, in sicer COBISS Ela in literarnih del po delih z učencem zanimivo/privlačno tematiko.

sBiblos. Podoben način dela lahko uporabimo tudi za (obvezno)

domače branje, ki nemalokrat učencev ne pritegne in se ga

3.2.1 COBISS Ela izogibajo ali pa površno preberejo (pogosto zgolj obnove na

Platforma COBISS Ela na spletni strani [7] nudi pregledna spletu). Digitalno tehnologijo lahko v tem primeru uporabimo

navodila, katere naprave in aplikacije za branje podpirajo tudi za refleksijo o prebranem in analizo prebranega, in sicer v

izposojo in branje e-knjig. Kot prednost platforme gre izpostaviti, obliki e-dnevnika, lastne e-knjige, e-diagrama ali e-table. Ob tem

da lahko e-knjige beremo na mobilnih napravah, ki temeljijo na je potrebno učencem le še dati nekaj usmeritev/navodil, kako se

Androidu in tudi na sistemu iOS. Zraven je navedeno tudi ime lotiti dela.

aplikacije, ki si jo je potrebno namestiti. Navodilom so dodani Vključevanje e-knjig v šolsko prakso pa se vidimo kot

tudi e-vodiči za posamezne korake. koristno tudi za možnost večjega poudarka na lastni izbiri

E-knjige lahko beremo tudi na osebnih računalnikih ter literarnega dela v obliki e-knjige za branje, saj lahko učenec

bralnikih e-knjig in bralnikih inkBOOK. Preprost je tudi način zlahka dostopa do druge e-knjige, če ga tematika ene ne pritegne.

iskanja želene e-knjige, saj je brskalnik enak iskalniku tiskanih To zagotovo deluje zelo motivacijsko pri današnjih generacijah,

knjig v knjižnicah. To je zagotovo velika prednost, saj ni ki so vešče dela z digitalnimi viri.

potrebnega “novega učenja” iskanja želene e-knjige.

V bazi e-knjig najdemo večino sodobnih slovenskih SKLEP

mladinskih romanov, ki najbolj motivirajo mlade bralce, saj jih Digitalizacija v šolstvu je pogosto videna kot nekaj

nagovarjajo z njim aktualnimi vsebinami/temami. E-knjigo si negativnega, vendar je potrebno pristopiti v smeri iskanja

lahko izposodimo na dveh različnih napravah, na voljo pa nam je priložnosti in pozitivnih vidikov. Šole in učiteljev računalniki (in

21 dni od dneva izposoje. Naenkrat si lahko uporabnik izposodi naraščajoča uporaba umetne ingteligence) ne bodo mogli

5 e-knjig. To je lahko uporabno za učence, saj lahko e-knjige nadomestiti, lahko pa učitelji tehnologijo izkoristimo sebi in

berejo na dveh različnih napravah, na eni npr. v šoli in na eni namenom pouka v prid. Glede na to, da so naši učenci in dijaki

doma. Prav tako lahko hkrati berejo več različnih knjig, jih med “digitalin pionirji” in “digitalin domorodci”, je potrebno

seboj primerjajo (iščejo podobnosti in razlike, analizirajo didaktične pristope prilagajati njim in značilnostim njihovega

posamezne motive ipd.). učenja. S tem vsekakor ne mislimo na popolno opustitev

3.2.2 sBiblos “tradicionalnih” pristopov in načinov dela, ampak tehnologijo

Beletrinin sBiblos prav tako deluje na mobilnih napravah z izkoristiti za motivacijo tistih področij/dejavnosti, ki jih

operacijskima sistemoma Android in iOS. Na spletni strani [9] zaznavamo kot šibkejše oz. manj učinkovite.

so podrobna navodila, kako po korakih namestiti potrebne V prispevku smo v skladu s tem namenom želeli prikazati,

aplikacije za dostop do e-knjig. Zelo uporabni so tudi e-vodiči in kako en segment pri pouku slovenščine, tj. motivacija za branje

slikovno gradivo, ki pomaga pri nameščanju in uporabi bralnika. (morda tudi domače ali vsaj skupno dolgo branje v razredu)

E-knjige lahko uporabniki berejo na osebnih računalnikih, skušati razvijati skladno z značilnostmi šolajočih generacij, da bo

mobilnih napravah in e-bralnikih. čim učinkoviteje in da bodo uspeli uresničiti (vsaj del) ciljev

Do storitev sBiblos lahko učenci dostopajo z AAI pouka.

uporabniškimi podatki, v kolikor ima šola urejen dostop do

storitve Moja Knjižnica in podpisano pogodbo o obdelavi VIRI IN LITERATURA

osebnih podatkov z IZUM-om. [1] Alenka Kepic Mohar. 2021. Nevidna moč knjig:

V zbirki e-knjig najdemo večino literarnih del, ki se branje in učenje v digitalni dobi. Zbirka Bralna

obravnavajo v šoli in tudi novejša dela z učencem in dijakom znamenja. Cankarjeva založba, Ljubljana.

privlačnimi temami. Tudi sBiblosovi zbirki pedstavljajo navečji [2] Alenka Žbogar. 2022. Branje kot potovanje. Branje v

del literarnega opusa romani in kratka proza. Učenci in dijaki si slovenskem jeziku, literaturi in kulturi. 58. seminar

lahko naenkrat izposodijo 3 e-knjige, na mesec 10 e-knjig, vsaka slovenskega jezika, literature in kulture. Založba

knjiga pa jim je na voljo 30 dni od dneva izposoje. V primerjavi Univerze v Ljubljani, Ljubljana. 5–6. DOI:

s COBISS Elo je čas izposoje e-knjig uporabnikom prijaznejši, https://doi.org/10.4312/SSJLK.58.2386-058X.

je pa na voljo manj hkrati izposojenih knjig. [3] Alojzija Zupan Sosič. 2014. Literarno branje. Jezik in

slovstvo 59, 4. 47–65.



[4] Anne Mangen idr. 2013. Reading linear texts on [19] Sonja Pečjak. 2022. Psihološka perspektiva učinkov

paper versus computer screen: Effects on reading digitalizacije na procese branja in učenja pri ljudeh.

comprehension. International Journal od Educational Branje v slovenskem jeziku, literaturi in kulturi. 58.

Research 58. 61–68. DOI: seminar slovenskega jezika, literature in kulture.

https://doi.org/10.1016/j.ijer.2012.12.002. Založba Univerze v Ljubljani, Ljubljana. 61–68. DOI:

[5] Boža Krakar Vogel. 2020. Didaktika književnosti pri https://doi.org/10.4312/SSJLK.58.2386-058X.

pouku slovenščine. Zbirka Učiteljeva orodja. Založba [20] Tjaša Pucko in Sonja Pečjak. 2016. Primerjava

Rokus Klett, Ljubljana. učinkovitosti e-branja in branja s papirja pri učencih v

[6] Britannica: e-book. DOI: srednji šoli. Pedagoška obzorja 31, 3–4, 55–69. DOI:

https://www.britannica.com/technology/e-book. https://www.dlib.si/listalnik/URN_NBN_SI_doc-

[7] COBISS Ela. DOI: 4Q2O9TK8/index.html.

https://www.cobiss.si/bralci/ela/#2. [21] Učni načrt. Program osnovna šola. Slovenščina.

[8] Gradniki bralne pismenosti: teoretična izhodišča. 2011. Ministrstvo za šolstvo in šport, Zavod

2020. Univerzitetna založba Univerze v Mariboru, Republike Slovenije za šolstvo, Ljubljana. DOI:

Maribor. DOI: https://www.gov.si/assets/ministrstva/MVI/Dokument

https://press.um.si/index.php/ump/catalog/book/515. i/Osnovna-sola/Ucni-

[9] Kako deluje sBiblos? DOI: nacrti/obvezni/UN_slovenscina.pdf.

https://www.sbiblos.si/kako-deluje. [22] Učni načrt. Slovenščina. Gimnazija: splošna,

[10] Kindle ali Kobo? Kateri e-bralnik izbrati? klasična, strokovna gimnazija. 2008. Ministrstvo za

DOI: https://racunalniske-novice.com/e-bralniki- šolstvo in šport, Zavod Republike Slovenije za

amazon-kindle-e-bralniki-kobo-e-knjige/. šolstvo, Ljubljana. DOI:

[11] Marjan Kodelja. 2023. Zakaj in kako izbrati e- https://eportal.mss.edus.si/msswww/programi2018/pr

bralnik? DOI: ogrami/media/pdf/un_gimnazija/un_slovenscina_gim

https://vsebovredu.triglav.si/tehnologija/zakaj-in- n.pdf.

kako-izbrati-e-bralnik. [23] Wikivir. DOI:

[12] Marjan Povše. 2022. Zvočne knjige v slovenskem https://sl.wikisource.org/wiki/Glavna_stran.

jeziku. DOI: https://blog.cobiss.si/2022/07/04/zvocne- [24] Wikipedija, prosta enciklopedija: elektronska knjiga.

knjige-v-slovenskem-jeziku/. DOI:

[13] Matic Razpet. 2017. Biblos. E-gradiva kot bližnjica https://sl.wikipedia.org/wiki/Elektronska_knjiga.

do uspeha? E-zbornik Bralnega društva Slovenije od [25] Zbirka slovenskih leposlovnih besedil. DOI:

12. strokovnem posvetovanju 8. septembra 2017. https://lit.ijs.si/leposl.html.

Bralno društvo Slovenije, Ljubljana. 71–76. DOI:

https://www.bralno-drustvo.si/wp-

content/uploads/2013/06/E-

GRADIVA_KOT_BLIZNJICA_DO_USPEHA_BDS

20171.pdf.

[14] Miran Hladnik. 2009. Za elektronsko knjigo. Šolska

knjižnica 19, 2–3. 126–133. DOI:

https://lit.ijs.si/zaeknjigo.html.

[15] Petra Javrh. 2017. V digitalnem svetu beremo

drugače. E-gradiva kot bližnjica do uspeha? E-

zbornik Bralnega društva Slovenije od 12.

strokovnem posvetovanju 8. septembra 2017. Bralno

društvo Slovenije, Ljubljana. 43–48. DOI:

https://www.bralno-drustvo.si/wp-

content/uploads/2013/06/E-

GRADIVA_KOT_BLIZNJICA_DO_USPEHA_BDS

20171.pdf.

[16] Simona Tancig. 2016. Od Prousta do Twitterja –

nevroedukacijske raziskave bralne pismenosti v

digitalni dobi. Bralna pismenost kot izziv in

odgovornost. Pedagoška fakulteta Univerze v

Ljubljani, Ljubljana, 9–26. DOI: https://www.pef.uni-

lj.si/wp-content/uploads/2022/07/Bralna-

pismenost_Posvet-PeF-2016.pdf.

[17] Slovar slovenskega knjižnega jezika. DOI:

https://fran.si/.

[18] Sonja Pečjak in Ana Gradišar. 2002. Bralne učne

strategije. Zavod Republike Slovenije za šolstvo,

Ljubljana.



Zlorabe pri starejših osebah1



Tanja Repič Slavič



Teološka fakulteta, Poljanska 4, Ljubljana, katedra za Zakonsko in družinsko terapijo



ter psihologijo in sociologijo religije



Frančiškanski družinski inštitut, Prešernov trg 4, Ljubljana



e-mail: tanja.repic@teof.uni-lj.si



Abstract / Povzetek UVOD

Zlorabe starejših predstavljajo resen in pogosto prikrit

Staranje prebivalstva je globalni fenomen, ki postavlja

družbeni problem, ki ima lahko hude posledice za žrtve.

nove izzive, zlasti v smislu varnosti in zaščite starejših

Gre za vsakršno obliko fizične, psihične, finančne in

oseb. Po podatkih Svetovne zdravstvene organizacije

spolne zlorabe ter zanemarjanja, ki se dogaja v različnih

(WHO) [1] se predvideva, da se bo svetovno prebivalstvo,

okoljih, vključno z domovi za ostarele, družinami in

staro 60 let in več, več kot podvojilo, z 900 milijonov leta

skupnostmi. Starejše osebe postajajo z leti in glede na

2015 na približno 2 milijardi leta 2050, saj se v številnih

zdravstvene težave vedno bolj odvisne od drugih, imajo

državah prebivalstvo hitro stara. Število „najstarejših“

manjšo možnost fizične obrambe in so zato bolj

odraslih, starih 80 let ali več, narašča hitreje kot število

izpostavljene in ranljive. V teoretičnem prispevku želimo

starejših odraslih. Do leta 2050 se bo na primer število

na podlagi pregleda raziskav predstaviti posamezne vrste

„najstarejših“ več kot potrojilo, in sicer s 125 milijonov

zlorab in njihovo prevalenco, dejavnike tveganja, izzive

leta 2015 na 434 milijonov [2]. Poleg tega imajo ženske v

pri zaznavanju, posledice in strategije za preprečevanje.

povprečju daljšo pričakovano življenjsko dobo kot moški,



Keywords / Ključne besede zato predstavljajo 61,6 % oseb, starejših od 80 let. Ženske

in „najstarejši“ starejši bodo verjetno tudi v prihodnosti

zlorabe starejših, prevalenca, dejavniki tveganja, ostali največja starostna skupina v ustanovah za

posledice, preprečevanje. dolgotrajno oskrbo [2]. Starejši trenutno predstavljajo tudi

največji delež odraslega prebivalstva, ki živi v ustanovah

Abstract

za odrasle z motnjami v duševnem razvoju [3]. Prebivalci

Elder abuse is a serious and often hidden social problem teh ustanov imajo pogosteje več oblik motenj, vključno z

that can have serious consequences for victims. It includes duševnimi, telesnimi ali vedenjskimi motnjami ter

any form of physical, psychological, financial, or sexual invalidnostjo. Zato so prebivalci v institucionalnih

abuse and neglect that takes place in a variety of settings, ustanovah zaradi svoje šibkosti običajno bolj odvisni od

including nursing homes, families, and communities. oskrbe drugih in so lahko bolj izpostavljeni nevarnosti

Older people become increasingly dependent on others zlorabe in zanemarjanja kot starejši v drugih oblikah

because of their age and health problems, they are less able skupnosti [4]. Razširjenost zlorab je lahko veliko večja,

to physically protect themselves, and are therefore more kot se poroča, saj je lahko poročanje povezano z

exposed and vulnerable. In this theoretical paper, we aim nezmožnostjo poročanja pri žrtvah ali s strahom pred

to review the research done on the topic and present the posledicami in povračilnimi ukrepi [5]. different types of abuse as well astheir prevalence, risk

factors, challenges in detection, consequences, and

prevention strategies. Približno 1 od 6 oseb, starih 60 let in več, je v zadnjem

letu v skupnosti doživela eno od oblik zlorabe [1].

Keywords / Ključne besede Raziskava, opravljena v Združenih državah Amerike na

elder abuse, prevalence, risk factors, consequences, 897 starejših osebah, ki so se med pandemijo zadrževale

prevention. doma, kaže, da se je med pandemijo COVID-19 stopnja

zlorab starejših še povečala O njej je poročala ena od petih



1 Permission to make digital or hard copies of part or all of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for third-party components of this work must be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.dem.4

starejših oseb (21,3 %), kar je 83,6-odstotno povečanje v zloraba, finančno zlorabo pa lahko spremlja zanemarjanje

primerjavi z ocenami razširjenosti pred pandemijo. Pred ali fizična zloraba [14]. Fizična zloraba vključuje namerne

pandemijo COVID-19 je zloraba starejših vsako leto telesne poškodbe starejših oseb, kot so udarci, potiski ali

prizadela enega od 10 ameriških starejših odraslih. druga škodljiva dejanja. Pregled fizičnih poškodb pri

Predpostavlja se, da je pandemija s seboj prinesla porast zlorabi starejših je pokazal, da se dve tretjini poškodb

zlorabe starejših zaradi posameznikov, ki jim je bilo pojavita na zgornjih okončinah in maksilofacialnem

odrejeno, da ostanejo doma, skupaj s povečanimi področju [15]. Nekatere raziskave kažejo, da je verjetnost

medosebnimi stresorji. reviktimizacije večja pri ženskah, ki so ovdovele, imajo

diagnozo demence in so se vrnile na dom, kjer je storilec

živel ali jih obiskoval [16]. Verjetnost prijave je v takih



1 OBLIKE ZLORAB PRI STAREJŠIH odvisnost od tistega, ki skrbi za žrtev, ki naj bi bila primerih zelo redka, saj je problem dokazovanje in

Zloraba starejših je enkratno ali ponavljajoče se dejanje ali istočasno hvaležna za skrb, saj sicer ne bi preživela, po

pomanjkanje ustreznega ukrepanja, ki se zgodi v katerem drugi strani pa je deležna nasilja in zlorab in ujeta v

koli odnosu, v katerem se pričakuje zaupanje, in ki starejši brezizhodni situaciji. osebi povzroči škodo ali stisko. Ta vrsta nasilja pomeni

kršitev človekovih pravic in vključuje fizično, spolno,

psihološko in čustveno zlorabo, finančno in materialno 1.2 Psihološka zloraba



zlorabo, zapuščanje, zanemarjanje ter resno izgubo Psihološka ali čustvena zloraba vključuje uporabo dostojanstva in spoštovanja [1]. Poleg delitve na vrsto ali groženj, ponižanja, ustrahovanja, preklinjanja in drugega obliko v splošnem, lahko zlorabe razvrstimo tudi na to, verbalnega vedenja in (ali) katere koli druge oblike kdo zlorabo izvaja - družinski člani, neformalni in duševne krutosti, ki povzroči duševno ali fizično stisko. formalni skrbnik ali znanec; ali okolje, v katerem se zgodi Študija, ki je zajela pregled 49 raziskav, je ugotovila - v skupnosti in instituciji [7]. V institucionalnem okolju razširjenost psihične zlorabe med starejšimi v Evropi, ki pa lahko zlorabo na splošno razvrstimo na zlorabo med znaša od 10,4 % do 29,7 %, pri čemer je višja v severnih stanovalci ali zlorabo osebja med stanovalci [4]. državah (Nemčija, Litva, Švedska) v primerjavi z južnimi

Zloraba starejših je pomemben javnozdravstveni problem. državami (Portugalska, Italija, Španija in Grčija) [22].

V pregledu 52 študij iz 28 držav iz različnih regij iz leta Ugotovitve so presenetljive, saj bi pričakovali, da bo na

2017 je bilo ocenjeno, da je bila v zadnjem letu vsaka šesta severu nižja tolerance do nasilja in boljši življenjski

oseba (15,7 %), stara 60 let in več, izpostavljena neki standard. Lako pa je prav to razlog, da je povečana

obliki zlorabe [8], kar pomeni približno 141 milijonov senzitivnost za vsako vrsto zlorabe in je zato poročanje o

ljudi. Čeprav so natančni podatki omejeni, pregled vsebuje pojavnosti višje v primerjavi z južnimi državami.

ocene razširjenosti deleža starejših ljudi, ki so bili deležni Dejavniki, ki so bili povezani z visokim tveganjem za

različnih vrst zlorabe. Meta analiza je pokazala, da je o psihično zlorabo, so bili: nizka socialna podpora, življenje

psihološki zlorabi poročalo 11.6% starejših ljudi, o fizični v najemniških stanovanjih, uživanje alkohola, pogosta

zlorabi 2.6%, o finančni 6.8%, o spolni zlorabi 0.9% in o uporaba zdravstvenih storitev ter pogosta anksioznost in

zanemarjanju 4.2% starejših oseb. somatske težave.



Podatki o obsegu problema v ustanovah, kot so bolnišnice, 1.3 Finančna zloraba

domovi za ostarele in druge ustanove za dolgotrajno Finančna zloraba vključuje izkoriščanje starejših oseb za

oskrbo, so pomanjkljivi. Vendar pa pregled nedavnih finančno korist, kot so goljufije, kraje ali prisiljevanje v

študij o zlorabi starejših v institucionalnih ustanovah kaže, nepoštene finančne transakcije. Raziskave kažejo, da so

da je 64,2 % zaposlenih samoporočalo, da je v zadnjem starejši pogosto tarča finančne zlorabe, zlasti če so odvisni

letu storilo neko obliko zlorabe [9]. od drugih za upravljanje svojih financ. Čeprav se slabo



državah Amerike je pokazala, da je 40 % osebja priznalo s precej večjo verjetnostjo depresije, posttravmatske psihično in 10 % fizično zlorabo v zadnjem letu [10]. stresne motnje in splošno anksioznostjo. Pri tem je Sistematični pregled zlorab v institucijah je pokazal, da se pomembno poudariti, da se tveganje za negativne Raziskava osebja v domovih za starejše v Združenih kontekstu čustvenih in zdravstvenih posledic, je povezano ravnanje s finančnimi sredstvi pogosto ne obravnava v



fizična zloraba pogosto pojavlja kot oblika maščevanja posledice še toliko bolj poveča, če so storilci finančne osebja nad fizično agresivnimi stanovalci [11]. Prav tako zlorabe družinski člani ali prijatelji [18]. je osebje poročalo, da je bolj verjetno, da bo agresivnim



stanovalcem odreklo izbiro česarkoli [12]. V drugi 1.4 Spolna zloraba nacionalni študiji v Združenih državah Amerike je 1,5 %

Spolna zloraba, opredeljena kot „kakršni koli spolni stiki

osebja samo poročalo, da je zagrešilo krajo [13].

brez privolitve“, naj bi bila najbolj skrita oblika zlorabe

starejših ter najmanj priznana in prijavljena vrsta slabega

ravnanja s starejšimi. Pregled je pokazal, da je bila večina

1.1 Fizična zloraba žrtev spolne zlorabe v domovih za ostarele kognitivno

Zloraba starejše osebe pogosto ni omejena le na eno oslabljenih (demenca, Alzheimerjeva bolezen, možganska

obliko; fizično nasilje na primer pogosto spremlja psihična kap in poškodbe možganov), s psihiatrično diagnozo in/ali



telesno šibkih (invalidski voziček, postelja, paraliziranost Nekateri so se znašli v kazenskih evidencah, nekaj pa jih

in omejena mobilnost) ter s somatskimi boleznimi. je bilo obsojenih.

Najstarejši bolniki (stari 79-99 let) so bili pogosteje

izpostavljeni spolni zlorabi. To bi lahko potrjevalo Ko gre za žrtve zlorab, so bile premeščene na drug oddelek

mnenje, da so v najstarejši starostni skupini poleg spola znotraj doma za ostarele, spremljale so jih ali pa sploh niso

pomembnejši tudi drugi dejavniki. To je poudarjeno tudi dobile pomoči. Zelo malo jih je bilo po napadu deležnih

v poročilu Svetovne zdravstvene organizacije, kjer je na zdravstvene in psihološke obravnave. Mnoge žrtve so bile

podlagi študij razširjenosti v skupnosti ugotovljeno, da so še naprej izpostavljene tveganju za nove zlorabe [19].



starejši moški v približno enakem deležu kot ženske Pogoste posledice spolnega nasilja za žrtve so težave z izpostavljeni zlorabi s strani zakoncev, odraslih otrok in duševnim zdravjem, vključno s samomorilnim vedenjem. drugih sorodnikov. Najpogostejši storilci so bili zaposleni Znane so smrtne žrtve, povezane s posilstvom. Podatki v domu za ostarele in sostanovalci. Drugi najpogostejši kažejo, da je 11 od 20 pacientov v domovih za starejše storilec je bil družinski član. Obiskovalci, ki so bili žrtvi umrlo v enem letu po spolnem napadu. Ni mogoče reči, ali znani ali neznani, so bili manj pogosti. V številnih so bile te smrti posledica napada ali ne, saj je bilo veliko primerih je imel storilec kognitivne motnje, psihiatrično žrtev starih in slabotnih. Kar zadeva prijavo zlorabe diagnozo, zlorabo snovi, kriminalno preteklost ali organom, sta bila najpogostejša razloga, zakaj primeri niso predhodno storjeno spolno zlorabo [14]. Študija Burgessa bili prijavljeni, pomanjkanje dokazov in zdravstveno s sodelavci [19] je pokazala, da so vse storilce vodili stanje žrtve [19]. različni motivi. Dva skupna imenovalca za vse storilce sta

bila, da so imeli nizke ocene na področju socialnih Zlorabe starejših, ne glede na vrsto, imajo lahko resne

kompetenc in da so iskali žrtve, ki so bile krhke in posledice za telesno in duševno zdravje, finančno in

nezaščitene. družbeno stanje, telesne ali psihične poškodbe,



Spolna zloraba starejših oseb presega tradicionalne meje pravic, depresijo, upad kognitivnih sposobnosti, finančno prezgodnjo umrljivost, izgubo ali kršitev človekovih

med spoloma, kulturami in vlogami žrtev in storilca.



Obstaja tveganje, da se spolne zlorabe starejših oseb ne kakovost življenja starejše osebe. Pri starejših ljudeh so opustošenje, namestitev v domove za ostarele in slabšo



obravnavajo resno, saj se jih zaradi starosti lahko lahko posledice zlorabe še posebej hude, okrevanje pa obravnava kot aseksualne. Zato mnogi menijo, da je malo lahko traja dlje [23, 24]. verjetno, da bi prišlo do spolne zlorabe. Zdravstveni

delavci se izogibajo spraševanju starejših oseb o spolnem



zdravju in nočejo verjeti obtožbam o spolni zlorabi ali se 3 Dejavniki tveganja nanje odzvati [20].

Raziskava [18], ki je preučevala dejavnike tveganja, ki



1.5 Zanemarjanje povečujejo možnost, da bo nekdo postal žrtev ali storilec

v primeru zlorab med starejšimi, je pokazala:

Zanemarjanje starejših ljudi je globalno vprašanje javnega

zdravja in človekovih pravic, ki ogroža zdravje in varnost - da so bile žrtve pogosteje ženske, starejših ljudi. Običajno se nanaša na zavračanje ali

nezmožnost zagotavljanja nege in zaščite na področjih - da so bile starejše belke bolj ogrožene v primerjavi z

hrane, vode, oblačil, higiene, zdravil, življenjskega okolja drugimi etničnimi skupinami, čeprav so nekatere študije

in varnostnih ukrepov. Lahko se kaže s strani drugih oseb, omejene na vzorce belcev,

ki naj bi skrbele za starejšo osebo ali kot

samozanemarjanje [21]. Nenamerno zanemarjanje se - da so bile starejše ženske z nižjim izobrazbenim statusom

večinoma pojavi zaradi malomarnosti, na primer, ko in nižjimi dohodki bolj izpostavljene tveganju

negovalec nima potrebnih veščin ali znanja, se ne zaveda viktimizacije na vseh področjih,



razpoložljivih vrst podpore ali ne more zagotoviti oskrbe, - da so starejši ljudje s telesnimi in/ali duševnimi ker je sam bolan. Običajne spremembe staranja in boleznimi ali odvisnostmi bolj izpostavljeni večjemu simptomi, povezani z boleznijo, se lahko prekrivajo z tveganju kot tisti brez bolezni, pri čemer se je pokazala znaki zanemarjanja in vodijo do premajhnega odkrivanja. največja viktimizacija v starostni skupini med 60 in 70 let. Zanemarjanje je lahko neposreden vzrok smrti ali prispeva

k smrtnemu izidu s poslabšanjem obstoječih stanj [22]. Dejavniki tveganja in značilnosti storilcev pa so pokazali:

- da je večina storilcev moških,



2 Posledice zlorab - da je v večini primerov storilec pogosto veliko mlajših

Burgers s sodelavci [19] piše v svoji študiji na temo od žrtve (več kot 30 let),



spolnih zlorab, da je bilo nekaj storilcev zlorab - da so storilci žrtvam običajno znani in kaznovanih in napadi niso imeli pravnih posledic, čeprav

so bile v več primerih zlorabe prisotne priče. V primerih, - da so nekateri med njimi bili že v preteklosti obsojeni

ko so bili bolniki storilci, so bili premeščeni na druge zaradi spolnih ali drugih kaznivih dejanj [18].

oddelke ali v druge domove za starejše. Kadar so bili

storilci člani osebja, so jim odpovedali delovno razmerje, Na ravni odnosov sta vrsta odnosa (npr. zakonec/partner

jim dali dopust ali jih premestili na drugo delovno mesto. ali otrok/starš) in zakonski stan lahko povezana s



povečanim tveganjem za zlorabo, vendar se ti dejavniki vplivi na njihovo vsakodnevno življenje. V tem kontekstu

razlikujejo glede na državo in regijo. Dejavniki na ravni je pomembno tudi spodbujanje medgeneracijskega

skupnosti in družbe, povezani z zlorabo starejših, lahko sodelovanja in solidarnosti, saj to dodatno okrepi socialno

vključujejo predsodke do starejših in nekatere kulturne povezanost ter pripomore k zmanjšanju diskriminacije na

norme (npr. normalizacija nasilja). Po drugi strani pa naj podlagi starosti. Na ta način lahko kot družba poskrbimo

bi bila socialna podpora in samsko življenje povezano z za boljše varstvo in spoštovanje pravic starejših, saj bomo

zmanjšanim tveganjem za zlorabo starejših [25]. nenazadnje prej ali slej vsi nekega dne stari.



References / Literatura

4 Strategije za preprečevanje ali preventiva [1]. World Health Organization. 2017. Elder Abuse Fact Sheet 357.

V nekaterih državah je zdravstveni sektor prevzel vodilno Geneva: World Health Organization.

[2] United Nations. 2017. Revision of World Population Prospects

vlogo pri opozarjanju javnosti na zlorabe starejših, v [Internet]. Available at: https://esa.un.org/unpd/wpp/

drugih pa je vodilno vlogo prevzel sektor socialnega (8 September 2024 sneto z interneta).

varstva. Na svetovni ravni je o zlorabi starejših in njenem [3] World Health Organization Regional Office for Europe. 2018.

Regional Report on Institutions Providing Long-term Care for

preprečevanju premalo znanega, zlasti v državah v Adults with Psychosocial and Intellectual Disabilities in the

razvoju. European Region. Copenhagen.

[4] McDonald, L., Beaulieu, M., Harbison, J., Hirst, S., Lowenstein,



Za preprečevanje in odzivanje na zlorabe starejših so bile A., Podnieks, E., and Wahl, J. 2012. Institutional abuse of older

adults: what we know, what we need to know. Journal of Elder

uporabljene številne strategije, vendar so dokazi o Abuse & Neglect, 24(2), 138-160.

učinkovitosti večine teh ukrepov trenutno omejeni. Med [5] World Health Organization. 2008. A Global Response to Elder

Abuse and Neglect: Building Primary Health Care Capacity to

najbolj obetavne strategije sodijo intervencije za Deal with the Problem Worldwide: Main Report. Geneva: WHO

oskrbovalce, ki nudijo storitve za razbremenitev bremena Press.

oskrbe, programi za upravljanje denarja za starejše [6] Chang, E. S., and Levy, B. R. 2021. High prevalence of elder abuse

during the COVID-19 pandemic: Risk and resilience factors. The

odrasle, ki so izpostavljeni finančnemu izkoriščanju, American Journal of Geriatric Psychiatry, 29(11), 1152-1159.

telefonske linije za pomoč in zavetišča za nujne primere [7] Gorbien, M. J., and Eisenstein, A. R. 2005. Elder abuse and

neglect: an overview. Clin Geriatr Med, 21, 279–92.

ter multidisciplinarne skupine, saj so potrebni odzivi [8] Yon, Y., Mikton, C. R., Gassoumis, Z. D., and Wilber, K. H. 2017.



pogosto povezani z več sistemi, vključno s kazenskim Elder abuse prevalence in community settings: a systematic review and meta-analysis. The Lancet Global Health , 5 (2), e147-

pravosodjem, zdravstvom, duševnim zdravjem, službami e156.

za zaščito odraslih in dolgotrajno oskrbo [25]. [9] Yon, Y., Ramiro-Gonzalez, M., Mikton, C. R., Huber, M., and

Sethi, D. 2019. The prevalence of elder abuse in institutional

Poleg tega raziskave [22] poudarjajo pomen podrobne settings: a systematic review and meta-analysis. European

journal of public health, 29(1), 58-67.

dokumentacije ter sodelovanja med kliničnimi zdravniki, [10] Pillemer, K., and Hudson, B. 1993. A model abuse prevention

sorodnimi zdravstvenimi delavci, organi pregona in program for nursing assistants. The Gerontologist, 33(1), 128-

medicinskimi forenzičnimi službami za izboljšanje 132.

[11] Lindbloom, E. J., Brandt, J., Hough, L. D., and Meadows, S. E.

rezultatov bolnikov in zmanjšanje tveganja za nadaljnje 2007. Elder mistreatment in the nursing home: A systematic

zlorab. review. Journal of the American medical directors

association, 8(9), 610-616.

[12] Meddaugh, D. I. 1993. Covert elder abuse in the nursing

home. Journal of Elder Abuse & Neglect, 5(3), 21-38.



Zaključek [13] Harris, D. K., and Benson, M. L. 1999. Theft in nursing homes: an overlooked form of elder abuse. Journal of Elder Abuse &

Zlorabe starejših so resen in kompleksen problem, ki Neglect, 11(3), 73-90.

zahteva večplastno obravnavo. Z naraščajočim številom [14] Malmedal, W., Iversen, M. H., and Kilvik, A. 2015. Sexual abuse

of older nursing home residents: A literature review. Nursing

starejših v svetovnem prebivalstvu je nujno povečati research and practice, 2015(1), 902515.

ozaveščenost o zlorabah ter izboljšati sisteme za [15] Murphy, K., Waa, S., Jaffer, H., Sauter, A., and Chan, A. 2013. A literature review of findings in physical elder abuse. Canadian

odkrivanje in preprečevanje teh dejanj. Zlorabe se Association of Radiologists' Journal, 64(1), 10-14.

pojavljajo v različnih oblikah, od fizičnih in psihičnih do [16] Mosqueda, L., Burnight, K., Gironda, M. W., Moore, A. A.,

finančnih in spolnih, pri čemer so žrtve pogosto odvisne Robinson, J., and Olsen, B. 2016. The abuse intervention model:

A pragmatic approach to intervention for elder

od svojih oskrbovalcev, kar povečuje njihovo ranljivost. mistreatment. Journal of the American Geriatrics Society, 64(9),

Potrebna je večja vključenost zdravstvenega in socialnega 1879-1883. [17] Macassa, G., Viitasara, E., Sundin, Ö., Barros, H., Torres sektorja, saj so prav ti sektorji ključni pri zaščiti starejših Gonzales, F., Ioannidi‐Kapolou, E., ... and Soares, J. J. 2013.

oseb. Prav tako je pomembno izboljšati usklajevanje med Psychological abuse among older persons in Europe: a cross‐

institucijami, da bi zagotovili celovitejši odziv na zlorabe. sectional study. Journal of aggression, conflict and peace

Razvoj programov za izobraževanje osebja v domovih za research, 5(1), 16-34. [18] Acierno, R., Watkins, J., Hernandez-Tejada, M. A., Muzzy, W.,

starejše ter skrbnikov na domu je ključen korak k Frook, G., Steedley, M., and Anetzberger, G. 2019. Mental health

zmanjšanju zlorab in izboljšanju obravnave starejših oseb. correlates of financial mistreatment in the National Elder

Mistreatment Study Wave II. Journal of Aging and Health, 31(7),

In nenazadnje, učenje mladih generacij z zgledom, kako 1196-1211.

lahko sodelujemo s starejšimi generacijami – mladost ima [19] Burgess, A. W., Prentky, R. A., and Dowdell, E. B. 2000. Sexual

več moči, starost pa več modrosti. Tako eno kot drugo je predators in nursing homes. Journal of psychosocial nursing and

mental health services, 38(8), 26-35.

potrebno, da lahko uspešno opravljamo svoje poslanstvo. [20] Connolly, M. T., Breckman, R., Callahan, J., Lachs, M., Ramsey-

Pomembno je tudi vzpostaviti mehanizme, ki omogočajo Klawsnik, H., and Solomon, J. 2012. The sexual revolution's last

frontier: how silence about sex undermines health, well-being,

starejšim osebam, da varno prijavijo morebitne zlorabe and safety in old age. Generations, 36(3), 43-52.



brez strahu pred posledicami ali dodatnimi negativnimi [21] Dong, X. 2017. Elder self-neglect: research and practice. Clinical interventions in aging , 949-954.

[22] Altendorf, A., Draper, B., Wijeratne, C., Schreiber, J., and



Kanareck, D. 2020. Neglect of older people: Touching on forensic

and pathophysiological aspects. The Gerontologist, 60(6), e449-

e465.

[23] Lachs, M. S., Williams, C. S., O'brien, S., Pillemer, K. A., and

Charlson, M. E. 1998. The mortality of elder

mistreatment. Jama, 280(5), 428-432.

[24] World Health Organization. 2002. The World Report on Violence

and Health, World Health Organization. Geneva, Switzerland.

[25] Pillemer, K., Burnes, D., Riffin, C., and Lachs, M. S. 2016. Elder

abuse: global situation, risk factors, and prevention strategies. The

Gerontologist, 56(Suppl_2), S194-S205.



[Vnesite besedilo]



Matične knjige pravoslavne župnije



sv. Cirila in Metoda v Ljubljani

Registry Books of the Serbian Orthodox Parish

St. Cyril and Methodius in Ljubljana



izr. prof. dr. Mari Jože Osredkar

Univerza v Ljubljani

Teološka fakulteta

Poljanska 4, 1000 Ljubljana,

Slovenija Tel: +386 1 434 58 46; fax:

+386 1 43458 54

mari.osredkar@teof.uni-lj.si



1 Uvod



POVZETEK 1.1 Demografski položaj balkanskih držav



Kljub nizki rodnosti in izseljevanju, Slovenija ohranja število Balkanske države se soočajo z zmanjševanjem prebivalstva. Albanija je zaradi izseljevanja od leta 2011 izgubila več kot prebivalcev zahvaljujoč priseljevanju iz južnih držav Balkana, 400.000 oz. 14% prebivalcev (1); v istem obdobju se je tudi na kar se močno odraža v islamskih in pravoslavnih verskih Hrvaškem prebivalstvo zmanjšalo za isto številko, pri njih to skupnostih v Sloveniji. V Islamski verski skupnosti v Sloveniji, pomeni 9,25% (2). Število prebivalcev Srbije se je v zadnjih Makedonski pravoslavni Cerkvi in v Srbski pravoslavni Cerkvi desetih letih zmanjšalo za skoraj pol milijona ali za 6,7 % (3). v Sloveniji se zelo povečuje število vernikov. Namen članka ni Severna Makedonija ima 9,2% manj prebivalcev kot pred predstavitev številk, temveč na vzorčnem – reprezentativnem dvajsetimi leti (4), tako, da je Makedoncev v tej državi manj od primeru ljubljanske srbske pravoslavne župnije pokazati, kako se dveh milijonov. Tudi na Kosovu je 200.000 prebivalcev manj kot je v Srbski pravoslavni Cerkvi v zadnjih 30 letih povečala pred petimi leti (5). Demografi v teh državah kot glavni razlog dejavnost pri podeljevanju zakramentov krsta in poroke ter pri za zmanjševanje števila prebivalstva navajajo izseljevanje. Edino opravljenih pogrebih. V prispevku so predstavljene matične Črna Gora nima manjka (6). knjige ljubljanske župnije sv. Cirila in Metoda Srbske

pravoslavne Cerkve, iz katerih je to dejstvo razvidno.

1.2 Privilegij Slovenije

Ključne besede: demografija, rodnost, Srbska pravoslavna

župnija sv. Cirila in Metoda v Ljubljani, matične knjige. Medtem, ko se število prebivalcev vseh samostojnih držav, ki so

nastale na ozemlju bivše Jugoslavije, drastično zmanjšuje,

Slovenija ohranja število prebivalcev (trenutno jih ima nekoliko

več od dveh milijonov in 124.000) (7) oz. jih kljub nizki nataliteti

ABSTRACT in izseljevanju celo nekoliko povečuje. Slovenija ima namreč

veliko boljše gospodarstvo od južnih držav Balkana, veliko

Despite the low birth rate and emigration, Slovenia maintains its možnosti zaposlovanja in v primerjavi z njimi predvsem boljše

population thanks to immigration from the southern countries of plače. Vsekakor pa je v Sloveniji dobra kvaliteta življenja. Vsi ti

razlogi privabljajo Hrvate, Bošnjake, Srbe, Makedonce in

the Balkans, which is strongly reflected in the Islamic and

Albance v Slovenijo. Posledično se veča število pripadnikov v

Orthodox religious communities in Slovenia. In the Islamic

islamskih in pravoslavnih verskih skupnostih v »deželi pod

religious community in Slovenia, the Macedonian Orthodox Alpami«. Church and the Serbian Orthodox Church in Slovenia, the

number of believers is increasing greatly. The purpose of the

article is not to present these numbers, but rather to show, using 1.3 Rast islamskih in pravoslavnih skupnosti v Sloveniji exemplary - representative examples of the Serbian Orthodox

parish in Ljubljana, how activity in the Serbian Orthodox Church Medtem, ko je ob zadnjem uradnem štetju prebivalstva, ki je

has increased over the past 30 years in the granting of the vsebovalo vprašanje verske pripadnosti, leta 2002, v Sloveniji

sacraments of baptism and marriage, as well as funerals. The živelo nekoliko več kot 48.000 muslimanov, jih je danes v naši

paper presents the registry books of the Ljubljana parish of St. državi, po oceni vodstva skupnosti, več kot 100.000. »Razloga

Cyril and Methodius of the Serbian Orthodox Church, from za to sta predvsem dva: nadaljevanje priseljevanja

which this fact is evident. muslimanskega prebivalstva v Slovenijo in nove generacije

muslimanov, ki so že rojene v Sloveniji« (8). Pred

Keywords: demography, birth rate, Serbian Orthodox Church, petindvajsetimi leti je v Sloveniji deloval en imam, danes jih

Parish of St. Cyril and Methodius in Ljubljana, parish registers. dvajset skrbi za pastoralno oskrbo vernikov. Poleg njih ima

skupnost za pastoralne potrebe (verouk, administracija)

zaposlenih še deset oseb. Makedonska pravoslavna Cerkev išče

pastoralne prostore v Mariboru, ker se je tam povečalo vernikov

močno povečalo. Ravno tako se v Sloveniji veča število



2

Information Society 2024, 12. oktober 2024, Ljubljana, Slovenia

pripadnikov Srbske pravoslavne Cerkve. Nobena od teh treh strani 99. Skupno lahko torej na 99 straneh najdemo 931 oseb, ki

verskih skupnosti ne ve točno, koliko članov ima. Mnogi verniki so se z zakramentom krsta včlanile v Srbsko pravoslavno Cerkev

se sicer udeležujejo obredov v Sloveniji, ampak naša domovina (v veliki večini so bili krščeni otroci) in podrobne podatke o

je za njih le prehodna in ko se ponudi priložnost, nadaljujejo njihovih starših in botrih. Krstna matična knjiga št. VIII ima

svojo pot na zahod Evrope; drugi verniki se udeležujejo obredov vpisane krste od septembra 2000 do decembra 2006 (97 strani)

le občasno in svoje prispevke za versko skupnost raje oddajajo v (11). Krstna knjiga št. IX ima vpisane krste od januarja 2007

matični domovini. Nihče torej sistematično ne vodi evidence novembra 2012 na 103 straneh (12). Krstna knjiga št. X ima

članstva. vpisane krste od februarja 2013 do decembra 2018 (na 197

Namen tega prispevka ni ugotavljanje števila pripadnikov, straneh) (13). Krstna knjiga št. XI ima vpisane krste od januarja

temveč na vzorčnem primeru analize matičnih knjig pokazati na 2018 do januarja 2023 na 99 straneh (14) in Krstna knjiga št. XII

povečanje aktivnosti podeljevanja zakramentov in opravljenih ima vpisane krste od februarja 2023 do danes (15). Šest knjig, ki

pogrebov v Srbski pravoslavni župniji sv. Cirila in Metoda v smo jih dobili v vpogled, so v dobrem stanju, pisava v cirilici je

Ljubljani, ki je največja in hkrati reprezentativna za Srbsko čitljiva, so pa pogoste napaka pri zaporednih številkah oz. pri

pravoslavno Cerkev v Sloveniji. Krst, poroka in pogreb so numeriranju knjižnih strani. Razen na platnicah, kjer so

namreč trije pokazatelji, da so priseljenci »pognali korenine« v uporabljene rimske številke strani, so v knjigah za številčenje

slovenskem okolju. uporabljene »arabske številke«!



2 Zgodovina SPC v Sloveniji 3.2. Poročne matične knjige



Začetke Srbske pravoslavne Cerkve v Sloveniji najdemo v letih označena z letnicami 1952-2007. Na 99-tih straneh vsebuje Pregledali smo tri poročne matične knjige. Poročna knjiga I je

po prvi svetovni vojni, ko je bila ustanovljena kraljevina Srbov podatke o vseh porokah na župniji od leta 1954 do 2007 (16).

Hrvatov in Slovencev. Prva pravoslavna župnija je bila Navedena so imena ženinov in nevest, ki so sklenili cerkveno

ustanovljena v Ljubljani leta 1919. V Mariboru dve leti kasneje. poroko v ljubljanski pravoslavni cerkvi in hkrati njihovi

V Celju je v tistem času bivalo okoli 200 orožnikov pravoslavne podrobni podatki. Poročna knjiga II ima na 49-tih straneh

vere s svojimi družinami. Duhovno oskrbo zanje je prevzel vpisane poroke od leta 2008 do 2015 (17). Poročna knjiga III pa

ljubljanski vojaški duhovnik protojerej Dimitrije Janković, ki je od 2015 danes (18). Vse tri knjige so v dobrem stanju, pisava v

imel veliko zaslug, da je bila leta 1921 ustanovljena cerkvena cirilici je čitljiva in vse numeriranje je korektno. Razen na

občina, in nato leta 1928 tudi župnija v Celju(9). Na začetku platnicah, kjer so uporabljene rimske številke strani, so v knjigah

druge svetovne vojne so okupatorji Nemci porušili cerkvi v za številčenje uporabljene »arabske številke«! Celju in v Mariboru in s tem onemogočili župnijsko versko

življenje. Po koncu vojne je duhovnik iz Ljubljane oskrboval 3.3. Pogrebne knjige srbske pravoslavne vernike po vsej Sloveniji. Postopoma pa so

se ponovno ustanavljale župnije: V Celju leta 1968, v Kopru leta Pogrebne matične knjige so tri. Pogrebi I, na straneh od 1 do 100

1976, dve leti kasneje v Kranju, v Novi gorici leta 1980, v Novem vsebuje seznam vseh pogrebov od leta 1952 do 2004 (19). V

Mestu 1983; zadnji dve sta bili ustanovljeni 2005 (ponovno) v knjigi Pogrebi II pa so vpisane osebe in njihovi podatki, ki so jih

Mariboru in 2011 v Postojni. Leta 1968 je deloval v Sloveniji en duhovniki pospremili na pokopališče in opravili verski obred pri

duhovnik Srbske pravoslavne Cerkve, ob vsaki ustanovitvi grobu od leta 2005 do 2021 (20). V zadnji, Pogrebi III so vpisani

župnije se je pridružil tudi nov duhovnik; na nekaterih župnijah pogrebi od leta 2021 do danes (21).

deluje več duhovnikov, tako, da je danes v Sloveniji na osmih V spodnji razpredelnici smo navedli število krstov, porok in

župnijah aktivnih 15 duhovnikov Srbske pravoslavne Cerkve. pogrebov od leta 1952 do danes. Skupnost se v zadnjih desetletji močno širi oz. se povečuje

število njenih vernikov. Priseljujejo se iz Srbije in iz Republike 3.1 Tabela 1: Podatki iz matičnih knjig srbske iz Bosne in Hercegovine. Točnega števila vernikov ne ve

nihče povedati, po oceni duhovnika Aleksandra Obradovića je na

vseh njihovih župnijah v Sloveniji aktivnih približno 100.000 Leto Krsti Poroke Pogrebi



vernikov. Omenjeni duhovnik in njegov kolega Rade Despotović 1952 6



Ljubljane vključuje tudi Domžale, Škofljico in Kamnik, aktivnih ocenjujeta, da je v letu 2024 v ljubljanski župniji, ki poleg mesta 1953 6

približno 2000 vernih družin, kar pomeni 8 do 10 tisoč vernikov. 1954 4 5



1955 2 4

3 Predstavitev matičnih knjig

1956 0 5



Župnija sv. Cirila in Metoda, srbske pravoslavne Cerkve v 1957 1 3



Ljubljani, katere cerkev se nahaja na Prešernovi cesti 35, hrani 1958 0 6 tri vrste matičnih knjig: krstne (Knjiga krštenih), poročne

(Knjiga venčanih) in pogrebne (Knjiga umrlih). 1959 2 1



1960 1 7

3.1 Krstne matične knjige

1961 3 4



Za našo raziskavo nam je bil omogočen vpogled v Krstne 1962 2 7 matične knjige od številke VII do XII. Krstna matična knjiga št.

VII ima na sprednji platnici napisane letnice 1994 do 2000 (10). 1963 0 11



Toda prvi vpisani krsti so iz januarja leta 1995; zaključi pa se s 1964 1 11 krstom, ki je bil podeljen 9. septembra 2000 in je zabeležen na



Information Society 2020, 5–9 October 2020, Ljubljana, Slovenia



1965 5 10 2009 176 27 49



1966 1 11 2010 172 24 46



1967 3 9 2011 165 34 69



1968 2 11 2012 160 29 69



1969 1 11 2013 159 37 48



1970 2 7 2014 184 30 36



1971 3 15 2015 170 20 55



1972 5 18 2016 181 22 54



1973 7 15 2017 215 23 53



1974 5 15 2018 195 35 77



1975 0 10 2019 173 29 68



1976 2 15 2020 106 14 78



1977 5 16 2021 158 14 105



1978 3 6 2022 188 32 92



1979 4 13 2023 218 21 82



1980 6 10



1981 6 10 4 Krsti

1982 2 13



1983 Število krstov v ljubljanski srbski pravoslavni župniji nam v 15 14



1984 21 2 znan podatek o spreminjanju (rasti) števila članov v skupnosti, bistvo kaže na položaj rodnosti v tej verski skupnosti. Čeprav ni

1985 nam prav število krstov orientacijsko razodeva demografske 27 14



1986 spremembe v skupnosti. Ko želimo primerjati rodnost v župniji 29 16

in v Sloveniji na splošno, nam koeficient 1,5, ki ga statistični

1987 30 10 urad objavlja za Republiko Slovenijo, nič ne koristi, ker



1988 koeficienta za župnijo Srbske pravoslavne Cerkve v Ljubljani ne 15 5

moremo izračunati. Ne poznamo namreč števila žensk v rodni

1989 28 14 dobi. Da bomo lahko primerjali s slovensko rodnostjo, bomo



1990 16 15 obdobja v župniji in v Sloveniji. primerjali povprečje prvih treh in zadnjih treh let obravnavanega

1991 Število krstov se od 1995 do 2023, torej v zadnjih tridesetih letih, 19 9



1992 na prvi pogled ni bistveno spremenilo, oz. opažamo zgolj majhno 19 19

rast. Če pa upoštevamo »odcepitev« župnije Postojna v letu

1993 2011, vidimo, da se je takrat krivulja na grafu št. 1 obrnila 24 25



1994 navzdol, potem pa se zopet povzpela. Od 1995 do 1997 lahko iz 31 17

matičnih knjig razberemo, da je bilo na župniji v povprečju

1995 200 43 37 163,66 krstov na leto. V triletju 2021 do 2023 pa se je to



1996 povprečje dvignilo na 188 krščenih otrok. S preprostim računom 142 24 26

ugotovimo, da se je rodnost od zadnjega triletja v prejšnjem

1997 tisočletju do danes povečala za 14,87%. Ko pa to konstantno 149 34 23



1998 rodnost primerjamo rodnosti v Sloveniji, pa pridemo do 140 27 30

zanimivih rezultatov. Da bo primerjava korektna, bomo tudi tu

1999 primerjali povprečji prvih in zadnjih treh let obravnavanega 191 28 37



2000 obdobja. V Sloveniji se je po podatkih Statističnega urada (22) 165 35 34

leta 1995 rodilo 18.980 živorojenih otrok, leta 2023 pa 16.989.

2001 155 45 45 Od leta 1995 do 1997 se je v Sloveniji povprečno na leto rodilo



2002 18.644 živorojenih otrok. Od 2021 do 2023 pa je to povprečje 100 41 52

17.866 rojstev. To pomeni zmanjšanje natalitete za 4,17%. V

2003 158 38 38 Sloveniji v prvem desetletju tretjega tisočletja do sredine drugega



2004 desetletja sicer beležimo povečano število rojstev, od 2017 pa do 177 30 44

2023 pa je drastičen padec, ki se zaključi z najnižjim številom

2005 155 24 52 rojstev v zgodovini vodenja statistike v Sloveniji. V ljubljanski



2006 župniji Srbske pravoslavne Cerkve se je rodnost torej povečala 164 39 49

za nekaj manj od 15%, Slovenci pa beležimo zmanjšanje za nekaj

2007 150 36 50 več od 4%. Razlika je skoraj 20%.



2008 145 36 46



4

Information Society 2024, 12. oktober 2024, Ljubljana, Slovenia

V spodnjem grafu nam krivulja kaže gibanje krstov na

obravnavani župniji. Število porok v dvajsetih letih 21. stoletja ostaja identično številu

iz devetdesetih let. Vendar, kakor smo omenili zgoraj, je

4.1 Graf 1: Število krstov potrebno upoštevati nov način skupnega življenja. Vsekakor pa

je ohranitev števila porok pokazatelj velikega števila parov oz.

družin. Če namreč primerjamo s katoliškimi župnijami, kjer je



Krsti število zelo upadlo, je je ohranitev števila na pravoslavni župniji znamenje velikega števila vernikov.

250

200 6 Pogrebi



150

Najboljši pokazatelj gibanja aktivnosti na župniji svetih bratov

100 Cirila in Metoda je pogrebna matična knjiga. Pri izseljencih, ki

50 imajo močne vezi s svojo matično domovino, torej s svojimi

sorodniki in prijatelji, je značilno, da svoje pokojne pokopljejo

0 na pokopališču v domačem kraju, v Srbiji ali Republiki srbski.

Šele, ko »poženejo korenine« v novi domovini, torej v Sloveniji,

1995 1997 1999 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2019 2021 2023 se odločijo in kupijo grob na primer na ljubljanskih Žalah in tu

Krsti pokopljejo svoje umrle. Dobili smo vpogled opravljenih

pogrebov od leta 1952 do danes. Prvih deset let število ni

preseglo deset pogrebov na leto. Nato se številka veča; od 1963

do 1993 se giblje med deset in dvajset. Upoštevati moramo, da

se številka nekoliko zaustavi v letih, ko so nove župnije

5 Poroke postopoma zmanjševale teritorij in število vernikov ljubljanske

župnije: 1968, ko je bila (ponovno) ustanovljena župnija Celje,



Odgovorni na župniji so nam omogočili vpogled v poročne 1976, ko je bila ustanovljena župnija Koper, 1978 z ustanovitvijo župnije v Kranju, 1983 v Novem Mestu in 2011 v Postojni. Po matične knjige od leta 1954 dalje. Do 1983 je število porok na

leto zelo majhno. Prvih deset let se število giblje od 0 do 4, leta letu 1995 je bilo nad trideset pogrebov v povprečju na leto. Z

1980 jih je bilo 6; leta 1985 najdemo vpisanih petnajst parov, ki letom 2020 pa številka pogrebov strmo naraste. Upoštevati

so se poročili v cerkvi sv. Cirila in Metoda. V obdobju od 1983 moramo večjo umrljivost, ki jo je povzročil virus Korone leta

do 1994 se število giblje od 15 pa vse do 31, od 1995 pa do 2002 2021. V letih 2022 in 2023 se namreč število pogrebov zmanjša.

pa je število med 30 in 45. Tudi tu je potrebno upoštevati, da je

bila leta 2011 ustanovljena nova pravoslavna župnija v Postojni, 6.1 Graf 3: Število pogrebov ki je od ljubljanske župnije »odtrgala ozemlje od Vrhnike do

Postojne. Zaradi manjšega števila vernikov, je v tem obdobju



zabeležen padec porok v ljubljanski župniji. Od 34 poročenih Pogrebi parov v letu 2011 vidimo padec števila (z izjemo 2013) porok vse

do leta 2018, ko se zviša. V zadnjih šestih letih pa se število zopet

manjša, tako, da imajo v letu 2023 zabeleženih le 21 porok. Ta 150

upad je po izjavi duhovnika Aleksandra Obradovića povezan s 100

tendenco, ki je v Sloveniji »moderna«, da mladi pari ne sklepajo

uradnih zakonskih zvez ampak zgolj živijo skupaj »brez obljub 50



oz. obveznosti. V pogovoru z avtorjem tega prispevka je 0

duhovnik, ki je odgovoren za vso Srbsko pravoslavno Cerkev v



na ljubljanski župniji opravijo, zaprosijo starši, ki niso 1952 1959 1968 1977 1986 1995 2002 2005 2008 2011 2014 2017 2020 2023 Sloveniji pojasnil, da v zadnjih letih polovica krstov otrok, ki jih

(cerkveno) poročeni.



5.1 Graf 2: Število porok Sklepamo na konstantno priseljevanje, le ob ustanovitvi novih

župnij je padec opravljenih pogrebov. V zadnjih tridesetih letih



Poroke se je število pogrebov povečalo za štiri do pet krat, to je od 20 na

90 pogrebov.

60



40 7 Analiza gibanja številk



20

Kljub temu, da je ljubljanska župnija v osemdesetih letih

0 prejšnjega stoletja z ustanovitvijo novo meške župnije izgubila

Grosuplje in Kočevje, leta 2011 pa ozemlje od Vrhnike do

1954 1963 1972 1981 1990 1999 2005 2008 2011 2014 2017 2020 2023 Postojne, ki se je pastoralno osamosvojila, se je število krstov na

župniji v zadnjih tridesetih letih povečalo za slabih 15%, število



Information Society 2020, 5–9 October 2020, Ljubljana, Slovenia



porok je sicer ostalo enako, število pogrebov pa se je povečalo (7)Statistični urad Slovenije:

za skoraj 5 krat. Pri analizi dejavnosti v Srbski pravoslavni https://www.stat.si/statweb/News/Index/12998

župniji svetih bratov Cirila in Metoda v Ljubljani moramo

omeniti še dve dejstvi. V začetku devetdesetih let je na župniji (8) Jeglič, Urška. Neujemanja z islamskimi predpisi pokopa pri

deloval en duhovnik, danes jih deluje 5, ki poučujejo tudi verski muslimanih v Republiki Sloveniji. Bogoslovni vestnik : glasilo

1 Teološke fakultete v Ljubljani. [Tiskana izd.]. 2020, letn. 80, [št.]

nauk 120 otrok . Po izpovedi duhovnika Aleksandra je bilo leta

2013 v Ljubljani samo 40 vero učencev. Torej se je število 4, str. 876

duhovnikov v zadnjih tridesetih letih povečalo za 5 krat, število

(9) Cvelfar, Bojan, Srbska pravoslavna Cerkev na Slovenskem

vero učencev pa v zadnjih desetih letih 3 krat. med obema vojnama, Ljubljana: Ištitut za novejšo zgodovino,

Če te ugotovitve združimo, lahko napišemo, da so se verske 2017, 318-319. aktivnosti v zadnjih tridesetih letih na predstavljeni župniji

močno povečale. Iz tega lahko sklepamo, da se je tudi povečalo (10) Knjiga krštenih VII, Parohija ljubljanska u Ljubljani.

število vernikov oz. prišlekov iz Srbije in Republike srbske, ki

izpovedujejo Srbsko pravoslavno vero. (11) Knjiga krštenih VIII, Parohija ljubljanska u Ljubljani.



(12) Knjiga krštenih IX, Parohija ljubljanska u Ljubljani.

8 Sklep

(13) Knjiga krštenih X, Parohija ljubljanska u Ljubljani.

Pri migracijah smo ljudje najbolj občutljivi na vrednote, ki jih

prinesejo migranti s seboj in se razlikujejo od naših vrednot. (14) Knjiga krštenih XI, Parohija ljubljanska u Ljubljani.

Najprej gre za kulturo, potem jezik in nenazadnje tudi versko

prepričanje. Odnos med prišleki in lokalnim prebivalstvom se od (15) Knjiga krštenih XII, Parohija ljubljanska u Ljubljani.

države do države razlikuje. Evropska unija zagovarja in predlaga

model integracije. Integrirani migranti se bodo lahko vključili v (16) Knjiga venčanih I, Parohija ljubljanska u Ljubljani.

slovensko kulturo in se naučili slovenskega jezika. Vera pa je

nekaj tako intimnega, da človek ohrani versko prepričanje in (17) Knjiga venčanih II, Parohija ljubljanska u Ljubljani.



prebivalci drugačnih veroizpovedi. Za skupno bivanje je torej (18) Knjiga venčanih III, Parohija ljubljanska u Ljubljani. pripadnost svoji verski ustanovi tudi v okolju, kjer živijo



vključuje sprejemanje in spoštovanje drugačnosti, kar je (19) Knjiga umrlih I, Parohija ljubljanska u Ljubljani. nujno potreben medkulturni in medverstveni dialog. Ta pa

pogojeno s poznanjem veroizpovedi in verske prakse, v našem

primeru, srbske pravoslavne skupnosti v Sloveniji. Bojan (20) Knjiga umrlih II, Parohija ljubljanska u Ljubljani.



Cvelbar piše v svoji knjigi, da so v Celju v letih po prvi svetovni (21) Knjiga umrlih III, Parohija ljubljanska u Ljubljani. vojni katoličani raztrgali celo oglase, ki so Srbe vabili na

ustanovni sestanek Srbske pravoslavne občine. Danes se kaj (22) Statistični urad, SiStat:

takega ne dogaja več. V Sloveniji imamo zgleden odnos med https://pxweb.stat.si/SiStatData/pxweb/sl/Data/-

katoličani, pravoslavnimi in muslimani. Dokaz tega /05J1002S.px/table/tableViewLayout2/

prijateljskega dialoga so srečanja in skupne izjave o družbeno

pomembnih vprašanjih.



Reference in viri



(1)https://forbes.n1info.si/novice/zdaj-je-izmerjeno-balkanska-

drzava-se-je-skrcila-za-vec-kot-400-tisoc-ljudi/



(2) https://www.rtvslo.si/svet/evropa/v-10-letih-se-je-stevilo-

prebivalcev-na-hrvaskem-zmanjsalo-za-skoraj-400-000/608576



(3)https://www.slobodnaevropa.org/a/srbija-popis-

stanovnistvo/32187570.html



(4)https://vecer.com/svet/popis-prebivalstva-v-severni-

makedoniji-vsi-rinejo-v-skopje-10279088



(5)https://reporter.si/clanek/svet/na-kosovu-16-milijona-

prebivalcev-se-pred-desetletjem-je-tam-zivelo-18-milijona-

ljudi-1284458



(6)https://balkans.aljazeera.net/news/balkan/2024/1/25/crna-

gora-prvi-rezultati-popisa-minimalan-rast-broja-stanovnika



1 Na vseh osmih župnijah v Sloveniji obiskuje verouk približno 400 otrok.

Poučevanje je v srbskem jeziku, čeprav uporabljajo tudi pravoslavni molitvenik v

slovenskem jeziku.





EMIGRATIONS FROM KOSOVA IN THE FIRST TWO




DECADES OF THE 21st CENTURY- TREND AND



CHALLENGES



Dr.sc. Sami Behrami

Kosova Environmental Protection Agency

Luan Haradinaj Str. (former Media Palace), 14th floor, 10000 Pristina, KOSOVA

E-mail: samibehrami@gmail.com,



Abstract:

1. Introduction



In the first two decades of the 21st century, Kosova is Population (people) are the greatest asset of a country characterized as a territory with a very high rate of



emigration, a rapid decrease in births and natural increase, Population emigration from Kosova to European countries followed by changes in the structure of the population by is not a new phenomenon. Throughout the 20th century age, which have consequently caused challenges in many and the beginning of the 21st century Kosova has segments of life in Kosova. remained a typical migratory territory. The main causes High rate of emigration is the main factor that today Kosova that forced the population to emigrate were and still is facing radical changes in its demographic profile. The remain: the low situation is so unfavorable that it can be defined as a level of economic development, the destruction caused by demographic crisis. wars, the unfavorable political situation, the creation of the The demographic problems that Kosova is facing can be system based on political party criteria and nepotism, prevented or mitigated through population policies that avoiding meritocracy and values in the creation of the would enable the creation of conditions for sustainable system, the dissatisfaction of the citizens of Kosova with demographic development, which is achieved by creating the perspective of the socio-economic future and with the conditions for people, particualrly for the youth, to have general level of well-being, which is manifested by a high confidence in building a future in their country. rate of unemployment (especially among young people) If the population, particularly young people, have access and a high percentage of people living in poverty, etc. to quality education and health care, if they are in stable Therefore, the economic issues related to unemployment employment and receive support for raising children, etc., and poverty are the main concerns of Kosova citizens. they will be less likely to emigrate and more likely to decide Recent opinion polls show that these two economic to have children. Eliminating the causes that push young parameters are also national security threats that Kosova people to emigrate is the key to cease the deepening of faces with. According to these data, about 70% of the demographic crisis in Kosova. respondents believed that the high rate of unemployment The perspective of future development made it necessary is the biggest threat to Kosova (1) . Considering the high to take measures to control demographic processes, rate of unemployment and the large number of people especially emigration and births, after the forecasts of living below the poverty level, emigration is seen as an Kosova’s population show that in the future Kosova will opportunity by the citizens of Kosova to find a better social face many challenges as a result directly from the and economic perspective abroad, turning emigrations into unfavorable demographic developments, which in the a risk for demographic, socio-economic and security following years and decades will only deepen further. developments in the country. As a result of the high migratory flux, Kosova nowadays is Key words: emigration, population, demographic facing radical changes in its demographic profile: on the sustainability, challenges, Kosova etc. one hand, there is a drastic reduction in the number of births - natural increase, while on the other hand, as a Permission to make digital or hard copies of part or all of this work for result of selective emigration, the population is aging with personal or classroom use is granted without fee provided that copies are accelerated trends.. not made or distributed for profit or commercial advantage and that Since emigrations are the most important factor in the copies bear this notice and the full citation on the first page. Copyrights decline or stagnation of the population of Kosova, it is for third-party components of this work must be honored. For all other

uses, contact the owner/author(s). necessary for the state to act in order to affect the

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia reduction of the causes that encourage emigration.

© 2024 Copyright held by the owner/author(s). Creating the conditions for people, particularly for the

youth to contribute to building a future in their country, is

the key to stopping emigration and solving the

demographic crisis in Kosova.



501



2. External (international) emigrations of the population The situation is so unfavorable that it can be considered a

demographic crisis.

. For a period of over a century, Kosova has been Since the end of the 1999 war, strong waves of emigration

characterized by a negative migration balance, which from Kosova have been associated with the second

means that the number of emigrants has been greater than decade of the 21st century, with a culmination during the

the number of returnees-immigrants. years 2014 and 2015, which are also associated with the

In the period 2000-2021, over 420,000 inhabitants left economic crisis and global uncertainty. While the reasons

Kosova (according to KAS statistics), with an average of for emigration, in the beginning, were of an economic

over 21,000 inhabitants per year, although there was nature, now the loss of hope and uncertainty about the

previously the conviction that upon the freedom and future in Kosova has become the main reason for

independence, Kosova would also recover from this emigration.

century-old wound. st The emigrations of the two decades of the 21 century are

External emigration is one of the main reasons why the most unfavorable waves to date. This trend is

Kosova nowadays is characterized by unfavorable developing in circumstances of reduced births, decline in

demographic trends such as: stagnation of the total natural increase, stagnation with decreasing trends in the

number of the population with a downward trend, rapid total number of the population and the accelerated process

trends of decreasing the birth and natural increase, of deformation of the population structures, particularly that

deterioration in the structures of the population by age - by the age towards the aging of the population.

demographic age, as well as other population structures.





Graph 1. Trends in births, natural increase, emigrations and external migration balance in Kosova 2011-2021( 2)



40000



20000



0

2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021



-20000



-40000



-60000



-80000

F NI E MB





2.1. Trends of external emigrations of the population 2011-2021



Like in many countries of the world, also in Kosova, population emigration in Kosova 2011-2021, from 2011 to

emigration statistics are incomplete because many 2015 there is a constant increase in the number of

residents do not deregister their residence in Kosova emmigrants. The year 2014 and particularly 2015,

before they leave or they do so long after moving. represent the years with an extreme increase in the

However, even such incomplete data show that in the number of emmigrants when the number of emmigrants

period 2011-2021 more than 309 thousand inhabitants compared to 2011 had increased by 323.9% or 4.2 times

moved abroad, this number is much higher (almost 190 (2014) and 598.8% or close to 7 times (2015). Excluding

thousand more) than in the period 2001-2010, when more the mass exodus and deportation of about 1 million

than 119 thousand of them moved. Albanians by the Milosevic’s regime in 1999, between the

Although the emigration statistics are incomplete, they are end of 2014 and the beginning of 2015, an unprecedented

the only official data on the number of emigrants presented phenomenon of illegal emigration occurred in Kosova (3).

in table 1. As it can be seen from the table and graph on In the period 2017-2021, there was also a continuous



502



increase in emigration, with exception of the year 2020 making this period with the most movement restrictions in

when Kosova was characterized by a positive migration the history of international travel (4). However, in 2021, the

balance. number of emmigrants was much higher than in 2020 (4.8

The main reason for this is the pandemic of the disease times higher) and close to 2014, so it can be expected that

COVID-19, due to which many countries closed their after 2021 there will be no decrease in the number of

borders, so the employment of foreign nationals in these emmigrants.

countries became difficult or impossible. Thus, more than

150 countries had imposed such restrictive measures,



Graph 2. Number of emigrants from Kosova 2011-2021(5)



80000



70000



60000



50000



40000



30000



20000



10000



0

2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021



Number of emigrants



That emigrations have reached alarming proportions is economic development, unemployment, prospects for a

also proven by the fact that during 2021, 119 people left better and safer, more dignified life, etc.

Kosova within a day, or 1 person every 12 minutes. Based The situation in terms of emigrations may even deteriorate

on the preliminary data, during 2022 the situation only in the case of visa liberalization and the union of Kosova

worsened even more. with the EU, which may trigger an increasing wave of

Even more worrying is the fact that based on various emigration, given the fact that Kosova lags behind the EU

surveys carried out in the country, about 60% of the countries in the socio-economic sphere, since one of the

respondents have expressed the desire to leave Kosova, main factors of migrations is the difference in incomes

mainly the new generations (20 to 40 years old) who make between the countries of origin and the hosting countries.

up the main demographic and birth contingent, the Such a tendency is also based on the practices of various

contingent of work, security, etc., causing a decline in countries which, after liberalization and membership in the

“human capital”, as the main carrier of socio-economic EU, have faced an increased migratory flux.

development.

Despite the fact that Kosova faces a very high emigration 2.2. External emigrations from Kosova by

flux, more than half of Kosova’s citizens (55.4%) agree that municipalities

emigration will have a negative impact on future

developments in the country, 15.3% agree that there will The analysis of emigration at the level of municipalities

not be any impact, 16.9% do not know what impact they provides us with information about the spatial dimension of

will have, while 11.6% agree that emigration will have a this phenomenon in Kosova, as well as the ascertainment

positive impact (6). whether there have been significant changes in certain

Taking into account the current progress but also what is periods of time.

expected in economic development, employment In order to ascertain the spatial change and the fact that

opportunities and other unfavorable circumstances, for there have been changes in the intensity of external

some more time there should be expected migratory emigration, there will be used as the indicator - the rate of

movements of the population with undiminished intensity. emigration, which represents the ratio between the number

Unfortunately, Kosova’s institutions have done ver y little in of emigrants and the number of the resident population in

addressing the emigration driving factors such as: 2021 in the respective municipality (per 100 inhabitants).

The number of emigrants in relation to the number of



503

resident population in Kosova differs greatly among 19% - from 7 municipalities, 20-29% - from 6

municipalities. The data from the 2011 population census municipalities, 30-39% - from 8 municipalities and over

show that in municipalities with a small number of resident 40% - from 11 municipalities. In relation to the resident

population, such as: Shtërpca, Graçanica, Ranillug and population, Suhareka, Obiliq, Gjilan, Mitrovica, Vushtrri,

Mamusha, the emigration was lower, below 10%. etc. had the largest number of population involved in

In this census, Kosova had 21.4% of the population living emigration.

abroad, whereas between 1-9% of the population abroad

was from 9 municipalities, 10-19% from 12 municipalities, 2.3. Emigration by Municipalities 2011-2021

20-29% from 13 municipalities, 30-39% from 3



number of emigrants in the 2011 census, nearly 47% of the continued with considerable intensity during the period 2011-2021. During this period, almost all municipalities in population lived abroad, Klina 34.1%, Mitrovica 33.2%, Kosova were characterized by a negative migration Gjakova 31.4%, etc. to the resident population, Suhareka had the largest Migration, Migration, especially external emigration, municipalities and over 40% from 1 municipality. In relation



emigration at the municipal level, between the 2011 pandemic (when movements, especially abroad, were restricted), only the municipalities of Prishtina and Kaçanik The data and the difference (increase) in the data on (7) balance, except for the year 2020 when, due to the

show that the phenomenon of emigrations in many had a negative migration balance; all other municipalities population census and the emigration statistics of 2021

municipalities has taken on worrying proportions. in 2020 had a positive balance.

Based on the statistics of the Kosova Agency of Statistics Data (8) published by the Kosova Agency of Statistics

in 2021, the percentage of the overall Kosova population show that at the Kosova level, in the period 2011-2021, an

living abroad reached 36.6% from 21.4% in the 2011 average of 15.9 residents (16 residents per 100) were

population census, while there was a significant increase involved in migration compared to the total number of

in this percentage in its municipalities. Thus, between 1- residents in 2021

9% of the population abroad was from 6 municipalities, 10-

.



Map 1 and 2: The number of emigrant population compared to the resident population by municipalities in 2011 and

2021, expressed as a percentage.





. As in the 2011 population census and during the 2011- be due to incomplete statistics (non-inclusion) and the fact



2021 period, in municipalities with a small number of that these municipalities are already experiencing

resident populations, migration was lower, such as in demographic aging (lack of migration potential).



Elezit, North Mitrovica, and Deçan, municipalities that had During the 2011-2021 period, in relation to the resident Partesh, Gračanica, Zvečan, Zubin Potok, Kllokot, Hani i

fewer than 5 emigrants per 100 residents. However, it population, the municipalities with the highest number of

should be noted that low migration in some municipalities emigrants (per 100 residents) were: Obiliq 45.4%, Gjilan

(especially those with a majority Serb population) may also 40.3%, Kamenica 38.6%, Vushtrri 31.3%, Fushë Kosova



504



28.3%, Suhareka 26.6%, Ferizaj 22.2%, Mitrovica 20.4%, pupils by 16.1% in primary and lower secondary

Shtime 20.4%, etc. It should also be noted that external education, 11.2% in upper secondary education,

migration is the dominant form of migration throughout the and 21.3% in higher education has been

observed period and in almost all municipalities in Kosova. observed in the last five years);

 National security: (In the 21st century, the

3. The Consequences of emigration concept of “security” has transformed

significantly and is no longer equated only with

military security but also includes other forms

The unfavorable economic situation in Kosova, the decline such as demographic, economic, ecological,

in the standard of living for many citizens, the lack of political, and other types of security);

opportunity to find work even in one's profession, long  And more.

waits for employment, insufficient wages considering

qualifications, an unfavorable climate, and the loss of Due to space constraints, only the consequences of

prospects for a better life, especially among the youth, are migration on demographic developments in Kosova are

some of the most important economic factors driving addressed in this paper.

individuals to decide to emigrate temporarily or

permanently from Kosovo. 3.1. Demographic Consequences

In the UNDP Public Pulse survey of 2022, the unfavorable Demographic issues are of crucial importance for the

economic situation—unemployment (36.4%) and poverty development of any state because development cannot be

(21.1%)—are the main problems Kosova currently faces achieved under unfavorable demographic conditions

(9 where migration plays a significant role. ). Unemployment, poverty, and the search for a better

future for their families remain the primary reasons for Migration affects the size of the population by causing the

migration among Kosovars, especially among the youth. loss of that segment of the population which constitutes

Among the youth, the primary reason for considering development potential, birth rates (with a decline in births),

emigration was the lack of job opportunities in Kosova, natural population growth, and the population structures

according to 67% of respondents (55% in 2018). Another (demographic, economic, social, ethnic, etc.), leading to

45% cited the lack of personal prospects as a reason (32% natural depopulation, brain drain, and population aging.

in 2018), and 39% mentioned the unfavorable economic As a result of the decrease in the number of births from

situation in their families/low income levels (38% in 2018). 38,687 (2000) to 22,830 (2021), representing a decline of

The findings from focus groups further emphasized the 15,857 births or nearly a halving of births (-41%), and the

issues of unemployment, financial difficulties, and lack of increase in the number of deaths, natural increase is

prospects as possible reasons for migration (10). characterized by an even faster decline compared to birth

Over the last twenty years, Kosova has been characterized rates. Thus, compared to 2000 when the natural increase

by massive emigration, especially among highly educated was 34,296 people, it decreased to 9,811 in 2021,

youth. This is evidenced by the increase in the participation representing a reduction of 24,485 people, or about 3/4 (-

of emigrants with higher education (before emigration) 71.4%). The natural increase in 2021 is similar to that of

from 17.8% in the 1990s to around 31.5% (11) the second half of the 1930s. in the first

decade of the 21st century. The scale of the decline in the number of births is illustrated

Particularly concerning is the recent trend of increased by the fact that in 2008, an average of 94 babies per day

emigration not only of individuals, but also of entire families were born in Kosova, while in 2021 this number decreased

(12). to 62 births per day, and regarding natural increase, from Many of the young people who emigrate do not leave

alone; in most cases, they move with their families, and 76 new residents to only 27 residents.

many others start families in the host country, significantly The significant impact of migration on the decline in birth

impacting Kosovo's demographic landscape. rates and natural increase is also evidenced by the

The consequences of migration manifest in all segments increase in the percentage of births and natural increase

of life: that occurred abroad. From 2011 to 2021, the percentage

 Demographic: of births outside Kosova increased from 19.3% to 31.8% (the size of the population,

causing the loss of the segment of the population of the total number of births, meaning that migration has

that constitutes development potential, birth also taken away a significant portion of the birth rate—

rates, natural population growth, and the about 1/3.

structure of the population—demographic, The rapid decline in natural increase and selective

economic, social, ethnic, etc.—leading to natural migration (ages 20-39 constitute nearly half of the



 aging. For example, while in 2011 people aged 60 and Socio-economic: (low per capita GDP growth, over constituted 9.6% of the population, in 2021 this age depopulation, brain drain, and population aging); emigrants) have accelerated the process of population



 the social system and pensions, emigration of 0-19 years has shown a trend of decreasing participation from 38.1% to 32.3%. Population aging has negative qualified individuals, etc.); consequences for demographic, economic, social Education system: decreased competitiveness, negative impact on group increased to 13.3%. At the same time, the age group slowed economic growth, loss of labor force,

of pupils and students as a result of emigration developments, etc. (the reduction in the number



and declining natural growth—over the last seven To prevent population aging due to the predominance of young age groups in migration flows, it is necessary to years in Kosova, a reduction in the number of



505



discourage external migration of the population at a young However, if measures are not prompt and effective, merely

age through improvements in conditions, opportunities, reducing emigration and implementing the measures

access, and integration of youth into professional training, outlined above will not be sufficient. It is also necessary to

vocational education, university education in market- consider strategies for attracting emigrants from other

demanded fields, and the labor market. countries, which is a process fraught with many

Alongside the concerning negative demographic trends of challenges.

today, recent migration trends are increasingly impacting Today, due to the lack of policies, emigration and the

other areas of life, such as the labor market, destabilization diaspora may be considered more of a loss and threat than

of educational and health systems, and in the coming an asset and partner, although in reality, the diaspora is a

years will also affect the economic, social, and pension valuable asset in the establishment of the state of Kosova.

systems, inter alia, for an extended period. In the long

term, the old structure of the workforce, negative natural

increase, and the departure of youth from Kosovo will

threaten the functioning of pension and disability insurance

funds and health insurance, which will lack sufficient funds References / Literatura

to cover future obligations.

To reduce the negative socio-economic effects of



migration, it is necessary to adopt a series of measures 1) Behrami, S. (2019). “Emigrations! Is Kosovo at Risk from This and strategies for population policies, measures for Phenomenon?” Kosova Sot. Accessible at: Kosova Sot migration sustainability, and a series of economic reforms (accessed August 2022). https://www.kosova- that would improve the business environment, stimulate sot.info/opinione/431875/emigrimet-a-po-rrezikohet-kosova- job creation, favor investments, and restore optimism nga-ky-fenomen/ (e qasur në gusht 2022) among youth, which would mitigate and reduce the 2) Population Assessment - Various Years. Accessible at: emigration contingent from Kosova, transforming https://ask.rks-gov.net/sq/agjencia-e-statistikave-te- migration from a negative process into a positive one. kosoves/sociale/vleresimi-i-popullsise .

3) Islami, H. (2018). “Push-Pull Factors in External Migrations:

Theoretical Approaches and the Case of Kosovo.” Social

4, Conclusion Studies 4. Academy of Sciences and Arts of Kosovo. Section

of Social Sciences. Prishtina. p. 33. Accessible at:



Mass emigration from Kosova may now be considered a https://ashak.org/botime/studime-shoqerore-4/ 4) Behrami, S. (2020). “Population Emigrations and the threat to the demographic, economic, and social prosperity Pandemic.” On the Occasion of December 18 - International of Kosovar society. Therefore, an imperative for Kosova’s Migrants Day. Accessible at: current and future development is that the emigration of its https://www.albinfo.at/emigrimet-e-popullsis-dhe-pandemia/ population becomes sustainable. Sustainable migration 5) Behrami S (2023) “ Kosova facing the challanges of implies proactive demographic, economic, social, emigration”. Institute for Development Policies- INDEP. technological, and regional policies that will ensure that https://indep.info/kosova-perballe-sfidave-te-emigrimit/ 6) “Women and Migration in Kosovo: Exploring the Data.” Kosova remains an attractive place to live, primarily for its Accessible at: residents. https://www.undp.org/sites/g/files/zskgke326/files/2022-Since migration is the most significant factor in the decline 07/PPDataGenderStory.pdf . or stagnation of Kosova’s population, it is necessary for the 7) Kosovar Migration (2014). ASK. Prishtina. p. 19. Accessible state to act to influence the reduction of the causes driving at: https://ask.rks-gov.net/media/1379/migrimi-kosovar-emigration. This can be achieved by increasing efforts to 2014.pdf reduce economic and psychological insecurity, 8) Population Assessment - Various Years. Accessible at: accelerating economic and social development, creating https://ask.rks-gov.net/sq/agjencia-e-statistikave-te-kosoves/sociale/vleresimi-i-popullsise an atmosphere of prosperity, and supporting health and 9) Public Pulse XXII, USAID, UNDP. Prishtina, April 2022. p. 20. educational programs. If the causes driving the population Accessible at: to emigrate are not addressed, other measures aimed at https://www.undp.org/sites/g/files/zskgke326/files/2022-increasing birth rates may simply result in more people 06/Public%20Pulse%202022%20-%20ALB_0.pdf leaving the country. Creating conditions for people, 10) “Challenges and Perspectives of Youth in Kosovo” (2021). especially youth, to have confidence in building a future in UNDP. Prishtina. p. 28. Accessible at: their own country is key to stopping the demographic crisis https://www.undp.org/sites/g/files/zskgke326/files/migration/k in Kosova. If the population—particularly young people— s/YouthChallengesandPerspectivesinKosovoALB.pdf 11) “Human Development Report in Kosovo 2014 - Migration as hve access to quality education and healthcare, stable a Development Force.” UNDP. Prishtina, 2014. p. 24. employment, and support for child-rearing, they will be less 12) Behrami, S. (2022). “Are Population Emigrations Turning into likely to emigrate and more likely to decide to have a General Threat for Kosovo?” Accessible at: children. https://telegrafi.com/po-shnderrohet-emigrimi-popullsise-ne-

kercenim-te-pergjithshem-per-kosoven



506



Pravica do groba - možnosti pokopavanja na pokopališčih



v Sloveniji



The Right to Grave - Possibilities of Burying in the



Cemeteries in Slovenia



Drago Čepar

Ljubljana, Slovenija

drago.cepar@gmail.com



Povzetek raziskava, ker država ne razpolaga s podatki, ki bi omogočili njeno preverbo. V Sloveniji zakon dopušča pokop v krsti in upepeljevanje



pokojnikov. Enoletna raziskava je pokazala, da je treba zavrniti Vsebina raziskave sodi v temo “pojavi in odnosi v sodobni

ničelno hipotezo, da ni diskriminacije tiste manjšine, ki si želi družini«. Zakaj? Ker so rojstvo, poroka, smrt - vključno z vsemi,

pokopa v krsti. Lokalni predpisi tak pokop ponekod z njimi povezanimi občutji, obredi, obveznostmi - ključne točke

prepovedujejo, onemogočajo ali otežujejo, država pa z zakonom družinskega življenja in družino ter odnose v njej močno

ne varuje pravice do groba. določajo; tudi odločanje o številu otrok, s čimer smo že pri



Ključne besede družine, njenih korenin in vrednot ter odločanje za življenje so demografiji. Spoštovanje, odnos in ljubezen do živih članov

Diskriminacija, družina, pokopališče, pokopavanje, pravica do močno povezani s spoštovanjem mrtvih teles sorodnikov,

groba, upepelitev posebej prednikov, načinom zadnjega slovesa od njih in

odnosom do njihovih grobov kot točk izražanja ljubezni do njih

Abstract in točk spoštljivega srečevanja družinskih članov. Spoštovanje

In Slovenia law allows for cremation and for burying of body in družinskih korenin, povezano z željo, da se rod nadaljuje, in bodo

a coffin. In a one-year research it was shown, that the null vnuki in pravnuki ponosni na svoje dede in pradede ter se bodo

hypothesis, that there is no discrimination of the minority, nanje ljubeče spominjali ob spomenikih ter na grobovih, v

preferring burial in a coffin, has to be rejected. Such burial is in katerih počivajo – to spoštovanje in te želje so težko združljive

some places prohibited, disabled or made difficult by local na primer z odločitvijo, da bi pokojnega družinskega člana

regulations; the state in its legislation does not protect the right poslali v postopek, v katerem mu bodo zdrobili lobanjo, ali da bi

to grave. njegove ostanke »pokopali« z raztrosom, ki za njim zabriše

vsako sled in ne ohranja mesta spomina nanj. Rezultati raziskave

Optional: Keywords so lahko dragocena izhodišča za preučevanje pojavov in odnosov

Discrimination, family, cemetery, burying, right to grave, v sodobni družini, njenih ekonomskih, antropoloških in

cremation socioloških vidikov ter ideoloških vplivov nanjo.



1 Uvod

V Sloveniji zakon dopušča tri oblike pokopa: pokop s krsto, 2 Potek raziskave

pokop z žaro in raztros pepela [1]. Pri prvi obliki gre za pokop Raziskava ima dva dela. V prvem delu sem vsako od 212

telesa, pri drugi in tretji pa najprej za sežig telesa in zdrobitev slovenskih občin 24. 5. 2022 zaprosil za sedem podatkov: število

kosti, potem pa pokop v žari ali raztros. Jedro članka je poročilo vseh pokopališč v občini ter števila pokopališč, na katerih: ni

o raziskavi možnosti pokopa na pokopališčih v Sloveniji, ob mogoče dobiti novega prostora za klasični grob, je pa mogoč

ničelni hipotezi, da ni diskriminacije tiste manjšine, ki si želi klasični pokop v obstoječi grob, če mu je potekla mirovalna

pokopa v krsti. Za testiranje te hipoteze je bila potrebna posebna doba; klasični pokop sploh ni mogoč, opravljajo pa se druge vrste

pokopi; ni mogoče dobiti novega prostora za žarni grob oziroma

nišo, je pa mogoč pokop žare v obstoječi klasični grob, žarni grob

∗Drago Čepar ali nišo; žarni pokop sploh ni mogoč, opravljajo pa se druge vrste

Permission to make digital or hard copies of part or all of this work for personal or pokopi; niso možni pokopi z raztrosom pepela, so pa možni

classroom use is granted without fee provided that copies are not made or distributed drugačni pokopi; ni mogoč nikakršen pokop. for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s). Tridesetim občinam je bilo treba ponovno poslati, ker nisem

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia poslal na pravi naslov. Polovica vseh je odgovorila takoj,

© 2024 Copyright held by the owner/author(s).

http://doi.org/10.70314/is.2024.dem.7 polovica preostalih po drugem pošiljanju; po tretjem 5. 10. 2022

jih je ostalo nekaj deset – od teh sem podatke dobil posamično,



nekaj tudi po telefonu. Ponekod so pred odgovorom želeli Drugi del raziskave je pokazal, da je pravica do groba

podrobnejšo obrazložitev namena raziskave. Izkazalo se je, da ne oziroma, kot pravi Ustavno sodišče, »posmrtna zaščita osebnosti

le država ampak tudi mnoge občine nimajo teh podatkov, saj pokojnika in njegovega dostojanstva ter želja, da bodo njegove

mnogo pokopališč upravljajo krajevne skupnosti. Za taka odločitve, ki so bile sprejete za časa življenja, spoštovane tudi po

pokopališča (kakšna občina jih ima tudi več deset) je bilo treba njegovi smrti« [2] grobo kršena tudi pri pokopu oseb, ki jih

poiskati kontakte krajevnih skupnosti in podatke dobiti od njih, občina pokoplje na svoje stroške. Med 21 nenačrtno

večinoma po telefonu. Konec aprila 2023 sem imel vse pregledanimi občinami - nobena ni mestna - sem v dvanajstih

odgovore. primerih v odloku o pokopališki in pogrebni dejavnost ter

urejanju pokopališč našel določbo o obveznem sežigu; od teh

V drugem delu sem pregledal nekatere občinske odloke o dvanajstih je v šestih primerih določen raztros, v treh žara, v treh

pogrebni in pokopališki dejavnosti in v njih poiskal z ničelno pa raztros ali žara; dve občini sta zapisali vse tri možnosti in ena

hipotezo povezane določbe. od njih daje prednost pokopu v krsti; ena navaja krsto ali žaro;

šest odlokov ne določa ničesar. Od dvanajstih mestnih občin jih

šest določa sežig ( tri raztros, tri pa žaro ali raztros), v šestih pa

to v odloku ni določeno. V tistih šestih, ki določajo sežig, ni

3 Rezultati določbe, da je treba vprašati sorodnike ali upoštevati željo

Prvi del raziskave pove, da je vseh pokopališč 1 231. Število pokojnega; da si bo občina skušala povrniti strošek pokopa iz

pokopališč ene občine se giblje od nič v dveh občinah do 45 v pokojnikove zapuščine, pač. Pregledal sem torej 33, oziroma

eni občini. Žare povsod pokopavajo v klasične grobove, nekatera dobrih 15 odstotkov vseh občinskih odlokov. To je le del celotne

pokopališča pa imajo za žare niše v žarnem zidu, žarna polja, ali populacije 212 občin, za razliko od prvega dela raziskave, v

posamične žarne grobove. Med skupno 212 občinami jih 67 na kateri sem zbral podatke o celotni populaciji pokopališč. Vendar

vseh svojih pokopališčih: zagotavlja pokope v krstah in v žarah; to zadošča za ugotovitev diskriminacije. Koliko te vrste

prostora za raztros nimajo; na voljo pa je prostor tako za najem diskriminacije bi našel, če bi pregledal vse občinske odloke,

novih klasičnih grobov kot tudi žarnih grobov oziroma niš. Na težko ocenim, saj vzorec 33 občin ni bil narejen po pravilih

treh pokopališčih ne pokopavajo v krstah - tudi če je na voljo slučajnega vzorčenja. Ob tem je treba navesti, da so ti odloki v

klasičen grob, ki mu je potekla mirovalna doba. Na 50 konkretnih primerih lahko protizakoniti, saj področni zakon

pokopališčih je sicer pokop v krsti mogoč v grob, ki mu je določa, da se pogrebna slovesnost opravi v skladu z voljo

potekla mirovalna doba, ni pa na voljo novih prostorov za pokojnika in na način, določen s pokopališkim redom, če pa

klasične grobove. Eno pokopališče ima prav zanimivo ureditev. pokojnik ni izrazil svoje volje o načinu pokopa in pogrebni

Parcele za klasične in žarne grobove dobiš zastonj. Prostora je slovesnosti, odloča o tem naročnik pogreba [3].

dovolj za ene in druge. Če zahtevaš nov klasični grob in nimaš Do tu navedena diskriminacija temelji na predpisih, obstaja

svojega (družinskega), grob dobiš. Če pa zahtevaš klasični pokop pa še diskriminacija pri izvajanju sicer nediskriminatornih

in imaš svoj grob, se najprej opravi sondiranje. Če se pokaže, da predpisov. Ko je na papirju mogoče, dejansko pa nemogoče ali

je pokop v ta grob mogoč, se to naredi, če pa ni mogoč, stranka skoraj nemogoče.

ne more dobiti novega prostora, ampak mora dati pokojnika v Gospa, ki mi je potožila, da je bila zelo prizadeta, ker leta

žaro, čeprav je prostora za grobove dovolj. Na drugem 2018 za njeno sestro ni bilo klasičnega groba na domačem

pokopališču imajo na enem delu pravilo, da se v klasični grob, pokopališču in jo je morala dati upepeliti, je iz občine, ki mi je v

ki mu je potekla mirovalna doba, sme v krsti pokopati samo odgovoru na vprašalnik sporočila, da na vseh njihovih

ožjega sorodnika osebe, ki je bila zadnja pokopana. pokopališčih pokopavajo tudi v krstah in da imajo prostor za

nove klasične grobove. Sama je bila takrat govorila z

Pokopališča, kjer ne bi bil mogoč žarni pokop, ni. Na 150 odgovornim pri pogrebnem podjetju in z delavcem na

pokopališčih, sicer ni mogoče dobiti nove niše ali žarnega groba, pokopališču; oba sta povedala, da nove grobne parcele ni. O tem

je pa mogoč pokop žare v klasični grob. Prostor za raztros je neskladju informacij sem obvestil odgovornega pri tem podjetju.

urejen na 169 pokopališčih v 123 občinah in sicer v 89 občinah Pogledal je v kataster grobov in povedal, da so še prosti grobovi,

na enem, v 25 na dveh, v sedmih na treh, v eni na štirih in v eni da pa se bo še pozanimal, ker takrat še ni bil v tej službi. Poskrbel

na petih pokopališčih. Enajst pokopališč je v opuščanju in se na sem, da se je o tem pogovoril neposredno s prizadeto gospo.

njih ne pokopava več. Povedal ji je, da je šele prvo leto na tem mestu, a ima informacije

svojih ljudi in tudi iz katastra grobov je razvidno, da so grobne

parcele sedaj - in so vedno bile - na voljo, in je možno sedaj,

kakor je bilo tudi v preteklosti, dobiti nov grob. Dejansko sta res

4 Zaključki raziskave dva opuščena groba zravnana in v katastru zavedena kot prosta,

Ničelno hipotezo o nediskriminaciji je treba zavrniti. Tisti, ki vendar sta tako utesnjena, da je grob zelo težko izkopati; delavec

si želijo pokopa v krsti, so očitno diskriminirani: na treh bi moral sproti odnašati zemljo, ker ob grobu ni prostora zanjo.

pokopališčih, kjer sploh ni pokopa v krsti; na 50 pokopališčih, ki Tako da se delavci, s katerimi pride naročnik pokopa v stik,

nimajo prostora za nove grobove; na pokopališču, kjer domačin izogibajo kopanju teh grobov in rečejo, da to ni mogoče; med

ne more dobiti novega groba, čeprav prostor je; tam, kjer sme v ljudmi v vasi zato že leta velja, da prostih klasičnih grobov na

krsti sme pokopati samo ožjega sorodnika osebe, ki je bila zadnja tem pokopališču ni. Drugje na primer se delavec – neposredni

pokopana. izvajalec – ob zahtevi za klasičen pokop zgrozi (“a potem bo pa

treba kopat”) in pove, da bo v tem primeru pogreb lahko šele več



dni pozneje kot bi bil žarni. Ko stranka vztraja pri klasičnem 5 Razmislek o vzrokih pokopu, se izkaže, da lahko dobi isti termin kot pri žarnem.

Leta 2022 smo tu opozorili na neverjetno povečanje števila

žarnih pokopov, ki ga ni mogoče pojasniti s podatki s področij

Zaradi večjega napora pri kopanju klasičnega groba, morebiti

gospodarjenja s prostorom, niti z ekonomskimi, etičnimi,

pa tudi iz drugih razlogov, neposredni izvajalci marsikje - v

verskimi in ekološkimi razlogi [4]. Ob ekoloških razlogih (Med

časovni in duševni stiski sorodnikov umrlega marsikdaj uspešno

upepelitvijo človeškega telesa nastane na primer 160, po

- prepričujejo naročnike, naj se odločijo za upepelitev. Na papirju

nekaterih virih pa celo več kg toplogrednega plina CO2[5]), hudi

torej lahko dobiš grob, vendar Bog ne daj, da bi vztrajal, da ga



res izkopljejo. razlogi govorijo v prid klasičnemu pokopu v krsti. Na koncu smo energetski potratnosti upepeljevanja, posebej etični in verski



Raziskava torej ugotavlja diskriminacijo naročnikov pokojnikov. ostali brez razlage tako hitre rasti števila in deleža upepelitev

klasičnih pokopov tako s predpisi kot tudi z izvajanjem

predpisov:

Po ugotovitvi, da lokalne in državne oblasti diskriminirajo

- s krajevnimi predpisi na več deset pokopališčih, ki

tiste, ki se odločajo za pokop v krsti, in torej pospešujejo

prepovedujejo klasičen pokop v krsti ali ne

upepeljevanje, se še ostreje zastavlja vprašanje o pravih vzrokih

zagotavljajo prostora za nov klasični grob,

tega pospeševanja ter skupinah ali osebah, ki stojijo za njim.

- z občinskimi predpisi v mnogih občinah, ki z odlokom



določajo upepelitev oseb, ki jih občina pokoplje na Zahvala

svoje stroške,

- Zahvaljujem se več sto prijaznim ljudem, ki so mi posredovali z državnimi predpisi, saj področni zakon ne

prepoveduje navedene diskriminacije na krajevni in dragocene podatke in izkušnje ter mi potrpežljivo razložili, kar

občinski ravni, ne zagotavlja enakih možnosti za je bilo treba.

izbiro načina pokopa in na državni ravni ne določa

zbiraja podatkov o možnostih pokopa, References / Literatura

- [1] Zakon o pogrebni in pokopališki dejavnosti (ZPPDej), UL RS, št. 62/16 in s konkretnim izvajanjem sicer nediskriminatornih

predpisov, ko uslužbenci pogrebnih podjetij ob 3/22 – ZDeb , 17. člen.

[2] Odločba US RS št. U-I-223/16 z dne 23. 4. 2020.

naročanju pogreba odvračajo naročnike od izbire [3] Zakon o pogrebni in pokopališki dejavnosti (ZPPDej), UL RS, št. 62/16 in

klasičnega pokopa. 3/22 – ZDeb, 14. člen.

[4] Drago Čepar, 2022, POLITIČNA KOREKTNOST IN SPOŠTOVANJE

TELES UMRLIH, Zbornik 25. mednarodne multikonference

INFORMACIJSKA DRUŽBA-IS 2022, zvezek F, Demografske in

družinske analize, J. Malačič, M.J. Osredkar, M. Gams (ur.), str. 15-17,

Ljubljana, (https://is.ijs.si/?page_id=15435).

[5] I. Zobavnik, M. Pristavec Đogić, Upepeljevanje pokojnikov in sežiganje

biološkega materiala, str. 7, DZRS, 2016





Ocena bremena telesne neaktivnosti v Sloveniji




Physical inactivity burden assessment in Slovenia



Daša Farčnik† Petra Došenović Bonča Marko Pahor Maja Zalaznik

School of Economics and School of Economics and School of Economics and School of Economics and

Business Business Business Business

University of Ljubljana University of Ljubljana University of Ljubljana University of Ljubljana

Ljubljana, Slovenia Ljubljana, Slovenia Ljubljana, Slovenia Ljubljana, Slovenia

dasa.farcnik@ef.uni-lj.si petra.d.bonca@ef.uni-lj.si marko.pahor@ef.uni-lj.si maja.zalaznik@ef.uni-lj.si



Daša Pruš Armin Paravlić Janez Vodičar

Faculty of Sport Faculty of Sport Faculty of Sport

University of Ljubljana University of Ljubljana University of Ljubljana

Ljubljana, Slovenia Ljubljana, Slovenia Ljubljana, Slovenia

dasa.prus@fsp.uni-lj.si armin.paravlic@fsp.uni-lj.si janez.vodicar@fsp.uni-lj.si



Povzetek between 13,300 and 17,600 cardiovascular patients, between



Prispevek ocenjuje breme telesne neaktivnosti v Sloveniji v letu patients. Furthermore, we estimate that in 2019, physical 7,000 and 13,000 diabetic patients, and around 1,500 dementia

2019, in sicer na podlagi epidemiološkega pristopa osnovanega inactivity and consequently cardiovascular diseases, diabetes,

na populaciji, in sicer na podlagi relativnega tveganja prevalence and dementia resulted in the deaths of between 188 and 252

treh vrst bolezni (srčno-žilnih bolezni, sladkorne bolezni in individuals, most due to cardiovascular diseases. Due to physical

demence) zaradi telesne neaktivnosti in populacijskega inactivity, a total of 156,036 disability adjusted life years were

pripisljivega deleža. Na podlagi tega pristopa ocenjujemo, da je lost due to cardiovascular diseases, diabetes, and dementia.

bilo zaradi telesne neaktivnosti v Sloveniji leta 2019 med 13 in



17,5 tisoč srčno-žilnih obravnavanih pacientov, med 7 in 13 tisoč Keywords obravnavanih sladkornih pacientov in okrog 1.500 obravnavanih

pacientov z demenco. Nadaljnje ocenjujemo, da je v letu 2019 Physical inactivity, burden of physical inactivity, mortality

zaradi telesne neaktivnosti in posledično srčno-žilnih bolezni,

sladkorne bolezni in demence umrlo med 188 in 440 oseb, največ

1 Uvod

zaradi srčno-žilnih bolezni. Zaradi telesne neaktivnosti in

posledično srčno-žilnih bolezni, sladkorne bolezni in demence je Redna in dovolj intenzivna telesna aktivnost je ključna za

bilo izgubljenih 156.036 zdravih let življenja. ohranjanje dobrega zdravja in dolgoživosti. Raziskave jasno

kažejo, da telesna aktivnost znatno znižuje verjetnost za

Ključne besede predčasno smrt in deluje kot preventivni ukrep proti številnim

kroničnim nenalezljivim boleznim [1], [2]. Po podatkih Svetovne

Telesna neaktivnost, breme telesne neaktivnosti, umrljivost

zdravstvene organizacije je telesna neaktivnost četrti vodilni

Abstract dejavnik tveganja za nenalezljive bolezni po vsem svetu in naj bi

letno povzročila med 3,2 in 5 milijonov smrti na globalni ravni

This paper assesses the burden of physical inactivity in Slovenia [3]. V Sloveniji po podatkih Svetovne zdravstvene organizacije

in 2019 based on an epidemiological top-down approach, 73% odraslih dosega priporočene ravni telesne aktivnosti, ki

grounded on the population and thus a prevalence approach using vključujejo vsaj 150 minut zmerne intenzivne aerobne dejavnosti

population attributable fraction. Using this method, we estimate ali 75 minut visoko intenzivne dejavnosti na teden, 27% pa ne

that due to physical inactivity in Slovenia in 2019, there were [4]. Temu vsled se povečuje tveganje za nenalezljive bolezni,

predvsem za srčno-žilne bolezni, sladkorno bolezen, različne

Permission to make digital or hard copies of part or all of this work for personal or vrste raka, debelost, hipertenzijo, osteoporozo, osteoartritis,

classroom use is granted without fee provided that copies are not made or distributed depresijo in kognitivni upad. for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must V tem prispevku se zato osredotočamo na tri vrste bolezni, ki

be honored. For all other uses, contact the owner/author(s). so posledica telesne neaktivnosti: izbrane srčno-žilne bolezni,

© 2024 Copyright held by the owner/author(s). sladkorno bolezen in demenco. Namen prispevka je oceniti Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

breme telesne neaktivnosti, ki izhaja iz bremena teh treh

navedenih bolezni. Tako na podlagi relativnega tveganja za



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Farčnik et al.



nastanek posamezne vrste bolezni izračunamo populacijski Nato je izračunan pripisljiv delež med izpostavljenimi (v

pripisljivi delež, ki ga uporabimo za oceno prevalence teh nadaljevanju PDI), in sicer z uporabo enačbe 3, ki temelji na

bolezni zaradi telesne neaktivnosti ter oceno umrljivosti in razliki med verjetnostjo (P) tveganja bolezni med

izgubljenih zdravih let življenja. izpostavljenimi in verjetnostjo tveganja bolezni med

neizpostavljenimi, glede na tveganje bolezni med

izpostavljenimi.

2 Metodologija in podatki



Breme telesne neaktivnosti je ocenjeno na podlagi pristopa, ki 𝑃(𝐵|𝐴)−𝑃(𝐵|𝐴) ̅̅̅ PDI - 𝑝𝑟𝑖𝑝𝑖𝑠𝑙𝑗𝑖𝑣 𝑑𝑒𝑙𝑒ž 𝑚𝑒𝑑 𝑖𝑧𝑝𝑜𝑠𝑡𝑎𝑣𝑙𝑗𝑒𝑛𝑖𝑚𝑖 = (3) 𝑃(

𝐵|𝐴)

posnema že obstoječe poskuse ocene bremena bolezni [5], [6].



Tako uporabimo epidemiološki pristop, pristop od zgoraj Na podlagi izračunov IP in PDI za srčno-žilne bolezni in



approach) in prevalenčni pristop. Pristop temelji na naslednjih za srčno-žilne bolezni znaša 4,6 %, za sladkorno bolezen pa 8,3 navzdol, ki je osnovan na populaciji (angl. population-based sladkorno bolezen je izračunan PAF na podlagi enačbe 4. PAF

korakih: 1) Ocenjena je prevalenca telesne neaktivnosti; 2)

%. Izračun PAF za demenco na podlagi razpoložljivih podatkov

Ocenjeno je relativno tveganje za nastanek bolezni zaradi telesne ni bil mogoč, zato je uporabljen zgolj PAF iz literature. neaktivnosti. Relativno tveganje je razmerje prevalence (števila

primerov bolezni v opazovanem obdobju) med izpostavljenimi 𝑃𝐴𝐹 = 𝐼𝑃 ∙ 𝑃𝐷𝐼 (4) in neizpostavljenimi (npr. tveganje za opazovani dogodek v



testni skupini/tveganje za opazovani dogodek v kontrolni Podatki o telesni neaktivnosti izhajajo iz Nacionalne

skupini). Izpostavljena skupina je tista skupina, ki ni bila dovolj raziskave o EHIS, ki je bila nazadnje izvedena v letu 2019.

telesno aktivna – to so telesno neaktivni posamezniki. Anketo je izvedel NIJZ na reprezentativnem vzorcu

Neizpostavljena skupina je tista skupina, ki je bila dovolj telesno posameznikov starih 15 let in več, in sicer 9.900 oseb. Eden

aktivna – to so telesno aktivni posamezniki. 3) Izračunan je izmed sklopov EHIS ankete se nanaša tudi na telesno

populacijski pripisljivi delež (angl. PAF – population attributable aktivnost/gibanje. Telesna aktivnost in neaktivnost je definirana

faction). PAF predstavlja delež, ki ga je mogoče pripisati na podlagi vprašanja: »Ali ste dnevno povprečno telesno dejavni

populaciji. Meri sorazmerno zmanjšanje populacijske bolezni ali vsaj 30 minut ali skupno 150 minut na teden? [Upoštevajte

umrljivosti, ki bi nastala, če bi izpostavljenost dejavniku tveganja aktivnosti pri delu, vsakodnevnih opravilih in v prostem času.]«

zmanjšali na alternativni idealni scenarij izpostavljenosti [5], [7]. z možnim odgovorom DA ali NE. Ker meja 150 minut na teden

4) Na podlagi PAF je ocenjena prevalenca bolezni zaradi telesne ustreza priporočilom Svetovne zdravstvene organizacije je za

neaktivnosti, število umrlih in izgubljena zdrava leta življenja. potrebe tega prispevka telesna neaktivnost definirana z

V oceni je upoštevano relativno tveganje (RR) iz dveh virov. odgovorom NE na to vprašanje. Slika 1 prikazuje delež telesno

Prvi vir relativno tveganje povzema po obstoječih študijah [5], neaktivnih po starostnih razredih in po spolu v letu 2019.

drugi vir pa so podatki Nacionalnega inštituta za javno zdravje

(v nadaljevanju NIJZ), in sicer iz Nacionalne raziskave o zdravju

in zdravstvenem varstvu (European Health Interview Survey – 45,0

EHIS, v nadaljevanju EHIS). Temu vsled uporabimo in 40,0

izračunamo tudi dva različna populacijska pripisljiva deleža. [5] 35,0

30,0

poroča o relativnem tveganju za srčno-žilne bolezni v višini 0,75 25,0

ter za sladkorno bolezen v višini 0,61. Na podlagi relativnega 20,0

tveganja in deleža neaktivnih in uporabi formule (1) poročajo o 15,0

populacijskem pripisljivem deležu - PAF za Slovenijo kot: 3,5% 10,0

za srčno-žilne bolezni ter 4,4% za sladkorno bolezen. Demenca 5,0

v študiji ni upoštevana, zato je uporabljen PAF iz študije [8] in 0,0

15-24 25-34 35-44 45-54 55-64 65-74 75-84 85-94

znaša 4,5%.

𝑃𝐴𝐹 = 𝑃𝑒(𝑅𝑅−1) Moški Ženske ∙ 100 , 𝑃 = 𝑑𝑒𝑙𝑒ž 𝑡𝑒𝑙𝑒𝑠𝑛𝑜 𝑛𝑒𝑎𝑘𝑡𝑖𝑣𝑛𝑖ℎ (1)

𝑃 𝑒 𝑒

(𝑅𝑅−1)+1



Na podlagi podatkov NIJZ pa smo po vzoru [9] izračunali Slika 1: Delež telesno neaktivnih (v %) po starostnih

delež izpostavljenih primerov (v nadaljevanju IP), in sicer z kategorijah, 2019 uporabo enačbe 2. Delež izpostavljenih primerov primerja dve

množici (skupini ljudi), in sicer tisto, ki je bila izpostavljena Osredotočamo se na tri vrste bolezni, ki so posledica telesne

dejavniku tveganja (v tem primeru telesni neaktivnosti) in je neaktivnosti: izbrane srčno-žilne bolezni, sladkorno bolezen in

obolela, to je množica A, in tisto, ki ni bila izpostavljena demenco. Natančneje, izbrane srčno-žilne bolezni (s kodami po

dejavniku tveganja (niso bili telesno neaktivni) in so vseeno Mednarodni klasifikaciji bolezni MHB-10-AB, verzija 6):

oboleli, to je množica B. IP je izračunan posebej za srčno-žilne Esencialna (primarna) arterijska hipertenzija (I10),

bolezni in sladkorno bolezen. Hipertenzivna bolezen srca (I11), Angina pektoris (I20), Akutni

𝐼𝑃 = 𝑑𝑒𝑙𝑒ž 𝑖𝑧𝑝𝑜𝑠𝑡𝑎𝑣𝑙𝑗𝑒𝑛𝑖ℎ 𝑝𝑟𝑖𝑚𝑒𝑟𝑜𝑣 = 𝑛(𝐴∩𝐵) (2) miokardni infarkt (I21); Naslednji miokardni infarkt (I22);

𝑛(𝐵) Kronična ishemična bolezen srca (I25); Ateroskleroza (I70) ter

𝐴 … 𝑚𝑛𝑜ž𝑖𝑐𝑎, 𝑘𝑖 𝑗𝑒 𝑖𝑧𝑝𝑜𝑠𝑡𝑎𝑣𝑙𝑗𝑒𝑛𝑎 𝑑𝑒𝑗𝑎𝑣𝑛𝑖𝑘𝑢 𝑡𝑣𝑒𝑔𝑎𝑛𝑗𝑎 𝑖𝑛 𝑗𝑒 𝑜𝑏𝑜𝑙𝑒𝑙𝑎

𝐵 … 𝑚𝑛𝑜ž𝑖𝑐𝑎, 𝑘𝑖 𝑛𝑖 𝑖𝑧𝑝𝑜𝑠𝑡𝑎𝑣𝑙𝑗𝑒𝑛𝑎 𝑑𝑒𝑗𝑎𝑣𝑛𝑖𝑘𝑢 𝑡𝑣𝑒𝑔𝑎𝑛𝑗𝑎 𝑖𝑛 𝑗𝑒 𝑜𝑏𝑜𝑙𝑒𝑙𝑎 Ateroskleroza arterij udov (I70.2) ter Srčno popuščanje (I50). Za

𝑛 … š𝑡𝑒𝑣𝑖𝑙𝑜 𝑝𝑜𝑠𝑎𝑚𝑒𝑧𝑛𝑖𝑘𝑜𝑣 demenco so upoštevane naslednje bolezni: Demenca pri

Alzheimerjevi bolezni (F00); Vaskularna demenca (F01);



Ocena bremena telesne neaktivnosti v Sloveniji Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



Demenca pri drugih boleznih (F02); Neopredeljena demenca sladkorne bolezni in demence je v letu 2019 umrlo med 188 in

(F03); Alzheimerjeva bolezen (G30). Za sladkorno bolezen pa: 252 posameznikov. To sicer predstavlja en oziroma dva odstotka

Sladkorna bolezen tipa 1 (E10); Sladkorna bolezen tipa 2 (E11); vseh umrlih, vendar zaradi telesne neaktivnosti in posledičnih

Druge vrste sladkorna bolezen (diabetes) (E13); Sladkorna bolezni nastajajo znatni stroški [13]. Ti vključujejo neposredne

bolezen neopredeljenega tipa (E14). in posredne stroške, med njimi neposredne zdravstvene in

Podatki o številu obravnavanih bolnikov izhajajo iz nezdravstvene stroške ter posredne stroške, ki so posledica izgub

naslednjih virov: za srčno-žilne bolezni [10], za sladkorne produktivnosti zaradi bolniškega staleža (absentizma),

bolnike [11], za demenco pa [12]. Vsi podatki so za leto 2019, prezentizma, pa tudi prezgodnje umrljivosti, prezgodnjega

razen za demenco, za katero so podatki za leto 2018. Podatki o upokojevanja, delovne neaktivnosti in stroškov podjetij, ki

številu umrlih so pridobljeni s strani NIJZ, o izgubljenih morajo zaradi izgube zaposlenih iskati in usposabljati nove

zdravih letih življenja pa iz Inštituta za metrike in ocenjevanje zaposlene.

zdravja ("Institute for Health Metrics and Evaluation"- IHME). Čeprav Slovenija velja za športen narod, podatki kljub temu

Inštitut poroča o izgubljenih zdravih letih življenja za vse srčno- kažejo, da se delež telesno neaktivnih med starostnimi razredi

žilne bolezni (ne samo za izbrane in predstavljene predhodno), razlikuje in je visok predvsem pri starejših. Zato prispevek

za diabetes tipa 2 ter Alzheimerjevo bolezen in druge oblike ponuja tudi nekaj vhodnih informacij za snovalce nacionalnih

demenc. preventivnih programov, v sinergiji z dosedanjimi aktivnostmi

krovnih institucij, ki pripravljajo programe in politike

športno/gibalnih aktivnosti.

3 Rezultati

Ocena števila bolnikov izbranih bolezni zaradi telesne

neaktivnosti kaže, da se število bolnikov razlikuje od Zahvala

populacijskega pripisljivega deleža (Tabela 1). Ocenjujemo, da Prispevek je financiran s strani Javne agencije za raziskovalno

je bila večina pacientov obravnavanih zaradi izbranih srčno- dejavnost Republike Slovenije št. V5-2246. Za posredovane

žilnih bolezni, in sicer med 13.300 in 17.600 bolnikov (ponovno podatke se iskreno zahvaljujemo Nacionalnemu inštitutu za

odvisno po uporabljenega populacijskega pripisljivega deleža), javno zdravje.

število obravnavanih pacientov zaradi sladkorne bolezni je bilo

med slabih 7 tisoč in 13 tisoč, obravnavanih pacientov z demenco References / Literatura

pa okrog 1.500. [1] M. Nelson et al., ‘Physical Activity and Public Health in Older Adults:

Podobno je tudi ocena števila umrlih zaradi telesne Recommendation From the American College of Sports Medicine and the

neaktivnosti odvisna od populacijskega pripisljivega deleža. American Heart Association’, Circulation, vol. 116, no. 9, pp. 1094–1105,

Aug. 2007.

Ocenjujemo, da je bilo število umrlih zaradi izbranih bolezni, ki [2] W. Haskell et al., ‘Physical Activity and Public Health: Updated

so posledica telesne neaktivnosti, med 188 in 252, največ zaradi Recommendation for Adults From the American College of Sports

Medicine and the American Heart Association’, Circulation, vol. 116, no.

srčno-žilnih bolezni. Ocenjujemo, da je bilo število umrlih zaradi 9, pp. 1081–1093, Aug. 2007.



telesne neaktivnosti, ki so umrli zaradi sladkorne bolezni precej [3] WHO, ‘Global Health Estimates: Life expectancy and leading causes of death and disability’. Accessed: Sep. 18, 2023. [Online]. Available: nižje, in sicer med 17 in 31 umrlimi, podobno tudi v primeru https://www.who.int/data/gho/data/themes/mortality-and-global-health-

demence, ko je bilo število umrlih zaradi telesne neaktivnosti 16. estimates

Število izgubljenih zdravih let življenja pa je zaradi srčno- [4] WHO, ‘Global status report on physical activity 2022: country profiles’.

Accessed: Sep. 03, 2024. [Online]. Available:

žilnih bolezni znašalo 115 tisoč, dodatno še dobrih 20 tisoč zaradi https://www.who.int/publications/i/item/9789240064119



sladkorne bolezni in 20.200 zaradi demence, skupaj torej 156 [5] D. Ding et al., ‘The economic burden of physical inactivity: a global analysis of major non-communicable diseases’, The Lancet , vol. 388, no. tisoč let življenja. 10051, pp. 1311–1324, Sep. 2016, doi: 10.1016/S0140-6736(16)30383-X.

[6] A. C. Santos, J. Willumsen, F. Meheus, A. Ilbawi, and F. C. Bull, ‘The

cost of inaction on physical inactivity to public health-care systems: a

Tabela 1: Ocena bremena telesne neaktivnosti population-attributable fraction analysis’, Lancet Glob. Health, vol. 11,

no. 1, pp. e32–e39, Jan. 2023, doi: 10.1016/S2214-109X(22)00464-8.

[7] I.-M. Lee, E. J. Shiroma, F. Lobelo, P. Puska, S. N. Blair, and P. T.



Število obravnavanih Katzmarzyk, ‘Effect of physical inactivity on major non-communicable Število umrlih DALYs pacientov diseases worldwide: an analysis of burden of disease and life expectancy’,

Bolezni PAF1 The Lancet, vol. 380, no. 9838, pp. 219–229, Jul. 2012, doi: PAF2 PAF1 PAF2

10.1016/S0140-6736(12)61031-9.



Izbrane srčno- [8] R. C. Vergara, P. Zitko, A. Slachevsky, C. San Martin, and C. Delgado, 155 204 115.062 ‘Population attributable fraction of modifiable risk factors for dementia in žilne bolezni 13.340 17.608 Chile’, Alzheimers Dement. Diagn. Assess. Dis. Monit. , vol. 14, no. 1, p. Sladkorna 6.894 12.967 17 31 20.680 e12273, Feb. 2022, doi: 10.1002/dad2.12273. bolezen [9] B. Siegerink and J. L. Rohmann, ‘Impact of your results: Beyond the

Demenca 1.536 16 20.294 relative risk’, Res. Pract. Thromb. Haemost., vol. 2, no. 4, pp. 653–657,

Oct. 2018, doi: 10.1002/rth2.12148.

Skupaj 188 252 156.036 [10] K. Janša and D. Gavrić, ‘Kazalniki zdravstvenih storitev bolezni obtočil.

Poročilo za obdobje 2019-2023’. Zavod za zdravstveno zavarovanje

Slovenije, 2024.

[11] P. D. Bonča, D. Gavrić, K. Janša, and J. Sambt, ‘Direct diabetes-related

4 Zaključek healthcare expenditures in Slovenia: recent evolution and future

projections based on population-level data’, Expert Rev. Pharmacoecon.

V prispevku ugotavljamo, da zaradi telesne neaktivnosti v Outcomes Res., vol. 24, no. 3, pp. 427–436, Mar. 2024, doi:

Sloveniji nastaja breme, ki se med drugim odraža v številu 10.1080/14737167.2024.2302423.

[12] J. Georges, C. Bintener, and O. Miller, ‘Dementia in Europe Yearbook

obolelih, umrlih in izgubljenih zdravih letih življenja. Na primer, 2019: Estimating the prevalence of dementia in Europe | Alzheimer

zaradi telesne neaktivnosti in posledičnih srčno-žilnih bolezni, Europe’, Jun. 2024. Accessed: Sep. 03, 2024. [Online]. Available:



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Farčnik et al.



https://www.alzheimer-europe.org/resources/publications/dementia- [13] D. Farčnik et al., ‘Vpliv športne/gibalne aktivnosti na ekonomske in europe-yearbook-2019-estimating-prevalence-dementia-europe družbene stroške zdravja’. Končno poročilo projekta, Ekonomska

fakulteta, 2024.



Cohort Fertility Trends in Post-Yugoslav States1



Ivan Čipin

Department of Demography

Faculty of Economics & Business, University of Zagreb

icipin@efzg.hr



Abstract



This study provides an in-depth analysis of cohort fertility, parity progression, and family size in the

former Yugoslav countries, which include Bosnia-Herzegovina, Croatia, Kosovo, Montenegro, North

Macedonia, Serbia, and Slovenia. Though politically united for much of the 20th century, these

countries exhibit diverse fertility patterns due to their distinct socio-economic and cultural histories.

The analysis utilizes census data to compute completed cohort fertility rates (CFR) and parity

progression ratios (PPR), revealing the different paces of fertility decline. Slovenia, Croatia, and Serbia

form a low-fertility cluster, with cohort fertility stabilizing at around two children per woman by the

1930 cohort. Bosnia-Herzegovina, North Macedonia, and Montenegro constitute a high-fertility cluster

where the fertility transition occurred more gradually. Kosovo remains unique with persistently high

fertility, though it has experienced substantial declines in recent cohorts. The findings highlight how

socio-economic transformations, cultural norms, and historical legacies have shaped fertility

behaviours in each country, providing insights into long-term fertility trends.



Keywords: Cohort fertility, parity progression, family size, former Yugoslavia



Introduction



The former Yugoslav countries are an interesting case study for demographic analysis due to their

shared historical and political past, yet vastly different socio-cultural trajectories. The region has

experienced a complex process of demographic transition, influenced by varying degrees of

industrialization, urbanization, and shifts in family norms and gender roles. Yugoslavia’s political

dissolution in the 1990s further differentiated the trajectories of its successor states, as each country

embarked on distinct political and economic paths. Fertility decline, an integral component of

demographic transition, occurred at different pace across the region, influenced by both economic

factors and deeply rooted cultural traditions (Breznik, 1991; Rašević, 1971).



The countries that made up Yugoslavia were heterogeneous in terms of religion, ethnic composition,

and socio-economic development. For example, Slovenia, with its history of Austro-Hungarian

influence, had long experienced lower fertility rates and higher economic development compared to

Kosovo, which was marked by a deeply agrarian, patriarchal society (Erlich, 1966). The question this

study addresses is whether the fertility behaviour of the former Yugoslav countries has converged or

if the region continues to exhibit significant diversity in cohort fertility trends. This analysis is important



1 This presentation draws on research previously published in Comparative Population Studies,

presenting key insights from the article Cohort Fertility, Parity Progression, and Family Size in Former

Yugoslav Countries by Ivan Čipin, Kryštof Zeman, and Petra Međimurec.



for understanding the demographic evolution of these countries in the context of their socio-economic

and cultural histories.



Fertility in Context: Socioeconomic and Cultural Influences



The fertility transition in the former Yugoslav countries was influenced by economic, cultural, religious,

and institutional factors, leading to regional differences in fertility patterns. Yugoslavia’s complex

historical and political background shaped the socio-economic trajectories of its republics, affecting

their fertility behaviours. Yugoslavia was historically split between Western (Austro-Hungarian) and

Eastern (Ottoman) cultural spheres, influencing marriage patterns and family structures. Slovenia and

Croatia, with more Western European influences, transitioned to lower fertility rates earlier due to

urbanization, industrialization, and expanded education for women. By contrast, Bosnia-Herzegovina

and Kosovo, with strong Ottoman legacies, maintained traditional patriarchal family structures for

longer, resulting in higher fertility (Erlich, 1966; Breznik, 1991). Religious diversity across the region

also shaped fertility norms. Catholic regions like Croatia and Slovenia experienced earlier fertility

decline, in part due to more progressive attitudes toward birth control, while predominantly Muslim

Kosovo adhered to traditional norms, leading to higher fertility rates (Rašević, 1971; Breznik, 1991).

Women's education and labour force participation were critical factors in fertility change. In Slovenia

and Croatia, higher education led to delayed marriage and smaller family sizes, while in Bosnia-

Herzegovina and Kosovo, lower educational attainment for women delayed fertility transition

(Woodward, 1985; Rašević, 1971).



The socialist era brought further changes, promoting gender equality and female workforce

participation, particularly in urban areas, which contributed to declining fertility. However, traditional

values persisted in rural areas, slowing this transition. Institutional factors also played a role. Unlike

many socialist states, Yugoslavia did not implement strong pronatalist policies and allowed for regional

variation in family planning. Slovenia facilitated access to contraception and abortion, while Kosovo,

influenced by its traditional values, maintained higher fertility levels. (Breznik, 1991; Drezgić, 2010).

The economic instability of the 1990s further reduced fertility as large families became less viable

(Calic, 2019; Kapor-Stanulović & David, 1999).



Overall, the fertility transition in former Yugoslav countries were shaped by a combination of socio-

economic modernization, cultural norms, and institutional policies. While modernization spurred

fertility decline in some regions, strong religious and cultural traditions delayed it in others, reflecting

the complex socio-political dynamics of the region.



Data and Methods



The study analyses census data covering women born between 1868 and 1973, offering a

comprehensive look at fertility trends over more than a century. Completed cohort fertility rates (CFR)

and parity progression ratios (PPR) were calculated for the seven successor states of Yugoslavia. The

census data used for this analysis span multiple decades, from the first post-war census in 1948 to the

censuses conducted in the 2010s. The CFR measures the average number of children born to women

by the end of their reproductive years, while the PPR shows proportion of women who progress from

one parity to the next. These indicators provide a detailed understanding of how family size and fertility

behaviour have evolved in the region.



To examine the similarities and differences between the countries, the study also analyses the parity

composition of families, looking at the proportion of women with different numbers of children. The

results are grouped into clusters based on similar fertility patterns. This methodological approach



allows for the identification of key trends in fertility decline and provides a basis for understanding the

factors contributing to these patterns.



Results



The completed cohort fertility of women born between 1868 and 1973 in the former Yugoslav

republics shows high fertility rates before the demographic transition. In the 1870 cohort, fertility

ranged from 4.3 to 6.0 children per woman. Serbia, Slovenia, and Croatia experienced early fertility

declines, while Bosnia-Herzegovina, North Macedonia, and Montenegro maintained higher fertility

until around 1900 cohort. By the 1930 cohort, fertility in Serbia, Croatia, and Slovenia had dropped

below 2 children per woman, while Bosnia-Herzegovina, North Macedonia, and Montenegro had

declined to around 3.5 children per woman. Kosovo, however, retained high fertility levels at 5.7

children per woman.



The analysis reveals three distinct clusters of cohort fertility behaviour. Croatia, Slovenia, and Serbia

represent a low-fertility cluster, where fertility declined rapidly during the first half of the 20th century

and stabilized at around two children per woman by the 1930 cohort. This early fertility decline is

attributed to factors such as industrialization, urbanization, rising female education, and changing

family norms. For instance, Slovenia, the most economically advanced of the Yugoslav republics,

exhibited fertility patterns similar to those seen in Western Europe, with high levels of childlessness

and smaller family sizes becoming the norm by the mid-20th century.



The second cluster consists of Bosnia-Herzegovina, North Macedonia, and Montenegro, where fertility

remained high until the mid-20th century but began to decline more gradually. These countries saw a

slower fertility transition due to their more traditional social structures and slower socio-economic

modernization. In these regions, large families remained more common, and the fertility decline was

less pronounced until the 1960s and 1970s cohorts. The persistence of higher parity progression ratios

to third and subsequent children in these countries illustrates the slower pace of change compared to

the low-fertility cluster.



Kosovo, meanwhile, stands out as a unique case with persistently high cohort fertility throughout much

of the 20th century. Fertility in Kosovo remained at around six children per woman until the 1940

cohort, only beginning to decline significantly in later cohorts. By the 1970 cohort, fertility had fallen

to just under three children per woman, but this remains exceptionally high compared to other

European countries. The high fertility in Kosovo is linked to its agrarian economy, patriarchal family

structures, and lower levels of female education and labour force participation (Breznik, 1991).

Traditional family norms and the strong influence of religion contributed to the slower adoption of

birth control and smaller family sizes in Kosovo.



Conclusion



The findings of this study highlight the significant diversity in fertility patterns across the former

Yugoslav countries, underscoring the importance of historical, socio-economic, and cultural contexts

in shaping fertility behaviour. The clustering of countries into three distinct groups—low-fertility, high-

fertility, and Kosovo as an outlier—demonstrates how varying levels of modernization, economic

development, and cultural traditions have influenced cohort fertility trends over the past century.

While some convergence in fertility behaviour is evident, particularly in the gradual decline of fertility

in the high-fertility cluster, significant differences remain.



Slovenia, Croatia, and Serbia transitioned to below-replacement fertility levels earlier due to socio-

economic modernization and changing family norms. In contrast, Bosnia-Herzegovina, North



Macedonia, and Montenegro experienced a more gradual decline, influenced by cultural factors that

supported larger families for a longer period. Kosovo, with its persistently high fertility rates,

represents a case where traditional family structures and slow socio-economic development delayed

the fertility transition.



These results contribute to the broader understanding of fertility transition in the region and suggest

that while economic and structural changes play a significant role in fertility decline, cultural and

religious factors can slow or accelerate this process. The study provides a foundation for further

research into the complex interplay of socio-economic and cultural factors that continue to shape

fertility behaviour in the former Yugoslav countries.



References



Breznik, D. (1991) Stanovništvo Jugoslavije. Titograd: Chronos, Konzorcijum instituta društvenih nauka.



Calic, M. J. (2019) A History of Yugoslavia. Central European Studies. West Lafayette, Indiana: Purdue

University Press.



Drezgić, R. (2010) Policies and practices of fertility control under the state socialism. In: The History of

the Family 15,2: 191-205 [https://doi.org/10.1016/j.hisfam.2009.11.001].



Erlich, V. (1966) Family in Transition: A Study of 300 Yugoslav Villages. Princeton, NJ: Princeton

University Press.



Kapor-Stanulović, N.; David, Henry P. (1999) Former Yugoslavia and Successor States. In: David, Henry

P. (Ed.): From Abortion to Contraception: Resources to Public Policies and Reproductive Behavior in

Central and Eastern Europe from 1917 to the Present. Westport, CT: Greenwood Press: 279-315.



Rašević, M. (1971) Determinante fertiliteta stanovništva u Jugoslaviji. Beograd: Institut društvenih

nauka, Centar za demografska istraživanja.



Sklar, J. L. (1974) The Role of Marriage Behavior in the Demographic Transition: The Case of Eastern

Europe around 1900. Population Studies 28,2: 231-247.



Woodward, S. L. (1995) Socialist Unemployment: The Political Economy of Yugoslavia 1945-1990.

Princeton, NJ: Princeton University Press.



Pregled planetarnih in lokalnih demografskih projekcij



Overview of Demographic Global and Local Projections



Matjaž Gams †

Inteligentni sistemi

Jozef Stefan Institute

Ljubljana, Slovenija

Matjaz.gams@ijs.si



1 Uvod

Abstract / Povzetek



Prispevek obravnava globalne in lokalne demografske Demografske spremembe predstavljajo enega ključnih

spremembe, ki vključujejo zmanjševanje rodnosti, staranje izzivov današnjega časa, saj se svet sooča s prehodom iz obdobja

prebivalstva in upad delovno aktivnih prebivalcev. Globalna rast eksponentne rasti prebivalstva v fazo stagnacije in potencialnega

prebivalstva se je skoraj ustavila, z izjemo Afrike, kjer se rodnost upadanja. Razvite države, vključno s Slovenijo, že beležijo

sicer še povečuje, a tudi tam počasi upada. Evropa, vključno s negativne demografske trende, ki se kažejo v nizki rodnosti,

Slovenijo, se sooča s stagnacijo in celo upadom prebivalstva, kar staranju prebivalstva in zmanjšanju delovno aktivnega

povzroča dolgoročne socialne in gospodarske izzive. Prispevek prebivalstva. Takšne spremembe vplivajo na vse družbene

poudarja potrebo po strateških rešitvah, kot so spodbujanje podsisteme, od gospodarstva do zdravstva in socialnih sistemov

rodnosti in ustrezna migracijska politika, da bi preprečili [1]. Po podatkih Rimskega kluba bo svetovno prebivalstvo v

dolgoročno demografsko krizo in ohranili stabilno rast naslednjih desetletjih verjetno upadlo na šest milijard, kar

prebivalstva. prinaša nove izzive na globalni ravni [2].

Demografski premiki so rezultat kompleksnih dejavnikov,

Ključne besede kot so izboljšan dostop do izobraževanja, večja udeležba žensk



Demografske spremembe, rodnost, staranje prebivalstva, na trgu dela, urbanizacija in dostopnost kontracepcije. Wolfgang Lutz [3] v svoji študiji navaja, da so ti dejavniki bistveno globalna rast, Slovenija prispevali k padcu rodnosti, predvsem v državah z visokim

življenjskim standardom. Ob tem pa ne moremo zanemariti tudi

vplivov, kot so spremembe v družinskih vrednotah in zakasnitev

rojstev. Phillip Longman [4] poudarja, da imajo te spremembe

Abstract dolgoročne posledice na gospodarski razvoj, saj manjše število

The paper discusses global and local demographic changes, mladih vpliva na inovativnost in produktivnost, hkrati pa

focusing on declining birth rates, population aging, and the povečuje pritisk na socialne sisteme.

shrinking working-age population. Global population growth has Vse bolj aktualna postaja tudi razprava o okoljskih posledicah

nearly halted, except in Africa, where birth rates are still rising zmanjševanja prebivalstva. Medtem ko manjše število ljudi

but slowly declining as well. Europe, including Slovenia, is potencialno zmanjšuje pritisk na naravne vire, pa upad

facing population stagnation or even decline, creating long-term prebivalstva prinaša tudi tveganje urbanega propadanja in

social and economic challenges. The paper highlights the need zanemarjanja infrastrukture, kar izpostavlja Alan Weisman [5] v

for strategic solutions, such as promoting birth rates and adopting svoji knjigi The World Without Us. Prav tako se države soočajo

appropriate migration policies, to prevent a long-term z izzivom, kako oblikovati ustrezne politike za naslavljanje teh

demographic crisis and ensure stable population growth. sprememb. Nekatere države se zatekajo k spodbujanju rojstev,

druge pa k priseljevanju, kar pa odpira etična in politična

Keywords vprašanja, povezana z integracijo priseljencev in ohranjanjem



Demographic changes, Birth rates, Population aging, global prepričanju, da prihaja eksponentna rast prebivalstva, ampak ta growth, Slovenia kulturne identitete [6]. Mediji in množice še vedno živijo v

se je že iztekla, nastavki za upad in izumiranja posameznih

narodov pa so že nastavljena [7].

∗Article Title Footnote needs to be captured as Title Note Zato je nujno, da se na demografske spremembe odzovemo

†Author Footnote to be captured as Author Note celostno, s strateškimi politikami, ki upoštevajo tako

Permission to make digital or hard copies of part or all of this work for personal or gospodarske, socialne kot okoljske vidike. Učinkovite rešitve

classroom use is granted without fee provided that copies are not made or distributed morajo temeljiti na poglobljenem razumevanju vzrokov in

for profit or commercial advantage and that copies bear this notice and the full posledic teh sprememb ter na sprejemanju inovativnih pristopov

citation on the first page. Copyrights for third-party components of this work must za dolgoročno trajnostno upravljanje prebivalstva. be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia V prispevku prikažemo projekcije in tabele, ki potrjujejo

© 2024 Copyright held by the owner/author(s). uvodne misli: svet bo iz eksponentne rasi prešel v (skoraj) https://doi.org/10.70314/is.2024.dem.10

http://doi.org/DOI_RECEIVED_AFTER_REVIEW

ekonomski upad, ki bo morda ogrozil tudi rast človeške svetovno prebivalstvo v prihodnjih desetletjih začelo upadati, kar

civilizacije. Pregled demografskih projekcij je narejen pretežno bo prineslo pomembne gospodarske, družbene in okoljske

iz dveh virov: Eurostat, Worldometer, delno Wikipedije. spremembe. Avtorja poudarjata, da se bo zaradi tega svet soočal

Pregled po svetu je predstavljen v sekciji 2, nato je v sekciji z izzivi, kot so zmanjšanje delovne sile, upad inovacij in

3 predstavljen pregled po Sloveniji in na koncu sledi diskusija v naraščajoč pritisk na socialne sisteme. Po drugi strani pa bo ta

zadnji sekciji. upad prinesel tudi olajšanje glede pritiskov na naravne vire in

okolje. A po mnenju Elona Muska, skladno z mnenjem Matjaža

Gamsa [8] bo ta upad lahko pomenil zaton človeške tehnološke

2 Svet civilizacije.





Slika 1: Število prebivalcev na zemlji od začetka štetja do

danes. Vir: World population, Worldometer



Svetovno prebivalstvo je po precej stabilnem stanju, ki je



trajalo več kot tisočletje, okoli leta 1500 prešlo v obdobje Slika 2: Rodnost kot število otrok na žensko od leta 1955 do

eksponentne rasti. Ta nenaden porast prebivalstva je bil danes. Vir: Worldometer. posledica izboljšane prehrane, napredka v medicini in



povečanje rodnosti. Kljub temu pa se ta eksponentna rast v pomeni, da je povprečna ženska v tem obdobju rodila pet otrok. sodobnem času počasi ustavlja. Po zadnjih podatkih je bilo 17. To obdobje visoke rodnosti je bilo značilno za države po vsem industrijski revoluciji, ki so omogočili zmanjšanje umrljivosti in Rodnost je bila leta 1950 približno 5 otrok na žensko, kar

septembra 2024 na svetu 8,177,068,100 prebivalcev (Slika 1 in svetu, zlasti v manj razvitih regijah, kjer so bili otroci pomemben

Tabela 1). Napovedi za prihodnost nakazujejo, da bo svetovno vir delovne sile in opora za starejše generacije. Vendar pa se je

prebivalstvo doseglo vrhunec okoli sredine tega stoletja, nato pa sredi 20. stoletja začela globalna preobrazba, ki je postopoma

bo začelo upadati, predvsem zaradi zmanjšane rodnosti in privedla do zmanjšanja rodnosti.

staranja prebivalstva. Od leta 1950 dalje je rodnost začela upadati, saj so države v

različnih fazah razvoja uvedle politike, ki so spodbujale uporabo

Tabela 1: Rast števila prebivalcev od 1955 dalje. Vir: kontracepcije, povečale dostop do izobraževanja (zlasti za

Worldometer. ženske) in podprle večjo udeležbo žensk na trgu dela. Do danes

se je globalna rodnost znižala na približno 2.2 otroka na žensko,



kar je le nekoliko nad trajno rodnostjo 2.1, ki je potrebna za



ohranjanje stabilnega prebivalstva (Slika 2).

Zmanjševanje rodnosti je kompleksen pojav, ki vključuje

tako gospodarske, družbene kot kulturne dejavnike. Wolfgang

Lutz [3] v svoji študiji The End of World Population Growth

navaja, da so višje stopnje izobrazbe, predvsem pri ženskah,

močno vplivale na zmanjšanje rodnosti, saj so izobražene ženske

bolj nagnjene k odlaganju rojstev in imajo na splošno manj otrok.

Poleg tega je dostop do kontracepcije in izboljšana zdravstvena

oskrba omogočila boljšo reproduktivno izbiro, kar je še dodatno

zmanjšalo rodnost.

Tudi Phillip Longman [4] v svoji knjigi The Empty Cradle

opozarja na gospodarske posledice nizke rodnosti, saj zmanjšano

število otrok v razvitem svetu pomeni manj delovne sile v

prihodnosti in večje breme za socialne sisteme, ki se soočajo s

Različne študije poudarjajo, da smo priča prelomnemu staranjem prebivalstva. Po drugi strani pa Alan Weisman [5] v

trenutku v demografskih spremembah. Na primer, po podatkih, knjigi The World Without Us izpostavlja okoljske prednosti nižje

ki jih navaja Darrell Bricker in John Ibbitson v svoji knjigi Empty rodnosti, saj zmanjšano prebivalstvo pomeni manjši pritisk na

Planet: The Shock of Global Population Decline (2019) [7], bo naravne vire in manjše emisije ogljikovega dioksida.

Hkrati z zmanjševanjem rodnosti se je povečevala povprečna stanje. Tabela 2 ponazarja ta trend rasti in vključuje ključne

življenjska doba na našem planetu, z 20-tih let leta 1970 je zrasla podatke, ki kažejo na upočasnjevanje rasti v kasnejših letih,

na 30 let leta 2024 (Slika 3), to je za 50%. Tudi pričakovana predvsem zaradi nižje rodnosti in staranja prebivalstva. Kljub

življenjska doba je rasla od leta 1959 do leta 2024, ko znaša v temu ostaja vprašanje, kako se bodo družbeni, gospodarski in

povprečju 73.3 leta, za moške 70.7 let in za ženske 76.0 (Slika 4). okoljski sistemi prilagodili na tako obsežno število prebivalcev.

Razlika v dolgosti življenja med moškimi in ženskami je opazna

od začetka merjenja do danes, ko znaša 7%. Tabela 2: Predvidena rast prebivalcev po svetu.





Po celinah so prebivalci razporejeni glede na Tabelo 3 in po

državah glede na Tabelo 4.



Tabela 3: Prebivalstvo in drugi osnovni podatki po celinah.



Tabela 3 prikazuje razporeditev prebivalcev po posameznih

Slika 3: Povprečna starost, vir Worldometer, od leta 1970 (20 let) celinah, Slika 5 pa grafično pokaže ta razmerja. Slika 6 kaže rast

stalno narašča in je v letu 2024 30 let. prebivalstva po celinah, pri čemer jasno izstopa rast prebivalstva

v Afriki, medtem ko se v Evropi in Severni Ameriki pričakuje

Pričakovana življenjska doba se je iz leta 1950 (46 let) do leta stagnacija oziroma upad. Ti podatki kažejo na pomembne

2024 dvignila na 73 let. To podaljševanje življenjske dobe in regionalne razlike, ki bodo vplivale na globalno demografsko

zamik (ljudje živijo cca 40 let po obdobju rojstev) povzroča vtis, sliko v prihodnjih desetletjih, pri čemer bo rast v nekaterih delih

da se demografski padec še ni začel, medtem ko se v praksi že je. sveta močno presegla rast v drugih.

Nova študija Rimskega kluba [2] je napovedala, da bo svetovno Te demografske spremembe bodo imele daljnosežne

prebivalstvo v najbolj verjetnem scenariju verjetno upadlo v posledice za mednarodno politiko, gospodarstvo in trajnostni

naslednjem stoletju na 6 milijard. razvoj, saj bodo države z največjo rastjo morale obvladovati vse

večje pritiske na svoje vire in infrastrukturo, hkrati pa bodo



zaradi večje populacije imele vse večjo vlogo v svetu.





Slika 5: Število prebivalcev po celinah.



Slika 4: Rast življenjske dobe. Vir: Worldometer.





Število prebivalcev Zemlje naj bi se v prihodnjih desetletjih

še naprej povečevalo, pri čemer naj bi doseglo vrhunec med

letoma 2050 in 2060, kot je prikazano v Tabeli 2. Projekcije

kažejo, da bo globalna populacija do takrat dosegla približno

deset milijard ljudi, kar predstavlja znatno rast glede na trenutno





Slika 8: Razmerja med verniki različnih verstev.



Slika 6: Napovedi razmerij med celinami.

3 Evropa in Slovenija

Slika 7 prikazuje gostoto prebivalstva po celinah, pri čemer

Kitajska in Indija močno izstopata kot najbolj gosto poseljeni

regiji na svetu. Prav tako so opazne višje stopnje poseljenosti v

nekaterih delih Združenih držav Amerike in Evropske unije, kjer

koncentracija prebivalstva vpliva na infrastrukturne in

gospodarske izzive.





Slika 7: Gostota prebivalstva po celinah.



Tabela 4: Največje države sveta.





The EU population is projected to decrease by



1.3 % between 2019 and 2050



Source: Eurostat (proj_19np) and (proj_19rp3)



V Tabeli 4 so prikazane največje države sveta glede Slika 9: Projekcije Evrope do leta 2050.

prebivalstva, pa tudi drugi osnovni podatki kot površina. Indija

je prehitela Kitajsko, skupaj pa imata 3 milijarde prebivalcev od Slika 9 predstavlja rezultate najnovejših demografskih

8 na svetu. Na tretjem mestu bi morala biti EU, a je v vseh projekcij regionalnih vzorcev prebivalstva v 31 državah (27

statistikah prikazana kot posamezne države. Zato so tretje ZDA, držav članic Evropske unije (EU) in štirih članic Evropskega

sledi Indonezija, Pakistan, Nigerija itd. združenja za prosto trgovino (EFTA). Ponuja strnjen pregled

Morda je zanimiv prikaz števila vernikov po posameznih rezultatov za regije na ravni NUTS 3, uporabljenih pri zadnjih

državah, kot je to razvidno iz Slike 8. evropskih demografskih projekcijah, temelječih na podatkih

EUROPOP2019.

Te regionalne projekcije prebivalstva zajemajo časovno starejših prebivalcev, predvsem tistih nad 65 let, vztrajno

obdobje do konca stoletja, pri čemer se rezultati, predstavljeni povečuje, kar postaja vse bolj izrazit trend v mnogih razvitih

tukaj, nanašajo na leto 2050. Projekcije temeljijo na podatkih o državah. Posledično se zmanjšuje delež delovno aktivnega

prebivalstvu in življenjskih dogodkih do leta 2018, uporabljena prebivalstva, kar ustvarja pritisk na pokojninske in zdravstvene

pa je bila klasifikacija NUTS-2016, ki deli ozemlja teh 31 držav sisteme, ki morajo zagotavljati skrb za starajoče se prebivalstvo.

na 1.216 geografskih enot na ravni NUTS 3. Za enostavnejšo Ta demografski premik vodi tudi do sprememb v družbenih

razlago "EU+" v tem primeru zajema vseh 31 držav, vključno s strukturah, saj postajajo starejši vse bolj odvisni od mlajših

štirimi članicami EFTA (Islandija, Lihtenštajn, Norveška in generacij in javnih storitev, kar bo dolgoročno zahtevalo

Švica). Za predstavitev podatkov EUROPOP2019 na nacionalni prilagoditve v politikah in socialnih sistemih.

ravni se sklicuje na članek o demografskih projekcijah v EU.

Slovenske podatke začenjamo z rastjo pričakovane

življenjske dobe na Sliki 10, ki je skozi desetletja postopoma

naraščala zaradi izboljšane zdravstvene oskrbe, življenjskih

pogojev in dostopa do izobraževanja. Na desnem robu slike je

jasno razviden upad v pričakovani življenjski dobi, ki ga je

povzročila pandemija COVID-19, kar je posledica povečane

umrljivosti, predvsem pri starejših prebivalcih in rizičnih

skupinah. Ta upad je pokazatelj, kako lahko nepričakovani

zdravstveni krizni dogodki začasno prekinijo pozitivne trende v

izboljševanju življenjske dobe, vendar se pričakuje, da bo z

obvladovanjem pandemije pričakovana življenjska doba

ponovno naraščala.





Slika 12: Stopnja rodnosti in število novorojenih v Sloveniji.



Slika 12 prikazuje stopnjo rodnosti, ki je že vrsto let precej

pod mejo 2,1 otroka na žensko, kar je minimalna stopnja za

ohranjanje stabilnega prebivalstva. Nizka stopnja rodnosti

pomeni, da prebivalstvo ni sposobno nadomestiti izgubljenih

generacij, kar vodi v dolgotrajen demografski upad. Če se ta

trend nadaljuje, bi lahko določene populacije kot slovenska v

nekaj stoletjih postopoma izumrle, saj brez zadostne rodnosti ni

mogoče zagotoviti zadostne reprodukcije in nadaljnje rasti

prebivalstva. Takšni scenariji postavljajo družbe pred številne

izzive, saj bodo morale sprejeti ukrepe za spodbujanje rodnosti,



Slika 10: Rast pričakovane življenjske dobe prebivalcev genetsko izumrtje. da bi preprečile dolgoročno demografsko krizo in posledično

Slovenije – razviden je učinek kovida na desnem robu.



4 Diskusija





V prispevku so predstavljene ključne demografske razmere in



prihodnje projekcije. Svetovna rast prebivalstva se je skoraj

povsem ustavila, pri čemer afriške države še vedno beležijo rast,

vendar tudi tam stopnja rodnosti počasi upada. V Evropi in

nekaterih drugih delih sveta pa že prihaja do stagnacije ali celo

upadanja prebivalstva. Slovenija se sooča s podobnim izzivom,

saj nizka rodnost in majhnost države nakazujeta dolgoročno

demografsko krizo, ki bi lahko vodila v znatno zmanjšanje

prebivalstva. Kljub tem opozorilom javnost pogosto napačno

razume situacijo, pri čemer prevladuje napačno prepričanje, da

je na svetu preveč ljudi. V resnici se pričakuje, da bo čez nekaj

sto let svetovna populacija bistveno nižja kot danes, medtem ko

Slika 11: Delež starejših kot 65 let. bi lahko Slovenci povsem izginili, če ne pride do temeljite

spremembe v politikah in odnosu do demografskih vprašanj.

Vizionarji, ki že desetletja opozarjajo na te trende in se

Daljša življenjska doba v kombinaciji z nizko rodnostjo, kot zavzemajo za trajnostne demografske rešitve, bi morali biti

je to značilno za Slovenijo (Sliki 11 in 12), pomeni, da se delež cenjeni in priznani, namesto da so pogosto tarča kritik.



References / Literatura [5] Weisman, A. (2008). The World Without Us. Picador. [6] Saunders, D. (2017). Maximum Canada: Why 35 Million Canadians Are

[1] Gams, M., & Malačič, J. (2019). Bela knjiga slovenske demografije. Not Enough. Alfred A. Knopf Canada.

URN:NBN:SI. [7] Bricker, D., & Ibbitson, J. (2019). Empty Planet: The Shock of Global

[2] Turner, B. (2023). World's population could plummet to 6 billion by the Population Decline. Robinson.

end of the century, study suggests. Live Science. [8] Kapitanovič, P. (2024). V nekaj stoletjih bomo izumrli, Delo,

[3] Lutz, W., Sanderson, W. C., & Scherbov, S. (Eds.). (2004). The End of https://www.delo.si/razno/v-nekaj-sto-letih-bomo-izumrli.

World Population Growth in the 21st Century: New Challenges for Human

Capital Formation and Sustainable Development. Routledge.

[4] Longman, P. (2004). The Empty Cradle: How Falling Birthrates Threaten

World Prosperity and What to Do About It. Basic Books.





Pokritost strokovne in znanstvene produkcije v Sloveniji v




področju epidemiologije depresije kot dela demografije



zdravja v COBISS.SI



Coverage of professional and scientific production in



Slovenia in the field of epidemiology of depression as part



of health demography in COBISS.SI



Tvrtko-Matija Šercar1 Romana Muhvič Šumandl2



Institut informacijskih znanosti Institut informacijskih znanosti

Maribor (IZUM) Maribor (IZUM)

Maribor Slovenija Maribor Slovenija

tvrtko.sercar@ext.izum.si romana.muhvic@izum.si



Povzetek other time it is part of statistics, etc.. Only scientists and experts in the field of demography can authoritatively assume the Informacijsko znanost in informacijsko tehnologijo uporabljajo responsibility for creating a knowledge map of demography.The vse druge znanosti kot uporabno epistemologijo. Podana je MIRA national mental health program is the first strategic spodbuda k premisleku o izdelavi sheme védenja demografije, ki document of its kind that connects the mental health services of naj bi zajemala demografijo zdravja in epidemiologijo na all responsible sectors of health, social care and education for the koherenten način, saj se v sedanji znanstveni praksi povezuje period 2018-2028 in the Republic of Slovenia. Targeted demografija različno s drugimi znanstvenimi področji. Enkrat coverage is checked by searching COBISS.SI. predstavlja z ekonomsko teorijo posebno podklaso v klasi



družbenih ved, drugič je del statistike itp. Odgovornost za Keywords



prevzamejo le znanstveniki in strokovnjaki s področja knowledge map of demography, healh demography, izdelavo sheme védenja demografije lahko avtoritativno

demografije. MIRA nacionalni program duševnega zdravja je epidemiology of depression, MIRA, COBISS.SI, SGC

prvi tovrstni strateški dokument, ki povezuje službe duševnega

zdravja vseh odgovornih sektorjev zdravstva, socialnega varstva,

1 Uvodna opomba

vzgoje in šolstva za obdobje 2018-2028 v Republiki Sloveniji.

Ciljana pokritost je preverjana z iskanjem v COBISS.SI. Lani sva s kolegom Šobotom predstavila rezultate komparativne

bibliografske raziskave o odzivu sistema COBISS.SI in

Ključne besede COBISS.SR znotraj mreže COBISS.net na zahteve uporabnikov

po informacijah o znanstveni produkciji v področju demografije

shema védenja demografije, demografija zdravja, epidemiologija

in sociologije prebivalstva v Republiki Sloveniji in Republiki

depresije, MIRA, COBISS.SI, SGC

Srbiji. [1]

Abstract Na vabilo k sodelovanju s prispevkom na letošnji konferenci

Demografske in družinske analize - IS2024 sva se odločila s

Information science and the information technology with which

kolegico Muhvič pripraviti prispevek o bibliografski pokritosti

it is fused are used by all other sciences as an applied

strokovne in znanstvene produkcije v Sloveniji v področju

epistemology. An incentive is given to consider the creation of a

socialnih, bioloških in genetskih razsežnosti epidemiologije

knowledge map of demography, which should cover health

psihiatričnih motenj kot dela demografije zdravja (Health

demography and epidemiology in a coherent way, since in

demography).

current scientific practice, demography is connected differently

with other scientific fields. One time it represents a special



subclass in the class of social sciences with economic theory, the 2 Shema védenja demografije – spodbuda k

1 Dr.sc. Tvrtko-Matija Šercar, upokojen sodelavec IZUM premisleku

2 Romana Muhvič Šumandl, Vodja sektorja Upravljanje servisov, IZUM

Permission to make digital or hard copies of part or all of this work for personal or classroom use »Na začetku je bila informacija, beseda je prišla kasneje.« [2]

is granted without fee provided that copies are not made or distributed for profit or commercial Informacijsko znanost in informacijsko tehnologijo uporabljajo

advantage and that copies bear this notice and the full citation on the first page. Copyrights for vse druge znanosti kot uporabno epistemologijo. Sprememba

third-party components of this work must be honored. For all other uses, contact the

owner/author(s). paradigme je bila prepoznana v začetku leta 2000. Nova

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s). paradigma pomeni premik od uporabe tehnologij, računalništva

http://doi.org/10.70314/is.2024.dem.11

in inženirstva kot podpore raziskovalnem delu in znanju k • hierarhične (po stopnjah, predmeti so razdeljeni hierarhično,

integraciji tehnologij, računalništva in inženirstva v strukturo od najbolj splošnih do najbolj specifičnih) in

vseh drugih znanosti. Niti nekoč sila uspešen slogan »podpora • fasetno/analitično-sintetične (predmeti so razdeljeni na

znanju« več ne izraža bistva nove paradigme o integraciji! medsebojno izključujoče pravokotne fasete) klasifikacije.

Bilo je ugotovljeno, da se je l. 2002 začela »digitalna doba«. [3]

Naloga klasifikacije in kategorizacije dokumentov se sestoji iz



dodeljevanja dokumentov enemu ali več razredom ali

kategorijam. To je mogoče narediti "ročno" (ali "intelektualno")

in algoritemsko. Intelektualna klasifikacija dokumentov je

večinoma področje bibliotekarstva, algoritemska klasifikacija

dokumentov pa predvsem informacijske znanosti in

računalništva.

Obstajata dve glavni filozofiji predmetne klasifikacije

dokumentov:

Slika 1: Leto 2002 - začetek digitalne dobe [3] • pristop, ki temelji na vsebini (content-based), in

• pristop, ki temelji na zahtevi (request-based).

Oblikovati »knowledge map« ali shemo védenja pomeni določiti



meje znanstvenega področja, opredeliti njegove glavne dele po Vsebinska klasifikacija je klasifikacija, pri kateri teža, dodeljena

predmetu in odnose med njimi. posameznim temam v dokumentu, določa razred, v katerega je

Shema védenja demografije (demography knowledge map) ni dokument uvrščen.

dorečena na način, ki ga je pred dvadesetimi leti kot »standard« Na zahtevo usmerjeno razvrščanje (ali indeksiranje) je

postavil Zins na primeru sheme védenja informacijske znanosti. razvrščanje, pri katerem pričakovana zahteva uporabnikov

[4][5] vpliva na to, kako so dokumenti razvrščeni. Klasifikator se

Običajna definicija demografije ne vključuje ravno demografije sprašuje: "Pod katerimi deskriptorji naj se nahaja ta entiteta?" in

zdravja (Health demography). »Pomislite na vse možne poizvedbe in se odločite, za katere je

Po nekem pristopu sta demografija in epidemiologija deli zadevna entiteta ustrezna«. [6]

statistike. Statistika (običajni pomen popis številčnega stanja; Razlikujemo dodeljevanje dokumentov v razrede

izvirno iz nemščine Statistica Staatslehre, Staatswissenschaft ("klasifikacija") in dodeljevanje predmetov dokumentom

državoslovje, državna teorija, politična znanost, status nlat tudi ("predmetno indeksiranje"), vendar to razlikovanje ni plodno.

Staat, Status, država) se ukvarja z zbiranjem, organizacijo, »Te terminološke razlike so povsem nesmiselne in povzročajo

analizo, interpretacijo in predstavitvijo podatkov o gospodarskih, samo zmedo«. [7]

socialnih ipd. procesih. Obstajajo številne tehnike samodejnega (avtomatskega)

Z uporabo statistike so se razvile številne specialnosti v različnih razvrščanja dokumentov.

področjih. Narodna in univerzitetna knjižnica (NUK) je leta 1994 izdala

Demografija je preučevanje (study) vseh vrst populacij in je zelo navodila za oblikovanje predmetnih oznak Vsebinska obdelava

splošna znanost, ki jo je mogoče uporabiti za katerokoli v vzajemnem katalogu. Leta 2000 se je v IZUM-u začel razvoj

dinamično populacijo, torej tisto, ki se spreminja skozi čas ali Splošnega geslovnika COBISS.SI (SGC), ki sega na vsa

prostor. področja človeškega znanja in je namenjen predmetnemu

Epidemiologija je preučevanje dejavnikov, ki vplivajo na zdravje označevanju v sistemu COBISS.SI. [8] Na iskalno besedo

in bolezen prebivalstva, ter služi kot temelj posegov v interesu »demografija« dobimo 29 predmetnih točk dostopa. javnega zdravja in preventivne medicine.



Znane knjižnične klasifikacije niso sheme védenja neke znanosti. Zinsova študija »Knowledge Map of Information Science« je

Najzgodnejša znana knjižnična klasifikacijska shema je bila bila namenjena raziskovanju temeljev informacijske znanosti,

klasifikacija knjig Pinakesa Kalimahe, učenjaka v Aleksandrijski izvedena pa je bila z metodo Delphi in analizo področja (domain

knjižnici v tretjem stoletju pred našim štetjem. analysis, DA) v letih 2003–2005.

Knjižnični klasifikacijski sistemi se razlikujejo glede na Metodo Delphi je Zins uporabil za zbiranje podatkov in za

označevanje, zgodovino, uporabo številčenja, hierarhijo in fasete. kritično in strukturirano debato (s pomočjo treh vprašalnikov)

Razdelitve človeškega znanja po posameznih znanstvenih med ekspertom.

področjih je mogoče najti v univerzalnih shemah, kot so Izraz »Delfi« izvira iz starogrške mitologije. Stari Grki so verjeli,

Klasifikacija Kongresne knjižnice (LCC), Deweyjeva decimalna da so območje glavne prerokinje Pitije v preročišču Delfi in so

klasifikacija (DDC), Univerzalna decimalna klasifikacija (UDC) tam bile narejene prerokbe, ki so narekovale in usmerjale vitalne

in Klasifikacija Colon (CC), sheme za določeno področje (npr. državne zadeve. [9]

klasifikacija za medicino NLM), za posamezno državo (npr. Metodo analize področja je Zins uporabil za izdelavo sheme

SAB, švedski knjižnični klasifikacijski sistem), klasifikacijske védenja informacijske znanosti.

sheme v bibliografskih virih, informacijskih servisih in zbirkah Domena je družbeno in teoretično definirano znanje skupine

podatkov. znanstvenikov istega ontološkega in epistemološkega interesa.

Obstajajo [10]

• naštevalne (predmetne značnice (subject heading) so Domene so v bistvu sestavljene iz treh vrst teorij in konceptov:

navedene po abecedi, s številkami, ki so dodeljene vsaki

značnici po abecednem vrstnem redu),

• ontoloških teorij in konceptov o predmetih človeške Učinki razrednih odnosov na depresijo in anksioznost presegajo

dejavnosti; učinke SES (socioeconomic stratification), kar kaže na premalo

• epistemoloških teorij in konceptov znanja in načinov raziskane mehanizme v socialni epidemiologiji, npr. dominacijo

pridobivanja znanja, ki implicirajo metodološka načela o in izkoriščanje. [16][17]

načinih raziskovanja predmetov; in Po podatkih Eurostata za leto 2019 je 7,2 % prebivalcev držav

• članic Evropske unije poročalo o kronični depresiji. To je skoraj socioloških konceptov o skupinah ljudi, ki se ukvarjajo z

istimi predmeti. pol odstotne točke več kot pet let pred tem. Slovenija je v samem

Kot formalne predstavitve množice konceptov z nekega področja vrhu lestvice. Pri nas je takšnih namreč kar 15,1 odstotka

in razmerij med njimi imajo ontologije nepogrešljivo vlogo pri prebivalcev, kar je skoraj tri odstotne točke več kot na

razvoju informacijskih sistemov s semantično komponento. [11] Portugalskem, ki zaseda drugo mesto in skoraj osem odstotnih

[12][13][14] točk več od evropskega povprečja. Na tretjem mestu depresivcev

so Švedi z 11,7 odstotka takšnih, ki so poročali o kronični

Različne znanosti, ki proučujejo iste predmete in probleme iz depresiji. Slovenija se uvršča med države z najvišjim odstotkom

svoje perspektive, se udeležujejo vzajemenega delovanja. depresivnih moških (14,3 %).

Obstajata "mala" in "velika" interdisciplinarnost. Najpomembnejši strateški dokumenti Evropske unije za duševno

"Mala" interdisciplinarnost je povezovanje specialnosti znotraj zdravje so Evropski pakt za duševno zdravje in dobro počutje in

neke širše discipline, med katerimi je "oddaljenost" majhna. Sklepi Sveta o Evropskem paktu za duševno zdravje in dobro

"Velika" interdisciplinarnost je povezovanje različnih počutje.

znanstvenih disciplin, med katerimi je "oddaljenost" velika. [15] Državni zbor RS je sprejel leta 2018 Nacionalni program

Rinia je meril in vrednotil interdisciplinarno naravo raziskav in duševnega zdravja 2018–2028 MIRA.

prenos znanja med disciplinami z uporabo kvantitativnih analiz Program MIRA je Nacionalni program duševnega zdravja, ki kot

citiranja. Citiranje je način, na kateri znanost funkcionira kot prvi tovrstni strateški dokument v Republiki Sloveniji celovito

industrija znanja. Neki znanstveni članek najboljše naslavlja in ureja področje duševnega zdravja. Program povezuje

»indeksirajo« citati virov. Rezultati analize citiranja kažejo na obstoječe ter dodaja nove službe in strukture ter tako odgovarja

velik pretok znanja med disciplinami. Citatna analiza razkriva na potrebe ljudi na področju duševnega zdravja. Glavni poudarki

veliko dinamičnost sodobne znanosti, medtem ko so se programa so krepitev duševnega zdravja ter preprečevanje in

knjižnične klasifikacijske sheme (LCC, DDC, UDC in dr.) za celostna obravnava duševnih motenj. S tem namenom povezuje

organizacijo shranjevanja knjižničnega gradiva v prostoru službe vseh odgovornih sektorjev, torej zdravstva, socialnega

pokazale za dokaj statične. varstva, vzgoje in šolstva.

Cilj programa je vzpostaviti dobro podporno okolje na vseh

področjih varovanja duševnega zdravja. Službe in storitve



poskuša približati uporabnikom in jih spodbuditi k zgodnjemu

iskanju pomoči ter na tak način zmanjšati institucionalizacijo na

področju duševnega zdravja. [18]



4 Rezultati iskanj v področju demografije

zdravja in epidemiologije depresije

Slika 2: Primer police podrazreda Pokritost smo preverjali z iskanjem v COBISS.SI.



Iskalni niz Št.zadetkov

3 Program MIRA – Nacionalni program

duševnega zdravja v Republiki Sloveniji Demografija zdravja Slovenija 255

Duševno stanje demografija Slovenija 90

Socialna in psihiatrična epidemiologija je področje, ki preučuje Depresija demografija Slovenija 53

porazdelitev in dejavnike duševnega zdravja znotraj populacij. Duševno zdravje MIRA Slovenija 130

Raziskuje, kako družbeni dejavniki, kot so socialno-ekonomski Demografija MIRA 46

status, mreže skupnosti in življenjski dogodki, vplivajo na MIRA demografija Slovenija 31

incidenco in razširjenost psihiatričnih motenj. Ta veja Epidemiologija depresije 289

epidemiologije pogosto vključuje interdisciplinarne raziskave, ki

združujejo vpoglede iz socialne psihologije, sociologije, Slika 3: Iskanja v COBISS.SI po področjih

antropologije in javnega zdravja, da bi razumeli kompleksne

interakcije med družbenimi okolji in rezultati duševnega zdravja. Rezultati začetnih poizkusnih iskanj po ključnih besedah v

Duševne motnje spadajo med deset poglavitnih vzrokov COBISS+ v t.i. osnovnem iskanju (pomeni nediferencirano

oviranosti v svetu s hudimi družbenimi in gospodarskimi iskanje, ki ga uporabimo predvsem, kadar imamo dokaj popolne

posledicami za posameznike, družine in skupnosti. Slabo podatke; z uporabljenimi iskalnimi pojmi se sproži iskanje po

duševno zdravje tako pomeni izgubo najmanj 3 do 4 % bruto vseh podatkih znotraj enega bibliografskega zapisa, tudi po

domačega proizvoda, po nekaterih ocenah pa še več. Svetovna morebitnih povzetkih in celotnem besedilu, če je le-to dostopno

zdravstvena organizacija kot Evropska unija in Slovenija online):

namenjajo promociji in varovanju duševnega zdravja vedno več Analiza najdenih zadetkov nam pokaže:

pozornosti in sredstev.

• kako so se dela iz navedenega področja objavljala skozi • prevladujejo monografske publikacije nad članki

leta



demografija zdravja

20

demografija zdravja 4

46



50

40



20 14 13 14 15 10 8 9 9 6 8 8 231 30 23 16 19 20 17

10

članki monografije serijske

0

0 0 1 1 1 1 1 1 1 1 1 1 2 2 2 2 0 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4

-2 20 0 0 0 0 0 0 0 0 0 0 0 0 0 0 02

2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

397 l. l. l. l. l. l. l. l. l. l. l. l. l. l. l. l.

1 Slika 8: področje »demografija zdravja« in oblika gradiva l.



• avtorji del iz obravnavanih področij

Slika 4: »demografija zdravja« skozi leta



681

duš avtorji z 2 zadetki 30 evno zdravje MIRA

3

Hovnik-Keršmanc, Marjetka 3

3

20 18 3 Bradač Hojnik, Barbara 16 16



3

13 13

15 Temnik, Sanja 3 9 8 6

10 5 6 5 4 3 4 4 Zupančič, Agata 7 2 2 2 5 7

Dominkuš, Davor 7



0 7 4 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 7 0 0 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 Čobal, Nadja 0-2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 l. 2 l. l. l. l. l. l. l. l. l. l. l. l. l. l. l. Švab, Vesna 8 7 9 8 1

l. Maučec Zakotnik, Jožica 9

9

Makivić, Irena 9

0 100 200 300 400 500 600 700

Slika 5: »duševno zdravje MIRA« skozi leta



Slika 9: aktualni avtorji za obravnavana področja

epidemiologija depresije



45 5 Zaključni premisleki



50



20 12 13 14 15 13 12 7 8 modernem slovarju Daše Komac [19] v izrazu »demography« 10 sledi prevodnemu ustrezniku »demografija« razlaga »statistično 30 20 23 25 22 20 21 19 V splošnem angleško-slovenskem (in slovensko-angleškem) 40 0



9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 0 1 1 1 1 1 1 1 1 1 2 2 22 0 0 0 0 0 0 0 0 0 01 0 0 0 0 020 2 2 2 2 2 2 2 2 2 2 2 2 2 22 proučevanje sestava/rasti prebivalstva«, ki naj bi uporabnike -2 8 l. l. l. l. l. l. l. l. l. l. l. l. l. l. l. 8 9 vodila do »pravilne izbire besede, stalne zveze ali fraze«! 1 »Definitio fi(a)t per genus proximum et differentia specifica« l.

spada v filozofijo in je sholastična formulacija klasičnega

Slika 6: »epidemiologija depresije« skozi leta definicijskega pravila, ki sega vse do Aristotela. Klasičen primer

je opredelitev ljudi kot razumnih živali.

• V razlagi definicije demografije je »statistično proučevanje« kakšne tipologije so dela

(statistika) »genus proximum«, »differentia specifica«

proučevanje »sestava/rasti prebivalstva«.

demografija zdravja Takšne definicije so praviloma redke v pogovornem

(kolokvijalnem) jeziku. [20]

78



73 »Genus proximum« je »višji« rodovni pojem, vendar so odnosi



80



70 med znanostmi horizontalni in ne vertikalni in ne moremo trditi,

60

50 da je ena kot višja nadrejena drugi!

40 28

30 21 Z bibliometričnimi raziskavami se lahko meri in vrednoti prenos

20

7 12 15 10 11

10 znanja (knowledge transfer). med demografijo in drugimi

0 vedami. Tako odkrita interdisciplinarnost kaže na dele drugih

ved, ki ne sodijo v demografijo. [15]



Slika 7: področje »demografija zdravja« in tipologija del





[7] Lancaster, F. W. (2003). Indexing and abstracting in theory and practice.

Library Association, London. Indexing and abstracting in theory and

practice : Lancaster, F. Wilfrid (Frederick Wilfrid), 1933-2013 : Free

Download, Borrow, and Streaming : Internet Archive

[8] IZUM, 2024. Splošni geslovnik COBISS.SI Splošni geslovnik SGC (SGC)

SGC- Splošni geslovnik COBISS SGC- Splošni geslovnik COBISS

[9] Prashant Nasa, Ravi Jain, and Deven Juneja, 2021. Delphi methodology

in healthcare research: How to decide its appropriateness. World J

Methodol, Jul 20, 11(4), pp. 116-129. doi: 10.5662/wjm.v11.i4.116

Delphi methodology in healthcare research: How to decide its

appropriateness - PMC (nih.gov)

[10] Hjorland, B. and Hartel, J., 2003. Afterword: Ontological, epistemological

and sociological dimensions of domains. Knowledge Organization 30(3-

Slika 10: Shematski prikaz razmerja med vrstami 4), pp.:239-245

interdisciplinarnosti in vrstami inovativnosti (v poljih https://www.researchgate.net/publication/297845253_Afterword_Ontolo

podan tipičen primer) [15] gical_epistemological_and_sociological_dimensions_of_domains

[11] Šercar, T., 1988. 2. Problemske razine komunikacija: b) Semantička

teorija komunikacija. V: Šercar,T. Komunikacijska filozofija znanstvenih

V LCC so »družbene vede« razred H. Podrazred HA je časopisa. Zagreb-Ljubljana: ČGP DELO-Globus. Str. 80-85.

[12] Fensel, D., 2002. Ontologies: a silver bullet for knowledge management

»statistika«. Podrazred HB je »ekonomska teorija« in and electronic commerce. Ontologies: A Silver Bullet for Knowledge



»demografija«. Management and Electronic Commerce | SpringerLink [13] Šercar, T. M., 2002. Ontologies: a silver bullet for knowledge V Wikipediji, prosti enciklopediji, stran "Oris (Outline) management and electronic commerce (Dieter Fensel). Organizacija

demografije" ne obstaja. Obstaja le članek o demografiji kot znanja 07 (3-4) Ontologies: a silver bullet for knowledge management and

disciplini. Prav tako v Wikipediji ne obstaja stran "demografija electronic commerce (Dieter Fensel) - Organizacija znanja (cobiss.si)

[14] Bošnjak, Aleš, 2016. Analiza nadgradnja informacijskega sistema o

zdravja" (Health Demography«. Obstaja le članek o raziskovalni dejavnosti s sematično komponento [na spletu]. Doktorska



»epidemiologiji depresije« disertacija. Maribor : A. Bošnjak. [Dostopano 14 avgust 2024]. Pridobljeno s: https://dk.um.si/IzpisGradiva.php?lang=slv&id=63111 Proučevanja v področju epidemiologije depresije kot dela [15] Rinia, E. J., 2007. Measurement and evaluation of interdisciplinarity

demografije zdravja so pokazala, da je depresija glavni vzrok research and knowledge transfer (doktorska disertacija), Universiteit

obolevnosti in umrljivosti ljudi povsod po svetu! Leiden. Microsoft Word - EdRDissTotalDef.doc (universiteitleiden.nl)

[16] Seath, J. P., Bates, L. M., Keyes, K. M., Muntaner, C., 2015 Nov.

Anxious? Depressed? You might be suffering from capitalism:

Rezultati iskanja po bibliografskih bazah podatkov v sistemu Contradictory class locations and the prevalence of depression and anxiety

COBISS.SI predstavljajo: in the United States. Sociol Health Illn. 37(8): 1352-1372. Published

online 2015 Aug 3. doi: 10.1111/1467-9566.12315 Anxious? Depressed?

• podlago za dopolnitev SGC z dvema novima predmetnima You might be suffering from capitalism: Contradictory class locations and

oznakama »demografija zdravja« (health demography) in the prevalence of depression and anxiety in the United States - PMC

(nih.gov)

»epidemiologija depresije« iskalne besede »demografija«, [17] Šercar, T.M., 2024. Capitalism and the Depression. International Journal

kar naj bi upoštevali tudi pri izdelavi sheme védenja of English Literature and Social Sciences 9(3) Dostopno na Capitalism

and the Depression (ijels.com)

demografije (knowledge map of demography); [18] Nacionalni inštitut za javno zdravje (NIJZ), 2024. MIRA Nacionalni

• program duševnega zdravja. Dostopno na Opis programa - Program prispevek k nastajanju znanstvene podpore Nacionalnemu

Mira (zadusevnozdravje.si)

programu duševnega zdravja v Republiki Sloveniji MIRA [19] Komac, Daša, 2001. Splošni angleško-slovenski in slovensko-angleški



2018-2028. moderni slovar = English-Slovenian and Slovenian-English modern dictionary [urednica: Mojca M. Hočevar; zapis angleške izgovarjave:

Rastislav Šuštaršič] ,Ljubljana: Cankarjeva založba, 2001. – 910 str. ; 23

Odgovornost za izdelavo sheme védenja demografije cm. – (Zbirka Moderni slovarji).

(knowledge map of demography) lahko avtoritativno prevzamejo [20] Ogleznev, V.V. in Surovcev, V.A., 2015. Определение per genus

proximum et differentiam specificam и юридический язык Shole 9 (2),

le znanstveniki in strokovnjaki s področja demografije. 228-240 Vitaly Ogleznev, Определение per genus proximum et

differentiam specificam и юридический язык - PhilPapers



References / Literatura schole-9-2.pdf (core.ac.uk)

[1] Šercar, Tvrtko in Šobot, Pero, 2023. Demografija in bibliografija =

Demography and bibliography [Elektronski vir]. Demografske in

družinske analize [Elektronski vir] = Demographic and Family Analyzes :

Informacijska družba - IS 2023 = Information Society - IS 2023 : zbornik

26. mednarodne multikonference = proceedings of the 26th International

Multiconference : zvezek F = volume F : 9. oktober 2023, 9 October 2023,

Ljubljana, Slovenia (Str. 26-34) IS2023_Volume-F - Demografija.pdf

(ijs.si)

[2] Dretske, F.I., 1981. Knowledge and the Flow of Information. Oxford:

Blackwell

[3] Hilbert, M. in Lopez, P., 2011. The World’s Technological Capacity to

Store, Communicate, and Compute Information. Science,

332(6025),pp. 60-65 DOI: 10.1126/science.1200970

[4] Zins, Chaim (2007). Classification Schemes of Information Science:

Twenty-Eight Scholars Map the Field. JASIST (Journal of the American

Society for Infromation Science and Technology), March, 58(5): 645–672.

(PDF) Classification schemes of Information Science: Twenty-eight

scholars map the field (researchgate.net)

[5] Šercar, T. M. in Brbre, I., 2007. Prispevek k filozofiji knjižničarstva in

informacijske znanosti. Organizacija znanja 12 (3), str. 119-136

https://www.cobiss.si/oz/HTML/OZ_2007_3_final/22/ OZ 2007/3

(cobiss.si)

[6] Soergel, Dagobert (1985). Organizing information: Principles of data base

and retrieval systems. Orlando, FL: Academic Press. Organizing

information: Principles of data base and retrieval systems.





Ethical Consideration and Sociological Challenges in the Integration of Artificial




Intelligence in Mental Health Services



Saša Poljak Lukek*

sasa.poljaklukek@teof.uni-lj.si

University of Ljubljana, Faculty of Theology

Ljubljana, Slovenia



Abstract feedback and personalized treatment plans [4],

This article explores the transformative potential and developing of novel diagnose tools [2].

of artificial intelligence (AI) in the field of mental AI in mental health services is implemented

health, with a particular focus on ethical through models like chatbots, digital platforms,

considerations and social challenges. As AI and avatar therapy, enhancing accessibility and

tools become increasingly sophisticated, their treatment options. Chatbots provide therapy via

ability to support mental health interventions natural language processing [5], while digital

presents both opportunities and challenges. We platforms support online mostly cognitive

discuss the importance of a human-centered behavioral therapeutic interventions [6]. Avatar

approach to AI development and the need for therapy uses AI to help patients manage

comprehensive ethical guidelines to ensure conditions like dementia, autism spectrum

patient safety and well-being. In addition, this disorder, and schizophrenia [7].

paper explores key social trends such as the

evolving dynamics of modern  families, aging 1.2 The Prospect of artificial

population, migration and considers how AI can intelligence in mental health services

be integrated into these contexts to improve The future orientation underlines the importance

mental health care. of digital health in overcoming challenges such

as limited access to services, especially in

Keywords: underserved regions, and outlines measures to

Artificial Intelligence, Mental Health, Human- ensure equitable access to digital health

Centered Approach, Ethics, Modern Family solutions across the European region [8]. The

Dynamics, Aging Populations, Migration use of AI in mental health services raises

questions about the role of non-human



1.1 algorithms and the long-term impact on the Artificial intelligence in mental understanding of illness and the human health services 1 interventions, transparency in the use of Introduction

condition [9]. There are also concerns about

Research on the application of AI in mental

potential bias, gaps in ethical and legal

health care has shown some positive effects on

frameworks, and the possibility of misuse

the treatment of mental health problems [1],

[10,11].

including early detection [2,3], providing

However, there are at least two potentially

 positive effects of the use of AI in healthcare:

This Publication is a Part of the Research Program The Accessibility and personalization of services.

Intersection of Virtue, Experience, and Digital Culture: Ethical

and Theological Insights, financed by the University of AI offers new mechanisms to reach those who

Ljubljana. might not otherwise be served. AI-supported

Permission to make digital or hard copies of part or all of this tools can improve the early detection and

work for personal or classroom use is granted without fee diagnosis of mental disorders [12]. AI chatbots



provided that copies are not made or distributed for profit or have shown promise in increasing referrals to commercial advantage and that copies bear this notice and mental health services, especially for minority the full citation on the first page. Copyrights for third-party

components of this work must be honored. For all other groups who are blocked from accessing



uses, contact the owner/author(s). traditional care [13]. These technologies can Information Society 2024, 7 – 11 October 2024, Ljubljana, provide initial assessments, psychoeducation Slovenia

© 2024 Copyright held by the owner/author(s). and even treatment, expanding access to mental

https://doi.org/10.70314/is.2024.dem.12 health support [12]. AI-driven virtual assistants



and wearable devices enable continuous



monitoring and personalized care, which could 2.2 Aging Populations

improve patient outcomes [11,14]. AI offers promising solutions for supporting an

The integration of artificial intelligence into aging population, particularly in addressing

mental health services represents a promising cognitive decline and mental health challenges.

avenue for the development of personalized AI applications can monitor vital signs, health

treatment plans through the sophisticated indicators, and cognition, as well as provide

analysis of large datasets, enabling the support for daily activities [20]. With an

identification of optimal therapeutic strategies increasing number of elderly individuals, AI can

tailored to specific client profiles [15,16]. This support mental health care by providing

data-driven methodology enables the dynamic companionship through intelligent animal-like

adaptation of therapy to the evolving needs of robots (e.g., Paro, Harp seal) and assisting in

the client. monitoring and managing conditions like

dementia [21,22]. AI can also help in tracking

2 Overcoming Sociological cognitive health and providing timely

Challenges through the Integration of interventions to maintain mental well-being in

Artificial Intelligence in Mental Health older adults. These technologies have the

Services potential to enhance independent living and

quality of life for older adults and their families.



2.1 Modern Family Dynamics

Modern family trends show that family 2.3 Migration



structures and attitudes have changed Migrants often face mental health challenges

significantly in recent decades [17]. There is a due to displacement, cultural adjustment and

growing acceptance of different family forms, language barriers. AI can help migrants access

including unmarried cohabitation, same-sex mental health services by providing culturally

relationships and joint custody arrangements and linguistically relevant resources and

[18]. These changes reflect an expansion of support. Chatbots and AI-driven platforms can

developmental idealism and increasing support bridge gaps in care by providing immediate help

for individual freedom in family choice [17]. and continuity of care across different regions

On the other hand, there is a growing need for [23].

mental health services for families [19]. As the Recent research highlights the increasing role of

most vulnerable members of the family - the digitalization and artificial intelligence (AI) in

children - are usually also at risk, quick and migration and mobility systems, especially in the

effective action in family mental health is of context of the COVID-19 pandemic [24]. While

great importance. Many families are struggling these technologies offer opportunities for

with various psychological problems. Together improving human rights and supporting

with the changing family structure, this means a international development, they also bring

great burden for every family member. In challenges that require careful consideration of

addition, access to psychologists, psychiatrics design, development and implementation

and therapists is limited, leading to an acute aspects. The integration of AI into migration

shortage of mental health professionals processes requires a focus on human rights at

worldwide. all stages that goes beyond technical feasibility

The accessibility of services is probably the and companies' claims of inclusivity [24].

strongest argument for the integration of AI in

healthcare [12]. AI-powered conversational 3 Ethical Consideration in the

agents can improve the accessibility of mental Integration of Artificial Intelligence in

health services by being available online at all Mental Health Services

times and in underserved areas, being scalable, One of the main caveats to the use of AI in

reliable, fatigue-free, and providing consistent mental health is the introduction of new ethical

support, being culturally sensitive to adapt, and standards to ensure user safety. The approach to

helping with education and symptom integrating AI into services should therefore be

management. human-centered [25]. Any innovation should



therefore focus on people in their most essential, with a focus on dealing with potential

vulnerable position. It is important to assess all bias, especially among marginalized groups, the

risks with sufficient accuracy and avoid misuse risks associated with data privacy and security,

of AI as much as possible. The most important and the challenges posed by the lack of

areas for ethical consideration when integrating transparency of AI models.

AI into mental health services should be privacy,



bias, transparency, security. 4 Conclusion

Data privacy and security are critical in digital We propose to define AI as a new ethical entity

healthcare and require robust measures to in the field of mental health [30]. AI represents a

protect sensitive information and prevent novel artifact that changes interactions,

unauthorized access. Protecting privacy rights concepts, epistemic fields and normative

and ensuring informed consent are critical to requirements. This change requires a

maintaining trust and ethical standards in the redefinition of the role of AI, which lies on a

use of personal health data [11]. Combining spectrum between a tool and an agent. This shift

multiple data streams increases the risk of underscores the need for new ethical standards

unauthorized use, which exacerbates privacy and guidelines that recognize the unique status

issues. Ensuring informed consent and of AI as a distinct and influential actor in the field

maintaining transparency, especially in of mental health.

emergency operations, are critical to addressing The integration of AI into services can, on the

these ethical concerns and protecting the rights one hand, provide more efficient and faster

of participants [26]. solutions to some of the sociological challenges

The use of AI in mental health treatment raises of today's society, but on the other hand,

ethical concerns about bias, particularly among requires a precise and correct definition of the

marginalized populations who are already limits within which these models can be used.

discriminated against and lack access to mental These efforts aim to bridge the gap between

health care. It is uncertain whether AI-assisted technology and human-centered care and

psychotherapy can effectively address cultural ensure that AI complements, rather than

differences and close treatment gaps in diverse replaces, the therapeutic benefits of human

populations [27]. In addition, populations that interaction.

are traditionally marginalized in fields such as

psychology and psychiatry are most vulnerable

to algorithmic biases in AI and machine learning Literature

[27,28]. These biases limit the ability of AI to

provide culturally and linguistically appropriate [1] Sandhya Bhatt. 2024. Digital Mental Health: Role of

Artificial Intelligence in Psychotherapy. Annals of

mental health resources, exacerbating existing Neurosciences, 0,0, 1-11.

inequalities. The persistence of such biases in AI doi:10.1177/09727531231221612



systems not only risks increasing health [2] Sijia Zhou, Jingping Zhao and Lulu Zhang. 2022. Application of Artificial Intelligence on Psychological inequalities, but also exacerbates existing social Interventions and Diagnosis: An Overview. Frontiers in inequalities and raises critical ethical Psychiatry , 13(March), 1 – 7.

considerations [9]. https://doi.org/10.3389/fpsyt.2022.811665



The future of artificial intelligence in clinical [3] Klaudia Kister, Jakub Laskowski, Agata Makarewicz and Jakub Tarkowski. 2023. Application of artificial settings is affected by a significant ethical intelligence tools in diagnosis and treatmentof mental dilemma concerning the trade-off between the disorders. Current Problems of Psychiatry , 24, 1 – 18.

performance and interpretability of machine https://doi.org/10.12923/2353-8627/2023-0001



learning models [29]. The lack of transparency [4] Rachel L. Horn and John R. Weisz. 2020. Can Artificial Intelligence Improve Psychotherapy Research and in AI models makes it difficult to detect and Practice? Administration and Policy in Mental Health and correct biases. This underscores the need for Mental Health Services Research , 47, 5, 852 – 855.

greater transparency to ensure ethical and fair https://doi.org/10.1007/s10488-020-01056-9



clinical decision-making. [5] Kerstin Denecke, Alaa Abd-alrazaq and Mowafa Househ. 2021. Artificial Intelligence for Chatbots in Mental In summary, the integration of AI into mental Health: Opportunities and Challenges. In: Househ, M., health services requires the establishment of Borycki, E., Kushniruk, A. (eds) Multiple Perspectives on

strict ethical standards to protect the safety and Artificial Intelligence in Healthcare. 115– 128.

privacy of users. A human-centered approach is https://doi.org/10.1007/978-3-030-67303-1_10

[6] Elias Aboujaoude, Lina Gega, Michelle B. Parish and Century of Family Attitude Trends in the United States.

Donald M. Hilty. 2020. Editorial: Digital Interventions in Sociology of Development, 9, 1, 1–32.

Mental Health: Current Status and Future Directions. https://doi.org/10.1525/sod.2022.0003

Front. Psychiatry 11, 111. doi: 10.3389/fpsyt.2020.00111 [19] WHO. 2022. World mental health report: transforming

[7] Kay T. Pham, Amir Nabizadeh & Salih Selek. 2022. mental health for all. (June 2022) Retrieved August 20,

Artificial Intelligence and Chatbots in Psychiatry. 2024 from

Psychiatric Quarterly, 93, 1, 249–253. https://www.who.int/publications/i/item/978924004933

https://doi.org/10.1007/s11126-022-09973-8 8.

[8] WHO. 2022. Regional digital health action plan for the [20] Sara J.Czaja and Marco Ceruso. 2022. The Promise of

WHO European Region 2023–2030 (RC72). (July 2022). Artificial Intelligence in Supporting an Aging Population.

Retrieved August 20, 2024 from Journal of Cognitive Engineering and Decision Making,

https://www.who.int/europe/publications/i/item/EUR- 16, 4, 182– 193.

RC72-5 https://doi.org/10.1177/15553434221129914

[9] Amelia Fiske, Peter Henningsen and Alena Buyx. 2019. [21] Maria R. Lima. 2024. Home Integration of Conversational

Your Robot Therapist Will See You Now: Ethical Robots to Enhance Ageing and Dementia Care.

Implications of Embodied Artificial Intelligence in ACM/IEEE International Conference on Human-Robot

Psychiatry, Psychology, and Psychotherapy. Journal of Interaction, 115– 117.

Medical Internet Research, 21, 5, e13216. https://doi.org/10.1145/3610978.3638378

https://doi.org/10.2196/13216 [22] Wendy Moyle. 2019. The promise of technology in the

[10] Elizabeth C. Stade, Shannon Wiltsey Stirman, Lyle future of dementia care. Nature Reviews Neurology, 15,

Ungar, Cody L. Boland,H. Andrew Schwartz, David B. 6, 353–359. https://doi.org/10.1038/s41582-019-0188-y

Yaden, Joao Sedoc, Robert J. DeRubeis, Robb Willer and [23] Zahra Abtahi, Miriam Potocky, Zarin Eizadyar, Shanna L.

Johannes C. Eichstaedt. 2024. Large Language Models Burke, Nicole M. Fava. 2022. Digital Interventions for the

Could Change the Future of Behavioral Healthcare: A Mental Health and Well-Being of International Migrants:

Proposal for Responsible Development and Evaluation. A Systematic Review. Research on Social Work Practice,

Mental Health Res 3, 12. 33, 5, 518-529. Doi:10.1177/10497315221118854

https://doi.org/10.1038/s44184-024-00056-z [24] Marie McAuliffe, Jenna Blower and Ana Beduschi. 2021.

[11] David B. Olawade, Ojima Z. Wada, Aderonke Odetayo, Digitalization and artificial intelligence in migration and

Aanuoluwapo Clement David-olawade, Fiyinfoluwa mobility: Transnational implications of the covid-19

Asaolu and Judith Eberhardt. 2024. Enhancing mental pandemic. Societies, 11, 4, 135.

health with Artificial Intelligence: Current trends and https://doi.org/10.3390/soc11040135

future prospects. Journal of Medicine, Surgery, and [25] Luke Balcombe and Diego de Leo. 2022. Human-

Public Health, 3, 100099. Computer Interaction in Digital Mental Health.

https://doi.org/10.1016/j.glmedi.2024.100099 Informatics, 9,1, 14.

[12] Koki Shimada. 2023. The Role of Artificial Intelligence in https://doi.org/10.3390/informatics9010014

Mental Health: A Review. Science Insights 43, 5, 1119- [26] Nicholas C. Jacobson and Matthew D. Nemesure. 2021.

1127. doi:10.15354/si.23.re820 Using Artificial Intelligence to Predict Change in

[13] Max Rollwage, Johanna Habicht, Keno Juechems, Ben Depression and Anxiety Symptoms in a Digital

Carrington, Sruthi Viswanathan, Mona Stylianou, Tobias Intervention: Evidence from a Transdiagnostic

U. Hauser and Ross Harper. 2023. Using Conversational Randomized Controlled Trial. Psychiatry Research, 295,

AI to Facilitate Mental Health Assessments and Improve 113618.

Clinical Efficiency Within Psychotherapy Services: Real- https://doi.org/10.1016/J.PSYCHRES.2020.113618

World Observational Study. JMIR AI, 2, e44358. [27] Bennett Knox, Pierce Christoffersen, Kalista Leggitt, Zeia

https://doi.org/10.2196/44358 Woodruff and Matthew H. Haber. 2023. Justice,

[14] David D. Luxton. 2020. Ethical implications of Vulnerable Populations, and the Use of Conversational

conversational agents in global public health. Bulletin of AI in Psychotherapy. American Journal of Bioethics, 23,

the World Health Organization, 98, 4, 285–287. 5, 48–50.

https://doi.org/10.2471/BLT.19.237636 https://doi.org/10.1080/15265161.2023.2191040

[15] Leonard.Bickman. 2020. Improving Mental Health [28] Zoha Khawaja and Jean C. Bélisle-Pipon. 2023. Your

Services: A 50-Year Journey from Randomized robot therapist is not your therapist: understanding the

Experiments to Artificial Intelligence and Precision role of AI-powered mental health chatbots. Frontiers in

Mental Health. Adm Policy Ment Health, 47, 795–843. Digital Health, 5, 1278186. doi:

https://doi.org/10.1007/s10488-020-01065-8 10.3389/fdgth.2023.1278186

[16] Silvan Hornstein, Valerie Forman-Hoffman, Albert [29] Danilo Bzdok and Andreas Meyer-Lindenberg,. 2018.

Nazander, Kristian Ranta and Kevin Hilbert. 2021. Machine Learning for Precision Psychiatry: Opportunities

Predicting therapy outcome in a digital mental health and Challenges. Biological Psychiatry: Cognitive

intervention for depression and anxiety: A machine Neuroscience and Neuroimaging, 3, 3, 223–230.

learning approach. DIGITAL HEALTH, 7, 1-11. https://doi.org/10.1016/J.BPSC.2017.11.007

doi:10.1177/20552076211060659 [30] Jana Sedlakova and Manuel Trachsel. 2023.

[17] Josef Ehmer. 2021. A historical perspective on family Conversational Artificial Intelligence in Psychotherapy: A

change in Europe. In Norbert F. Schneider and Michaela New Therapeutic Tool or Agent? American Journal of

Kreyenfeld (eds). Research Handbook on the Sociology Bioethics, 23, 5, 4–13.

of the Family, 143– 161. https://doi.org/10.1080/15265161.2022.2048739

https://doi.org/10.4337/9781788975544.00018

[18] Keera Allendorf, Linda Young-Demarco and Arland

Thornton. 2023. Developmental Idealism and a Half-



Novejši razvoj rodnosti po redu rojstva v Sloveniji



Recent Development of Fertility in Slovenia by Birth Order



Janez Malačič



Ekonomska fakulteta Univerze v Ljubljani



Kardeljeva ploščad 17, 1000 Ljubljana, Slovenija



Elektronska pošta: janez.malacic@ef.uni-lj.si



POVZETEK ABSTRACT



Med demografi vse bolj prevladuje mnenje, da According to the opinion of more and more

rodnosti tako v Sloveniji kot drugod po Evropi demographers in developed world sub-



v prihodnjih letih in desetletjih ne bo mogoče replacement fertility in developed countries

povečati na raven, ki bi dolgoročno zagotavljala is here to stay. However, the idea that

enostavno obnavljanje prebivalstva. Kljub temu pronatalist economic and other population

pa je še zmeraj tudi živa ideja, da bi spodbujanje

policy measures can somehow help to

rojevanja drugega in tretjega otroka, morda pa

tudi četrtega otroka v družini, lahko dvignilo increase births of the second, the third or

rodnost na raven, ki bi skupaj z zmernimi neto even the fourth child in the family is still



pozitivnimi in selekcioniranimi migracijami alive too. Therefore, the analysis of the



zagotovila dolgoročno nemoteno obnavljanje fertility by birth order is of the special

prebivalstva. Zaradi tega je pomembna analiza interest. This paper shows that Slovenian



rodnosti po redu rojstva, ki smo jo opravili v sub-replacement fertility in the last three to

tem besedilu. Analiza za zadnja desetletja kaže four decades has been deeply founded in the

za Slovenijo, da v razmerah prenizke rodnosti prevalence of the births of the first and the



prevladujejo rojstva nizkih redov rojstva. To sta second birth orders as well as in the sharp

največkrat prva dva reda rojstva. Verjetnost

decline of the births of the higher orders.

povečanja družine, ki kaže, koliko družin z

dvema otrokoma se odloči za rojstvo tretjega The probability that the family with two

children will get the third one drastically

otroka, se po vseh naših tablicah drastično zniža

v primerjavi z verjetnostjo, da družina z enim decline in the comparison with the

otrokom dobi drugega otroka. Hkrati pa nekaj probability that the family with one child



večje verjetnosti povečanja družine za rojstva will get the second one is shown in the

višjih redov ne morejo kompenzirati tega fertility tables by birth order for Slovenia in

ključnega padca, ker je rojstev višjih redov v the years 2011 and 2021. Higher order



razmerah prenizke rodnosti zelo malo. Vse to pa births in Slovenia have been very low since

je skupaj s trdno uveljavljeno normo rojevanja the beginning of the 1980s. Fertility data by



le enega do dveh otrok v povprečju na družino birth order and our analysis show the

v Sloveniji zelo slaba popotnica za učinkovitost

prevalence of the very low reproductive

prebivalstvene politike, ki bi želela dvigniti

normo na povprečno dva do tri otroke na norms in Slovenia. The prospects for

družino. effective pronatalist policy and for the

increase of the number of births of all birth

orders and especially of the third and



eventually the fourth birth in Slovenia are enostavno obnavljanje prebivalstva in morda

weak and uncertain. celo niti ne na raven, ki bi skupaj s pozitivnimi

neto migracijami zagotovila ta cilj dolgoročne

demografske stabilnosti, pa se tako v stroki kot

v politiki iščejo tematike, kjer pa bi morda le bili

1 UVOD možni določeni pozitivni premiki. Ena od takih

Prebivalstvo Slovenije je podobno kot tematik je del področja rodnosti, ki obravnava

druga prebivalstva razvitih evropskih držav značilnosti in trende rojevanja po vrstnem redu

končalo demografski prehod po koncu rojstev. Pri tem je v ospredju upanje, da bi

začasnega izbruha rodnosti po drugi svetovni zasnovali, sprejeli in uresničevali politike, ki bi

vojni. To pomeni, da je zanj že nekaj več kot povečale rojevanje otrok drugega in še posebej

šestdeset let značilen moderni režim tretjega reda. Da bi pa bilo možno to uresničiti,

obnavljanja prebivalstva z nizko rodnostjo in je potrebno proučevati področje rojevanja po

nizko smrtnostjo. V zadnjih dveh tretjinah tega redu rojstev,

obdobja je splošna rodnost tako nizka, da ne

zagotavlja V tem besedilu bo avtor analiziral več dolgoročno nemotenega

novejše statistične podatke o rodnosti po redu

obnavljanja prebivalstva. Slovensko in cela

rojstev v Sloveniji. Pri tem je treba povedati, da

vrsta drugih evropskih prebivalstev lahko

počasi še naraščajo zaradi dovolj velikih so slovenski podatki vitalne statistike dobri,

večje težave pa se pojavijo pri podatkih, ki jih

pozitivnih neto migracij. Kjer pa teh ni, ali pa so

pridobivamo s popisi prebivalstva. V Sloveniji

prisotne celo bolj ali manj velike neto je bil zadnji klasični popis prebivalstva

emigracije, pa se prebivalstva takšnih evropskih

držav zmanjšujejo. Vendar tudi Slovenija in organiziran leta 2002, od leta 2011 naprej pa

Statistični urad Republike Slovenije (SURS)

podobne države z rastjo prebivalstva zaradi

večjega priseljevanja od odseljevanja ne morejo opravlja registrske popise. Težave s popisi

povzročajo raziskovalcem določene probleme,

računati na dolgoročno kompenzacijo prenizke

vendar so ti problemi pri raziskovanju rodnosti

rodnosti z dovolj velikimi neto priselitvami

po redu rojstva še bolj izraziti in bodo prisotni

zaradi prenizke ravni rodnosti, zelo majhnih

možnosti, da bi se raven rodnosti v zadostni tudi v analizi prikazani v tem besedilu.



meri dvignila in zaradi nepripravljenosti Besedilo je razdeljeno na običajne

velikega dela prebivalstev teh držav, da bi točke, kot so uvod, sklep ter literatura in viri.

sprejeli dovolj veliko število novih imigrantov. Razen tega pa vsebuje še dve osrednji vsebinski

Demografski modeli in projekcije kažejo, da bi točki. Prva med njima ima namen prikazati

že v nekaj desetletjih morale biti pozitivne neto najbolj osnovne statistike in značilnosti rodnosti

migracije na leto večje od letnega števila po redu rojstva v Sloveniji v zadnjih dveh do

živorojenih. To pa bi vodilo do politične treh desetletjih. Prikazanih bo tudi nekaj

nestabilnosti in veliko večjega nasprotovanja starejših podatkov, ki bodo pokazali razsežnosti

imigracijam, kot ga vidimo v Sloveniji in sprememb. Druga med njima pa je osrednja

Evropi v zadnjih letih. točka tega besedila in obravnava tablice

rodnosti po redu rojstva v Sloveniji v letih 2011

Tako zaostrene demografske razmere v

Sloveniji in v vse večjem številu evropskih in 2021.

držav vse bolj postavljajo demografsko

problematiko v ospredje družbenega zanimanja

in razprav o politikah, s katerimi bi vsaj omilili

procese, ki smo jih okvirno opisali v prvem

odstavku tega besedila. Ne glede na to, da med

demografi vse bolj prevladuje mnenje, da

rodnosti tako v Sloveniji kot drugod po Evropi

v prihodnjih letih in desetletjih ne bo mogoče

povečati na raven, ki bi dolgoročno zagotavljala



2 RODNOST PO REDU ROJSTVA V velike večine rojstev in s tem tudi rojstev nižjih

SLOVENIJI V MODERNEM REŽIMU redov je prišlo v prvih desetih letih trajanja

NIZKE RODNOSTI zakonske zveze. Le pri rojstvih višjih redov od

tretjega in naprej so bili leta 1991 v Sloveniji

Analiza rodnosti je tradicionalno deleži drugega desetletja trajanja zakonske

namenjala veliko pozornosti zakonski rodnosti. zveze še tudi pomembni, vendar je bilo teh

V prebivalstvih z visoko ravnjo poročenosti je

rojstev že takrat malo. (Malačič, 2000, str. 98-

ta rodnost določala značilnosti celotne rodnosti,

99) Navedeni podatki so transverzalni, saj nam

hkrati pa se je zelo razlikovala od zunaj

longitudinalni niso na voljo. Zelo verjetno pa se

zakonske rodnosti. V Sloveniji se je že pred več osnovne značilnosti vpliva trajanja zakonske

desetletji zelo znižalo absolutno in relativno

zveze na rodnost v naslednjih desetletjih niso

število sklenjenih zakonskih zvez. Zakonodaja

je z mnogih vidikov izenačila pravni položaj bistveno spremenile.

zakonskih in izven zakonskih skupnosti, število Znižanje števila in deležev živorojenih

slednjih pa se je izrazito povečalo. Zaradi tega višjih redov rojstva v Sloveniji se kaže tudi v

se je precej zmanjšala analitična vrednost obdelavi in objavljanju uradnih statističnih

študija zakonske rodnosti. To velja tudi za podatkov v statistični bazi SURS SISTAT. Več

rodnost po vrstnem redu rojstva. V razmerah, ko obdelav v tej bazi se preprosto konča pri redu

se v Sloveniji rodi okrog 56 % živorojenih na rojstva 4+, čeprav so se obdelave še v zadnjih

leto izven zakonske zveze, nam značilnosti desetletjih prejšnjega stoletja končevale z

zakonskih rojstev ne dovoljujejo posploševanja odprtimi razredi 6+, 7+, 8+ ali celo 9+. Zaradi

na celotno rodnost. Po drugi strani pa pogosto tega smo lahko vključili v tabelo 1 znatno manj

slišimo, da se izven zakonske skupnosti podatkov po redu rojstva za Slovenijo, kot bi si

obnašajo podobno kot poročeni pari, Ker pa želeli.

imamo za njih veliko manj in slabše statistične



statistike težje ali nemogoče analizirati. Zaradi podatke o številu živorojenih v Sloveniji po redu rojstva v izbranih letih obdobja 1991-2023. podatke, jih je na osnovi običajne uradne Tabela 1 prikazuje letne (transverzalne)

vsega navedenega bomo vpliv trajanja zakonske Red rojstva se konča z zgornjim odprtim

zveze na rodnost in red rojstev v Sloveniji

razredom 4 in več (4+). Izbrana leta upoštevajo

prikazali s podatki za leto 1991. Tega leta se je

30,0 % živorojenih v zakonski zvezi rodilo v logiko desetletnih med popisnih intervalov,

čeprav je SURS to logiko deloma porušil.

prvem letu trajanja zakonske zveze, 51,8 % pa

Dodano pa je še leto 2023, za katero imamo

se jih je rodilo v prvih treh letih trajanja

zadnje objavljene podatke. Prikazani so

zakonske zveze. Za starostne razrede trajanja

absolutni podatki in relativna struktura v

zakonske zveze 5-9, 10-14, 15-19 in 20+ so bili

odstotkih.

odstotki naslednji 21,5, 6,5, 1,8 in 0,4. Do



Tabela 1: Absolutno in relativno število živorojenih v Sloveniji po redu rojstva v izbranih letih obdobja

1991-2023. Vir: SISTAT, dostop 7.9. 2024



Leto 1991 2001 2011 2021 2023

Rojstva 1 Št. % Št. % Št. % Št. % Št. %

1 2 3 4 5 6 7 8 9 10 11

N 21.583 100,0 17.477 100,0 21.947 100,0 18.984 100,0 16.989 100,0

N1 10.210 47,3 8.599 49,2 10.637 48,5 8.557 45,1 7.884 46,4

N2 8.510 39,4 6.371 36,5 8.327 37,9 7.487 39,4 6.391 37,6

N3 2.179 10,1 1.865 10,7 2.210 10,1 2.244 11,8 2.094 12,3

N4+ 684 3,2 642 3,7 773 3,5 696 3,7 620 3,6

1 N je število živorojenih; N , … , N so zaporedoma števila rojstev od prvega reda do reda štiri in več.

1 4+



Število živorojenih v Sloveniji v tega reda. Takšne stopnje bomo podrobneje

izbranih letih, ki so prikazana v tabeli 1 precej obravnavali v naslednji točki tega besedila.

niha in je najnižje leta 2023, ki je zadnje leto, za

Stopnje rodnosti po redu rojstva glede

katerega imamo podatke. Bolj presenetljivo pa na ženske, pri katerih lahko pride do teh rojstev,

je, da se struktura živorojenih v Sloveniji po

pa je povezano s kazalci verjetnosti povečanja

redu rojstva v vseh teh letih zelo malo ali skoraj družine. Tak kazalec nam pove, kakšna je



rodilo največ otrok prvega reda, delež se giblje verjetnost, da se družina z danim številom otrok nič ne razlikuje. V vseh izbranih letih se je

poveča za dodatnega otroka. Verjetnosti



je rojstev prvega in drugega reda, delež teh dveh povečanja družine lahko računamo na med 45,1 in 49,2 odstotka. Velika večina rojstev



živorojenih tretjega reda je le malo večji kot osnovi. Prvi način se uporablja za kohorte, ki so končale svojo rodno obdobje. Pri njem redov se giblje med 84,0 in 86,7 odstotka. Delež longitudinalni (kohortni) in transverzalni

10,0 odstotkov, delež četrtega reda in več pa je

uporabljamo popisne podatke. Težava pa je v

nizek in se giblje med 3,2 in 3,7 odstotka.

Navedeni statistični podatki kažejo, da je tem, da imajo kazalci bolj zgodovinsko kot

praktično vrednost, saj nam kažejo, kaj se je

prebivalstvo Slovenije v obdobju 1991-2023 dogajalo v kohorti, ki je že končala svoje rodno



rojstva in rojeva le dokler ne doseže števila obdobje. Tega načina tukaj ne bomo tipično maltuzijansko prebivalstvo, ki načrtuje

rojstev na družino oziroma partnersko skupnost, obravnavali.



ki je bilo načrtovano. Pri načrtovanju še zmeraj Praktično bolj uporaben je transverzalni

pride do odstopanj, vendar so ta majhna. način izračunavanja verjetnosti povečanja

Majhne spremembe v strukturi živorojenih po družine. V literaturi je na voljo več načinov

redu rojstev pa kažejo še, da se je v Sloveniji v izračunavanja takšnih verjetnosti. (Malačič,

zadnjih desetletjih ustalila in verjetno utrdila ne 2006, str. 97-100) Prvi način uporablja podatke

samo nizka norma števila živorojenih otrok na o zakonski rodnosti po redu rojstev in modelske

eno žensko, ampak norma, ki je le eden do dva porazdelitve rojstev reda i glede na trajanje

otroka na žensko. razmaka do reda i-1. Tudi tega načina tukaj ne



Trditev v zadnjem stavku prejšnjega bomo obravnavali. Namesto tega bomo v

naslednji točki prikazali izračun verjetnosti

odstavka lahko podkrepimo tudi s podatki povečanja družine na osnovi transverzalnih

zadnjega na terenu izpeljanega popisa

podatkov po redu rojstva za vsa rojstva.

prebivalstva leta 2002. Popis je zbral podatke o

številu žensk starih 15 in več let po tem, ali je

ženska rodila ali ni rodila in po tem, koliko

otrok je rodila. Navedenega popisnega leta je 3 TABLICE RODNOSTI PO REDU

bilo v Sloveniji med 15 in več let starim ROJSTVA V SLOVENIJI V LETIH 2011 IN

ženskim prebivalstvom 26,5 odstotka žensk, ki 2021



niso ali še niso rodile. Deleži med njimi, ki so Za sestavo oziroma izračun tablic

rodile zaporedoma enega, dva, …., do 6 in več rodnosti po redu rojstva na osnovi letnih

živorojenih otrok pa so bili 19,5, 37,2, 11,1, 3,5, oziroma transverzalnih podatkov potrebujemo

1,3, in 1,1 odstotka. Iz teh podatkov izračunano podatke o številu živorojenih po redu rojstva za

povprečno število živorojenih otrok na eno izbrano leto, o srednji starosti matere ob rojstvu

žensko je bilo 1,55, kar je blizu stopnji totalne otroka i-tega reda ter podatke o specifičnih

(celotne) rodnosti v zadnjih letih. (SISTAT, stopnjah rodnosti po redu rojstva ri, ki jih

dostop 7. 9. 2024) izračunamo tako, da število rojstev i-tega reda



Podatki o številu živorojenih po redu delimo s številom žensk, ki se jim ti otroki lahko

rojstev nam omogočajo izračunati različne vrste rodijo. Če vzamemo za primer stopnjo rodnosti

stopenj rodnosti po redu rojstva. Med njimi so drugega reda, potem število rojstev drugega



še posebej zanimive stopnje, pri katerih reda delimo s številom žensk rodne starosti, ki

primerjamo število živorojenih danega reda so rodile enega otroka. Za rodno starost lahko

rojstva s številom žensk, ki lahko rodijo otroke vzamemo starostni razred 15-49 ali pa 15-44. V

našem primeru smo vzeli slednjega, ker Ko smo izračunali stopnje ri , lahko v

uporabljamo pri izračunu verjetnosti povečanja nadaljevanju izračunamo verjetnost povečanja

družine predpostavko, da ženske rojevajo do 45. družine po obrazcu, ki sta ga utemeljila Chiang,

leta. Dejansko število rojstev starejšim materam C.L. in van den Berg, B.J. (Chiang, C.L. and

od 45 let pa je zelo nizko. Na ta način smo van den Berg, B.J., 1982). Obrazec za verjetnost

stopnje ri izračunali po naslednjem obrazcu: ri = ai je: ai = (xw – xi) * ri / 1 + ((xw – xi+1) * ri). Pri

Ni/30Vf,i-1,15, pri čemer je i enak 1, 2, …, n+. V tem je xw enak 45,0, poševnica / je znak za

preteklosti smo lahko dobili podatke za deljenje, zvezdica * pa znak za množenje. Ko

imenovalec specifičnih stopenj rodnosti ri le za imamo na tak način vse potrebne podatke za

popisna leta. V Sloveniji, kjer imamo register sestavo tablic rodnosti po redu rojstva, jih

prebivalstva in registrsko statistiko, pa nam moramo še izračunati in zaključiti. Za stolpec 5

statistična baza SISTAT omogoča izračun tudi v tabeli 2 uporabimo izhodiščno (modelsko)

za druga leta za katera v bazi objavijo podatke. število 100.000 živih žensk, na katere

Med štirimi leti, za katera so na voljo podatki o apliciramo izračunane verjetnosti povečanja

ženskah starih 15 in več let po enoletnih in družine. Stolpec nam kaže število žensk brez

petletnih starostnih razredih po številu otrok, z najmanj enim otrokom in tako dalje.

živorojenih otrok (brez, z enim, z dvema Stolpec 6 dobimo kot razliko zaporednih

otrokoma itd.) smo za našo analizo izbrali leti vrednosti stolpca 5 deljeno s 100.000.

2011 in 2021.



Tabela 2: Tablice rodnosti na osnovi reda rojstev za Slovenijo za leti 2011 in 2021. Vir: Chiang, C.L.

and van den Berg, B.J., 1982; Malačič, J., 2006 in SISTAT, dostop 9. 9. 2024.



Red rojstva Srednja Specifične Verjetnost Število žensk Končna

starost ob stopnje povečanja z i in več razdelitev po

rojstvu rodnosti družine živorojenih redu (v %)

i xi ri ai li di/l0

1 2 3 4 5 6

Slovenija 2011

0 15,0 0,05527 0,87480 100.000 12,6

1 28,8 0,10785 0,70520 87.480 25,8

2 31,3 0,02211 0,24065 61.691 46,8

3 33,3 0,03203 0,28315 14.846 10,6

4+ 34,9 4.204 4,2

Slovenija 2021

0 15,0 0,04966 0,84419 100,000 15,6

1 29,6 0,11511 0,70360 84.419 25,0

2 31,8 0,02501 0,15589 59.397 44,2

3 33,4 0,03159 0,27914 15.199 11,0

4+ 35,1 4.243 4,2



V tabeli 2 smo prikazali tablice rodnosti dve tretjini družin z enim otrokom se odloči za

na osnovi reda rojstva za Slovenijo za leti 2011 drugega otroka. Naslednja verjetnost, da

in 2021. Najbolj zanimiv kazalec v tablicah je družina z dvema otrokoma dobi tretjega otroka,

verjetnost povečanja družine v četrtem stolpcu. pa drastično pade. V Sloveniji je še nadalje

Vrednosti tega kazalca sta dokaj visoki za padla med letoma 2011 in 2021. Leta 2021 se je

verjetnost, da družina brez otrok dobi prvega le 15,6% družin z dvema otrokoma odločilo še

otroka in da družina z enim otrokom dobi za tretjega otroka. Po podatkih SISTAT-a je bilo

drugega otroka. Vendar sta tudi ti dve vrednosti v Sloveniji v starostnem razredu žensk 50-54 let

med letoma 2011 in 2021 padli. Le malo več kot leta 2011 6,4 % in leta 2021 10,1 % žensk, ki



niso rodile. (SISTAT, dostop 9. 9. 2024) otrokom dobi drugega otroka. Hkrati pa nekaj

Razlika med tema podatkoma je presenetljivo večje verjetnosti povečanja družine za rojstva

velika in očitno precej niha med petletnimi višjih redov ne morejo kompenzirati tega

starostnimi razredi, ki so dokončali rojevanje in ključnega padca, ker je rojstev višjih redov v

rodno starost. razmerah prenizke rodnosti zelo malo. Vse to pa

je skupaj s trdno uveljavljeno normo rojevanja

Za primerjavo z zgornjima dvema le enega do dveh otrok v povprečju na družino

letoma navedimo, da so bile vrednosti v Sloveniji zelo slaba popotnica za učinkovitost

verjetnosti povečanja družine, prav tako



izračunane po prikazani metodologiji, za normo na povprečno dva do tri otroke na prebivalstvene politike, ki bi želela dvigniti

Slovenijo za leto 1961 v primeru zaporedja a0 družino. Vsi matematični demografski modeli

,a1 , … , a5 naslednje 0,93142, 0,80099, kažejo, da samo slednja norma zagotavlja

0,57079, 0,46080, 0,466523 in 0,48388, za leto dolgoročno nemoteno obnavljanje prebivalstva.

1981 pa zaporedoma 0,93811, 0,81473,

0,30978, 0,25493, 0,28310 in 0,32492.

(Malačič, 2006, str. 99) Leta 1961 je bilo v

LITERATURA IN VIRI

Sloveniji 28.955 in leta 1981 29.220



živorojenih. (SL-2013, str. 77-78) To število 1. Chiang, C.L. and van den Berg, B.J.

vpliva na velikost ri in tudi na dolžino vrste, ki (1982) Fertility Table fort he

smo jo lahko izračunali. Vrednosti ai so bile v Analysis of Human Reproduction.

obeh letih občutno večje kot tiste v tabeli 2, Mathematical Biosciences, 62.

čeprav so se tudi že v teh dveh letih znižale. Leta 2. Feichtinger, G. and Lutz, W. (1983)

1961 so bile še deloma pod vplivom končevanja Eine Fruchtbarkeitstafel auf

demografskega prehoda. Paritatsbasis. Zeitschrift fur

Bevolkerungswissenschaft, Jg. 9,

3/1983. Boldt Verlag Boppard.

4 SKLEP 3. Malačič, J. (2000) Demografija.

Teorija, analiza, metode in modeli,

Podatki o rodnosti po redu rojstva ter na



njihovi osnovi izračunane tablice rodnosti po Ljubljana. 4. izdaja, Ekonomska fakulteta, redu rojstva in še posebej verjetnosti povečanja

4. Malačič, J. (2006) Demografija.

družine za Slovenijo, ki so prikazani in

Teorija, analiza, metode in modeli,

analizirani v tem besedilu se praviloma

nanašajo na posamezna izbrana leta in so po 6. izdaja, Ekonomska fakulteta,

Ljubljana.

naravi transverzalni. Raziskovalci smo jih 5. Statistični letopis Republike največkrat prisiljeni uporabljati zaradi tega, ker

Slovenije (2013), SURS, Ljubljana.

tako kohortni oziroma longitudinalni pa tudi

6. SISTAT (2024) Statistični urad

transverzalni podatki za mnoga druga leta niso

Republike Slovenije, Ljubljana.

na voljo. Pogosto si pomagamo tudi z ocenami

in približki pa tudi s posebnimi metodami

analize za nepopolne in specifične podatke.



Ne glede na vse to, pa je naša analiza v

tem besedilu pokazala presenetljivo enotno in

dosledno sliko na področju, ki smo ga

analizirali. V razmerah prenizke rodnosti

prevladujejo rojstva nizkih redov rojstva. To sta

največkrat prva dva reda rojstva. Verjetnost

povečanja družine, ki kaže, koliko družin z

dvema otrokoma se odloči za rojstvo tretjega

otroka se po vseh naših tablicah drastično zniža

v primerjavi z verjetnostjo, da družina z enim



The aggregate loss of GDP resulting from premature



mortality: a methodological approach based on GDP



decomposition at occupational level



Tjaša Redek† Petra Došenović Daša Farčnik Tanja Istenič

School of Economics and Bonča School of Economics and School of Economics and

Business Business Business School of Economics and

University of Ljubljana University of Ljubljana University of Ljubljana Business

Slovenia Slovenia Slovenia University of Ljubljana

tjasa.redek@ef.uni-lj.si @ef.uni-lj.si Slovenia dasa.farcnik tanja.istenic@ef.uni-lj.si

petra.d.bonca@ef.uni-lj.si



Abstract only represents a personal and family loss, but also a significant Premature death (dying before the age of 65) represents not

This paper proposes a novel method for estimating the economic loss of human capital and consequently a loss of productivity and

burden of premature mortality in the working population. The GDP, amounting to the cumulative loss of value added reflecting

method is based on GDP (value added) decomposition at the the burden of premature mortality. The economic burden of

occupational level. The approach estimates the loss of value premature mortality can be estimated at the microeconomic or



integrating data on employed individuals, causes of death, and Health economics assesses the economic costs of disease company balance sheets, the results provide a decomposition of using different methods [2]. At the macroeconomic (societal) GDP losses by occupational groups. The results show significant level, the losses are both market losses (non-health consumption) differences in the value added lost between occupational groups, and economic welfare losses (which consider not only the market, with professionals and technicians contributing the most to the productivity of individuals in different occupations. By approach to estimate the economic burden of premature mortality at the macroeconomic level. added due to premature deaths in Slovenia by considering the macroeconomic level [2]. This paper proposes an alternative

but also non-market losses). While market losses are usually

total loss. This approach provides an alternative estimate of estimated using general or partial equilibrium models (both

economic losses compared to traditional methods by taking into simulation and regression based), the economic welfare losses

account the contribution of each occupational group to GDP. are estimated using full-income models. The WHO [2]

However, the study also acknowledges limitations, such as systematically summarizes the approaches proposed and used in

uncertainties regarding future retirement ages and a relatively the literature to estimate the macroeconomic consequences of

small sample size for certain occupations. This methodology diseases.

provides valuable insights for policymakers about the The Cost of Illness (CoI) approach focuses on estimating the

importance of addressing the economic burden of premature total economic burden of a particular disease by combining both

mortality in the labour force. direct and indirect costs [3], [4], [5]. Direct costs include medical

expenses, treatment, and travel costs, while indirect costs

Keywords consider the loss of productivity due to illness, absenteeism, or

Premature mortality, productivity loss, economic burden, value premature death. This method quantifies the overall economic



added decomposition, occupational level burden of a disease on society, often expressed as a percentage of GDP. Despite its usefulness in providing a broad estimate of

the financial burden, the CoI approach has limitations. It tends to



1 Introduction on medical expenditures and labor productivity losses, without oversimplify the broader economic effects by focusing primarily

Between 2009 and 2022, an average of just over one fully capturing long-term effects such as capital depletion or

thousand people of working-age died in Slovenia, which changes in labor supply. In addition, non-market impacts may

corresponds to around 0.11 % of the 900 thousand people in not be considered [2].

employment on average. During this period, the average age of Regression-based models estimate the impact of health

death was 49.4 years and the individuals retired at an average age indicators, such as life expectancy or mortality, on economic

of 59 years [1]. Not reaching the average retirement age means output, usually GDP. The method uses econometric models to

that each individual lost an average of 8 working years. identify statistical relationships between health and economic

†Corresponding author outcomes, relying on historical or cross-country panel data. The

Permission to make digital or hard copies of part or all of this work for personal or main advantage of regression models is their simplicity and

classroom use is granted without fee provided that copies are not made or relatively low data requirements. However, endogeneity issues—

distributed for profit or commercial advantage and that copies bear this notice and when health and income influence each other—can affect the

the full citation on the first page. Copyrights for third-party components of this results, and the models are sensitive to the specification of the

work must be honored. For all other uses, contact the owner/author(s). production function [2].

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Calibration models combine micro-level estimates of health

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.dem.14 impacts on income with macro-level data, such as GDP,

demographic data, and labour force statistics, to simulate how



health changes affect national economic output. The calibration without a constant to allocate aggregate variable values (e.g.,

approach allows for the decomposition of health impacts across household consumption) to individuals. Similarly, we regress

different countries or regions, providing flexibility in scenarios firm’s value added on the share of broad occupational groups.

where data is limited. However, the calibration models are The beta coefficients are provided in the appendix (Table A1).



effects is more challenging to capture accurately. each occupational group according to the following formula: The full-income model approach primarily sensitive to mortality impacts, while morbidity (illness) The beta coefficients are then used to calculate the weights for

goes beyond traditional



of lost production and the value of lives lost to illness. It uses 𝛽 𝑖 ∙ ∑ 𝑥𝑖 market-based valuations by estimating both the monetary value techniques such as the Value of Statistical Life (VSL) or 𝑤𝑒𝑖𝑔ℎ𝑡𝑖 = 10 ∑ 𝑖 𝛽𝑖 ∙ ∑ 𝑥𝑖

Willingness to Pay (WTP) to quantify social welfare losses [6],

[7]. Although full-income models capture the broader societal Where 𝑖 represents each occupational group i∈(1,10), 𝛽 is the

costs of illness, including non-market impacts, they involve calculated regression coefficient for each occupational group,

contentious assumptions, such as discount rates and VSL, which and 𝑥 is the number of employees in each occupational group.

can lead to extreme or difficult-to-interpret results [2]. The added value per employee (VAEreal,i) according to the

We extend existing methods by building on the occupational profile is calculated using the following formula, National

Transfer Accounts (NTA) approach to the decomposition of where TOT_ VAEreal,i is the total real value added of the

value added company: [8], which measures the economic impact of

premature death by assessing the reduction in GDP, that is, by

estimating the loss of productivity. This implies calculating the VAE real,𝑖 = TOT_ VAE𝑟𝑒𝑎𝑙 ∗ 𝑤𝑒𝑖𝑔ℎ𝑡𝑖

contribution that individuals would have made to GDP is

calculated, based on their specific sector and productivity level. In the next step, we calculate the number and average age of

This productivity loss is then multiplied by the number of years individuals by occupational groups who died prematurely. Based

the individual would have continued to work. The total losses are on the actual retirement age during the observed period (on



[9]. calculated (by subtracting the retirement age and the age of death by occupational group) and multiplied by the estimated value then aggregated to determine the overall impact on the economy average only 59 years), the number of lost productive years was

added of each occupational profile (VAEreal,i). By summing the



2 Methodology and data economic burden of premature mortality. lost value added due to lost productive years we estimate the

The methodology builds on the production loss approach due The analysis was prepared using micro-data from registries

to premature mortality and the resulting loss of GDP. However, covering the period between 2009 and 2022 [1]. The database

instead of estimating productivity loss using (sectoral) averages, integrates several different population data sources: (1) the

we recognise that the actual productivity of firms is highly Statistical registry of active population database from 2009 to

dependent on their human capital structure, which is the core of 2022, which contains annual data on 838,000 to 998,000

their intangible capital [10]. Intangible capital generates a large individuals, with data on age, occupation, and employment

part of the value added in a knowledge-based economy [11], [12]. sector; (2) cause-of-death data for the period 2009 to 2022, which

Therefore, we assess the loss of value added by assessing the contains the 4-digit code of the cause of death and death date; (3)

contribution of individuals through their occupation, which more company balance sheet data for again for 2009 to 2022. The

accurately captures the actual loss of value added due to merged database of employed individuals and companies was

premature death. used for the decomposition of value added by occupational group,

The methodological analysis is based on several steps and while the data on the number of deaths was used to estimate the

originates from the production function of a company, focusing loss of value added. Due to a small number of observations,

on the contribution of the individual occupations or occupational military occupations were excluded.

groups to value creation. First, the employees were categorized

into ten broad occupational groups according to the Standard

Classification of Occupations, which aligns with the 3 Results International Classification of Occupations. These broad



occupational groups are: (0) military professions, (1) legislators, in employment in Slovenia, averaging around 900,000 annually The analysis is based on a population of 12.7 million people

senior officials, managers, (2) professionals, (3) technicians and



people, (6) agricultural, forestry, fishery, and hunting representing 0.11% of all observed individuals [1]. occupations, (7) non-industrial labor, (8) machine operators, Table 1 presents the data on the cumulative number of industrial producers and assemblers, (9) elementary occupations. premature deaths among the employed population in Slovenia associate professionals, (4) clerks, (5) service workers, sales individuals from this group died, about 1,000 per year, from 2009 to 2022. During this period, approximately 14,000

Analyses show that the contributions of different occupational



sensible to account for the actual structure of employees and their some differences among occupational groups. For example, real contribution when calculating the loss of added value. managers and professionals retired at an older age, whereas the The productivity of employees, measured by value added per employees in elementary occupations retired younger. groups to value added vary [10], [13], [14], [15], making it age of 49.4. Individuals generally retired at 59 years, but with between 2009 and 2022. On average, the individuals died at the

employee, is calculated by decomposing real value added at The loss of productive years, calculated as the difference

company level for the period between 2009 and 2022 to the level between the average age at retirement and the average age at

of different occupations. The decomposition of value added death, was highest in the group of professionals, who lost an

follows the NTA methodology [8], which uses regression average of 9 years, and was lowest among skilled agricultural,



fishery and forestry workers, where the average loss was only 1 0,0 5,0 10,0 15,0 20,0 25,0 year.

Occupational decomposition of aggregate value added at 1 Managers

firm level allowed us to estimate the average contribution of each 2 Professionals

occupation group to the value added, considering also their 3 Technicians and Associate relative size. Using these estimates, and the data on the average Professionals

number of premature deaths, average age at death and average 4 Clerical Support Workers

retirement age by occupational groups, the estimated loss of 5 Service and Sales Workers

productive years was calculated (Table 1). 6 Skilled Agricultural, Forestry and

Fishery Workers

7 Craft and Related Trades Workers

Table 1: The number of employed individuals’ premature

deaths, average age at death and retirement, loss of 8 Plant and Machine Operators, and

Assemblers

productive years of life, 2009-2022 9 Elementary Occupations



Number Average Average Loss of % of total value lost

age at age at productive

of deaths death retirement years of life Figure 2: The contribution of occupations to total value

0 Armed Forces

71 44.2 55.8 11.6 added lost (in % of all value added lost)

Occupations

1 Managers 698 53.4 61.2 7.8

2 Professionals 1,678 52.3 61.3 9.0

3Technicians and 4 Discussion and conclusion

Associate 1,828 51.2 59.2 8.0 This study offers a novel approach to estimating the

Professionals economic burden of premature mortality in Slovenia by focusing

4 Clerical Support

Workers 966 50.2 58.3 8.1 on the loss of added value and consequently the potential loss of

5 Service and GDP. In contrast to traditional approaches that might rely on

Sales Workers aggregate data or income loss models, this study employs a more 1,476 49.8 57.9 8.1

6 Skilled detailed decomposition method based on the added value



Agricultural, generated by companies, the occupational structure, and the 836 57.1 58.2 1.0 Forestry, Fishery premature mortality rates. This approach allows for a more Workers precise estimate of the economic burden by considering the 7 Craft and specific contributions to value added of the different Related Trades 2,573 49.9 58.7 8.8 occupational groups, which differ significantly in their Workers productivity and value added in the economy. 8 Plant, Machine

Operators, 1,571 50.0 58.0 8.0 One of the key strengths of this methodology is the use of

Assemblers microdata, which provides a detailed overview of the economic



9 Elementary burden of premature mortality at the level of individual 1,562 49.8 57.8 8.0 Occupations occupational groups. By using data from multiple sources, Total 13,259 49.4 59.0 9.6 including the Statistical registry of the active workforce, Data: [1], own calculations. mortality data, and company financial statements, the study



The occupational group of professionals contributed the most as it acknowledges the differential impact of premature deaths achieves a high level of specificity. This level of detail is crucial,

to the total number of lost productive years (28%), as the number across different sectors and occupations, which is often

of premature deaths was highest in this occupational group. This overlooked in broader, less nuanced analyses.

was followed by technicians and associate professionals, for However, the study also has limitations that need to be

whom the average number of years lost was also high and the considered. One important limitation is the inability to accurately

relative group size was even larger. This group contributed about predict future retirement ages, especially given the trend towards

one fifth of years lost. increasing retirement ages in Slovenia and elsewhere. This

Figure 1 presents data on the contribution of each introduces a potential underestimation of economic loss, as

occupational group to the total value added lost due to premature individuals might work longer in the future than currently



thetotal value added lost. Followed by technicians and plant and size of premature deaths within certain occupational groups may affect the robustness of the results, and the lack of control for deaths. Professionals contributed the most, around a fifth of assumed in the model. In addition, the relatively small sample

machine operators. other risk factors may also bias the results. This also means that

the actual numbers could be larger. Furthermore, while the study

introduces an alternative methodology, the approach could

benefit from further refinement, particularly in the

decomposition method, which would take greater account of

worker heterogeneity.

While the proposed methodology provides a detailed and

occupation-specific approach to assessing the economic burden



of premature death, its applicability may be limited by certain [6] L. A. Robinson and J. K. Hammitt, “Valuing Reductions

assumptions and data limitations. Future research could enhance in Fatal Illness Risks: Implications of Recent Research,”

this model by incorporating more dynamic elements, such as Health Econ., vol. 25, no. 8, pp. 1039–1052, Aug. 2016,

evolving retirement patterns, changing productivity levels and doi: 10.1002/hec.3214.

broader demographic trends, to provide a more comprehensive [7] W. Viscusi, “The Value of Risks to Life and Health,” J.

picture of the economic impact of premature mortality. Despite Econ. Lit., vol. 31, no. 4, pp. 1912–46, 1993.

these limitations, the study contributes significantly to the [8] United Nations, “Measuring and Analysing the

literature by providing a more nuanced understanding of the Generational Economy: National Transfer Accounts

economic costs associated with premature mortality that could Manual,” United Nations Department of Economic and

inform both policy decisions and future research directions. Social Affairs, Report, 2013. Accessed: Jul. 12, 2024.

[Online]. Available:

Acknowledgments https://www.un.org/en/development/desa/publications/m

easuring-and-analysing-the-generational-economy.html

This work was partially financed by ARIS projects: projects P5- [9] P. Hanly, M. Ortega-Ortega, and I. Soerjomataram,

0128, P5-0441, J5-4575, J5-4540, P5-0117. “Cancer Premature Mortality Costs in Europe in 2020: A

Comparison of the Human Capital Approach and the



References Friction Cost Approach,” Curr. Oncol., vol. 29, no. 5, Art. no. 5, May 2022, doi: 10.3390/curroncol29050287.



[1] Statistični urad Republike Slovenije, “Registrski podatki, [10] C. Bloch, A. Protogerou, and N. S. Vonortas, Eds.,

pridobljeni v okviru pogodbe 9602_269_2023.” 2023. Intangible Assets, Productivity and Economic Growth:

[2] World Health Organization, “WHO guide to identifying Micro, Meso and Macro Perspectives, 1st edition.

the economic consequences of disease and injury,” Routledge, 2024.

[11] H. Piekkola, “Intangible capital: The key to growth in

World Health Organization, Geneva, 2009. Accessed: Europe,” Intereconomics, vol. 46, no. 4, pp. 222–228, Sep. 17, 2024. [Online]. Available:

2011, doi: 10.1007/s10272-011-0387-2.

https://www.who.int/publications/i/item/9789241598293

[3] T. J. Moore and J. P. Caulkins, “How cost-of-illness [12] H. Piekkola, “Broad-based intangibles as generators of

growth in Europe,” Econ. Innov. New Technol., vol. 27,

studies can be made more useful for illicit drug policy

analysis,” no. 4, pp. 377–400, 2018, doi:

Appl. Health Econ. Health Policy., vol. 5, no.

10.1080/10438599.2017.1376170.

2, pp. 75–85, 2006, doi: 10.2165/00148365-200605020-

[13] C. Bloch, H. Piekkola, M. Rybalka, C. Eklund, and K.

00002. van Criekingen, “Measuring intangible assets at the firm

[4] P. Hanly, M. Ahern, L. Sharp, D. Ursul, and G.

Loughnane, “The cost of lost productivity due to level – development of an occupation based approach,”

Aarhus University, Deliverable 4.3, 2021.

premature mortality associated with COVID-19: a Pan-

European study,” [14] T. Istenič, T. Redek, and D. Farčnik, “Gender and Age

Eur. J. Health Econ. HEPAC Health Wage–Productivity Gaps in Intangible and Non-

Econ. Prev. Care, vol. 23, no. 2, pp. 249–259, Mar. Intangible Work Occupations,” Econ. Bus. Rev., vol. 26,

2022, doi: 10.1007/s10198-021-01351-8. no. 1, pp. 1–12, Mar. 2024, doi: 10.15458/2335-

[5] Y. A. Kim et al., “The economic burden of breast cancer 4216.1332.

in Korea from 2007-2010,” Cancer Research and [15] H. Piekkola, C. Bloch, M. Rybalka, and T. Redek,

Treatment, vol. 47, no. 4. Korean Cancer Association, “Intangibles from innovative work – their valuation and

pp. 583–590, 2015. doi: 10.4143/crt.2014.143. technological change,” University of Vaasa, Deliverable

5.3, 2021.





Mitigating Adverse Health Effects of Chronic Illnesses on




Work Outcomes: A Conceptual Framework



Petra Došenović Mojca Bavdaž Nina Ponikvar Tjaša Redek

Bonča† School of Economics and School of Economics and School of Economics and

School of Economics and Business Business Business

Business University of Ljubljana University of Ljubljana University of Ljubljana

University of Ljubljana Slovenia Slovenia Slovenia

Slovenia mojca.bavdaz@ef.uni-lj.si nina.ponikvar@ef.uni-lj.si tjasa.redek@ef.uni-lj.si

petra.d.bonca@ef.uni-lj.si



Abstract



The paper offers a comprehensive framework for linking health

outcomes and their determinants to work outcomes through

understanding how health, workplace and institutional

environment affect work outcomes. Its goal is to identify the

complex determinants of work ability and both favourable and

adverse work outcomes. Work ability is determined by health

and quality of life outcomes, which are influenced by both

modifiable and non-modifiable health factors as well as macro,

mezzo and micro-level incentives to improve health outcomes

including job demands, and workplace support. Work outcomes-

oriented improvement strategies also support work retention and

return to work. Figure 1: Mitigating adverse health effects of chronic

illnesses on work outcomes: a conceptual framework



Keywords



health determinants, health outcomes, work ability, work 2 Determinants of Health, Adverse Health and

outcomes, health and work outcomes-oriented improvement Work Ability Effects, and Work Outcomes

incentives

2.1 Determinants of Health



1 Health status is fundamentally determined by the conditions Introduction

in which we are born, grow, live, and age [1]. These conditions,

The ability to retain employment or achieve faster return to known as health determinants, factors, or drivers, operate at

work (RTW), particularly for individuals with chronic illnesses various levels and in different contexts. In examining health

or disabilities, is an increasingly important topic in an aging determinants, we follow the WHO [2] framework, distinguishing

society, requiring prolonged labour market participation. The between individual, socioeconomic, sociocultural, and

increasing prevalence of long-term health conditions, an aging environmental factors.

workforce, and fluctuating job demands have highlighted the Individual factors include personal attributes that impact

importance of maintaining work ability. health, such as attitudes, knowledge, genetics, etc. Knowledge

The goal of this paper is to develop a conceptual framework about healthy food choices, recommended physical activity

that shows the links between health determinants, health levels, and the correct interpretation of food labels also plays a

outcomes, work ability and work outcomes, and their interplay

with incentives aimed at improving work outcomes that are critical role [3]. This knowledge is often associated with health

essential for extending productive working lives. This literacy that is an independent factor of self-assessed health, and

framework is useful for designing research on how individuals, low health literacy contributes to health disparities [1], [4].

organizations and policy makers can improve long-term Attitudes can also positively or negatively influence health,

employability and productivity in the face of health challenges. as argued by the theory of reasoned action [5], showing that

attitudes and subjective norms predict behavioural intentions,

†Corresponding author. which, in turn, predict actual behaviours. The value individuals

place on health is pivotal; those who prioritize health are more

Permission to make digital or hard copies of part or all of this work for personal or likely to make it a priority over other demands, such as time,

classroom use is granted without fee provided that copies are not made or distributed money, and competing life priorities [3], [6], [7]. for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must Genetics is another key individual factor affecting health.

be honored. For all other uses, contact the owner/author(s). Family history is recognized as one of the strongest risk factors

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia for common diseases, including cancer, cardiovascular disease,

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.dem.15 diabetes, autoimmune disorders, and psychiatric illnesses [8].



Social factors are sociocultural and socioeconomic. 2.2 Modifiable and Non-modifiable Risk Factors



risky behaviours. Peer influence operates similarly; social and religion (sociocultural), employment, education, income (socioeconomic) and geographic location and related access to support and integration are highly protective against mortality, health services (environmental). Some factors are more easily with their impact on health comparable to or exceeding that of changed than others. Also, the degree to which a person is able traditional behavioural risk factors such as smoking, high alcohol to change them varies from person to person [3]. Non-modifiable consumption, lack of exercise, and obesity. Traditional and health determinants include genetics (individual), and family, social media is increasingly recognized as a commercial media and culture (sociocultural). While you can’t change these determinant of health [9], influencing health through public factors, it’s important to be aware of them. relations, advertising, and economic pressures [10]. The degree of control individuals can exert over their health Socioeconomic factors include employment, education, and choices related to religion, friendships, and media exposure, which of the above explained health determinants are in or out of an individual’s control. Modifiable factors can be changed and particularly in childhood. Family also plays a crucial role in controlled to a certain extent, such as skills (individual), peers shaping attitudes towards health and decisions on protective and pressure, media, religion, and culture. Family heavily influences To improve health outcomes, it is crucial to understand Sociocultural factors include influences from family, peer



income. Employment status and occupation significantly impact varies across our lifespan [3]. Control increases until adulthood and decreases in older adulthood. The changing influence of health by influencing living and working conditions, levels of determinants through life stages varies from person to person physical activity, and stress [3]. People at work may face hazards [19]. Further, modifiable factors often require support, such as owing to chemicals, biological agents, physical factors, adverse access to healthcare or financial resources, to be effectively ergonomic conditions, allergens, safety risks, and psychosocial managed. When individuals lack such support, the potential to factors. More than 100 occupational diseases are classified [11]. modify their health outcomes diminishes, contributing to the Working conditions are linked to an increased risk of developing feeling of powerlessness in improving their health [20]. work-related illness [12] and depressive disorders [13]. New

psychosocial and physical stressors contribute to health



disparities. Evidence suggests that better job quality provides 2.3 Health Outcomes-oriented Improvement

protective effects against musculoskeletal disorders, mental

health issues, and general health problems [14], and enhances Incentives



labour productivity, notably when employees are healthy [15]. The term “morbidity” is commonly used to describe the

Income is also strongly associated with health outcomes [16]. burden of suffering, in terms of impairment or disability, caused

Low income can restrict access to health services, whereas higher by an illness or health condition. Addressing chronic illness

income allows for better living conditions and access to morbidity aims at the best achievable state of health that

healthcare. Higher income often sets expectations for subsequent encompasses all dimensions of physical, mental, and social well-

generations regarding education, employment, and income being [21].

although positive association between income and health may Strategies designed to help individuals live well should

flatten or even reverse at the highest income levels [17]. include a broad array of activities targeting primary, secondary,

Environmental factors encompass both geographical and tertiary prevention for all persons, with or without a chronic

location and access to health services and technology. Natural illness. Accordingly, public health mechanisms aimed at

and built environments play significant roles in health, with some preventing undesired health outcomes (primary prevention such

environmental characteristics contributing to chronic diseases as vaccination, tobacco cessation, physical activity promotion,

[3]. A review of 133 diseases by Prüss-Ustün [18] established healthful eating, injury prevention) can also help people with a

that nearly a quarter of the global disease burden could be chronic illness or disability to live more healthfully. Health

prevented by reducing environmental risks. monitoring using real-time data plays a significant role in

Individual, social, and environmental factors interact in influencing health outcomes, by enabling early detection of

influencing health outcomes. Education—a key socioeconomic health conditions, facilitating continuous management of chronic

factor—directly influences individual factors like knowledge and diseases, and improving overall patient care [22].

skills. Additionally, a person’s environment is often shaped by Health and long-term care mechanisms that prevent or delay

their employment and income, which are linked to the number of complications, build coping skills, improve function, or alleviate

hours spent at work. In some cases, the living environment can pain and suffering may serve a dual purpose: (i) reducing the

restrict educational and employment opportunities, further magnitude of illness burden over an individual’s remaining years

impacting health. People living in rural or remote areas are more of life; (ii) reducing and/or delaying the development of

likely to experience poorer health outcomes, due to limited additional complications or comorbidities in a way that serves to

access to health services and education and employment compress the period of morbidity until later in life [21]. Further,

opportunities. Sociocultural factors also play a significant role; occupational health services improve health outcomes of the

for instance, not speaking the native language can hinder active population. Progress in the field of occupational and

educational success and makes it more challenging to navigate environmental medicine has led to a better understanding of the

the healthcare system. This interconnectedness underscores the most efficient strategies for recognizing a wide variety of work-

importance of considering the combined effects of individual, related ailments [23]. social, and environmental factors in shaping health outcomes [3].



2.4 Impact of Health Outcomes on Functional addressing both organisational factors and individual health



Independence and Work Ability needs, RTW programs can prevent adverse work outcomes such as long-term absenteeism, presentism, unemployment and

All above-mentioned incentives are aimed at improving disability retirement [28], [29], [33], [36]. They can also mitigate

adverse health outcomes and thereby maintaining functional short-term adverse effects of chronic illnesses such as frequent

independence and, consequently, work ability. Functional shorter sick leaves and presenteeism [26].

independence can be viewed narrowly as the ability to perform While employer organisations and their support of

basic activities of daily living (eating, dressing, hygiene, occupational health programs play an important role in

transferring) [24]; or more broadly as the ability to perform improving work outcomes, other factors also mitigate adverse

instrumental activities of daily living (performing chores, effect of reduced work ability on work outcomes. Flexibility is

managing finances, taking medication, using transportation) [25] not important just in employer organizations, but also on the

that enable an independent life. Research shows that functional labour market at large, which involves labour code, industrial

independence has a strong impact on work outcomes [26], [27]. relations and active labour market policies (ALMPs) that include

Functional independence is strongly linked to an individual’s various measures, from training to job search assistance,

work ability defined as a holistic concept encompassing both subsidies, supported employment opportunities and programs to

personal health and the work environment’s role in enabling or support entrepreneurial activities. Such measures enable

hindering an individual's ability to work [28]. The concept refers occupational horizontal and vertical mobility [39], [40]. Other

to an individual’s capacity to remain employed and perform national employment policies, besides ALMP, may be important

work-related activities, especially in the context of chronic to facilitate (via regulation) and promote inclusive workplace

diseases, aging, or disabilities [29]. practices and ensure that individuals are not excluded from job

Work ability is influenced by a range of factors including opportunities [41]. Lastly, passive supports as well as social

health status, workplace accommodations, and social support security at large are relevant. Social security systems, including

systems [30]. Physical health (mobility, endurance) and mental sick leave policies and disability insurance, provide essential

well-being (cognitive functioning, emotional resilience) are the financial support to individuals who are temporarily or

most important factors to maintain work ability [31]. In addition, permanently unable to work due to illness [41]. job demands, its physical and psychosocial (cognitive,

emotional) demands, significantly influence an individual’s

ability to work effectively and productivity [32]. Workplace 3 Conclusion

factors (organisational support, ergonomics, flexibility) play a The paper proposes a conceptual framework that shows

crucial role in enhancing or diminishing work ability, especially complex links between health determinants, health outcomes,

for those with chronic conditions or disabilities [33]. Social and work ability and work outcomes and their interplay with various

occupational factors (age, training opportunities, work-life incentives aimed at improving both health and work outcomes.

balance) also impact work ability [34]. It highlights the significance of modifiable and non-modifiable

health factors that influence adverse health effects of chronic



2.5 Work Outcomes and Work Outcomes- illnesses and the resulting functional independence, work ability

and work outcomes. It emphasises the role of supportive work

oriented Improvement Incentives environments, health-oriented policies, and mechanisms that

Work ability has a notable impact on work outcomes. If facilitate retention and RTW. The research highlights that

chronic illnesses result in an individual’s prolonged sickness improving work outcomes requires a complex approach

absenteeism, unemployment or work inactivity, work retention involving individual, social, and environmental interventions, as

and RTW are favourable work outcomes. RTW refers to the well as policies that promote inclusive employment practices.

process of reintegrating employees back into the workplace after This framework serves as a valuable tool for understanding how

an absence due to illness or injury. Besides work ability, RTW health and workplace factors interplay to improve long-term

also depends on several patient’s pull and push factors, from employability, particularly in aging populations or individuals

meaningfulness of work to being offered an adjusted work with chronic conditions. It can serve as a foundation for future

environment [35]. Successful RTW initiatives focus on assessing empirical research, literature review and policy preparation.

and improving work ability, ensuring that the work environment

and job demands align with the employee’s current health status. Acknowledgments

RTW outcomes are enhanced by flexible work arrangements,

gradual reintegration programs, and workplace accommodations This work was partially co-financed by the following projects:

that mitigate the impact of health conditions [36], [37]. By P5-0096, P5-0128, P5-0441, P5-0117, J5-4575, J5-4540.



References

[1] D. Nutbeam and J. E. Lloyd, “Understanding and Responding to Health [4] D. Schillinger, “Social Determinants, Health Literacy, and Disparities:

Literacy as a Social Determinant of Health,” Intersections and Controversies,” HLRP Health Lit. Res. Pract., vol. 5, Jul. Annu. Rev. Public Health ,

vol. 42, pp. 159–173, Apr. 2021, doi: 10.1146/annurev-publhealth- 2021, doi: 10.3928/24748307-20210712-01. 090419-102529. [5] R. L. Heilbroner, I. Ajzen, M. Fishbein, and L. C. Thurow, Understanding

[2] Attitudes and Predicting Social Behavior. Prentice-Hall, 1980. World Health Organization, “Determinants of health.” Accessed: Sep. 20,

2024. [Online]. Available: https://www.who.int/news-room/questions- [6] S. P. J. Kremers, “Theory and practice in the study of influences on energy and-answers/item/determinants-of-health balance-related behaviors,” Patient Educ. Couns., vol. 79, no. 3, pp. 291–

[3] 298, Jun. 2010, doi: 10.1016/j.pec.2010.03.002. PDHPE, “HSC PDHPE,” HSC PDHPE. Accessed: Sep. 20, 2024.

[Online]. Available: https://pdhpe.net/pdhpe-net/hsc-pdhpe/



[7] S. Sumaedi et al., “Factors influencing intention to follow the ‘stay at [30] I. Mairhuber, E. Prammer, U. Waginger, and S. Brigitte, “Understanding

home’ policy during the COVID-19 pandemic,” Int. J. Health Gov., vol. employment participation of older workers (UEP),” Vienna:FORBA, Jan. 26, no. 1, pp. 13–27, Jan. 2021, doi: 10.1108/IJHG-05-2020-0046. 2014.

[8] Institute of Medicine (US) Committee on Assessing Interactions [31] M. Gamperiene, J. F. Nygård, I. Sandanger, B. Lau, and D. Bruusgaard,

Among Social, Behavioral, and Genetic Factors in Health, Genes, “Self-reported work ability of Norwegian women in relation to physical

Behavior, and the Social Environment: Moving Beyond the and mental health, and to the work environment,” J. Occup. Med. Toxicol., Nature/Nurture Debate. in The National Academies Collection: Reports vol. 3, no. 1, p. 8, Apr. 2008, doi: 10.1186/1745-6673-3-8. funded by National Institutes of Health. Washington (DC): National [32] T. I. van den Berg, S. J. Robroek, J. F. Plat, M. A. Koopmanschap, and A. Academies Press (US), 2006. Accessed: Sep. 20, 2024. [Online]. Burdorf, “The importance of job control for workers with decreased work

Available: http://www.ncbi.nlm.nih.gov/books/NBK19929/ ability to remain productive at work,” Int. Arch. Occup. Environ. Health,

[9] M. Zenone, N. Kenworthy, and N. Maani, “The Social Media Industry as vol. 84, no. 6, pp. 705–712, 2011, doi: 10.1007/s00420-010-0588-1.

a Commercial Determinant of Health,” Int. J. Health Policy Manag., vol. [33] V. L. Kristman, W. S. Shaw, C. R. L. Boot, G. L. Delclos, M. J. Sullivan, 12, no. Issue 1, pp. 1–4, Dec. 2023, doi: 10.34172/ijhpm.2022.6840. and M. G. Ehrhart, “Researching Complex and Multi-Level Workplace

[10] D. Even, S. M. Abdalla, N. Maani, and S. Galea, “News media as a Factors Affecting Disability and Prolonged Sickness Absence,” J. Occup.

commercial determinant of health,” Lancet Glob. Health, vol. 12, no. 8, Rehabil., vol. 26, no. 4, pp. 399–416, 2016, doi: 10.1007/s10926-016-pp. e1365–e1369, Aug. 2024, doi: 10.1016/S2214-109X(24)00191-8. 9660-3.

[11] T. K. Peckham, M. G. Baker, J. E. Camp, J. D. Kaufman, and N. S. Seixas, [34] C. Varianou-Mikellidou, G. Boustras, O. Nicolaidou, C. Dimopoulos, I.

“Creating a Future for Occupational Health,” Ann. Work Expo. Health, Anyfantis, and P. Messios, “Work-related factors and individual

vol. 61, no. 1, pp. 3–15, Jan. 2017, doi: 10.1093/annweh/wxw011. characteristics affecting work ability of different age groups,” Saf. Sci.,

[12] G. Papadopoulos, P. Georgiadou, C. Papazoglou, and K. Michaliou, vol. 128, p. 104755, Aug. 2020, doi: 10.1016/j.ssci.2020.104755.

“Occupational and public health and safety in a changing work [35] T. Lundberg and S. Melander, “Key Push and Pull Factors Affecting environment: An integrated approach for risk assessment and prevention,” Return to Work Identified by Patients With Long-Term Pain and General

Saf. Sci., vol. 48, no. 8, pp. 943–949, Oct. 2010, doi: Practitioners in Sweden,” Qual. Health Res., vol. 29, no. 11, pp. 1581– 10.1016/j.ssci.2009.11.002. 1594, Sep. 2019, doi: 10.1177/1049732319837227.

[13] R. Rugulies et al., “Work-related causes of mental health conditions and [36] V. L. Kristman, C. R. L. Boot, K. Sanderson, K. E. Sinden, and K.

interventions for their improvement in workplaces,” The Lancet, vol. 402, Williams-Whitt, “Implementing Best Practice Models of Return to Work,”

no. 10410, pp. 1368–1381, Oct. 2023, doi: 10.1016/S0140- in Handbook of Disability, Work and Health, U. Bültmann and J. Siegrist, 6736(23)00869-3. Eds., Cham: Springer International Publishing, 2020, pp. 589–613. doi:

[14] G. Henseke, “Good jobs, good pay, better health? The effects of job quality 10.1007/978-3-030-24334-0_32.

on health among older European workers,” Eur. J. Health Econ. HEPAC [37] S. L. Fallman, “Organizational flexibility and health among line managers Health Econ. Prev. Care, vol. 19, no. 1, pp. 59–73, Jan. 2018, doi: and employees in health care,” 2020, Accessed: Sep. 18, 2024. [Online].

10.1007/s10198-017-0867-9. Available: https://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-286040

[15] I. Arends, C. Prinz, and F. Abma, “Job quality, health and at-work [38] J. Harrison and L. Dawson, “Occupational Health: Meeting the Challenges

productivity,” OECD, Paris, Jun. 2017. doi: 10.1787/43ff6bdc-en. of the Next 20 Years,” Saf. Health Work, vol. 7, no. 2, pp. 143–149, Jun.

[16] E. Reche, H.-H. König, and A. Hajek, “The Relationship between Income 2016, doi: 10.1016/j.shaw.2015.12.004.

and Morbidity-Longitudinal Findings from the German Ageing Survey,” [39] J. Ilmarinen, “Promotion of Work Ability during Aging,” in Aging and Int. J. Environ. Res. Public. Health, vol. 18, no. 23, p. 12365, Nov. 2021, Work, CRC Press, 2002. doi: 10.3390/ijerph182312365. [40] G. Costa and S. Sartori, “Ageing, working hours and work ability,”

[17] E. Backlund, P. D. Sorlie, and N. J. Johnson, “The shape of the relationship Ergonomics, vol. 50, no. 11, pp. 1914–1930, Nov. 2007, doi:

between income and mortality in the United States: Evidence from the 10.1080/00140130701676054. National Longitudinal Mortality Study,” Ann. Epidemiol., vol. 6, no. 1, pp. [41] I. Seing, C. Ståhl, L. Nordenfelt, P. Bülow, and K. Ekberg, “Policy and 12–20, Jan. 1996, doi: 10.1016/1047-2797(95)00090-9. Practice of Work Ability: A Negotiation of Responsibility in Organizing

[18] A. Prüss-Ustün, J. Wolf, C. Corvalán, T. Neville, R. Bos, and M. Neira, Return to Work,” J. Occup. Rehabil., vol. 22, Jun. 2012, doi:

“Diseases due to unhealthy environments: an updated estimate of the 10.1007/s10926-012-9371-3. global burden of disease attributable to environmental determinants of

health,” J. Public Health Oxf. Engl., vol. 39, no. 3, pp. 464–475, Sep. 2017,

doi: 10.1093/pubmed/fdw085.

[19] D. F. Alwin and L. A. Wray, “A life-span developmental perspective on

social status and health,” J. Gerontol. B. Psychol. Sci. Soc. Sci., vol. 60

Spec No 2, pp. 7–14, Oct. 2005, doi:

10.1093/geronb/60.special_issue_2.s7.

[20] M. Williams, L. Perez, S. Siddiqi, N. Qureshi, J. Sousa, and A. Huttinger,

“Building the Evidence Base for Social Determinants of Health

Interventions,” ASPE. Accessed: Sep. 20, 2024. [Online]. Available:

https://aspe.hhs.gov/reports/building-evidence-base-social-determinants-

health-interventions

[21] Institute of Medicine, Board on Population Health and Public Health

Practice, and Committee on Living Well with Chronic Disease: Public

Health Action to Reduce Disability and Improve Functioning and Quality

of Life, Living Well with Chronic Illness: A Call for Public Health Action,

1st edition. National Academies Press, 2011.

[22] G. Paré, M. Jaana, and C. Sicotte, “Systematic Review of Home

Telemonitoring for Chronic Diseases: The Evidence Base,” J. Am. Med.

Inform. Assoc. JAMIA, vol. 14, no. 3, pp. 269–277, 2007, doi:

10.1197/jamia.M2270.

[23] L. S. Newman, “Occupational Illness,” N. Engl. J. Med., vol. 333, no. 17,

pp. 1128–1134, Oct. 1995, doi: 10.1056/NEJM199510263331707.

[24] I. Hartigan, “A comparative review of the Katz ADL and the Barthel Index

in assessing the activities of daily living of older people,” Int. J. Older

People Nurs., vol. 2, no. 3, pp. 204–212, 2007, doi: 10.1111/j.1748-

3743.2007.00074.x.

[25] M. P. Lawton and E. M. Brody, “Assessment of older people: self-

maintaining and instrumental activities of daily living,” The

Gerontologist, vol. 9, no. 3, pp. 179–186, 1969.

[26] A. M. Fouad, A. Waheed, A. Gamal, S. A. Amer, R. F. Abdellah, and F.

M. Shebl, “Effect of Chronic Diseases on Work Productivity: A Propensity

Score Analysis,” J. Occup. Environ. Med., vol. 59, no. 5, pp. 480–485,

May 2017, doi: 10.1097/JOM.0000000000000981.

[27] I. Grabovac and T. E. Dorner, “Association between low back pain and

various everyday performances,” Wien. Klin. Wochenschr., vol. 131, no.

21, pp. 541–549, Nov. 2019, doi: 10.1007/s00508-019-01542-7.

[28] J. Ilmarinen and K. Tuomi, “Past, Present and Future of Work Ability,”

People Work Res Rep, vol. 65, pp. 1–25, Jan. 2004.

[29] J. Ilmarinen and K. Tuomi, “Aging workers,” Occup. Environ. Med., vol.

58, pp. 546–52, Sep. 2001, doi: 10.1136/oem.58.8.546.





Ko tempo sestavina v stopnji celotne rodnosti izzveni




When tempo component in the total fertility rate fades



away



Jože Sambt

Katedra za matematiko, statistiko

in operacijske raziskave

Univerza v Ljubljani,

Ekonomska fakulteta

Ljubljana, Slovenija

joze.sambt@ef.uni-lj.si



Povzetek Ključne besede

Sredi 1980-ih let je začela povprečna starost matere ob rojstvu Rodnost, demografija, stopnja celotne rodnosti, obnavljanje

otroka hitro naraščati. To znižuje stopnjo celotne rodnosti, ki prebivalstva, Slovenija

kaže, koliko otrok bi rodila ena ženska, če bi dočakala konec

svoje rodne dobe in bi bila rodnost v posameznem njenem letu Abstract

starosti takšna, kot je v izbranem letu pri drugih ženskah v In the mid-1980s, the average age of mothers at childbirth began

posamezni starosti rodne dobe. Stopnja celotne rodnosti v takih to rise rapidly. This lowers the “total fertility rate” indicator,

razmerah sicer ustrezno odraža raven rodnosti v izbranem letu, which shows how many children a woman would have if she

vendar pa ne kaže ustrezno preferenc ljudi o številu otrok reached the end of her reproductive age and if fertility in each of

(omejenih s stvarnostjo). Število otrok, ki jih bodo ženske imele her years of age was the same as in the selected year for other

tekom svoje rodne dobe oziroma življenja, je v takih razmerah women of the same age. The total fertility rate in such

višje, le da ljudje odložijo rodnost v višjo starost in jih stopnja circumstances appropriately reflects the level of fertility in the

celotne rodnosti v tekočem letu ne zajame. V članku izračunamo selected year but does not show people’s preferences for the

prilagojeno stopnjo celotne rodnosti s pomočjo number of children, which are higher, only that people postpone

Bongaarts-Feeney metode in jo na ta način očistimo tempo childbirth to an older age, and the total fertility rate in the current

učinka (angl. tempo effect), da vidimo ločeno učinek ravni (angl. year does not capture them. In the article, we calculate the

quantum effect). Izkaže se, da tudi v letih okrog leta 2023, ko je adjusted total fertility rate using the Bongaarts-Feeney method

bila stopnja celotne rodnosti zgolj okrog 1,2, prilagojena stopnja by removing the ‘tempo effect’ from the total fertility rate to see

celotne rodnosti praktično ni upadla pod raven 1,6 otroka na the separate level effect (‘quantum effect’). It turns out that even

žensko. V zadnjih petih letih pa se je naraščanje povprečne in the years around 2023, when the total fertility rate was only

starosti mater ob rojstvu otroka povsem ustalilo na vrednosti 31,1 around 1.2, the adjusted total fertility rate practically did not fall

leta [1]. To pomeni, da negativnega »tempo« učinka na stopnjo below the level of 1.6 children per woman. In the last five years,

celotne rodnosti ni več, temveč je samo še »quantum« učinek. however, the increase in the average age of mothers at childbirth

Ker je stopnja celotne rodnosti (in s tem hkrati tudi prilagojena has completely stabilized at 31.1 years [1]. This means that the

stopnja celotne rodnosti) upadla z 1,64 v letu 2021 na 1,55 v letu negative “tempo” effect on the total fertility rate is no longer

2022 in nadalje na zgolj 1,51 v letu 2023, to pomeni najnižjo present, but only the “quantum” effect remains. Since the total

rodnost v zgodovini Slovenije, kar se tiče števila otrok, ki jih fertility rate (and thus also the adjusted total fertility rate) fell

ljudje imajo v svojem življenju. Samo upamo lahko, da se from 1.64 in 2021 to 1.55 in 2022 and further to only 1.51 in

rodnost ne bo na tej nizki ravni ustalila ali da bi se celo nadaljeval 2023, this means the lowest fertility in the history of Slovenia in

omenjeni trend upadanja iz zadnjih treh let. terms of the number of children people have in their lifetime. We

can only hope that fertility will not stabilize at this low level or

that the mentioned declining trend of the recent three years will

not even continue.



Permission to make digital or hard copies of part or all of this work for personal or Keywords classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full Fertility, demography, total fertility rate, population

citation on the first page. Copyrights for third-party components of this work must replacement, Slovenia be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.dem.16



1 prebivalstva ali ne, so starostno specifične stopnje rodnosti. Uvod

Enačba je praktično enaka kot enačba (2), le da izračunamo

Leta 1998 sta Bongaarts in Feeney [2] objavila članek, v katerem

stopnjo rodnosti posebej za vsako starost med 15. in 49. letom

sta želela ponuditi splošno sprejeto metodo za izločanje tempo

starosti:

učinka iz kazalnika stopnje celotne rodnosti. Pred tem so se že



pol stoletja demografi zavedali, da je stopnja celotne rodnosti kot 𝑁𝑥 𝑓 ustaljen demografski kazalnik popačena, če se starost žensk ob 𝑥 = ⬚ ⬚ ∙ 1000 , (3) 𝑉 ⬚ 𝑓,𝑥 porodu spreminja, vendar so čakali na ustrezno metodo za

izločitev tega dejavnika. Te vrednosti so zanimive same po sebi, saj kažejo, 1) koliko

Da bi lahko raven rodnosti primerjali med državami in skozi posamezni starostni razredi žensk prispevajo k reprodukciji

čas, je potrebno oblikovati ustrezen kazalnik. Primerjava prebivalstva, 2) so osnova za izračun tablic rodnosti, saj približno

absolutnega števila živorojenih med posameznimi državami in v izražajo verjetnost rojstva otroka v odvisnosti od starosti žensk

času ne bi imela smisla, saj je število živorojenih močno odvisno in 3) iz njih lahko izračunamo stopnjo celotne rodnosti, kar bomo

od števila prebivalcev. Če delimo število živorojenih s številom izvedli v nadaljevanju. prebivalcev, dobimo splošno stopnjo natalitete:



n 𝑁 = ∗ 1000 , (1) 2 Stopnja celotne rodnosti

𝑃

Stopnjo celotne rodnosti izračunamo kot vsoto vseh 35

pri čemer je N število živorojenih v nekem obdobju (po starostno specifičnih stopenj rodnosti:

navadi v enem letu), P pa je število prebivalcev v sredini tega



obdobja (ali povprečje v tem obdobju). Problem tega kazalnika ∑ 𝑓 𝑥 𝑇 𝑓 = , (4) 1000 je dvojen.



nima smisla, ker je pod vplivom različne starostne strukture dveh Prvič, primerjava tega kazalnika med različnimi državami pri čemer Tf označuje stopnjo celotne rodnosti, fx pa označuje starostno specifične stopnje rodnosti (pri čemer x predstavlja prebivalstev. To je še posebej problematično, če primerjamo vrednosti od 15. do 49. leta starosti), ki kažejo raven rodnosti v razvite in manj razvite države. Manj razvite države imajo mnogo posamezni starosti. Deljenje s konstanto 1000 povrne kazalnik v večji delež mladih oseb, starih npr. 3 leta, 7 let, 10 let itd. Ti osnovno obliko, ki kaže število otrok, medtem ko so vrednosti f x prebivalci znižujejo vrednost kazalnika n , čeprav ne morejo izražene na 1000 žensk – glej enačbo (3). vplivati na velikost števca (torej na število živorojenih, N ), saj Ideja torej je, da seštejemo stopnje rodnosti v posamezni starosti niso sposobni rojevati. Po drugi strani je v razvitih državah skozi celotno rodno dobo. To pomeni, da dobimo število otrok, mnogo večji delež starih oseb v starosti npr. 60 let, 70 let, 80 let ki bi jih rodila ena ženska v svoji rodni dobi, če bi imela v svoji itd. Tudi ti prebivalci niso sposobni rojevati, kar pomeni, da ne posamezni starosti takšno rodnost, kot so jo imele druge ženske morejo vplivati na velikost števca v enačbi (1), torej na število v posameznih starostih med 15. in 49. letom v analiziranem živorojenih. Se pravi, da neupravičeno znižujejo vrednost koledarskem letu. Hkrati je implicitno prisotna predpostavka, da kazalniku splošna stopnja natalitete ( n ) v razvitih državah z ta ženska dočaka konec rodne dobe – torej, da ne bo umrla do vidika ravni rodnosti, za katero ugotavljamo, ali je dovolj visoka konca svoje rodne dobe. Žensk, ki so že umrle v preteklosti, za obnavljanje prebivalstva ali ne. Drugič, konkretna vrednost namreč pri izračunu starostno specifičnih stopenj rodnosti v tega kazalnika sama po sebi ni zgovorna glede tega, ali dobljeni imenovalcu ne upoštevamo. rezultat predstavlja dovolj visoko ali nizko raven rodnosti, ki bo Kazalnik stopnja celotne rodnosti nam tako kaže število zagotavljala obnavljanje prebivalstva na dolgi rok ali ne. otrok, ki bi jih rodila ena ženska v svoji rodni dobi, če bi dočakala

konec rodne dobe in bi v njeni posamezni starosti vladale take

Naslednji kazalnik pri iskanju ustreznega kazalnika je

zakonitosti rojevanja, kot so veljale pri drugih ženskah v

stopnja celotne rodnosti:

posameznih starostih rodne dobe v koledarskem letu, za katero

𝑁 stopnjo celotne rodnosti računamo. Ta kazalnik je končno zelo

𝑓 = ∙ 1000, ⬚ ⬚ (2) 𝑉 zgovoren sam po sebi, ki ima tudi jasno mejo, ali se bo ob taki 35 𝑓,15

rodnosti (v odsotnosti selitev) število prebivalstva na dolgi rok

pri čemer je obnavljalo ali ne. Ta vrednost je za razvite države, kamor spada V f oznaka za število žensk, 15 je začetek

starostnega razreda, 35 pa širina starostnega razreda, ki je v tem tudi Slovenija, 2,1 otroka na žensko.

primeru celotna rodna doba. V števcu so torej namesto vseh Zakaj mora znašati rodnost nekoliko več kot 2 otroka na

prebivalcev tokrat samo ženske v starosti 15–49 let (tj. ženske v žensko, da bi se zagotavljala enostavna reprodukcija

rodni dobi). Gre torej za del prebivalstva, ki je v splošnem prebivalstva, kar pomeni, da bi se prebivalstvo na dolgi rok ravno

sposoben rojevati, kar pomeni, da lahko vpliva na velikost števca obnavljalo (ne pa naraščalo ali upadalo)? Če imamo pred očmi

(tj. na število živorojenih). Še vedno pa deloma ostane prisoten tipičen par moškega in ženske, ki ima dečka in deklico: zakaj ta

problem različne starostne strukture – tokrat starostne dva otroka nista dovolj za obnavljanje prebivalstva na dolgi rok?

porazdelitve žensk znotraj starostnega razreda 15–49 let. Ostane Fantek nadomesti očeta, deklica nadomesti mater in obnovitev

tudi problem (ne)zgovornosti samega kazalnika glede ravni prebivalstva za naslednji rod bi morala biti zagotovljena, a ne?

rodnosti, ali ta zagotavlja obnavljanje prebivalstva ali ne. Prvi razlog, da dva otroka nista dovolj, je to, da se med vsemi

Naslednji korak pri iskanju našega zgovornega kazalnika živorojenimi otroci rodi več fantkov kot deklic. Delež fantkov je

rodnosti, ki bi kazal, ali je rodnost dovolj visoka za obnavljanje okrog 51,5 %, medtem ko je delež deklic zgolj preostalih okrog



48,5 %. Ta deleža sta tako stabilna, da se razmerju med dečki in odčitamo, da so ženske, ki so bile v letu 2021 stare 50 let, tekom

deklicami ob rojstvu pravi celo demografska konstanta [3], str. svoje rodne dobe rodile v povprečju 1,77 otroka [6].

15. Ženska mora torej v povprečju roditi nekoliko več kot dva Če pa želimo čim bolj ažurne vrednosti, kaj se dogaja z

otroka, da v povprečju rodi eno deklico in s tem nadomesti samo rodnostjo, moramo namesto na »vzdolžni« kazalnik pristati na

sebe. »prečni« kazalnik, ki smo ga predstavili v enačbi (4). Problem pa

Drugi razlog, da stopnja celotne rodnosti 2,1 ni dovolj za nastane, če se ob tem povprečna starost mater spreminja. Kot

obnavljanje prebivalstva, pa je že prej omenjena umrljivost vidimo, je bilo temu v preteklosti dejansko tako. Slika 1

žensk, preden dočakajo konec rodne dobe. Čeprav bi ena ženska prikazuje, da se je povprečna starost matere ob rojstvu otroka od

rodila v povprečju eno deklico, to še ne bi bilo dovolj, saj ta leta 1984 do leta 2010 povišala s 25,3 let na 30,3 let. To pomeni,

deklica žal lahko umre v starosti npr. 0 let, 3 leta, 10 let itd., še da je naraščala približno 0,2 leta na leto.

preden se uspe reproducirati za naslednji rod. Večja kot je

umrljivost žensk pred koncem rodne dobe, večja mora biti 32



stopnja celotne rodnosti, da bo dovolj za obnavljanje roka 31 ot prebivalstva na dolgi rok. Zgolj iz podatkov o številu živorojenih vu 2010: 30,3 st in številu žensk v posamezni starosti tega ne moremo ugotoviti, 30



tistih žensk, ki so medtem že umrle. Ustrezen kazalnik, ki nam at →Vsako leto m 28 okrog 20 % manj t poda to informacijo, je število živih v začetku posamezne starosti rojstev aros saj podatki vsebujejo samo ženske, ki so še žive, ne vidimo pa ob roj Naraščanje v 29 e povprečju za 0,2 er leti na leto

( 27 lx ) iz tablic umrljivosti. Ta kazalnik prikazuje, kolikšen delež

st

žensk bi bil še živ v posamezni starosti, če bi bile tekom na

26

1984: 25,3

celotnega življenja žensk prisotne take razmere glede umrljivosti ovprečP 25

kot so bile prisotne v času oblikovanja tablic umrljivosti. Tablice 8 3 8 3 8 3 8 3 8 3 7 8 8 9 9 0 0 1 1 2 9 9 9 9 9 0 0 0 00 1

1 1 1 1 2 2 2 2 2

umrljivosti so namreč model, ki na (običajno) sto tisoč v danem Leto

trenutku živorojenih otrok s posameznimi kazalniki tablic Slika 1: Gibanje povprečne starosti matere ob rojstvu otroka

umrljivosti prikazujejo proces njihovega umiranja v odvisnosti v Sloveniji v obdobju 1978–2023. Vir: [5].

od starosti [3], str. 120. Najnovejše tablice umrljivosti za

Slovenijo kažejo, da bi ob razmerah iz leta 2022 več kot 99 % Iz tega sledi, da se je v tekočem letu realiziralo samo 80 %

žensk dočakalo starost 41 let [4], to je starost, v kateri je velika vseh rojstev, medtem ko so preostala rojstva preložena v

večina rodnosti pri ženskah že zaključena. V letu 2023 se je prihodnje koledarsko leto. Vendar pa z vidika »vzdolžne« oz.

namreč materam po tej starosti rodilo samo še 2,9 % vseh dokončne rodnosti, ki jo bomo lahko ugotavljali šele čez vrsto

živorojenih otrok [5]. Vidimo, da je umrljivost žensk v Sloveniji, let, to ne pomeni, da bo ženska imela v svoji celotni rodni dobi

zelo nizka in zato iz tega naslova ni treba, da bi bila vrednost število otrok na ravni 80 %, temveč na ravni 100 % – le da bo

stopnje celotne rodnosti višja od 2,1. V kaki afriški državi pa celo imela otroke v nekoliko višji starosti. Zaradi analize celotne

stopnja celotne rodnosti 2,3 ne bi bila dovolj, saj je umrljivost rodnosti na osnovi »prečne« analize in interpretaciji v smislu

žensk oz. deklic tam žal še vedno zelo visoka. »vzdolžne« analize torej močno podcenimo število otrok, ki jih

Stopnja celotne rodnosti naj bi torej v Sloveniji in drugih razvitih bo ženska imela v svoji rodni dobi.

državah kazala raven rodnosti in obnavljanje prebivalstva na

dolgi rok v odsotnosti selitev, pri čemer je potrebna vrednost za

obnavljanje prebivalstva 2,1 otroka na žensko. Za številne 3 (Tempo) prilagojena stopnja celotne rodnosti

demografske analize je to kljub vsemu povsem ustrezen Bongaarts-Feeneyeva metoda je preprosta in razumljiva metoda

demografski kazalnik, ki se odraža tudi v številu rojstev v za razmejitev učinka odlaganja rodnosti (»tempo« učinka) od

posameznem koledarskem letu, iz tako izražene rodnosti potem učinka ravni (»quantum« učinka). V tem našem primeru metoda

tudi izhaja starostna porazdelitev prebivalcev v prihodnje itd. postavi omenjenih 80 % kot novo osnovo, za večjo natančnost to

Problem pa nastane, če tako izračunano stopnjo celotne izvede ločeno po redih rojstva, hkrati pa za izločitev slučajnega

rodnosti razlagamo v smislu dokončne stopnje celotne rodnosti dejavnika uporabi drseče povprečje. Za podrobnejšo razlago glej

[3], str. 42 (angl. completed fertility), torej v smislu števila otrok, Bongaarts in Feeney [2].

jih bodo posamezne kohorte žensk rodile tekom svoje rodne dobe Na sliki 2 sta prikazani obe stopnji celotne rodnosti: običajna

oz. svojega življenja. stopnja celotne rodnosti in (tempo) prilagojeno stopnja celotne

Najbolj natančno sliko o tem, koliko otrok je posamezna rodnosti (angl. tempo adjusted total fertility rate), ki smo jo

kohorta žensk imela tekom svoje rodne dobe, dobimo, če kohorto izračunali na osnovi podatkov o številu žensk v posamezni

žensk spremljamo od začetka do konca njihove rodne dobe. Ta starosti in števila živorojenih otrok materam v posamezni

»vzdolžni« kazalnik rodnosti, ki ga, kot omenjeno, imenujemo starosti, ki so hkrati razčlenjeni po redu rojstva (1, 2, 3, in 4+).

dokončna stopnja celotne rodnosti, ima vsekakor svoje mesto v Pri tem smo uporabili podatke Statističnega urada RS (tudi

analizah, vendar pa je velika pomanjkljivost tega kazalnika interne podatke, saj vseh potrebnih podatkov na domači strani

njegova ažurnost. Čakati moramo namreč do 50. leta starosti Statistični urad RS več ne objavlja).

žensk neke kohorte, da lahko zaključimo, koliko otrok so te

ženske rodile v svoji rodni dobi. To pomeni, da ne bomo videli

neke ažurne slike rodnosti v državi, temveč bomo videli rodnost

za ženske, ki so se rodile pred 50 leti in so imele vrhunec svoje

rodne dobe pred približno 20 leti. Tako lahko npr. za Slovenijo



2,6 na samo 1,51 otroka. Ker je povprečna starost mater ob rojstvu

2,4 otroka že zadnjih pet let nespremenjena (31,1 let), to pomeni, da

2,2 gre za dejanski upad v ravni rodnosti (quantum učinek), kar se

2,0 tiče preferenc ljudi (skupaj z omejitvami glede njihove

žensko 1,8

na 1,6 realizacije). h

eni 1,4

1,2

ži 4 voroj 1,0 Diskusija

lo 0,8 Prilagojena stopnja celotne rodnosti



vi 0,6 V analizi smo ugotovili, da smo v zadnjih dveh letih (2022 Stopnja celotne rodnosti Šte 0,4 in 2023) priča najnižji rodnosti v zgodovini Slovenije z vidika

0,2 preferenc žensk in moških (skupaj z omejitvami glede njihove

0,0

8 2 6 0 4 8 2 6 0 4 8 2 6 0 4 8 2 realizacije) o številu otrok. Ta najnižja tempo prilagojena stopnja

5 6 6 7 7 7 8 8 9 9 9 0 0 1 1 1 2

9 9 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0

1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 celotne rodnosti ima nadalje še celo negativen trend, če si

Leto

dovolimo razmišljati o trendu na osnovi zgolj treh vrednosti

Slika 2: Gibanje stopnje celotne rodnosti in (tempo) (2021–2023). Upajmo, da sta vrednosti za leti 2022 in 2023 zgolj

prilagojene stopnje celotne rodnosti v Sloveniji v obdobju 1958– naključni nihljaj navzdol, ne pa raven, ki bi se v prihodnje

2023. Vir: [5] (različna leta); lastni izračuni. ohranjala ali da bi celo šlo za začetek trenda nadaljnjega

upadanja. V letih okrog leta 2003 so sicer bile vrednosti stopnje

Razlika med obema prikazanima stopnjama je začela vredno bolj celotne rodnosti bistveno nižje (zgolj okrog 1,2), ampak

naraščati od sredine 1980-ih let, ko je začela povprečna starost predvsem zaradi hitrega odlaganja rodnosti v višjo starost. Z

matere ob rojstvu otroka hitro naraščati (slika 1). Ko se je začela upoštevanjem tega popravka rodnost takrat praktično nikoli ni

rast upočasnjevati (med letoma 2005 in 2010), se je začela upadla pod raven 1,6 otroka. V zadnjih petih letih (2019 do 2023)

izračunana tempo prilagojena stopnja celotne rodnosti je tempo učinek izzvenel in te tolažbe ni več. Priča smo najnižji

približevati stopnji celotne rodnosti in se je v zadnjih letih prilagojeni stopnji celotne rodnosti v zgodovini Slovenije, ki se

povsem izenačila z njo. Naraščanje povprečne starosti matere ob je medtem zaradi neugodne starostne strukture žensk v rodni dobi

rojstvu otroka se je namreč povsem ustavilo – v zadnjih petih tudi že odrazila v najnižjem absolutnem številu živorojenih v

letih (od 2019 do 2023) namreč že ves čas znaša 31,1 let zgodovini Slovenije. V letu 2023 je število živorojenih namreč

[1].Tempo sestavina v stopnji celotne rodnosti je torej izzvenela. prvič upadlo pod 17 tisoč [1].

Kaj to pomeni za vsebinsko analizo rodnosti? Glede števila

živorojenih v posameznem koledarskem letu in posledično za



strukturo prebivalstva v prihodnje ipd. ostane vse enako. Gre za Zahvala prečni kazalnik, ki kaže trenutno raven rodnosti. Moramo pa biti

Delo v okviru tega članka je bilo deloma financirano s strani

previdni pri njegovi interpretaciji v duhu vzdolžnega kazalnika –

projekta SUSTAINWELL (Horizon Europe Research and

torej, koliko otrok bi ena ženska dejansko rodila tekom svoje

Innovation Programme), po pogodbi 101095175, ter s strani

rodne dobe ob trenutnih vzorcih rodnosti. V tem primeru bi npr.

Javne agencije za znanstvenoraziskovalno dejavnost RS (po

v letu 2003 napačno sklepali, da bo ženska tekom svoje rodne

pogodbah P5-0128, J5-4575 in V5-2264). Za uporabo podatkov

dobe imela v povprečju zgolj 1,2 otroka.



Po eni strani je ta naša ugotovitev dobra novica. To namreč zaključke, ki izhajajo iz podatkov, je v celoti odgovoren avtor po naročilu se avtor zahvaljuje Statističnemu uradu RS. Za vse



pomeni, da si ženske (skupaj z njihovimi partnerji) nikoli niso članka. želele (skupaj z omejitvami glede realizacije svojih želja) tako

zelo nizkega števila otrok, kot je to kazala stopnja celotne



rodnosti – npr. zgolj 1,2 otroka v letu 2003. Tudi v teh letih, ko References / Literatura je bila stopnja celotne rodnosti tako zelo nizka, so preference

glede števila otrok (skupaj z omejitvami glede njihove [1] Statistični urad RS, “Osnovni podatki o rojenih, Slovenija, letno,”

Podatkovna baza SISTAT. Accessed: Jun. 25, 2022. [Online]. Available:

realizacije) bile še vedno okrog 1,6 otroka (glej sliko 2). https://pxweb.stat.si/SiStatData/pxweb/sl/Data/-/05J1002S.px

Slabi novici pa sta ob tem dve. Prva slaba novica je, da so te [2] J. Bongaarts and G. Feeney, “On the Quantum and Tempo of Fertility,”

Popul. Dev. Rev., vol. 24, no. 2, pp. 271–291, 1998.

zelo nizke vrednosti stopnje celotne rodnosti (v primerjavi s [3] J. Malačič, Demografija: teorija, analiza, metode in modeli, 6th ed.



tempo prilagojenimi vrednostmi stopnje celotne rodnosti) vseeno Ljubljana: Ekonomska fakulteta, 2006. [4] Eurostat, “Life table by age and sex.” Accessed: Sep. 11, 2024. [Online]. vodile do nizkega absolutnega števila živorojenih v obdobju Available:

razkoraka med tema dvema kazalnikoma. V omenjenega četrt https://ec.europa.eu/eurostat/databrowser/view/demo_mlifetable/default/tab

stoletja, ko je povprečna starost mater ob rojstvu otroka naraščala le?lang=en&category=demo.demo_mor

[5] Statistični urad RS, “Živorojeni po starosti matere in spolu otroka, kohezijski

najhitreje (približno za 0,2 leti na leto), se je namesto okrog 25 regiji, Slovenija, letno,” Podatkovna baza SISTAT. Accessed: Sep. 01, 2024.

tisoč živorojenih letno rojevalo samo okrog 20 tisoč živorojenih. [Online]. Available: https://pxweb.stat.si/SiStatData/pxweb/sl/Data/-/05J1010S.px/table/tableViewLayout2/ Teh izgubljenih 125 tisoč živorojenih v samo teh 25 letih je trajno [6] Statistični urad RS, “Kazalniki rodnosti po starosti matere, kohezijski regiji,

izgubljenih in se ne pričakuje, da bi se ta rojstva kdaj naknadno Slovenija, večletno,” Podatkovna baza SISTAT. Accessed: Sep. 01, 2024.

realizirala, saj se ne pričakuje, da bi se povprečna starost mater [Online]. Available: https://pxweb.stat.si/SiStatData/pxweb/sl/Data/-

/05K1020S.px

ob rojstvu otroka v prihodnje začela nazaj zniževati.

Druga negativna novica pa je odločen upad stopnje celotne

rodnosti v zadnjih dveh letih. Iz ravni 1,64 otroka v letu 2021 se

je namreč v letu 2022 znižala na 1,55 otroka, v letu 2023 pa celo



Zbornik 27. mednarodne multikonference



INFORMACIJSKA DRUŽBA – IS 2024

Zvezek G



Proceedings of the 27th International Multiconference



INFORMATION SOCIETY – IS 2024

Volume G



Vzgoja in izobraževanje v informacijski družbi



Education in Information Society



Urednika / Editors



Uroš Rajkovič, Borut Batagelj



http://is.ijs.si



11. oktober 2024 / 11 October 2024

Ljubljana, Slovenia





PREDGOVOR




Pred vami je zbornik 27. konference Vzgoja in izobraževanje v informacijski družbi – VIVID

2024, ki se je odvijala 11. oktobra 2024 na Institutu Jožef Stefan v Ljubljani. Letos je bilo

sprejetih preko 60 prispevkov, ki pokrivajo širok spekter aktualnih tem s področja vzgoje in

izobraževanja v informacijski družbi.



S hitro napredujočo digitalno preobrazbo, ki vpliva na vsakdanje pedagoške procese, je rdeča

nit letošnje konference raziskovanje novih pristopov pri uporabi informacijsko-

komunikacijskih tehnologij, vključno s hibridnimi modeli izobraževanja, personalizacijo

učnih vsebin ter vključevanjem umetne inteligence v izobraževalne procese.



Skozi leta smo si prizadevali ustvariti okolje, kjer se prepletajo ideje, izkušnje in spoznanja z

različnih stopenj izobraževanja – od predšolske vzgoje do univerzitetnega in vseživljenjskega

učenja. Letos ni izjema. Še naprej razvijamo rešitve, ki ne le sledijo trenutnim trendom,

ampak postavljajo temelje za izobraževalno prihodnost, v kateri bo tehnologija podpirala

inovativne pristope k poučevanju in učenju.



Ta konferenca ni le prostor za izmenjavo najnovejših dognanj in praks, temveč tudi priložnost

za oblikovanje novih povezav, izmenjavo izkušenj in razvoj svežih zamisli, ki bodo krojile

prihodnost izobraževanja. Prispevki, predstavljeni na konferenci, so dragocen doprinos k

razvoju informacijske družbe in k premisleku o prihodnosti izobraževanja.



Naj vam ta zbornik služi kot vir navdiha in strokovnega znanja, ki bo spodbudil nadaljnje

raziskovanje ter inovacije na področju vzgoje in izobraževanja v digitalni dobi.



Uredniški odbor





FOREWORD




Here are the proceedings of the 27th conference Education in the Information Society –

VIVID 2024, which took place on October 11, 2024, at the Jožef Stefan Institute in Ljubljana.

This year, over 60 papers have been accepted, covering a wide range of current topics in the

field of education in the information society.



With the rapidly advancing digital transformation impacting everyday pedagogical processes,

the main theme of this year's conference is the exploration of new approaches to the use of

information and communication technologies, including hybrid education models,

personalized learning content, and the integration of artificial intelligence into educational

processes.



Over the years, we have strived to create an environment where ideas, experiences, and

knowledge from various levels of education—ranging from preschool to university and

lifelong learning—intertwine. This year is no exception. We continue to develop solutions

that not only follow current trends but also lay the foundation for an educational future where

technology supports innovative approaches to teaching and learning.



This conference is not only a platform for sharing the latest findings and practices but also an

opportunity to forge new connections, exchange experiences, and develop fresh ideas that will

shape the future of education. The papers presented at the conference are a valuable

contribution to the development of the information society and to the reflection on the future

of education.



May this proceedings serve as a source of inspiration and expertise, encouraging further

research and innovation in the field of education in the digital age.



The Editorial Board



PROGRAMSKI ODBOR / PROGRAMME COMMITTEE



Uroš Rajkovič (predsednik / Chair), Univerza v Mariboru, Fakulteta za organizacijske vede



Borut Batagelj (podpredsednik – Vice-Chair), Univerza v Ljubljani, Fakulteta za računalništvo in

informatiko



Zvone Balantič, Univerza v Mariboru, Fakulteta za organizacijske vede



Tadeja Batagelj, Svetovalni center za otroke, mladostnike in starše Maribor



Igor Bernik, Univerza v Mariboru, Fakulteta za varnostne vede



Mojca Bernik, Univerza v Mariboru, Fakulteta za organizacijske vede



Janez Bešter, Univerza v Ljubljani, Fakulteta za elektrotehniko



Uroš Breskvar, Elektrotehniško-računalniška strokovna šola in gimnazija Ljubljana



Andrej Brodnik, Univerza v Ljubljani, Fakulteta za računalništvo in informatiko in Univerza na

Primorskem, Fakulteta za matematiko, naravoslovje in informacijske tehnologije



Borut Čampelj, Ministrstvo za izobraževanje, znanost in šport RS



Dejan Dinevski, Univerza v Mariboru, Medicinska fakulteta



Tomi Dolenc, ARNES



Marjan Heričko, Univerza v Mariboru, Fakulteta za elektrotehniko, računalništvo in informatiko



Eva Jereb, Univerza v Mariboru, Fakulteta za organizacijske vede



Mirjana Kljajić Borštnar, Univerza v Mariboru, Fakulteta za organizacijske vede



Alenka Krapež, Gimnazija Vič



Branislav Šmitek, Univerza v Mariboru, Fakulteta za organizacijske vede



Srečo Zakrajšek, Fakulteta za medije



PROGRAMSKI SVET / STEERING COMMITTEE



Matjaž Gams (predsednik / Chair), Institut Jožef Stefan



Vladimir Batagelj, Univerza v Ljubljani, Fakulteta za matematiko in fiziko



Saša Divjak, Univerza v Ljubljani, Fakulteta za računalništvo in informatiko



Ivan Gerlič, Univerza v Mariboru, Fakulteta za naravoslovje in matematiko



Iztok Podbregar, Univerza v Mariboru, Fakulteta za organizacijske vede



Vladislav Rajkovič, Univerza v Mariboru, Fakulteta za organizacijske vede in Institut Jožef Stefan



Niko Schlamberger, Slovensko društvo Informatika



Franc Solina, Univerza v Ljubljani, Fakulteta za računalništvo in informatiko



Tomaž Skulj



Olga Šušteršič



Rado Wechtersbach



RECENZENTI / REVIEWERS



Alenka Baggia, Univerza v Mariboru, Fakulteta za organizacijske vede



Borut Batagelj, Univerza v Ljubljani, Fakulteta za računalništvo in informatiko



Tadeja Batagelj, Svetovalni center za otroke, mladostnike in starše Maribor



Andrej Brodnik, Univerza v Ljubljani, Fakulteta za računalništvo in informatiko in Univerza na

Primorskem, Fakulteta za matematiko, naravoslovje in informacijske tehnologije



Matija Lokar, Univerza v Ljubljani, Fakulteta za matematiko in fiziko



Marko Novaković, Univerzitetni klinični center Ljubljana



Uroš Rajkovič, Univerza v Mariboru, Fakulteta za organizacijske vede



Gašper Strniša, Šolski center Kranj, Strokovna gimnazija



Iva Strniša, Gimnazija Franceta Prešerna, Kranj



Iztok Škof, Osnovna šola Toma Brejca Kamnik



Marko Urh, Univerza v Mariboru, Fakulteta za organizacijske vede



Vladislav Rajkovič, Univerza v Mariboru, Fakulteta za organizacijske vede



Borut Werber, Univerza v Mariboru, Fakulteta za organizacijske vede



Jasmina Žnidaršič, Univerza v Mariboru, Fakulteta za organizacijske vede



Optimizacija praktičnega pouka digitalne tehnike z



avtomatskim testiranjem integriranih vezij



Optimization of Practical Lessons at Digital Electronics Using



Automated Integrated Circuit Testing



Jaka Albreht

Šolski center Kranj

Kranj, Slovenija

jaka.albreht@sckr.si

pouka. Dijaki običajno v primeru težav pokličejo učitelja. Le-ta

POVZETEK preveri delovanje in identificira kje je napaka. Pri tem mora biti

V članku je predstavljen problem porabe časa pri testiranju hiter in učinkovit, saj lahko na njegovo pomoč že čakajo drugi

integriranih vezij med praktičnim poukom digitalne tehnike. dijaki. S tem namenom je smiselno določene postopke

Učitelj mora ročno preveriti vse možne vhodne kombinacije na avtomatizirati in s tem izboljšati učinkovitost in hitrost učnega

vseh logičnih vratih znotraj integriranega vezja, kar je lahko procesa. V nadaljevanju je prikazan potek praktičnega pouka in

precej zamudno. Z namenom rešitve tega problema je bil težave s katerimi se soočamo. Predstavljen je sistem za

izdelan sistem za avtomatsko testiranje integriranih vezij. Le-ta avtomatsko testiranje integriranih vezij, ki je bil razvit z

nam omogoča, da v uporabniškem vmesniku izberemo oznako namenom izboljšanja in pohitritve procesa praktičnega pouka

vezja, ki ga želimo testirati. Integrirano vezje namestimo na digitalne tehnike.

ustrezno mesto na testni plošči, nato pa program sam preveri



vse možne kombinacije in nam preko virtualnih barvnih 2 POTEK PRAKTIČNEGA POUKA indikatorjev javi pravilnost delovanja posameznih logičnih vrat

znotraj integriranega vezja. Na ta način smo čas testiranja Vsebina praktičnega pouka digitalne tehnike je razdeljena na

skrajšali za približno 20-krat. tri dele. V prvem delu dijaki spoznavajo integrirana vezja, ki

vsebujejo logična vrata. Naučijo se kako poiskati

KLJUČNE BESEDE dokumentacijo integriranih vezij in iz nje razbrati pomembne

Integrirano vezje, avtomatsko testiranje, digitalna tehnika, podatke. Najprej na testni plošči preizkušajo delovanje

praktični pouk, Arduino, Labview, optimizacija posameznih čipov, kasneje pa izdelujejo tudi logična vezja, ki

so sestavljena iz več čipov. Glede na to, da se dijaki prvič



ABSTRACT srečajo z uporabo čipov, se verjetnost napak precej poveča.

Napake, s katerimi se pogosto soočajo so: oznaka na čipu ni

The article presents the problem of time consumption in the bila preverjena, odlomljen ali ukrivljen pin, nedelujoča vrata ali

testing of integrated circuits during the practical lessons of celoten čip, previsoka napajalna napetost, zamenjana polariteta

digital electronics. The teacher must manually check all napajalne napetosti, uničena svetleča dioda (LED), narobe

possible input combinations on all logic gates within the IC, obrnjena LED, slab stik …

which can be quite time-consuming. This allows us to select the Med učnim procesom dijak v primeru težav pokliče učitelja.

label of the circuit we want to test in the user interface. We Učitelj preveri ali so čipi pravilno povezani. Če kljub temu



board, and then the program itself checks all possible delujejo. Slednje je lahko zamudno, sploh, če vezje vsebuje combinations and informs us via virtual color indicators of the večje število čipov oz. logičnih vrat. Učitelj mora ročno place the integrated circuit in the appropriate place on the test vezje ne deluje, se ugotavlja kateri čip oz. vrata znotraj čipa ne



correct operation of the individual logic gates inside the preveriti vse možne vhodne kombinacije na posameznih integrated circuit. In this way, we reduced the testing time by logičnih vratih. Z namenom rešitve tega problema, smo prišli na about 20 times. idejo, da bi postopek preverjanja delovanja posameznih



KEYWORDS učitelj porabi, da ročno preveri ali čip deluje in čas, ki se ga logičnih vrat avtomatizirali. Za primerjavo vzemimo čas, ki ga

Integrated circuit, automated testing, digital electronics, porabi z uporabo sistema za avtomatsko testiranje čipov. V

practical lessons, Arduino, Labview, optimization prvem primeru je potrebno na posameznih vratih izvesti štiri

vhodne kombinacije (00, 01, 10, 11) in opazovati izhod preko



1 UVOD vrat v čipu. Običajno uporabljamo čipe s štirimi 2-vhodnimi LED. Nato je potrebno postopek ponoviti tolikokrat, kolikor je

Pri praktičnem pouku strokovnega modula Digitalna tehnika vrati. V drugem primeru pa namestimo čip na ustrezno mesto

dijaki izdelujejo vezja, ki so sestavljena iz integriranih vezij oz. na testni plošči, v aplikaciji preko spustnega seznama izberemo

čipov [1]. Pomembno je zagotoviti nemoten proces praktičnega kateri čip bomo testirali, kliknemo virtualni gumb “testiraj” in

čez manj kot sekundo dobimo rezultate. Vse skupaj nam vzame

največ 10 sekund časa na posamezen čip. V primeru ročnega

testiranja bi porabili približno tri minute. Torej smo čas, ki je

potreben za ugotavljanje napake, skrajšali za približno 20-krat.



3 SISTEM ZA TESTIRANJE VEZIJ

Sistem sestavlja mikrokrmilniška razvojna plošča Arduino

[2], ki je z žičkami povezana s testno ploščo, kamor namestimo

integrirano vezje (Slika 1). Sistem je povezan na računalnik z

USB-kablom, ki služi za napajanje in komunikacijo.

Uporabniški vmesnik je bil razvit v programskem okolju

Labview [3].





Slika 3: Integrirano vezje 7408 z logičnimi vrati IN (AND)



Logična funkcija AND (IN) na svojem izhodu vrne logično

1, če so vsi vhodi enaki 1. V našem primeru bo logično 1

predstavljala napetost 5 V, logično 0 pa napetost 0 V. Glede na

to, da imamo dva vhoda, bomo imeli štiri vhodne kombinacije

ničel in enic. Pravilnostna tabela prikazuje vse možne vhodne

kombinacije in pripadajoč izhod (Tabela 1) .



Tabela 1: Pravilnostna tabela funkcije AND (IN)



Vhod A Vhod B Izhod Y

0 0 0

0 1 0

1 0 0

1 1 1

Slika 1: Sistem za testiranje vezij

Pri testiranju logičnih vrat moramo na vhodih generirati vse

Aplikacija nam omogoča izbiro čipa, ki ga želimo testirati. Z štiri možne kombinacije in opazovati izhodno vrednost. Izhod

virtualnim gumbom “testiraj” sprožimo proces testiranja. V bo 1, če bosta oba vhoda enaka 1. V ostalih primerih mora biti

primeru pravilnega delovanja posameznih logičnih vrat se izhod 0.

indikator obarva zeleno, sicer pa rdeče. Izbirati je možno med Program, ki se izvaja na mikrokrmilniški plošči Arduino, na

petimi različnimi čipi (7400, 7402, 7432, 7486, 7408), ki svojih digitalnih izhodih generira vse možne kombinacije ničel

vsebujejo različna logična vrata (Slika 2). in enic. Nato preko digitalnih vhodov prebere izhodne vrednosti

posameznih logičnih vrat čipa. Rezultati se preko USB-

povezave posredujejo računalniku, kjer je nameščeno okolje



Labview. Le-to nam služi za prikaz rezultatov testiranja.



Dejanske izhodne vrednosti logičnih vrat se primerjajo z

želenimi in v primeru odstopanj se indikator obarva rdeče, kar

označuje nepravilno delovanje vrat. V našem primeru opazimo,

da četrta vrata ne delujejo pravilno, saj dobimo na izhodu same

enice (Slika 4). Pri vsakem izbranem čipu se prikaže tudi

ustrezna shema. Testiranje pa se prekine s klikom na gumb

“prekini testiranje”.



Slika 2: Izbira čipov v okolju Labview



Oglejmo si primer testiranja čipa z oznako 7408, ki vsebuje

štiri 2-vhodna IN (AND) vrata (Slika 3).



Uspelo nam je torej izdelati avtomatski sistem, ki je pohitril

proces iskanja napak, izboljšal potek izvajanja praktičnih vaj in

razbremenil učitelja.



VIRI

[1] Integrirano vezje. Dostopno na naslovu

https://sl.wikipedia.org/wiki/Integrirano vezje (17. 8. 2024)

[2] Arduino. Dostopno na naslovu www.arduino.cc (17. 8. 2024)

[3] Labview. Dostopno na naslovu www.ni.com (17. 8. 2024)



Slika 4: Testiranje čipa 7408



4 MOŽNOSTI NADGRADNJE

Sistem bi lahko nadgradili tako, da bi razširili nabor čipov.

Zanimivo bi bilo narediti samodejno razpoznavanje logičnih

vrat, brez, da bi vnaprej preko spustnega seznama izbrali čip.

Rezultati testiranja bi se lahko shranjevali v posebno tabelo ali

tekstovno datoteko. Alternativo uporabniškemu vmesniku bi

lahko predstavljalo vezje na testni plošči, kjer bi s tipkami

izbrali ustrezna vrata in LED, ki bi signalizirale pravilnost

delovanja. V tem primeru ne bi potrebovali računalnika s

programskim okoljem.



5 ZAKLJUČEK

Dijaki so pozitivno sprejeli sistem za testiranje. V primeru,

da se jim zdi, da je s čipom kaj narobe, ga prinesejo učitelju, ki

izvede avtomatsko testiranje. Na ta način hitro dobijo povratno

informacijo, kje je vzrok nepravilno delujočega vezja. Dostikrat

pa samo iščejo potrditev svojih opažanj. Npr. “Se mi je zdelo,

da četrta vrata ne delujejo.” oz. “Sedaj vem, da čip deluje.

Verjetno sem nekaj narobe povezal.”



Codeium kot podporni asistent v poučevanju



programiranja



Codeium as a Support Assistant in Teaching Programming



Davorin Babič

Srednja medijska in grafična šola

Ljubljana

Pokopališka ulica 33, 1000

Ljubljana davorin.babic@smgs.si



POVZETEK ABSTRACT

V prispevku predstavljamo orodje za programiranje v oblaku, In this paper, we present a cloud-based programming tool

ki temelji na vodilnih odprtokodnih modelih, podprto z based on leading open source models supported by generative

generativno umetno inteligenco, imenovano Codeium, za artificial intelligence, called Codeium, to support teachers and

podporo učiteljem in dijakom pri poučevanju in učenju osnov students in teaching and learning the basics of programming.

programiranja. Codeium v osnovi ponuja funkcije klepeta in Codeium basically provides chat features and calls to code to

poziva h kodi za generiranje programske kode. Ta generate software code. This functionality focuses on generating

funkcionalnost se osredotoča na generiranje kode iz besedilnih code from text inputs, allowing text-to-code generation within

vnosov, kar omogoča generiranje besedila v kodo znotraj the appropriate integrated development environment (IDE), in

ustreznega integriranega razvojnega okolja (IDE), v našem our case Visual Studio Code. By incorporating these advanced

primeru Visual Studio Code. Z vključitvijo teh naprednih code generation capabilities, the tool enhances learning streams

zmožnosti generiranja kode orodje izboljšuje učne tokove for teaching programming fundamentals, making the learning

poučevanja osnov programiranja, zaradi česar je učni proces bolj process more efficient and intuitive. In addition to enabling

učinkovit in intuitiven. Poleg omogočanja znanih interakcij, kot familiar interactions like those with ChatGPT, students can use

so tiste s ChatGPT-jem, lahko dijaki uporabljajo pametne smart templates to perform common actions such as adding

predloge za izvajanje običajnih dejanj, kot je dodajanje documentation to functions or refactoring program code. The

dokumentacije funkcijam ali preoblikovanje programske kode. code suggestion feature provides a viable option to complete and

Funkcija predlaganja kode ponuja izvedljivo možnost za generate code in various usage scenarios and offers additional

dokončanje in generiranje kode v različnih scenarijih uporabe in enhancements for an interactive coding experience. Code

ponuja dodatne izboljšave za interaktivno izkušnjo kodiranja. completion, commonly referred to as autocompletion, is an

Dokončanje kode, običajno imenovano samodokončanje, je invaluable feature in both learning and software development

neprecenljiva funkcija tako pri učenju kot pri razvoju that helps complete unfinished code segments. Code generation,

programske opreme, ki pomaga pri dokončanju nedokončanih on the other hand, involves the automatic generation of source

segmentov kode. Po drugi strani pa generiranje kode vključuje code from natural language input guided by user-defined

samodejno generiranje izvorne kode iz vnosa naravnega jezika, constraints. This capability increases the efficiency of the

ki ga vodijo uporabniško določene omejitve. Ta zmožnost development process by automating the generation of code based

povečuje učinkovitost razvojnega procesa z avtomatizacijo on user-supplied language specifications. The paper describes

ustvarjanja kode na podlagi jezikovnih specifikacij, ki jih the use of these functions in the learning process and their impact

zagotovi uporabnik. Prispevek opisuje uporabo teh funkcij v on the effectiveness of learning programming. Research has

učnem procesu in njihov vpliv na učinkovitost učenja shown that using Codeium significantly improves learning

programiranja. Raziskava je pokazala, da uporaba Codeiuma efficiency and promotes a deeper understanding of programming.

bistveno izboljša učinkovitost učenja in spodbuja globlje Finally, we present findings on the impact of Codeium on student

razumevanje programiranja. Na koncu predstavljamo ugotovitve performance and suggest directions for further research.

o vplivu Codeiuma na uspešnost dijakov in predlagamo smernice

za nadaljnje raziskave. KEYWORDS

Programming, Codeium, artificial intelligence, teaching,

KLJUČNE BESEDE learning process

Programiranje, Codeium, umetna inteligenca, poučevanje,

učni proces

1 UVOD

Programiranje je ustvarjalna dejavnost in kot ena ključnih

sestavin računalniškega mišljenja podpira kognitivne procese in



dokazuje računalniško usposobljenost učečih se deležnikov, programske kode in integrirana razvojna okolja (IDE), zaradi

hkrati pa ponuja platformo za prakso in izboljšanje reševanja česar so učni in delovni tokovi preprostejši in udobnejši. Na

problemov, kritično mišljenje, sodelovanje, in spretnosti osnovi integriranega klepeta je njegova uporabniška izkušnja

aktivnega učenja. Vendar se programiranja ni enostavno naučiti. podobna tisti, ki jo ponuja ChatGPT, vendar je integrirana v

Preobremenjenost zaradi nezadostnega predznanja in urejevalnike ali razvojna okolja, ki jih učitelji in dijaki

zapletenosti nalog kodiranja pri dijakih začetnikih povzroči uporabljajo, tako da lahko pogovorno komunicirajo z

nepričakovano veliko porabo časa za njihovo uspešno generativno umetno inteligenco znotraj teh istih orodij. To

razumevanje in reševanje [1]. Frustrirajoča izkušnja in sofisticirano orodje podpira zanimivo paleto več kot 70

ponavljajoči se neuspehi, zlasti na začetku, lahko pri dijakih programskih jezikov in zlasti zmanjša breme učenja osnov

zmanjšajo samoučinkovitost glede programiranja. programiranja na osnovi kodiranja. Codeium je pokazal izjemne

Za izboljšanje učne izkušnje v učni proces na začetku sposobnosti pri ustvarjanju človeških odzivov in zagotavljanju

vpeljemo pomožna programska okolja, ki pomagajo blažiti prilagojenih povratnih informacij, kar ga uvršča med potencialno

napačne predstave v sintaksi izbranega programskega jezika in orodje za preoblikovanje učne izkušnje v poučevanju in učenju

konceptualnem znanju. Poučevanje osnov programiranja v osnov programiranja. Transformativni potencial Codeium-a z

programskih okoljih, ki temeljijo na slikovnih delčkih, odpravi ustreznim načrtovanjem nudi učiteljem priložnost ustvarjati

sintaksne napake in dijakom omogoči delo na zasnovanih interaktivna in prilagodljiva učna okolja, ki lahko pritegnejo in

projektnih nalogah. Uporaba teh okolij znižuje oviro za vstop v motivirajo dijake za učenje. Tak pristop dijakom omogoča, da se

programiranje in dijakom omogoča, da se osredotočijo na učenje, bolj povežejo z učnim gradivom, kar posledično vodi do

kako oblikovati rešitev, ki jo z računalnikom lahko izvedejo. globljega razumevanja konceptov programiranja, jezikov in

Vendar se kmalu pojavi potreba po prehodu na besedilne postopkov, ki so bili abstrahirani skozi številne nivoje abstrakcij.

programske jezike, kar pa prinaša svoje izzive, hkrati pa odpira Pogovorni slog Codeium-a lahko ustvari bolj zanimivo in

vprašanja, s katerimi orodji ali tehnologijami v procesu prijetno učno izkušnjo, hkrati pa zmanjšuje dolgočasje in

poučevanja ta prehod čimbolj učinkovito izvesti. nezainteresiranost, ki se včasih pojavita pri tradicionalnem

Poučevanja in učenje osnov programiranja izboljša veščine poučevanju v obliki predavanj. To sodelovanje prispeva k bolj

programiranja v ustvarjalne namene, dijake pa seznani z angažiranemu in motiviranemu učenju, kar lahko pozitivno

osnovnimi koncepti programiranja in programskimi jeziki, kot so vpliva na uspeh dijakov v izobraževanju na področju medijske

Python, JavaScript in podobni. Dijaki pridobijo sposobnost produkcije.

pisanja preprostih programov z uporabo besedilnih programskih Učitelji z usklajevanjem zastavljenih učnih ciljev z odprto

orodij in se naučijo korakov v procesu reševanja problemov s uporabo umetne inteligence v ustrezno zasnovanem učnem

programiranjem. Ko poučevanje programiranja napreduje, se okolju dijake usmerjajo k uporabi umetne inteligence za

dijaki premikajo k bolj zapletenim programskim konceptom, kot opolnomočenje njihovega učenja, ne pa za njegovo nadomestitev.

so algoritmi, podatkovne strukture in programske paradigme. To Vključevanje umetne inteligence v sodelovalni učni proces lahko

dijakom zagotavlja trdne temelje v načelih programiranja in spremeni miselnost dijakov od zgolj uporabe umetne inteligence

tehnikah reševanja problemov ter poveča njihovo sposobnost za za hitre odgovore k uporabi umetne inteligence kot močnega

razvoj programske opreme za reševanje problemov iz resničnega zaveznika za nenehne izboljšave in razvoj učnih spretnosti. Pri

sveta [2]. Dijaki pridobijo tudi veščine testiranja, odpravljanja podpori orodij umetne inteligence pri poučevanju je še vedno

napak in vzdrževanja programske opreme, hkrati pa jim pomaga ključna vloga učitelja, ki se od poučevalca spreminja v

razumeti procese razvoja programske opreme za različna usmerjevalca učnega procesa. Učitelj mora zagotoviti pravilno

področja uporabe in jih pripravi na uspešno kariero v tehnologiji. uporabo orodij umetne inteligence, interpretirati rezultate in

V zadnjem času je prišlo do porasta priljubljenosti orodij za nuditi nadaljnjo razlago ter voditi dijake skozi učno snov. Prav

ustvarjanje ali dokončanje programske kode, ki temeljijo na tako je odgovoren za ustvarjanje ustrezne učne okolice, ki

umetni inteligenci: Github Copilot, Codeium, Tabnine, če spodbuja sodelovanje in interakcijo med dijaki. Najpomembneje

naštejemo le nekatere. Za razliko od tradicionalnih orodij za je, da mora učitelj poudariti veščine reševanja problemov in

samodejno dokončanje lahko ta nova generacija orodij za kritičnega mišljenja, ki jih kodiranje zahteva. Zato je potrebno

ustvarjanje ali dokončanje kode ponudi poljubne izrezke kode dijake spodbujati, da poglobljeno razmišljajo o problemih, ki jih

kot predloge, pri čemer upošteva kontekst kode, vključno s poskušajo rešiti, in da se jim približajo na kreativen in inovativen

komentarji, napisanimi v naravnem jezik. Čeprav vedno ne način. Prav tako je dijake potrebno poučiti tudi o etičnih

zagotavljajo pravilnosti predlogov, so lahko zelo koristni pri posledicah generativne umetne inteligence in strojnega učenja.

pospešitvi razvojnega procesa v razumevanju osnov Uvajanje inovativnih učnih tehnologij v pouk je vedno

programiranja. eksperimentalno in zahteva določeno zavedanje o potrebah

Za izboljšanje učinkovitosti poučevanja osnov programiranja učitelja in dijakov. Umetna inteligenca je začela iskati svojo pot

smo v učni proces pri kreativnem kodiranju vključili Codeium-a, v učilnice, šole in načine, na katere določene učne dejavnosti

že omenjeni uporabniški vmesnik klepetalnega robota, ki ga je dijakom pripeljemo domov preko projektnih nalog ali

razvilo tehnološko podjetje Exafunction. Codeium je napredno izvenšolskih dejavnosti. Čeprav te tehnologije ponujajo nove

orodje za ustvarjanje in dokončanje programske kode, izboljšano možnosti za poučevanje in učenje, prinašajo tudi nove izzive in

z umetno inteligenco, ki na novo definira tradicionalne prakse zahtevajo kritičen razmislek o osnovnih učnih strategijah in

kodiranja, saj je bolj iterativen in razčlenjuje rešitve vnešenih smernicah, ki oblikujejo njihovo uporabo. Pri poučevanju in

poizvedb korak za korakom. To ga naredi bolj didaktičnega in učenju kodiranja v takšnem učnem okolju morajo učitelji

razumljivejšega, hkrati pa omogoča večjo stopnjo globine spremljati aktivnosti, s katerimi se dijaki osredotočajo predvsem

odgovorov. Uspešno se vključuje v različne urejevalnike na inovativno reševanje problemov/izzivov in oblikovanje

elegantnih rešitev s pomočjo generativne umetne intelligence, izboljšanjem kode in povzemanjem kode. Vaje razumevanja

kar je prikazano v Tabeli 1. primarno zajemajo dejavnosti, kot je odkrivanje napak v

ustvarjeni programski kodi.



Tabela 1: Aktivnosti dijakov in njihova spremljava generativne umetne inteligence za programiranje pri učenju Pred raziskavo vpliva uporabe podpornega asistenta

programiranja so se dijaki že učili o osnovnih konceptih



Aktivnosti programiranja: spremenljivkah, operatorjih, izrazih, pogojih, Spremljava aktivnosti zaporedjih, vejitvah, zankah, funkcijah in seznamih znotraj

Reševanje Spodbujanje dijakov, da k izzivom slikovnega programiranja z delčki. Metodološki okvir so

problemov kodiranja pristopijo sistematično in da predstavljale praktično usmerjene vaje in so vključevale

kreativno razmišljajo o možnih dejanske primere uporabe. Da bi raziskali učinek uporabe

rešitvah. Codeiuma pri učenju osnov programiranja v python (Slika 1),

Računalniško Pomagati dijakom pri razvijanju smo učni proces uvedli hitro generiranja vsebin z relativno

mišljenje razumevanja delovanja računalnikov in enostavnimi problem, vezanimi za strokovno zahtevnost. Sistem

algoritmov. omogoča dijakom, da delajo na vrsti programskih nalog, ki so

Logično mišljenje Dijake naučiti uporabljati logičnega bile zasnovane za postopno uvajanje novih konceptov.

mišljenja za razdelitev kompleksnih



problemov na manjše, bolj obvladljive

dele.

Algoritmično Dijaki bi morali vedeti, kako ustvariti

mišljenje postopke za reševanje problemov po

korakih.

Ustvarjalnost Spodbujanje dijakov k razmišljanju

“out of the box” in ustvarjanju

projektov za zabavo in užitek.

Sodelovalnost Razvijanje sodelovalnega pristopa

učenja pri reševanju izzivov kodiranja

in medsebojnega deljenja pridobljenih

znanj in izkušenj.

Razhroščevanje in Dijake naučiti, kako prepoznati in

odpravljanje težav odpraviti napake na sistematičen način.

Etični vidiki Nudenje priložnosti dijakom, da

razpravljajo o etičnih posledicah

umetne inteligence in strojnega učenja

ter o tem, kako lahko njihova koda Slika 1: Uporabniški vmesnik Visual Studio Code: 1)



vpliva na družbo in svet okoli njih. besedilni vnos (poziv) problema v vdelanem oknu Codeiuma; 2) generator kode za vstavljanje kode, ki jo

ustvari umetna inteligenca glede na dijakov vnos, v



2 urejevalnik kode; 3) vnesena koda v pythonovski METODA DELA

urejevalnik z možnostjo dokončanja; 4) rezultat delovanja

Poučevanje in učenje programiranja s pomočjo generativne kode v terminalu urejevalnika. Vir: lastno delo

umetne inteligence je vključevanje tehnik in orodij strojnega

učenja v proces razvoja programske opreme [3] za izboljšanje 2.1 Učni proces z uporabo Codeiuma



interakcijo dijakov z generativno umetno inteligenco kaže programske kode, integriran v različne urejevalnike programske podobnosti z modernejšo obliko sodelovalnega dela, s kode in razvojna okolja (IDE), s svojimi funkcionalnostmi nalog računalniškega programiranja. Ta pristop v poučevanju z Codeium kot napredno orodje za generiranje in dokončanje



programiranjem v paru. Programiranje v paru opisuje aktivno omogoča prilagoditev učnega procesa na dijake osredotočenega prakso dveh dijakov v vlogi programerjev, ki skupaj izmenično poučevanja z naslednjimi možnostmi [4]: sodelujeta v neprekinjenem procesu do dokončanja naloge z

uporabo enega računalnika, tipkovnice in miške. En programer v • samodokončanje kode: Codeium samodejno predlaga

paru, voznik, izvaja kodiranje (tipkanje) in izvaja nalogo, dokončanje delov kode na osnovi konteksta nalog, kar

medtem ko drugi programer, krmar, pomaga pri načrtovanju, dijakom omogoča hitro in učinkovito pisanje kode;

pregledovanju, odpravljanju napak in predlaganju izboljšav in • generiranje kode iz naravnega jezika: dijaki lahko vnesejo

alternativ. Programiranje s pomočjo umetne inteligence v osnovi opis naloge v naravnem jeziku, orodje pa samodejno

nadomešča enega od dveh programerjev s podpornim asistentom generira ustrezno kodo;

generativne umetne inteligence. Učitelji s podpornim asistentom • interaktivna podpora: Codeium omogoča pogovorno

‘avtomatiziramo’ nabor vaj, ki jih je mogoče na splošno razvrstiti interakcijo, podobno kot ChatGPT, vendar znotraj

v dve zahtevnostni kategoriji: ustvarjanje in razumevanje. Vaje razvojnih okolij, kar omogoča enostavno integracijo v

ustvarjanja zajemajo dejavnosti, povezane z generiranjem učni proces (v obliki ustreznih razlag in pomoči, npr.

programske kode, dokončanjem kode, prevodom kode, dijaki lahko uporabijo Codeium za vprašanja, kot so



"Kako lahko izračunam povprečje števil iz nekega 2. Kvalitativna analiza: Izvedli smo anketo med dijaki,

številskega seznama?" ali " Kateri je najboljši način za kjer so ocenili uporabnost orodja, enostavnost uporabe ter

ocenjevanje uspešnosti dijakov?" in podobno, lahko njihovo zadovoljstvo z rezultati.

predlaga optimizacije, kot so izboljšanje besedilnega

vnosa, obravnavanje napak pri vnosu podatkov ali 2.5 Cilji raziskave

dodajanje funkcionalnosti za večjo robustnost programa). Med osnovne cilje raziskave smo uvrstili naslednje:



2.2 1) Oceniti vpliv uporabe podpornega asistenta v poučevanju Eksperimentalna zasnova

na učinkovitost učenja programiranja.

Za statistično analizo prispevka smo uporabili t-test, da smo

2) Raziskati, ali uporaba Codeiuma pripomore k boljšemu

preverjali vpliv uporabe orodja Codeium na uspešnost učenja

razumevanju in utrjevanju osnovnih programskih

programiranja pri dijakih. Raziskava je bila izvedena na skupini

konceptov in konstruktov.

53 dijakov v programu medijski tehnik ob koncu drugega

3) Preveriti, kako se dijaki odzivajo na uporabo orodij

ocenjevalnega obdobja, razdeljenih v dve skupini:

generativne umetne inteligence pri reševanju

eksperimentalno skupino (26 dijakov), ki je uporabljala Codeium,

problemov/izzivov z določenega strokovnega področja s

in kontrolno skupino (27 dijakov), ki tega orodja ni uporabljala.

programiranjem.

Analizirali smo dve glavni spremenljivki: čas dokončanja nalog



in število napak v kodi. 2.6 Hipoteze

Dijaki so imeli nalogo razviti enostavne aplikacije v

programskem jeziku python. Med izvajanjem nalog so Na podlagi ciljev raziskave oblikujemo naslednje hipoteze:

uporabljali funkcije samodokončanja in generiranja kode ter • Hipoteza 1 (H1): Povprečni čas dokončanja nalog bo

pametne predloge za izboljšanje svoje kode. Opazovali smo krajši pri dijakih, ki uporabljajo Codeium

njihovo interakcijo z orodjem in spremljali uspešnost pri (eksperimentalna skupina), v primerjavi s tistimi, ki tega

reševanju nalog, ob upoštevanju aktivnosti iz Tabele 1. Naloge orodja ne uporabljajo (kontrolna skupina).

so vključevale primere iz resničnega sveta, kar je omogočilo • Hipoteza 2 (H2): Povprečno število napak v kodi bo

dijakom, da so se soočili s praktičnimi izzivi, ki jih srečujejo manjše pri dijakih, ki uporabljajo Codeium, v primerjavi s

programerji v vsakdanjem delu. Primeri so bili skrbno izbrani, da kontrolno skupino. so se navezovali na osnovne koncepte, ki so jih dijaki že poznali,

a so hkrati vključevali tudi nekoliko zahtevnejše elemente, ki so

zahtevali dodatno razmišljanje in uporabo podpornega asistenta. 3 REZULTATI



2.3 Zahtevnostne kategorije nalog 3.1 Statistična analiza



Osnovna zahtevnost 1. Analiza časa dokončanja nalog (v minutah):

Naloge so se osredotočale na osnovne programske koncepte, • Neodvisna spremenljivka: skupina (eksperimentalna,

kot so deklaracija spremenljivk, matematični izrazi in pisanje kontrolna)

pogojnih stavkov. Na primer, dijaki so morali napisati program • Odvisna spremenljivka: čas dokončanja nalog (v minutah)

v obliki preproste aplikacije za izračun obsega in površine kroga • Test: neodvisni t-test (za primerjavo povprečij dveh

ter program za preverjanje, ali je vnešeno število pozitivno ali neodvisnih skupin)

negativno. 2. Analiza števila napak v kodi:

• Neodvisna spremenljivka: skupina (eksperimentalna,

Srednja zahtevnost kontrolna)

Naloge so vključevale delo z nizi, seznami, zankami in • Odvisna spremenljivka: število napak v kodi

slovarji. Ena naloga je zahtevala, da dijaki napišejo program, ki • Test: neodvisni t-test preveri, ali je dani niz palindrom, druga pa je vključevala izračun

povprečne vrednosti elementov seznama in prikaz elementov, ki 3.2 Izvedba t-testa so večji od povprečja.

Izvedli smo t-test za obe hipotezi. Na osnovi predhodno

Naprednejše naloge obdelanih rezultatov smo za izvedbo t-testa uporabili sledeče

Naloge so zahtevale uporabo regularnih izrazov, obdelavo podatke:

podatkov z uporabo knjižnice pandas in vizualizacijo podatkov s • Eksperimentalna skupina (Codeium):

pomočjo knjižnice matplotlib. Na primer, dijaki so morali • Povprečni čas dokončanja nalog: 45 minut, standardni

napisati program za preverjanje veljavnosti e-poštnih naslovov odklon: 5 minut.

in ustvariti stolpčni diagram, ki prikazuje razporeditev ocen • Povprečno število napak: 2, standardni odklon: 1.

dijakov. • Kontrolna skupina (brez Codeiuma):

• Povprečni čas dokončanja nalog: 55 minut, standardni

2.4 Analiza učinkovitosti odklon: 7 minut.



poučevanju programiranja je temeljila na dveh metodah: Analiza učinkovitosti Codeiuma kot podpornega asistenta v • Povprečno število napak: 5, standardni odklon: 2. Rezultati t-testa so bili naslednji (Tabela 2).

1. Kvantitativna analiza: Merili smo čas, ki so ga dijaki

porabili za dokončanje nalog, in število napak v kodi.



Tabela 2: Prikaz rezultatov t-testa veščin niso bili ocenjeni, saj so bili spremljani ob koncu drugega

ocenjevalnega obdobja. Poleg tega bi bilo koristno raziskati,

kako vključitev Codeiuma vpliva na učenje naprednejših

Čas dokončanja nalog Število napak

konceptov programiranja, kot so algoritmi in podatkovne

t-vrednost: −5.96 t-vrednost: −6.87 strukture, ter na razvoj projektov v skupinskem okolju.

p-vrednost: 2.33×10−7 p-vrednost: 8.89×10−9 Vendar pa je pomembno upoštevati, da lahko prekomerna

uporaba AI asistentov vodi do prekomerne odvisnosti, kar lahko



3.3 Kvantitativni rezultati probleme. Nadaljnje raziskave bi morale preučiti, kako uskladiti negativno vpliva na sposobnost dijakov, da samostojno rešujejo

Čas dokončanja nalog: p-vrednost je bistveno manjša od uporabo podpornih asistentov s tradicionalnimi metodami

običajnega praga statistične značilnosti (0,05), kar pomeni, da poučevanja, da bi zagotovili optimalne rezultate pri učenju

obstaja statistično značilna razlika med eksperimentalno in programiranja. Dijaki dojemajo Codeium kot dragocen učni vir

kontrolno skupino. Povprečni čas dokončanja nalog je bil programiranja in so od njega odvisni pri vodenju svojega učnega

bistveno krajši v eksperimentalni skupini, kar podpira Hipotezo procesa. Poleg tega dijaki izkazujejo določeno stopnjo zaupanja

1 (H1). v hiter odziv in natančnost Codeiuma, tako da se odločijo za

Število napak: p-vrednost je prav tako bistveno manjša od njegovo uporabo po neposrednem odpravljanju napak v kodi ali

0,05, kar kaže na statistično značilno razliko med skupinama. branju sporočil o napakah s konzole. Zato menimo, da je treba

Eksperimentalna skupina je imela bistveno manj napak v upoštevati naslednje vidike, ko dijaki uporabljajo podporne

primerjavi s kontrolno skupino, kar podpira Hipotezo 2 (H2). asistente za učenje programiranja. Prvič, v začetnih fazah

Rezultati so pokazali, da je eksperimentalna skupina naloge programiranja lahko dijaki v celoti izkoristijo pomoč podpornih

zaključila v povprečju 23% hitreje kot kontrolna skupina. Poleg asistentov, da poglobijo svoje razumevanje težav pri

tega so dijaki v eksperimentalni skupini naredili 28% manj napak programiranju in najdejo ustrezne rešitve. Drugič, dijaki bi

v primerjavi s kontrolno skupino. To kaže, da je uporaba morali postopoma povečevati pogostost neodvisnega urejanja

Codeiuma prispevala k učinkovitejšemu in natančnejšemu kode, da izboljšajo svoje sposobnosti programiranja in

pisanju programske kode. strokovnosti, ko učenje napreduje. Nazadnje, pomembna je

previdnost, da ne postanejo preveč odvisni od podpornih

3.4 Kvalitativni rezultati asistentov. Dijaki začetniki bi morali posvetiti več časa osvajanju

Anketa med dijaki je pokazala, da so podporni asistent ocenili osnovnih konceptov programiranja, preden začnejo uporabljati

kot zelo uporabnega. 85% dijakov v eksperimentalni skupini je orodja umetne inteligence, ne glede na njihovo splošno raven

izrazilo zadovoljstvo z orodjem in menilo, da jim je omogočilo uspešnosti. Podporne asistente je treba pravilno uporabljati kot

boljše razumevanje osnovnih konceptov programiranja. Dijaki pomožno orodje za učenje in reševanje problemov programiranja.

so izpostavili tudi, da je interaktivna narava orodja izboljšala



njihovo izkušnjo učenja in zmanjšala stres, povezan z učenjem 5 ZAKLJUČEK programiranja.

Vpeljava pomoči orodij generativne umetne inteligence v

poučevanje osnov programiranja učinkoviteje razjasni

4 RAZPRAVA nesporazume kot statični učbeniki ali različni podporni videi.

Naši rezultati potrjujejo, da je uporaba orodij, kot je Codeium, Codeium je odličen pri zagotavljanju prilagojenih in

lahko zelo koristna pri učenju osnov programiranja. Codeium potrpežljivih razlag programskih konceptov, ponujanju primerov

omogoča, da se dijaki osredotočijo na reševanje problemov, kode in rešitev, prilagojenih individualnim potrebam dijakov.

medtem ko sintaktične in strukturne podrobnosti prepustijo Njegov interaktivni pogovorni vmesnik spodbuja dijake k

orodju, kar zmanjša frustracije in poveča motivacijo za učenje. dialogu, utrjuje njihovo razumevanje z aktivnim sodelovanjem in

Pri nalogah osnovne zahtevnosti so dijaki naloge opravili hitreje povratnimi informacijami. Dijaki lahko težave, ki nastanejo pri

in z manj napakami, kar kaže na to, da Codeium učinkovito kodiranju, predstavijo na preprost način, hkrati pa strani

podpira razumevanje osnovnih konceptov. Pri nalogah srednje Codeiuma prejmejo izčrpno razlago po korakih, ki razjasni

zahtevnosti je Codeium pomagal dijakom pri reševanju temeljna načela skozi celoten učni proces. Takšna dinamična

kompleksnejših problemov, kar je privedlo do boljših rezultatov pomoč ponuja ključno podporo in premošča vrzeli, ko učetelj ni

v primerjavi s tradicionalnimi metodami učenja. Pri naprednejših na voljo. Uporaba Codeiuma lahko okrepi sposobnosti dijakov

nalogah, kjer so se dijaki soočali z najzahtevnejšimi nalogami, je za oblikovanje preprostih algoritmov in pisanje kode, kar

Codeium olajšal uporabo sofisticiranih tehnik in omogočil zmanjša stres, ki pogosto spremlja seznanjanje z uvodnimi

hitrejše dokončanje nalog z višjo kakovostjo kode. pristopi reševanja izzivov z besedilnim programiranjem.

Poleg tega, da je uporaba Codeiuma pozitivno vplivala na Nenazadnje se je opazilo povečano navdušenje nad učenjem

uspešnost dijakov pri reševanju nalog, je hkrati učiteljem osnov programiranja med številnimi dijaki po izpostavitvi

omogočala, da so se osredotočili na mentorstvo in usmerjanje poučevanju s Codeiumum. Kot učitelji moramo biti na tekočem

dijakov pri razvoju kreativnih in inovativnih rešitev. z najnovejšim tehnološkim razvojem in še naprej prilagajati

Kljub pozitivnim rezultatom ima naša raziskava nekaj svoje metode poučevanja, da bi naše dijake čim bolje pripravili

omejitev. Izvedena je bila na relativno majhnem vzorcu dijakov, na svet, ki je pred nami.

kar omejuje možnost posploševanja rezultatov. Poleg tega Ne glede na opisano je vključevanje generativne umetne

dolgoročni učinki uporabe Codeiuma na razvoj programskih inteligence v učni proces programiranja s pomočjo orodja

Codeium pokazala pozitivne učinke na uspešnost in motivacijo



dijakov. Z uporabo naprednih funkcionalnosti, kot so

samodokončanje in generiranje kode, Codeium zmanjšuje

obremenitve, povezane z učenjem programiranja, ter omogoča

bolj osredotočeno in ustvarjalno učenje. Prihodnje raziskave bi

morale raziskati dolgoročni vpliv takšnih orodij na razvoj

programskih veščin in njihovo učinkovitost v različnih

izobraževalnih okoljih.



LITERATURA IN VIRI

[1] K. Benda, A. Bruckman, and M. Guzdial, “When life and learning do not

fit: Challenges of workload and communication in introductory computer

science online,” ACM Transactions on Computing Education (TOCE),

vol. 12, pp. 1–38, 2012.

[2] C. Gordon, R. Lysecky, and F. Vahid, “Programming learners struggle as

much in Python as in C++or Java,” in 2022 ASEE annual conference &

exposition, 2022.

[3] H. Mozannar, G. Bansal, A. Fourney, and E. Horvitz, “Reading between

the lines: Modeling user behavior and costs in AI-assisted programming,”

in Proceedings of the CHI Conference on Human Factors in Computing

Systems, 2024, pp. 1–16.

[4] Codeium: The Ultimate AI Coding Assistant, [Online]. Dostopno prek:

https://www.romainberg.com/codeium-your-ultimate-ai-coding-assistant/

[Dostop: 16. 8. 2024].

[5] E. Chen, R. Huang, H.-S. Chen, Y.-H. Tseng, and L.-Y. Li, “GPTutor: a

ChatGPT-powered programming tool for code explanation,” 2023.



Uporaba spletnega orodja H5P za formativno spremljanje



poučevanja računalniškega mišljenja



Using the H5P Online Tool for Formative Assessment of



Computational Thinking Learning



Davorin Babič

Srednja medijska in grafična šola

Ljubljana

Pokopališka ulica 33, 1000

Ljubljana davorin.babic@smgs.si



dijakov razvoj in na sistematičen razvoj znanja in veščin na

POVZETEK njegovem profesionalnem področju.

Izhodišča: Vključevanje digitalnih orodij v izobraževanje je

spremenilo učni proces, zlasti na področjih, povezanih z KLJUČNE BESEDE

računalništvom, kjer je računalniško mišljenje temeljna veščina. H5P, formativno spremljanje, računalniško mišljenje,

Pogosto uporabljen programski jezik, python, je izbran za digitalne kompetence, spletna učna orodja

uvajanje dijakov v računalniško mišljenje zaradi svoje

preprostosti in vsestranskosti. Vendar pa učinkovito poučevanje ABSTRACT

računalniškega mišljenja zahteva ne le poučne vsebine, temveč Background: The integration of digital tools into education

tudi aktivno vključevanje in neprekinjeno spremljanje has changed the learning process, especially in fields related to

razumevanja dijakov. computer science, where computational thinking is a

Namen: Namen raziskave je ugotoviti, kako uporaba spletnih fundamental skill. A commonly used programming language,

orodij za formativno spremljanje učnega procesa, zlasti python, is chosen to introduce students to computational thinking

platforme H5P, vpliva na poučevanje računalniškega mišljenja because of its simplicity and versatility. However, effective

pri programiranju v pythonu. Raziskava ocenjuje učinkovitost teaching of computational thinking requires not only

H5P kot orodja za ustvarjanje interaktivnih in formativnih instructional content, but also active involvement and continuous

spremljevalnih dejavnosti, ki vključujejo dijake in izboljšujejo assessment of students' understanding.

njihovo razumevanje računalniškega mišljenja. Purpose: The purpose of the research is to determine how the

Metode: V raziskavo je bilo vključenih 140 dijakov tretjega use of online tools for formative assessment of the learning

letnika medijske produkcije iz programa medijski tehnik, ki so process, especially the H5P platform, affects the teaching of

sodelovali pri ocenjevanju vpliva formativnih dokazov z computational thinking in python programming. The research

orodjem H5P na njihove učne rezultate. Uporabljen je bil anketni evaluates the effectiveness of H5P as a tool for creating

vprašalnik, s katerim smo želeli ugotoviti, v kolikšni meri orodja interactive and formative follow-up activities that engage

H5P prispevajo k izboljšanju razumevanja in ohranjanja veščin students and improve their understanding of computational

računalniškega mišljenja pri dijakih pri soočanju s praktičnimi thinking.

problemi. Dijaki so z vprašalnikom ocenjevali povratne Methods: 140 third-year media production students from the

informacije o svojih izkušnjah in zaznavah glede formativnih media technician program were included in the research, who

dokazov, ki temeljijo na virih orodja H5P. Podatke smo obdelali participated in evaluating the impact of formative evidence using

s pomočjo spletne programske oprema za statistiko DATAtab, ki the H5P tool on their learning outcomes. A survey questionnaire

se izvaja v oknu brskalnika, s pomočjo standardnih statističnih was used, with which we wanted to determine to what extent the

metod smo izračunali povprečne vrednosti in standardni odklon. H5P tools contribute to improving students' understanding and

Rezultati in diskusija: Rezultati so pokazali, da je H5P retention of computational thinking skills when dealing with

učinkovito orodje za formativno spremljanje veščin practical problems. Students used a questionnaire to assess

računalniškega mišljenja pri programiranju v pythonu. feedback on their experiences and perceptions of formative

Pomembno izboljšanje dokazov o znanju dijakov po ocenah evidence based on H5P resources. We processed the data with

poudarja vrednost interaktivnih in takojšnjih mehanizmov the help of the DATAtab online statistics software, which is run

povratnih informacij, ki jih zagotavlja H5P. Hkrati pa rezultati in a browser window, and with the help of standard statistical

ankete kažejo tudi na visoko zadovoljstvo dijakov z nalogami methods, we calculated the average values and the standard

H5P, kar podpira njegovo nadaljnjo uporabo v izobraževalnih deviation.

okoljih. Vse to dokazuje, da z vzpodbujanjem in poučevanjem Results and discussion: The results showed that H5P is an

veščin računalniškega mišljenja učitelji lahko vplivajo na effective tool for formative assessment of computational

thinking skills in python programming. The significant



improvement in evidence of student knowledge across poročali o višji motivaciji in boljši uspešnosti. Izobraževalne

assessments highlights the value of the interactive and immediate digitalne tehnologije lahko omogočijo neposredno spremljanje

feedback mechanisms provided by H5P. At the same time, the dijakovega napredka, izmenjavo povratnih informacij ter

results of the survey also show high student satisfaction with the izobraževalcem zagotovijo načine za vrednotenje in prilagajanje

H5P tasks, which supports its further use in educational njihovih strategij poučevanja. V množici digitalnih orodij, ki so

environments. All this proves that by encouraging and teaching na voljo učitelju, je pomembno, da se pred uporabo, med njo in

computational thinking skills, teachers can influence the po njej učitelj nasloni na kriterije za vrednotenje digitalnih orodij

student's development and the systematic development of glede na njihovo funkcionalnost, prilagodljivost, praktičnost in

knowledge and skills in his professional field. uporabnost. Glede na izobraževalne koristi digitalne tehnologije

lahko to organiziramo v tri skupine [4]:

KEYWORDS

H5P, formative assessment, computational thinking, digital 1) tehnologije kot učno orodje, kar vključuje pregled

competences, online learning tools uporabe tehnologije v razredu – strojna oprema,

programska oprema,

2) računalniško podprto poučevanje, prilagojeni učni

1 UVOD programi, internet.

3) Tehnologije za zagotavljanje učenja, kar vključuje

1.1 Formativno spremljanje učnega procesa tehnologije izobraževanja na daljavo, kot so spletno

Formativno spremljanje je bistvenega pomena za izboljšanje učenje, mobilno učenje, kot tudi podporne tehnologije.

poučevanja in učenja, sodobni izobraževalni digitalni viri 4) Tehnologije za podporo učenju, kar vključuje

(različna digitalna učna okolja, aplikacije in platforme ter orodja razvijajoče se podpore za tehnologijo kot učno orodje in

vizualizacij) pa zagotavljajo velik potencial za njegovo zasnovo za zagotavljanje učenja – odprti izobraževalni viri, veliki

in izvajanje. Sodobni digitalni izobraževalni viri so večinoma podatki, učna analitika in umetna inteligenca.

interaktivni, kar omogoča viru, da se odzove na dejanja dijakov Nabor učnih strategij, ki jih uporabljamo učitelji za podporo

bodisi z izbiro pravilne učne poti, zagotavljanjem povratnih vključevanja sodobnih izobraževalnih digitalnih virov v učni

informacij dijakom ali izvajanjem dejanj, ki jih običajno izvaja proces, je v veliki meri bolj osredotočen na dijake, ki uporabljajo

učitelj. Poleg tega, da so interaktivni, izobraževalni viri pogosto računalnik kot učno orodje, učitelj je v vlogi moderatorja, ki

vključujejo različne vrste elementov z namenom, da dijakom podpira projektno učenje in neodvisno raziskovanje.

zagotovijo optimalno učno izkušnjo. Vključiti je mogoče

videoposnetke, zvočne posnetke, kvize (ki dijakom omogočijo, 1.2 Spletno okolje H5P

da vnesejo svoje odgovore) – vendar izobraževalni vir lahko Za ustvarjanje strokovnega izobraževalnega vira smo

vključuje tudi veliko različnih vrst elementov hkrati. Z uporabili specializirano spletno platformo, ki nam je zagotovila

omenjenimi viri zagotavljamo dragocene povratne informacije vsa potrebna orodja za ustvarjanje, upravljanje in skupno rabo

ne samo učiteljem, ampak tudi dijakom, predvsem v usmerjanju lastnih izobraževalnih virov. H5P (HTML5 Package) [5] je

izboljšav pri poučevanju in učenju za optimizacijo učnih brezplačno in odprtokodno ogrodje za sodelovanje pri vsebinah,

rezultatov dijakov. Praksa v razredu je formativna do te mere, da katerega namen je vsem olajšati ustvarjanje, skupno rabo in

dokaze o dosežkih dijakov pridobijo, interpretirajo in uporabljajo ponovno uporabo interaktivnih vsebin. Še posebej je dragoceno

učitelji, dijaki in njihovi vrstniki. Ti deležniki kasneje sprejemajo za formativno spremljanje, saj zagotavlja takojšnje povratne

odločitve o svojih naslednjih korakih pri poučevanju in učenju, informacije, ki dijakom omogočajo, da razmislijo o svojem

ki bodo verjetno boljše ali bolje utemeljene od odločitev, ki bi jih učenju in prepoznajo področja za izboljšave [6]. Lahko ga

sprejeli, če ne bi bilo pridobljenih dokazov. Povratne informacije implementirano v WordPress in premore veliko zanimivih

so ključna sestavina v procesu učenja dijakov. Ker je napredek v funkcij, vključno z interaktivnimi videoposnetki (Interactive

tehnologiji omogočil sprejetje digitalnih učnih okolij z Videos), predstavitvami tečajev (Course Presentations),

možnostmi ocenjevanja, so se postopoma spremenili tudi scenarijem razvejanja (Branching Scenario), povleci in spusti

pogostost, oblika dostave in pravočasnost povratnih informacij, (Drag and Drop), pogovornimi karticami (Dialog Cards),

ki izhajajo iz pridobljenih informacij o (pred)znanju. Povratna navideznim ogledom (Virtual Tour 360), slikovnimi vročimi

informacija je eden najvplivnejših delov učnega procesa [1] in je točkami (Hotspots), bliskovnimi karticami (Flashcards), povleci

opredeljena kot: informacije, ki jih zagotovi posrednik (npr. besede (Drag the Words), razvrsti odstavke (Sort the Paragraphs),

učitelj, vrstnik, knjiga, starš, jaz, izkušnje) v zvezi z vidiki časovnico (Timeline), nizom vprašanj (Question Set), vprašanji

posameznikove uspešnosti ali razumevanja. Po [1] je z več izbirami (Multiple Choice Questions), izpolnjevanjem

računalniško podprta povratna informacija učinkovita. V praznin (Fill the Blanks), True or False, zaporedje slik (Image

primerjalni študiji primera [2] so sodelujoči dijaki računalniško Sequencing), aritmetični kviz (Arithmetic Quiz), označevanje

podprto povratno informacijo ocenili višje kot tradicionalno besed (Mark the Words), narek (Dictation) in esej (Essay).

pisno povratno informacijo. Pomembni dejavniki pri tej Nastale interaktivne in adaptivne vsebine so vizualno privlačne

ugotovitvi so: jasna struktura povratnih informacij, njihova in dodatno motivirajo dijake za njihovo učenje. Spletišče

čitljivost, informacije o vrzelih v razumevanju dijakov in Wordpres pa omogoča delo na različnih napravah, od namiznih

prepoznavanje dobrih vidikov dijaških prispevkov. Poleg tega so računalnikov preko tabličnih računalnikov do pametnih

bile računalniško podprte povratne informacije dijakom telefonov. Izbrano orodje smo v pedagoški proces vključili za

zagotovljene hitreje kot tradicionalne povratne informacije. Tudi motivacijo za pripravo na utrjevanje znanja, za ponavljanja in

raziskava [3] je pokazala podobne rezultate, pri čemer so dijaki

utrjevanje znanja pred ocenjevanjem znanja in za sklepanje s praktičnim pristopom. Glede na omejen čas v učnem

samopreverjanje (Slika 1). procesu je izziv, kako dijake učinkovito naučiti smiselnega

nabora spretnosti, ki jim omogoča, da sami usmerjajo svoje



prihodnje učenje. Poleg tega, ker strokovni modul vključuje

sumativno ocenjevanje, potrebujemo druga sredstva za

spremljanje in merjenje učenja dijakov ter usmerjanje razvoja

samega modula. Te izzive obravnavamo tako, da združimo

kratka predavanja s kratkimi kvizi za formativno spremljanje in

učnemu procesu dodamo še eno učno dejavnost. Ugotavljamo,

da so v odsotnosti sumativnega ocenjevanja kratki, pogosti kvizi

s takojšnjimi povratnimi informacijami odlično orodje za

spremljanje učenja razreda kot celote. Dijaki poročajo, da so

kvizi, čeprav zahtevni, izboljšali njihovo razumevanje konceptov

programiranja, jih osveščali o morebitnih napakah in bili

zabavna učna izkušnja. Poleg tega rezultati iz tega prispevka

ponazarjajo, kako je mogoče novega programskega jezika učiti

Slika 1: Primerek spletišča v Wordpressu na Arnesovem dijake brez predhodnega znanja programiranja v kratkem času.

strežniku splet.arnes.si, znotraj katerega so bili dijaki vpeti V raziskavi je sodelovalo 140 dijakov, ki so imeli različne ravni

v formativno spremljanje učnega procesa predhodnih izkušenj s programiranjem in veščinami

računalniškega mišljenja. Povzemamo naše izkušnje, pridobljene

1.3 O računalniškem mišljenju pri oblikovanju in vključevanju interaktivnih virov v spremljanje

Bistveni del v poučevanju računalniškega mišljenja je učnega procesa programiranja.

osnovno razumevanje programiranja. Računalniško mišljenje pri

poučevanju opisuje nabor orodij metod, ki jih je mogoče 2.1 Oblikovanje nalog za formativno

uporabiti za oblikovanje, analizo in obdelavo problemov/izzivov, spremljanje učnega procesa

tako da jih lahko reši algoritemska logika. Temelji na miselnih Da bi pridobili ustrezne dokaze o dijakovem razumevanju

procesih in konceptih, kot so modeliranje problemov/izzivov z veščin računalniškega mišljenja, smo zasnovali nabor nalog s

uporabo abstrakcije (pomaga razumeti in oceniti zahtevne in pomočjo interaktivnega ustvarjalca vsebine H5P v platformi

zapletene probleme), dekompozicije (razdelitev problemov na WordPress. Te naloge vključujejo naslednje veščine

podprobleme), načrtovanje rešitev skozi zaporedne korake računalniškega mišljenja: dekompozicija, prepoznavanje

(logično in algoritmično mišljenje) in prepoznavanje vzorcev. V vzorcev, abstrakcija, algoritmi in odpravljanje napak. Naloge so

zadnjem obdobju je računalniško mišljenje postalo orodje za bile oblikovane z uporabo različnih virov z različnimi

reševanje problemov na skoraj vseh strokovnih področjih značilnostmi in strukturami ter didaktičnimi priporočili, kot je

srednješolskega izobraževanja. Poučevanje in učenje veščin podrobno opisano spodaj: računalniškega mišljenja tako postane ključnega pomena ne

samo za inženirje in računalnikarje, temveč tudi za dijake na Prva naloga: Dekompozicija z interaktivnim videom

področjih zunaj pretežno naravoslovno-tehniških ved, med Orodje H5P: Interaktivni video vmesnik (Interactive Video)

katere sodi tudi medijska produkcija, kjer je delovno-ustvarjalni Opis: Dijaki si ogledajo videoposnetek, ki razlaga

tok močno odvisen od omenjenih veščin. kompleksen problem, in se morajo ustaviti na določenih točkah,

da problem razdelijo na manjše dele.

1.4 Raziskovalna vprašanja Didaktično priporočilo: Dijake spodbujamo, da razpravljajo o



naslednja raziskovalna vprašanja: V opravljeni raziskavi smo si prizadevali odgovoriti na razčlenjenih delih v parih, hkrati pa jih spodbujamo za sodelovalno reševanje problemov.



 Druga naloga: Prepoznavanje vzorcev s funkcijo povleci in Kako učinkovita so orodja H5P pri izboljšanju dijakovega spusti razumevanja veščin računalniškega mišljenja pri Orodje H5P: Povleci in spusti (Drag and Drop) programiranju v pythonu? Opis: Dijakom je predstavljena vrsta izrezkov programske  Katera so najučinkovitejša orodja H5P za formativno kode in morajo prepoznati vzorce tako, da povlečejo in spustijo spremljanje poznavanja veščin računalniškega mišljenja? podobne strukture kode v kategorizirana polja.  Kako dijaki dojemajo uporabo formativnega spremljanja, ki Didaktično priporočilo: S to nalogo utrjujemo pomen temelji na H5P, v svojem učnem procesu? prepoznavanja vzorcev v kodi, ki je ključnega pomena za  Kakšen vpliv imajo formativni dokazi uporabe H5P na optimizacijo algoritmov. angažiranost dijakov za učenje in dosežene učne rezultate?

Tretja naloga: Abstrakcija s kvizom z več možnimi

odgovori

2 METODA DELA Orodje H5P: Kviz z možnimi izbirami (Multiple Choice

Poučevanja programiranja in z njim povezanimi veščinami Quiz)

računalniškega mišljenja postopoma postaja bistvena veščina na Opis: Dijaki odgovarjajo na vprašanja z več možnimi

različnih strokovnih področjih medijske produkcije in je odgovori, ki od njih zahtevajo, da prepoznajo temeljna načela za

pomemben del v odprtem kurikulu, ki zajema temeljno znanje in

kompleksnimi problemi, pri čemer učinkovito izvajajo ali majhnih skupinah, s čimer spodbujamo spretnosti

abstrakcijo. sodelovalnega reševanja problemov), ponudili takojšnje

Didaktično priporočilo: Zagotavljamo podrobne povratne povratne informacije (uporaba funkcije povratnih informacij

informacije za vsako možnost, da dijakom pomagamo razumeti, platforme H5P, da dijakom zagotovimo takojšnje odzive na

zakaj so določene abstrakcije ustrezne ali neustrezne. njihova dejanja in jim pomagamo, da se učijo iz napak v realnem

času) in poskrbeli za različnost nalog (prilagodimo kompleksnost

Četrta naloga: Algoritmi z izzivom interaktivnega

nalog glede na raven znanja dijakov, da ohranimo angažiranost

kodiranja

in izziv).

Orodje H5P: Razvrsti odstavke (Sort the Paragraphs)



Opis: Dijaki razvrstijo po ustreznem zaporedju premetane 2.2 Izvedba in zbiranje podatkov

stavke pythonovskega programa za implementacijo algoritma za

rešitev danega problema (Slika 2). Naloge so bile izvedene znotraj strokovnega modula

Didaktično priporočilo: Dovolimo večkratne poskuse in kreativno kodiranje s poudarkom na programiranju v pythonu.

zagotovimo namige, ki bodo dijake vodili k pravilni rešitvi, kar Po opravljenih nalogah so dijaki izpolnili anketo, da smo zbrali

bo okrepilo iterativno naravo kodiranja. povratne informacije o njihovih učnih izkušnjah in dojemanju

formativnega spremljanja, ki temelji na platformi H5P.





2.3 Oblikovanje ankete

Anketa je bila sestavljena iz kvantitativnih in kvalitativnih

vprašanj. Namen raziskave je bil izmeriti zaznano učinkovitost

orodij H5P, stopnjo angažiranosti in vpliv na njihove učne

rezultate. Anketna vprašanja so vključevala naslednje kazalnike:



 Kako bi ocenili interaktivnost nalog spletnega okolja H5P,

ki vam pomagajo razumeti veščine računalniškega

mišljenja?

Kazalniki: 1 (zelo nizko) do 5 (zelo visoko)

 Kateri vir H5P se vam je zdel najbolj učinkovit za učenje

konceptov programiranja v pythonu?

Kazalniki: več možnosti z možnostmi: interaktivni video,

povleci in spusti, kviz z več možnimi izbirami, vrsta vsebine

interaktivne kode, dostopna točka

 Kako so takojšnje povratne informacije, ki so jih zagotovile

naloge, izdelane v H5P, vplivale na vaš učni proces?

Kazalniki: 1 (sploh niso) do 5 (pomembno)

 Ali menite, da je uporaba orodij H5P izboljšala vaše

sposobnosti reševanja problemov v Pythonu?

Kazalnika: Da/Ne

 Kateri vidiki nalog, izdelanih v H5P, so se vam zdeli najbolj

koristni ali zahtevni?

 Kazalnik: Odgovori esejskega tipa



2.4 Statistična analiza

Podatki, zbrani v raziskavi, so bili analizirani s statističnimi

Slika 2: Primer naloge z orodjem H5P, razvrsti odstavke, ki metodami z ustrezno programsko opremo DATAtab

od dijaka zahteva razvrstitev premešanega zaporedja (https://datatab.net) [8, 9] za oceno učinkovitosti orodij H5P.

stavkov v logično in algoritmično ustrezno zaporedje za Rezultati ankete so bili analizirani s t-testom, da bi ugotovili, ali

pravilno izvedeno rešitev. obstaja statistično pomembna razlika v zaznavah dijakov pred in



Naloga 5: Odpravljanje napak z vročimi točkami Glede na odgovore ankete so bile testirane naslednje po uporabi orodij H5P za formativno spremljanje.

Orodje H5P: Slikovne vroče točke (Hotspots) hipoteze:

Opis: Dijaki prepoznajo in popravijo napake v ponujenem Ničelna hipoteza (H0): uporaba orodij H5P ne vpliva bistveno

delčku pythonove kode s klikom na napačne dele. na angažiranost dijakov in razumevanje veščin računalniškega

Didaktično priporočilo: S to nalogo poudarimo pomen mišljenja.

odpravljanja napak kot kritičnega koraka v procesu razvoja Alternativna hipoteza (H1): uporaba orodij H5P bistveno

programske opreme. izboljša sodelovanje dijakov in razumevanje veščin

računalniškega mišljenja.

Pri izvajanju nalog z orodji H5P smo učitelji zagotovili jasna Uporabljena formula t-testa je bila:

navodila (da dijaki razumejo cilje vsake naloge in kako ravnati z

orodji H5P), spodbujali sodelovalno učenje (dijaki delajo v parih



𝑋 ̅̅̅ − 𝑋 ̅̅̅ takojšnjo povratno informacijo spodbuja globlje razumevanje in 1 2



𝑡 = večjo angažiranost dijakov. Vsestranskost orodij H5P omogoča 2 2 𝑆 1 𝑆 2 učiteljem, da oblikujejo raznolike in prilagodljive naloge, ki √ + 𝑛 1 𝑛 2 lahko poskrbijo za različne učne stile in ravni znanja.

kjer so so ̅̅̅ 𝑋 1 in ̅̅̅ 𝑋 2 aritmetični sredini obeh skupin (pred in po

formativnem spremljanju), 2 2 𝑆 in 𝑆 varianci, 𝑛 in 𝑛 pa

velikosti vzorcev (v našem primeru kar celotno število dijakov). 1 2 1 2 5 ZAKLJUČEK

Opisna statistika je bila uporabljena za povzetek odgovorov Formativno ocenjevanje je bistvenega pomena pri

na anketo, analiza s t-testom pa je bila izvedena za identifikacijo izobraževanju programiranja za merjenje razumevanja dijakov in

odnosov med stopnjami angažiranosti dijakov in njihovimi njihovo vodenje skozi učni proces. V prispevku smo raziskvali

učnimi rezultati [7]. uporabo veščin računalniškega mišljenja skozi osnovne koncepte

in konstrukte v programiranju – kot so zaporedja, spremenljivke,

tipi podatkov, vejitve, zanke in funkcije – v pythonu za

3 REZULTATI ustvarjanje učinkovitih formativnih dokazov. Z vključitvijo teh



3.1 Opisna statistika razumevanju posredovanega in naučenega ter oblikujejo temeljnih elementov lahko učitelji spremljajo napredek v

Dobljeni rezultati so pokazali visoko stopnjo angažiranosti smiselne povratne informacije, ki ne samo ocenjujejo znanja,

dijakov pri nalogah, izdelanih z orodji H5P. Večina dijakov je ampak tudi krepijo sposobnosti reševanja problemov in logično

ocenila interaktivnost nalog kot "visoko" ali "zelo visoko", s mišljnje dijakov. Metodologije, obravnavane tukaj, zagotavljajo

povprečno oceno 4,3 na 5-stopenjski lestvici kazalnikov. Orodja praktične primere in strategije platforme H5P kot učinkovitega

Interactive Code Content Type in Drag and Drop so bila orodja za vključevanje formativnega spremljanja v učni proces

prepoznana kot najučinkovitejša za učenje konceptov poučevanja osnov programiranja v pythonu. Pozitiven vpliv na

programiranja v pythonu, pri čemer je 65 % dijakov dalo sodelovanje in razumevanje dijakov poudarja pomen

prednost tem orodjem. vključevanja takih interaktivnih orodij v kurikul za izboljšanje

digitalne kompetence.

3.2 Analiza s t-testom

t-test je pokazal znatno izboljšanje angažiranosti in VIRI IN LITERATURA

razumevanja dijakov po uporabi orodij H5P za formativno [1] J. Hattie and H. Timperley, “The power of feedback,” Rev. Educ. Res.,

spremljanje. Povprečne ocene za angažiranost so se povečale s vol. 77, no. 1, pp. 81–112, 2007.

[2] P. Denton, J. Madden, M. Roberts, and P. Rowe, “Students’ response to

3,1 na 4,2 (p < 0,05), za razumevanje pa z 2,9 na 4,0 (p < 0,05). traditional and computer‐assisted formative feedback: A comparative case

Ti rezultati kažejo, da so orodja H5P učinkovita pri izboljšanju study: Computer-assisted formative feedback,” Br. J. Educ. Technol., vol.

39, no. 3, pp. 486–500, 2008.

učnih izkušenj dijakov pri računalniškem mišljenju. [3] J. Gillespie, K. Winn, M. Faber, and J. Hunt, “Implementation of a

mathematics formative assessment online tool before and during remote

3.3 learning,” in Artificial Intelligence in Education. Posters and Late Kakovostne povratne informacije

Breaking Results, Workshops and Tutorials, Industry and Innovation

Kakovostne povratne informacije so pokazale, da so dijaki Tracks, Practitioners’ and Doctoral Consortium, Cham: Springer

International Publishing, 2022, pp. 168–173.

cenili takojšnje povratne informacije, ki jih zagotavljajo naloge [4] F. A. Inan, D. L. Lowther, S. M. Ross, and D. Strahl, “Pattern of classroom



z orodji H5P, kar jim je pomagalo hitro prepoznati in popraviti activities during students’ use of computers: Relations between instructional strategies and computer applications,” Teach. Teach. Educ., svoje napake (Tabela 1). Vendar se je nekaterim dijakom zdelo, vol. 26, no. 3, pp. 540–546, 2010.

da je naloga odpravljanja napak zahtevna zaradi zapletenosti [5] “The basics,” H5p.org. [Online]. Available:

predloženih delčkov programske kode. https://h5p.org/documentation/for-authors/the-basics. [Dostopno: 18. 8.

2024].

[6] A. Vihavainen, M. Luukkainen, and J. Kurhila, “Multi-faceted support for

MOOC in programming,” in Proceedings of the 13th annual conference

Tabela 1: Ustreznost kakovostnih povratnih informacij on Information technology education, 2012.

[7] M. Guzdial, "Education: Paving the way for computational thinking,"

Communications of the ACM, vol. 51, no. 8, pp. 25-27, 2008.

Anketno [8] D. Team, Descriptive and inferential statistics, Descriptive and Inference Povprečna ocena Standardni odklon

vprašanje Statistics • Simply explained - DATAtab. 2023.

[9] D. Team, Level of measurement, Level of measurement • Simply

Interaktivnost 4,3 0,6 explained - DATAtab. 2023.

naloge

Učinkovitost 4,1 0,7

povratne

informacije

Izboljšanje učenja 4,2 0,5



4 RAZPRAVA

Ugotovitve kažejo, da so orodja spletne platforme H5P

dragocena za formativno spremljanje učnega procesa in

napčredka dijakov pri izobraževanju o računalniškem mišljenju.

Zdi se, da interaktivna narava nalog H5P v kombinaciji s





Utrjevanje snovi preko interaktivne sobe pobega




Consolidating Knowledge through an Interactive Escape Room



Simona Bežan

OŠ Komenda Moste

Glavarjeva cesta 37

Komenda, Slovenija

bezan.simona@oskomenda-moste.si



In this paper, I present how an interactive escape room was

POVZETEK used to consolidate knowledge about the history of the students'

V sodobni šolski situaciji se učitelji, učenci in drugi hometown. I created these escape rooms using Google Forms and

udeleženci učnega procesa srečujemo z vedno več ponudbami Deck Toys, where students could only "escape" by

različnih aplikacij, ki spodbujajo vključevanje učencev v učni demonstrating their knowledge of the local history they had

proces. learned during a guided walk and through the use of provided

V kolikšni meri bomo kot učitelji aplikacije vključili v pouk, literature.

je odvisno od naše pripravljenosti raziskovanja in preizkušanja

novosti. Velikokrat se pri uvajanju novih aplikacij pojavi, vsaj KEYWORDS

pri nas starejših učiteljih, strah pred neizkušenostjo, saj so učenci Information and communication technology, interactive

spretnejši pri uporabi le-teh. escape room, Komenda, collaborative learning, student

Učenci imajo radi sobe pobega, pri katerih je treba ob pomoči engagement

različnih ugank in namigov najti izhod iz zaprtega prostora v

omejenem času. Na takšen način namreč največkrat praznujejo

rojstne dneve. Zato sem tudi sama razmišljala, da bi ustrezno 1 UVOD

zasnovano sobo pobega lahko uporabila kot učno metodo pri Sodobne tehnologije in mediji se odražajo na vseh področjih

pouku. našega življenja. Razpoložljivost informacijskih in

V prispevku bom predstavila, kako je potekalo utrjevanje komunikacijskih tehnologij (IKT) je postala norma, ki kaže

snovi o zgodovini domačega kraja preko interaktivne sobe napredek in razvoj nekaterih družb. IKT se je iz udobja doma

pobega. Za učence sem v Google Forms in Deck Toys pripravila preselila tudi v šolske prostore, kjer spodbuja motivacijo

interaktivni sobi pobega. Iz sob so lahko pobegnili le tako, da so učiteljev, olajša dostop do informacij in podpira sodobne

pokazali znanje iz zgodovine domačega kraja, ki so ga dobili na pristope k poučevanju. IKT tudi olajša poučevanje otrok s

sprehodu po njem in s pomočjo pripravljene literature. posebnimi potrebami in učiteljem omogoča pristop k učencem z

različnimi učnimi stili. [1]

KLJUČNE BESEDE Večji del svojega učiteljskega dela poučujem v petem razredu.

Informacijsko-komunikacijska tehnologija, interaktivna soba Posebej se rada pripravljam na ure, ko pri predmetu družbe

pobega, Komenda, sodelovalno učenje, učenci spoznavamo Slovenijo in domači kraj. Čeprav imajo učenci pred

seboj zemljevid in pri urah ob razlagi uporabljam različne

ABSTRACT pripomočke, opažam, da jih še najbolj privlačijo virtualni.

In today’s educational environment, teachers, students, and



other participants in the learning process are increasingly 2 INFORMACIJSKO-KOMUNIKACIJSKA

presented with various applications designed to engage students TEHNOLOGIJA V ŠOLI more deeply. The extent to which we, as teachers, incorporate

these applications into our lessons depends largely on our »Učiti se« pomeni spoznavati nekaj novega, pridobivati novo

willingness to explore and experiment with new technologies. vedenje, nove izkušnje, razvijati nove spretnosti, odkrivati

For many older teachers, there may be some apprehension about neznana področja, razumeti sebe in svet, v katerem živimo, ter se

adopting new applications due to a lack of experience, especially v njem čim bolje znajti. Učimo in izobražujemo se tudi mi učitelji,

since students are often more adept at using them. še posebej tisti, ki smo starejši in ki nismo odraščali z digitalno

Students are particularly fond of escape rooms, where they tehnologijo, s katero se srečujemo zadnja leta.

solve puzzles and follow clues to find their way out of a confined In čeprav smo pri svojem delu sicer avtonomni, iščemo

space within a limited time. This activity is a popular way for različne metode in oblike dela, s katerimi bodo učenci prišli do

them to celebrate birthdays. Inspired by this concept, I znanja in osvojili cilje učnega načrta.

considered using a well-designed escape room as a teaching Ob različnih metodah in oblikah dela, ki jih vključujem v

method in the classroom. poučevanje, se trudim, da uporabljam tudi informacijsko-

komunikacijsko tehnologijo (IKT). Dnevno uporabljam že

izdelano interaktivno gradivo založb, pripravljam preproste le nekateri znali prepoznati in poimenovati posamezno

Power Point predstavitve, lotim pa se tudi kakšnega znamenitost, več pa ne.

interaktivnega kviza. Po skupnem obhodu in ustavljanju ob pomembnih objektih,

Hiter napredek in razvoj tehnologije nam starejšim učiteljem ob katerih povem o vsakem nekaj zanimivosti, spoznamo tudi

večkrat prinaša nelagodje. Tehnologijo lahko na ustrezen in pomembne može, ki so nekoč živeli v Komendi.

smiseln način vključim v pedagoški proces le, če poznam Seveda omenim tudi Tadeja Pogačarja, svetovno znanega

raznovrstno IKT in jo znam tudi učinkovito uporabljati. Znati kolesarja in pisatelja Ivana Sivca (ki sedaj sicer živi v Mengšu),

moram torej izbrati primerno metodo v okviru IKT na podlagi a je še vedno med prvimi najbolj branimi avtorji, posebej

vsebine, ki jo poučujem, in jo nato na ustrezen način vključim v mladinskih knjig.

pouk. Ustavimo se ob spomenikih, pomembnih zgradbah in kipih.

Zgodi se mi, da včasih dobim odpor do njene uporabe, saj mi Učni sprehod po Komendi opravim vsako šolsko leto, a so si

le-ta predstavlja izziv, ki ga nisem vedno pripravljena sprejeti. A tokrat učenci še z večjim zanimanjem zapisovali podatke, kot leta

podpora s strani mlajših kolegic mi da dodatno motivacijo. Nove poprej, saj sem jih predhodno spodbudila, da bomo po končanem

tehnologije v pouk tako uvajam postopoma, kot pripomoček za poučnem sprehodu v razredu snov ponovili in utrdili preko

učenje. zanimivega virtualnega popotovanja.

V prispevku bom predstavila, kako sem pri utrjevanju snovi Prejšnja šolska leta smo namreč po prihodu v razred izvedli

v razredu uporabila spletne obrazce Google Forms in program kviz, ki sem ga pripravila po vzoru oddaje oz. družabne igre Moja

Deck Toy. Z obema sem ustvarila interaktivno sobo pobega. Slovenija.

Priprava na tak pouk je za učitelja bolj zahteven, pri uri s

pripravljenim IKT gradivom pa je opazovalec in usmerjevalec 2.2 Soba pobega

dela, aktivni pa so učenci. Pomembno je tudi, da s takim delom Večkrat sem že poslušala učence, ki so v šoli z zanimanjem

spodbujamo, krepimo in ustvarjamo sodelovalno učenje pripovedovali, kako so praznovali svoj oz. rojstni dan prijatelja



2.1 Ura družbe Zakaj ne bi poskusili nekaj podobnega še v razredu? v sobi pobega.

Pri uri družbe z učenci raziskujemo tudi domači kraj in Zavedam se, da tehnologijo lahko na ustrezen in smiseln

obravnavamo njegovo zgodovino. Vsako šolsko leto znova način vključim v pedagoški proces le, če jo znam učinkovito

ugotavljam, da učenci svojega kraja in znamenitosti ne poznajo. uporabljati. Zato moram znati izbrati primerno metodo v okviru

Še dobro, da imamo kolesarja Tadeja Pogačarja, saj je učencem IKT na podlagi vsebine, ki jo poučujem, in jo nato na ustrezen

znan le roza oz. rumeno obarvan Pogačarjev rondo v Komendi. način vključiti v pouk.

Tako si v razredu najprej preko Power Point predstavitve Za to je potrebna volja in seveda pomoč mlajših kolegic.

ogledamo glavne znamenitosti Komende. V zvezek si zapišemo Prvo sobo pobega sem naredila v Google Forms. Za delo v

miselni vzorec (izdelan v Canvi – oblikovale smo ga skupaj v tem okolju je potreben le Google račun, ki je za uporabnike

aktivu). brezplačen. Učenci za reševanje ne potrebujejo računa in se jim

ni potrebno prijavljati. To orodje mi je všeč, ker lahko k



vprašanjem dodamo slike, posnetke in tako naredimo reševanje

preglednejše, nazornejše in zanimivejše.

V Google Forms sem sestavila vprašanja na temo zgodovine

domačega kraja. Vsako vprašanje je na nek način zaklenjeno in

učenci lahko z naslednjim nadaljujejo šele, ko pravilno

odgovorijo na trenutno vprašanje. To se mi zdi koristno, saj so

učenci tako primorani priti do pravega odgovora, sicer ne morejo

nadaljevati.

Stvar je še nekoliko otežena, saj morajo pri odgovorih paziti

tudi na slovnični zapis besede, sicer program odgovora ne zazna

kot pravilnega. Na to sem jih opozorila tudi na prvi strani, kjer

sem jim napisala navodila za reševanje nalog.

Ta način mi je všeč tudi zato, ker lahko za učence napišem

tudi namig, če napišejo napačen odgovor in jih tako usmerim na

Slika 1. Miselni vzorec o Komendi (VIR: lasten, zajem pravo pot. Nekatere skupine so imele težave z vprašanjem, v

zaslonske slike) katerem stoletju se je rodil Glavar, saj so pozabili, kako se določa

stoletje. V namigu sem jih spomnila na način določanja stoletja.

Naslednjo uro gremo skupaj z učenci na sprehod po Komendi Drugo sobo pobega sem ustvarila s programom Deck Toys.

in raziskujemo njeno zgodovino. Že pred sprehodom jih Tudi pri tem reševanju učenci ne morejo nadaljevati z delom,

spodbudim, naj pozorno poslušajo, mogoče tudi kaj zabeležijo v dokler pravilno ne opravijo naloge. Program je deloma

beležke, ki jih nesejo s seboj, saj jih bo na povratku v šoli čakalo brezplačen, za dodatne možnosti pa je potrebno doplačilo.

razburljivo virtualno potovanje. Menim, da je za učitelje osnovna verzija za popestritev pouka

Predvidevala sem, da so bodo učenci zapomnili nekatere od povsem zadostna.

znamenitosti, o katerih smo se pogovarjali predhodno uro, zato Ta program je na začetku nekoliko težji in zamudnejši, ko pa

jih poskušam na sprehodu čim bolj vključiti. Žal ugotovim, da so ga enkrat usvojiš, ti pomaga ustvariti zanimive naloge za učence.





Slika 4. Začetna stran sobe pobega (VIR: lasten, zajem

Slika 2. Primer navodil za učence (VIR: lasten, zajem zaslonske slike)

zaslonske slike)





Slika 5. Primer naloge (VIR: lasten, zajem zaslonske slike)





Ta program ima slabost, saj ne upošteva šumnikov, kar zna



biti pri reševanju nekoliko moteče. Na to sem učence opozorila

na začetni strani z navodili. Zdi se mi prav in koristno, da se

učenci naučijo natančno brati navodila, kar opažam, da v zadnjih



Slika 3. Namig pri enem izmed vprašanj (VIR: lasten, zajem zaslonske slike) letih poučevanja to postaja vedno pogostejša težava.



Omogoča ustvarjanje različnih tipov nalog: križanke, spomin, 2.3 Delo v razredu

sestavljanke, naloge povezovanja … Z različnimi tipi nalog tako Z učenci smo po prihodu v učilnico najprej pogovorili o sobi

lahko približamo učenje različnim tipom učencev. Z reševanjem pobega. Veliko jih je sobo pobega že obiskalo in so navdušeno

teh nalog so tako utrjevali snov in hkrati ugotavljali, koliko pripovedovali, kako so morali ob pomoči različnih ugank in

znanja so odnesli od pouka. namigov najti izhod iz zaprtega prostora v omejenem času.

Povem jim, da bomo v nadaljevanju ure preko virtualne sobe

pobega ponovili, kar smo na učnem sprehodu po Komendi videli

in slišali.

Učence sem razdelila v heterogene skupine. Upoštevala sem

znanje učenca, sposobnosti in socialno spretnost ter jih

spodbudila k sodelovalnemu učenju. Sodelovalno učenje je

oblika učenja v majhnih skupinah, v katerih učenci skušajo

doseči zastavljene učne cilje. Temelji na njihovi medsebojni



soodvisnosti in odgovornosti, pomembno pa je, da vsak doda Otroci, ki v teh časih hodijo v šolo, so ves čas vpeti v sodobno

svoj prispevek k skupni nalogi. tehnologijo, zato jim vsekakor moramo priti naproti in

Učenci se tudi bolje spoznajo med seboj, razvijajo skupinsko posodobiti način poučevanja.

identiteto, doživijo medsebojno podporo. Naučijo se spoštovati Kot učitelji moramo najti pravo ravnovesje med uporabo IKT

razlike med člani skupine in razvijajo sinergijo v skupini. [2] in uporabo drugih metod, kar pa pogosto ni prav lahko delo.

Skupinam sem razdelila tablice. Povezavo do obeh virtualnih IKT ne bo nadomestila učitelja, saj se ni sposobna vživeti in

sob sem jim pripela v Microsoft Teams, kamor so vključeni vsi prilagoditi vzdušju v razredu, pa tudi odzivnosti na posamezne

učenci. S klikom na povezavo so se jim odprla vrata do virtualne učne strategije ji manjka, lahko pa dinamično in všečno obogati

sobe, iz katere so poskušali čim prej pobegniti. Najprej so se lotili učiteljevo razlago in ostale faze njegovega dela.

sobe pobega v MS Forms, nato še drugo, narejeno v Deck Toys.

S skupnimi močmi so učenci poskušali razvozlati vprašanje ZAHVALA

in se tako prebiti do naslednjega. Opazovala sem jih, kako so Hvaležna sem svoji sodelavki Marjeti Kuhar, ki prav tako kot

brskali po svojih zapiskih in pripravljeni literaturi. Ni jim bilo jaz poučuje v 5. razredu. S svojim znanjem na računalniškem

težko prebrati niti daljšega besedila, pomembno je bilo, da so področju ter stalno pripravljenostjo pomagati, me spodbuja k

našli pravi odgovor. preizkušanju in uporabi novih aplikacij.

Ko je vsem skupinam uspelo pobegniti iz sob, smo skupaj



opravili refleksijo njihovega dela. Učenci so svoje znanje LITERATURA

označili za uspešno, prav tako zmožnost pridobivanja informacij [1] DUH, Maruša. (2020). IKT v šoli in doma. Magistrsko delo. Maribor. iz pripravljenih gradiv. Objavljeno na spletnem mestu:

https://dk.um.si/IzpisGradiva.php?lang=slv&id=78153, dne 16. 7. 2024

[2] Kocjančič Kuhar, V. (2016). Sodelovalno učenje pri utrjevanju



3 matematike. Murska Sobota. Objavljeno na spletnem mestu: PREDNOSTI IN SLABOSTI

https://www.zrss.si/kupm2016/wp-content/uploads/sodelovalno-

Prednost je zagotovo sodelovalno učenje. V kvizu, ki sem ga ucenje.pdf, dne 16. 7. 2024

izvajala prejšnja leta, je bil razred razdeljen v dve skupini in

čeprav so bile naloge različnega tipa, vsi učenci v skupini niso

sodelovali. Tokrat so bili bolj motivirani za delo, samostojno so

raziskovali in brskali po pripravljeni literaturi.

Pozorno so morali prebrati vprašanja, sicer so bili zapisani

odgovori prepoznani kot nepravilni in niso mogli nadaljevati z

delom.

Učenci so bili motivirani za delo in številni so nalogo opravili

tudi doma. Tako so snov še enkrat ponovili in poskrbeli, da si

boljše zapomnijo podatke o zgodovini svojega domačega kraja.

Slabost je ta, da moram vnaprej pripraviti tablice, ki ne

delujejo vedno , kot bi si želeli. Največkrat nas razočarajo tik

pred uporabo. Da pričnejo pravilno delovati, je potreben dodaten

čas in če še nisi strokovnjak na tem področju, potrebuješ

dodatnega časa še več. Tudi tokrat je bilo tako. Dve tablici

namreč nista delovali.

Prav tako je slabost tudi ta, da sem morala oblikovati

vprašanja, na katera so učenci odgovarjali z eno besedo ali

besedno zvezo. Zato sem potem učencem postavljala še dodatna

ustna vprašanja, da kljub vsemu niso bili prikrajšani za kakšen

podatek.

Paziti so morali tako na pravopisne kot na tipkarske napake,

saj se je zgodilo, da so našli pravilni odgovor, a so se zatipkali in

zato niso mogli naprej. Zato so si vtipkane odgovore zapisovali

na list in če niso prišli do naslednjega vprašanja, so me poklicali,

da sem preverila njihov zapisani odgovor. Če je bil zapisan pravi

odgovor, sem jih opozorila na tipkarsko napako in spremljala

njihov vnos, da je bil pralni odgovor tudi ustrezno vpisan.



4 ZAKLJUČEK

Čeprav učencem tokrat ni bilo potrebno najti izhoda iz

zaprtega prostora v omejenem času, kot je to v fizični sobi

pobega, ki so jo obiskali že kdaj prej, so tudi pri reševanju iz

virtualne pokazali zanimanje in motivacijo za ponavljanje, kar pa

je bistvo za uspešno delo in učenje.





Pouk matematike z IKT




Mathematics Lessons with ICT



Blanka Bizjak

Ekonomska šola Ljubljana

Slovenija

blanka.bizjak@presernova.si



POVZETEK

V današnjem svetu je poznavanje informacijsko- 1 UVOD



komunikacijske tehnologije (IKT) in povezovanje z ostalimi Učitelji potrebujemo vedno več znanja in izkušenj, da bi bili

znanji nujno potrebno. sposobni razumeti in se prilagoditi vsakodnevnim situacijam, ki

Učenje s pomočjo sodobne tehnologije dijake motivira in jim se pojavljajo v učnem procesu.

omogoča hitrejše in kakovostnejše doseganje učnih ciljev. Ko Pri predmetu matematika je učiteljeva IKT pismenost

IKT povežemo še s sodelovalnim učenjem, dijaki pri tem pomemben dejavnik. Prav tako kot razumevanja določene teme

razvijajo tudi druge pomembne kompetence 21. stoletja. je treba dijake naučiti uporabljati tehnologijo pri reševanju

V srednji šoli ima risanje geometrijskih likov, teles in funkcij matematičnih problemov. Pri vsakodnevnem delu v razredu

velik pomen pri poučevanju matematike. Pomembno je, da opažamo, da so sicer digitalno pismeni, a je njihova sposobnost

dijaki razumejo, kako liki, telesa in funkcije nastanejo in kako se zbranosti in sprejemanja informacij zelo kratka, zato jih je treba

spreminja njihova velikost in oblika pri različnih spremenljivkah. ves čas motivirati.

V prispevku je predstavljen pomen in primeri uporabe IKT V nadaljevanju bo predstavljeno, kateri IKT se uporablja pri

pri pouku matematike. Predstavljeno je delo dijakov, ki so pouku matematike. Natančneje bodo prikazane dejavnosti, ki so

samostojno ali v skupinah reševali problemsko strukturirane dijake spodbudile k sodelovalnemu delu, reševanju problemov,

naloge, pri čemer so morali pridobljeno teoretično znanje raziskovanju in obenem k aktivni rabi IKT v vseh fazah učenja.

uporabiti, da so naloge najprej rešili, nato pa jih kritično preverili Namen prispevka je prikazati, kako se lahko s tehnologijo

in raziskovali s pomočjo uporabe različne IKT. izboljša poučevanje in razumevanje matematike.



KLJUČNE BESEDE

2 IKT PRI POUKU MATEMATIKE

IKT, matematika, skupinsko delo

Kompetentnost učitelja je »sposobnost učinkovitega



ABSTRACT delovanja v številnih situacijah, ki sicer temelji na pridobljenem znanju, vendar ni omejeno s tem znanjem« [1]. In today's world, the knowledge of information and Spoznanje, da strokovno znanje le delno zagotavlja uspešno communication technology (ICT) and the connection with other delo v šolstvu, privede učitelja v nenehno skrb za osebnostni skills is absolutely necessary. razvoj. In sicer tudi na področju poučevanja z uporabo Learning by means of modern technology motivates students sodobnega IKT. and enables them to achieve learning goals faster and with better Pri matematiki učitelji uporabljamo IKT z namenom quality. When we connect ICT with collaborative learning, izvajanja matematičnih postopkov in obravnave kompleksnejših students also develop other important 21st century competences. matematičnih problemov, hkrati pa se na ta način omogoča tudi In high school, drawing geometric shapes, solid figures and kompenzacijo primanjkljajev v znanju in sposobnostih dijakov. functions is of great importance in teaching mathematics. It is Z uporabo IKT učitelji s poudarkom na osmišljanju important for students to understand how geometric shapes, solid matematičnih pojmov matematiko približamo dijakom na figures and functions are created and how their size and shape zabavnejši način. change according to different variables. Pri pouku matematike IKT uporabljamo, da bi: The article presents the importance and the examples of the

use of ICT in mathematics lessons. We presented the work of  razvijali matematične pojme;

students who solved problem-structured tasks independently or  raziskovali in reševali probleme ter modelirali;

in groups, where they had to use the acquired theoretical  avtomatizirali določene postopke;

knowledge to first solve the tasks, and then critically check and  shranjevali, razvrščali, urejali, predstavljali

research them, using various ICTs. rezultate dela;

 preverjali znanje. [2]

KEYWORDS V nadaljevanju bodo predstavljeni primeri uporabe IKT pri

ICT, mathematics, group work pouku matematike.

2.1 Primer 1: MICROSOFTOVO ORODJE

ONENOTE

Interaktivna tabla omogoča, da se med urami uporablja

Microsoftovo orodje OneNote. V njem je za posamezen oddelek

oblikovan digitalni zvezek (slika 1), namenjen sprotnemu

sledenju pouka matematike. To orodje omogoča vnaprejšnjo

pripravo določenih gradiv, kar pripomore k večji strukturiranosti

pouka. Zapiske, ki nastajajo med poukom na koncu ure, je možno

deliti z dijaki. Prav tako se lahko pogleda katerakoli pretekla

snov oz. poljubna ura, s čimer se brez večjih težav ponovi snov,

na kateri temelji nadaljnja razlaga. Tak način poučevanja

omogoča, da je uporabljena tehnologija, ki je v danem trenutku

najbolj primerna, da bi dijaki lažje usvojili snov. Prednost

uporabe Onenote je tudi ta, da dijaki v enakem času rešijo več

nalog ter da jim nudi boljšo prostorsko predstavljivost snovi.

Vnaprej pripravljeno gradivo pomeni, da med poukom ostane

več časa za nadzor dela, komunikacijo in sodelovanje z dijaki ter



dodatno razlago, če česa niso razumeli. Vse, kar naredimo na Slika 2: Prikaz rešenih nalog z aplikacijo Geogebra na tabli, lahko shranimo, delimo ali uporabimo ob ponavljanju ter telefonu pripravi na preizkuse znanja.



2.3 Primer 3: RAČUNALNIŠKE



PREGLEDNICE IN GRAFI



Pri učni temi statistika dijaki uporabljajo program Excel za



izračun mediane, modusa in povprečne vrednosti. Dijaki, ki na

koncu opravljajo poklicno maturo, pa na ustnem delu uporabljajo

tudi program Excel pri določenih matematičnih situacijah, s

čimer pokažejo znanje prikazovanja podatkov s pomočjo tabele

in grafov (tabela 1 in slika 3).



Slika 1: Prikaz tabelne slike v digitalnem zvezku Tabela 1: statistika - primer uporabe





2.2 Primer 2: PROGRAM GEOGEBRA

Na poučevanje geometrijskih konceptov pozitivno vpliva

uporaba programov za dinamično geometrijo. Eden takšnih je

GeoGebra [3]. Gre za prosto dostopno aplikacijo, ki združuje

komponente geometrije, algebre in analize. Možnosti uporabe

GeoGebre zajemajo številne matematične vsebine. Med drugim

lahko z njeno pomočjo ponazarjamo geometrijske elemente in

obravnavamo njihove lastnosti, rišemo grafe funkcij in

obravnavamo njihove lastnosti, statistično obdelujemo podatke

ter ponazarjamo koncepte s področja verjetnosti. Program nam

omogoča spreminjanje lastnosti objekta in opazovanje teh

sprememb, s čimer pri dijakih spodbujamo vizualizacijo.

Prikazana je uporaba aplikacije GeoGebra pri pouku



matematike v drugem letniku srednje šole pri temi kvadratna

funkcija (slika 2). Aplikacija je brezplačna in dostopna tako na

Androidu kot na sistemu iOS in ima več kot 10 milijonov

prenosov.



Slika 3: Uporaba Excela

2.4 grafov ter matematičnih funkcij. Uporablja se lahko za različne Primer 4: PROGRAM PHOTOMATH



sistemu Android kot tudi na iOS. Dijaki uporabljajo Photomath Photomath [4] je aplikacija, ki deluje tako na operacijskem matematične naloge, kot so risanje grafov, reševanje enačb, analizo funkcij in drugo. Desmos je zasnovan za enostavno uporabo in je dostopen na za preverjanje svojih rešitev. Za prikaz rešitve se nalogo spletu ali v aplikaciji. fotografira, mobilna programska oprema PhotoMath prikaže Dijaki so imeli za nalogo, da s pomočjo matematičnih funkcij rešitev ali pa omogoči prikaz vseh korakov, ki so bili uporabljeni v Desmosu izdelajo črki svojega imena in priimka (slika 6). V za izračun. spodaj prikazanem primeru so potrebovali 20 različnih linearnih Prikazan je primer (slika 4), ko so dijaki preverjanje znanja funkcij z intervali. opravili v parih. Dijaka sta pred uporabo IKT izvedla najprej

medvrstniško vrednotenje. Med seboj sta pregledata rešitve in ob

tem razvijala še pozitivno komunikacijo, ki je nekonfliktna,

spoštljiva, hkrati pa sta uspela odločno predstaviti napake in

možnosti izboljšave.





Slika 4: Prikaz rešenih nalog z aplikacijo Photomath





2.5 Slika 5: Zrcaljenje zaslona Primer 5: ZRCALJENJE ZASLONA



Pri preverjanju znanja se uporablja zrcaljenje zaslona (ang.



Screen mirroring). To je tehnologija, ki omogoča prenos in

prikazovanje zaslona mobilne naprave na zaslon, v našem

primeru na projektor (slika 5). Na ta način lahko vsem dijakom

hkrati podamo kriterije uspešnosti. Prihranimo čas, ki ga

namenimo komunikaciji z dijaki in preverjanju, da so dijaki snov

res usvojili. Prikazan je primer, ko so dijaki najprej samostojno

rešili nalogo, sama pa sem medtem z mobilnim telefonom slikala

pravilno rešitev pri enem dijaku in jo projicirala na tablo.



2.6 Primer 6: MS TEAMS

Slika 6: Ustvarjanje v Desmosu

Za grajenje komunikacije z dijaki uporabljamo pri pouku tudi

MS Teams. Ekipe, ki jih ustvarimo za vsak razred posebej,



omogočajo prenos informacij in drugih gradiv vsem dijakom.

Klepet pa dijaki uporabljajo za individualni pogovor z učiteljem.

V MS Teams dijaki občasno oddajo tudi domačo nalogo in tako

dobijo povratno informacijo.



2.7 Primer 7: UPORABA PROGRAMA

DESMOS

Program Desmos [5] je razvil Eli Luberoff, profesor

matematike in fizike na Univerzi Yale. Uporablja se lahko kot

spletna ali mobilna aplikacija. Uporabniki programa si lahko

ustvarijo uporabniški račun, v katerega shranjujejo grafe, ki jih Slika 7: Igranje v Desmusu lahko kadar koli uporabijo. Desmos je interaktivno matematično

orodje, ki omogoča uporabnikom ustvarjanje in raziskovanje



Dijaki so poleg črk svojega imena, za zabavo, lahko pripravili

karkoli. In tako je pri domačem delu nastala slika 7.



3 ZAKLJUČEK

Kljub časovni stiski pri matematiki je pri individualnem delu

ali delu v skupinah in uporabi različnega IKT še bolj pomembno

zavedanje učitelja, da ne poda prehitro povratne informacije in

pusti, da dijaki sami popravijo svoje napake in imajo možnost za

izboljšanje.

Pri vsaki uporabi IKT dijaki niso usvojili le določene snovi,

temveč so se naučili tudi, kako preveriti in raziskati postopke

reševanja. Predvsem pa so se s pomočjo uporabe IKT naučili,

kako kritično analizirati svoje rezultate. Na ta način so bili bolj

motivirani, večkrat so tudi ponovili postopke reševanja in se učili

iz svojih in napak sošolcev.



LITERATURA IN VIRI

[1] Eurydice. 2002. Key Competencies: A developing concept in general

compulsory education.

[2] Sirnik Andreja, Bone Jerneja, 2016. Smernice za uporabo IKT pri

matematiki. Zavod za šolstvo

[3] Matematični program Geogebra. Dostopno na naslovu:

https://www.geogebra.org/

[4] Matematični program Photomath. Dostopno na naslovu:

https://photomath.com/

[5] Matematični program Desmos. Dostopno na naslovu:

https://www.desmos.com



Do boljšega razumevanja fizikalnih količin s pomočjo



temeljnih znanj računalništva in informatike



Towards a Better Understanding of Physical Quantities



through Fundamental Knowledge of Computer Science and



Informatics



Roman Bobnarič Lenka Keček Vaupotič

Gimnazija Ormož Gimnazija Ormož

Ormož, Slovenija Ormož, Slovenija

roman.bobnaric@gimnazija-ormoz.si lenka.kecek-vaupotic@gimnazija-ormoz.si



POVZETEK

V prispevku je predstavljena izvedba laboratorijskih vaj iz 1 Medpredmetno povezovanje malo drugače



fizike s pomočjo računalniške tehnologije na nekoliko drugačen Dijaki želijo delati praktično, kar dejansko pomeni, da

način. Velika praznina v pouku postaja razdeljenost med moramo povezovati znanje različnih predmetov. Vendar kljub

področji znanja in znanosti, ki je mladostniki sami ne zmorejo temu ostajamo pri pouku ločeni po predmetnih področjih. To se

zapolniti. Zato je bil projekt medpredmetne povezave zastavljen dogaja kljub medpredmetnemu povezovanju, kjer običajno en

tako, da dijaki spoznajo tako fiziko kot informatiko kot dve predmet prevzame vodilno vlogo, drugi pa ostane v vlogi

enakovredni znanosti, ki ena drugo nadgrajujeta in se povezujeta. podpornega predmeta.

Opisana je ideja, ki je pripeljala pouk od pogovora in priprave Zaradi motivacijskega učinka je uporaba fizičnega

gradiv do izvedbe vaj. Predstavljena je tudi izvedba v razredu in računalništva dobrodošla, ker dijakom nudi učinek takojšnjega

nekaj ključnih ugotovitev o izvedbi. odgovora in možnosti takojšnje povratne informacije. S tem je

poudarjena vloga povratne informacije, ki dijakom pomaga

KLJUČNE BESEDE razumeti, kaj delajo, učitelju pa predstavlja “pomočnika” pri

Računalništvo in informatika, fizika, fizično računalništvo, njegovem delu.

micro:bit Fizika je v osnovi znanost, kjer lahko dijaki dobijo znanja iz

zelo različnih področij, ker se s snovjo dotika praktično vsakega



ABSTRACT dela našega življenja. Ljudje to običajno začutimo, ko moramo

kaj narediti s svojimi rokami, ne le z razmišljanjem. Zato je

The article presents the implementation of physics lab pomemben del pouka laboratorijsko delo, s katerim dijaki

exercises using computer technology in a slightly different way. razvijajo ročne spretnosti in sposobnosti reševanja problemov.

A significant gap in education is the division between fields of Pri tem spoznavajo povezanost različnih področij znanosti, ki se

knowledge and science, which adolescents are unable to bridge prepletajo tudi s pomočjo sodobne komunikacijske in

on their own. Therefore, the project of interdisciplinary informacijske tehnologije. integration was designed so that students would understand both



physics and computer science as two equal sciences that 1.1 Enakost? complement and connect with each other. The idea that led the

teaching process from discussions and preparation of materials Med svojim delom učitelji, kakor tudi dijaki, spoznavamo, da

to the implementation of the exercises is described. The se zaradi časovnih okvirjev in zaporedja snovi posameznega

execution in the classroom and some key findings about the predmeta različni predmeti težko povezujejo med sabo. Tako se

implementation are also presented. običajno zgodi, da en predmet prevzame vlogo prvega oz. učnega

The conclusion of the article outlines plans for future work predmeta, drugi predmet pa je potem le vir podatkov ali metod

and the use of this approach in addressing other topics in physics dela. Na takšen način učitelj drugega predmeta pri pouku ne

and computer science, with the possibility of transferring it to doseže vseh ciljev, čas pa teče ne glede na to, kaj je bilo

other subjects. doseženega. Zato pride do razhajanja, ker imajo predmeti na

razpolago le določeno kvoto časa.



KEYWORDS Zaradi dela v skupini projekta MINUT NAPOJ

(http://naslokar.fmf.uni-lj.si/FMF/MINUT-NAPOJ/MINUT-

Computer science, physics, physical computing, micro:bit

NAPOJ-predstavitev.html#domov) sva se zavedala te pasti in sva

v osnovi skušala zmanjšati to razliko ter obema predmetoma

določila cilje kot izhodišče načrtovanja.



1.2 in preračun vrednosti izrazov. Ker so bile rešitve pridobljene Povezovanje znanja



S stališča fizike so laboratorijske vaje lahko namenjene zgolj programiranja, saj so videli, da je uporabno tudi za njihovo delo bliskovito hitro, je veliko dijakov začelo iskati pot do

izvajanju meritev ali pa so namenjene merjenju in razumevanju

in učenje. Tako je nastal krožek programiranja, kjer so dijaki

poti do meritev. V času nenehne dostopnosti do podatkov



postajajo poti do podatkov in znanja bolj pomembne kot podatki običajno zmanjkalo zagona in idej. In tako sva praviloma vsake spoznavali osnove programiranja. A ne za dolgo, ker jim je

sami. Zato so laboratorijske vaje odlična priložnost spoznati pot.



podatkov, smiselnost in uporabnost ter obseg podatkov in še kaj. sva vseeno oba našla pogosto izraženo željo dijakov, ki so vedno Žal to prinese tudi več razmišljanja izven območja udobja, ki ga znova izražali, da bi si želeli več uporabe računalnika oz. IKT pri dijaki ne sprejmejo kar tako. ostalih predmetih. In ne samo pot, tudi način pridobivanja podatkov, prenos informatike pri ostalih predmetih. V analiziranju svojega dela pa toliko časa ostala brez dijakov, ki bi jih zanimala uporaba



izmeriti. Z rezultati takih meritev je možno nadaljevati s Da bi ugotovili uspešnost dela, je bilo treba spremembo tudi V tem obdobju sva izvedela za seminar Izzivi poučevanja informatike v OŠ in SŠ na Fakulteti za matematiko in fiziko, ki primernim načrtovanjem, da se izognemo nesmotrni rabi časa in se izvaja že več let in ki sva se ga oba udeležila in kmalu postala dela ter izboljšamo uspešnost. Zato so dijaki izpolnili anketni del večjega projekta skupine učiteljev pod vodstvom dr. Andreja vprašalnik pred in po izvajanju aktivnosti. Z anketnim Brodnika in mag. Matije Lokarja – MINUT NAPOJ. Ta projekt vprašalnikom sva želela izmeriti odnos do obeh predmetov, je bil nadaljevanje prvotnega projekta NAPOJ. Osnovni cilj predznanje in predvsem spremembo, ki bi nastala med slednjega je bil sestaviti skupnost učiteljev, ki bi se dobivali, drugačnim načinom dela. izmenjevali svoje izkušnje, izmenjevali gradiva, se pogovarjali o

svojem pouku in izvedbi istih stvari pri drugih učiteljih. Pri



2 osnovnem projektu NAPOJ so sodelujoči spoznali, da je Vse že imamo, a ne uporabljamo

pomembna sestavina skupnosti skupna izdelava različnih učnih

Med opravljanjem laboratorijskih vaj kakor tudi ostalih gradiv. Na ta način so uspeli doseči, da projekt ni doživel konca

dejavnosti pri pouku fizike se pojavlja čedalje več stvari, ki kot veliko drugih projektov. Večina projektov se namreč po

nakazujejo, da bi morali v pouk vključevati uporabo uradnem zaključku konča in za njimi se običajno izgubijo sledi.

informacijske in komunikacijske tehnologije. To je izraženo na Tokrat se to ni zgodilo in rodilo se je nadaljevanje – MINUT

različnih ravneh – od želja dijakov, vodstva, učnih načrtov in še NAPOJ, kjer gre v bistvu za prenos idej mednarodnega gibanja

česa. Težava pa se pojavi takoj, ko to v resnici želimo narediti, STEM v izobraževanje in dopolnitev za naše šolsko delo.

ker so med dijaki zelo velike razlike v znanju in spretnostih. Ravno v tem projektu smo se prvo leto posvetili razmišljanju

V nadaljevanju bo prikazan način dela pri povezovanju in o tem, kako bi lahko povezovali predmete na način, da bi dijaki

spoznavanju osnovnih korakov fizičnega računalništva. Vsak dosegali cilje obeh (oziroma vseh) predmetov. Da ne bi bil eden

dan imajo dijaki večino časa pri sebi naprave, ki so polne zgolj vir podatkov, tabel, enačb in še česa za delo pri drugem. Ko

senzorjev, ki snemajo njihovo okolico, dogajanje in se na to, kar smo v začetku sestavljali zapise načrtov in predstavljali projekte

zaznavajo, tudi samostojno odzivajo. Ko smo se v uvodu kolegom, smo bili kot osnovnošolci, saj je marsikdo skoraj

aktivnosti pogovarjali o tem, so imeli dijaki ogromno idej, kako obupal, ko je po tri- ali večkrat popravljal svojo dokumentacijo.

uporabiti svoje naprave. Imajo ideje, s katerimi bi radi rešili svet, A se je vseeno splačalo.

a se zavedajo tudi, da ga ne morejo rešiti preko noči, ampak le z Na koncu smo dosegli, da smo jasno zapisali svoje ideje in

majhnimi koraki. Zato smo učitelji pozvani tudi k temu, da jim cilje. To nam je pomagalo, da smo lahko ocenili in uravnovesili

njihova vsakdanja orodja razsvetlimo v novi luči in jim svoj pouk za oba predmeta.

približamo načine, s katerimi lahko sami opazujejo svet okoli Kaj je boljšega od tega, ko se lahko pri pouku tudi igramo?

sebe tudi s pomočjo teh naprav in orodij, ki jih že poznajo in Tudi oba učitelja sva se igrala, ne le dijaki. Najprej sva morala

uporabljajo. Tako so npr. v šoli pri računalniških dejavnostih sama najti smisel, način in smoter uporabe teh “igračk”, ki jih

dijaki že uporabljali micro:bite, ki pa so služili zgolj učenju danes nosimo v žepih čisto vsi. programiranja.

Tako smo ugotovili, da lahko tudi micro:bit uporabimo za

merjenje velikega števila parametrov, s katerimi bi lahko 3 Ideja



opisovali spremembe okolja, kakor tudi spremembe našega V osnovi je bil cilj pri fiziki zastavljen precej višje, saj je življenja, ki je od teh sprememb odvisno. zvenelo zelo lepo, da imamo merilnik, ki ga lahko sami



2.1 sprogramiramo in uporabimo za vse meritve. Ko smo začeli Povezovanje predmetov

pregledovati tehnične karakteristike micro:bita, je kmalu fizik

Delo na povezovanju fizike in informatike se je pripravljalo doživel prizemljitev. Micro:biti delujo preko termometra, ki je na

kar nekaj časa, preden je bila aktivnost izvedena. Pred to izvedbo procesorju in ima očitno nekaj izolacije, ki onemogoča hitre

pouka sva se večkrat že dogovorila in izvedla kakšno dejavnost odzive, ne sme doživeti previsokih temperatur, ne sme se zmočiti

skupaj. Pri informatiki so dijaki pripravljali projektne naloge s in še kaj. Ima pa ta lastnost tudi stranski učinek – napako. V

tematiko drugih predmetnih področij, med katerimi je vsako leto vsakem primeru kaže termometer temperaturo, ki je nekoliko

bila prisotna fizika. Pri fiziki smo se z dijaki lotevali izvajanja drugačna od tiste v okolici. In to je bil namig, ki je rešil zagato.

poskusov z računalniškimi meritvami, kjer so potrebovali pomoč Poudarek ne bo na merjenju temperature, ampak na merjenju

informatike, ker niso imeli dovolj znanja. Pri fiziki so dijaki napake, pogovoru o napakah in o tem, kako napako odstraniti.

velikokrat srečali programiranje. Tu je bilo programiranje v To pomeni kalibrirati termometer.

jeziku Python uporabljeno za vnos podatkov za različne naloge

Zelo priročno orodje je v tem primeru računalnik, kjer lahko 3.3 Izvedba



drugega. Torej imamo skoraj idealno orodje za delo, merjenje in Izvedba pouka se je začela z načrtovanjem na srečanju izmerjene količine popravimo, preden jih uporabimo za kaj

učiteljev projekta MINUT NAPOJ, ko smo iskali primerne

učenje. Zdaj je treba le še najti prostor in čas, ko se lahko vse to

projekte za izvedbo v šoli in smo sestavljali načrte dela. Za tem

izvede.

je sledilo kar dolgo obdobje, ko sva oba učitelja pri svojih



3.1 predmetih opazovala dijake, si zabeležila kakšno posebnost in jih O projektu

že pred tem pripravljala na aktivnost, ki jo bodo izvedli. Pri fiziki

Projekt povezovanja fizike in informatike se je začel z smo izvajali laboratorijske vaje na klasičen način, saj je bil cilj

načrtovanjem ciljev in načina, kako izvesti meritve, s čim meriti, tudi izvedeti razliko in stališče dijakov do tega, kako pri pouku

kako dolgo in kaj izmeriti. Načrtovala sva, da bodo laboratorijske poskuse izvajati.

vaje izvedene v dveh šolskih urah, a je bilo treba dodati še eno Druga faza izvedbe pa se je začela nekaj dni pred izvedbo, ko

uro, da smo lahko (tako dijaki kot oba učitelja) rešili ankete pred sva oba dijake opozorila, da bomo tri ure skupaj izvajali

izvedbo in po izvedbi pouka. Tako je celoten projekt trajal tri laboratorijsko vajo s pomočjo IKT. Med dijaki je bilo zaznati

šolske ure. Vendar je bila dodatna ura koristno izrabljena. Ankete pozitivno pričakovanje, saj so že kakšen dan prej bili precej

smo namreč izvedli, da izmerimo, kaj se bo spremenilo in na kaj radovedni in so spraševali o poskusih ter o tem, kako bo delo

bo vplivalo spremenjeno delo pri pouku. potekalo …

Ker smo micro:bite imeli na šoli v zadostnem številu, sva se Manjša težava je bila v urniku dobiti dovolj ur za izvedbo

lotila programiranja in povezovanja ter celoten projekt, še preden takšne aktivnosti. Šola nama je v urniku omogočila premik

so se ga lotili dijaki, izvedla sama. Oba sva pri tem bila predmetov, da je nastalo zaporedje treh ur fizike in informatike,

prepričana, da bo časa premalo, a sva vseeno pustila časovno ko sva lahko izvedla pouk v računalniški učilnici. Pri tem sva

omejitev treh ur. Ob najinem izvajanju sva dobila še idejo, da bi pred izvedbo preverila tudi, ali bodo mize v računalniški učilnici

dijaki na micro:bite povezali še zunanji senzor. Na spletu smo na zadostovale in bo na njih dijakom uspelo zložiti, sestaviti in

pripravah projekta MINUT NAPOJ našli zunanji senzor z izvesti celoten poskus.

oznako BMP180, ki je meril temperaturo, tlak in vlažnost in je Izvedba se je začela z reševanjem anketnih vprašalnikov in s

bil idealen za projekt. A se je v ustreznem času dalo dobiti le tem meritvijo začetnega stanja. Po tem delu smo izvedli skupni

enega, pozneje pa več ni bil na voljo. Zato sva to idejo le uvod, kjer sva tako fizik kot informatičarka vsak iz svojega

zabeležila kot možnost nadgradnje projekta z boljšimi dijaki. zornega kota predstavila delo naslednjih treh šolskih ur. Dijake

Hkrati pa smo našli tudi komplet (Minode), ki omogoča sva pozvala, naj premislijo in se odločijo, ali si želijo delati

povezovanje micro:bitov na enostavnejši način. A je za namene samostojno ali v parih.

našega projekta komplet predrag, predvsem glede na to, da smo Sledila je predstavitev delovanja micro:bita in njegove

v šoli že imeli dovolj veliko število micro:bitov. funkcije, ki so jih dijaki spoznavali na enostavnih primerih



3.2 Operativni učni cilji naj se zgodi …). Dijaki so potrebovali nekaj časa in poskusov, uporabe (npr. izpiši svoje ime na zaslon, ob pritisku na gumb A

Za izvedbo sva si zastavila dokaj širok obseg operativnih da so spoznali način prenašanja kode na micro:bite, a je postopek

ciljev, ki smo jih v veliki meri dosegli: dovolj enostaven, da so se ga vsi naučili po že nekaj primerih

 programiranja (Slika 1). Dijak je sposoben uporabljati IKT tehnologijo pri izvajanju

fizikalnih poskusov in razvija sposobnosti za učinkovito in



uspešno pripravo, izvajanje in analizo poskusov.

 Dijak spozna in zna uporabljati računalniško tehnologijo in

programsko opremo ter s programiranjem rešiti preproste

probleme.

 Dijak zna posamezne strokovne teme obravnavati v luči

problemov ter se usposablja za učinkovito in konstruktivno

sodelovanje v skupini/paru.

 Dijak pozna pomen in vlogo digitalne/računalniške

tehnologije, jo spozna in zna uporabljati.

 Dijak razvija ustvarjalno mišljenje in sposobnost reševanja

problemov.

 Dijaki znajo izmeriti izbrane fizikalne količine.

 Dijaki znajo izmerjene vrednosti prikazati v tabele in grafe.



Cilji so postali vodilo za delo, na katere sva dodala še vsebino, Slika 1: Preizkušanje osnovnih funkcij micro:bitov.

ki je omogočala, da so dijaki lahko samostojno ali z vodenjem

posegli na novo področje dela in razmišljanja. Dijaki pred Naslednji korak je bilo spoznavanje merjenja fizikalne

izvedbo teme niso imeli izkušenj s programiranjem, kakor tudi količine, ki se je izpisovala na zaslonu micro:bita. V tej fazi so

ne s samostojnimi laboratorijskimi vajami. dijaki med pogovorom spoznali, da dobivajo različne rezultate,

čeprav se nahajajo v istem prostoru. Zato je takoj sledila razlaga

učitelja fizike, ki je prevzel zgodbo in dijake pozval, naj

poskušajo sami poiskati izvor teh napak. Ko smo s skupnimi

močmi ugotovili nekaj možnih virov napak, smo skupaj 4 Refleksija



pogledali še v tehnične karakteristike micro:bitov in ugotovili, da Po izvedbi določene učne teme na nov način se učitelju imajo določena odstopanja. Ko je učiteljica informatike pojasnila odpirajo številna vprašanja, ali je delo uspelo, kaj spremeniti in še vpliv porabe energije v procesorju med delovanjem če bi se to splačalo ponoviti. Čeprav se učitelji trudimo biti programov in toplote, ki se zato pri tem sprošča, smo imeli dovolj objektivni, načrtovati, evalvirati ves čas in tudi ocenjevati, pa razlogov, da se lotimo umerjanja. včasih pride tudi trenutek, ko se vmes postavi naš subjektivni V naslednjem koraku so dijaki spoznali namen uporabe pogled. Ravno zato, da se to ne zgodi, smo pod vodstvom spremenljivk (Slika 2). Z njimi smo si pomagali izračunati projekta MINUT NAPOJ [2] vsi izpolnjevali anketo – dijaki in razliko, ki so jo morali upoštevati v svojih meritvah. oba učitelja.

Ugotivila sva, da je šlo za zelo intenzivno izvedbo pouka, saj



sva morala dijake naučiti ne le izmeriti nekaj številk, ampak jih

tudi usmeriti na pot razmišljanja o globljih ciljih fizike in

informatike.

Večina dijakov je izrazila spremembo v smeri večjega

interesa za oba predmeta. Seveda to ni bil kvantni preskok, da bi

lahko trdili, da so zdaj vsi dijaki usmerjeni samo v fiziko in

informatiko, a se je opazil premik na celotni ocenjevalni lestvici

v smer pozitivnih izkušenj za oba predmeta. Presenetljivo je bilo,

da so dijaki izrazili, da jim je programiranje lažje, kot so

pričakovali. V prihodnje bi radi več takšnih aktivnosti. A je bilo

nekaj tudi sporočil, da ne smemo z IKT pretiravati.



Slika 2: Del programske kode 15



računalnik, kjer lahko z zabeleženimi podatki naredimo več kot Na koncu smo raziskali še možnost prenosa podatkov v v 10 ko ija le to, da jih prikažemo na prikazovalniku merilnih naprav (Slika d 5 3). Iz njih lahko sestavimo tabele in grafikone, se ob podatkih št.

lotimo drugačnih analiz. Morda celo primerjamo podatke

različnih merilnih postaj med sabo in iščemo vzroke ter 0

posledice. Odprtih je veliko možnosti. da delno ne

prej potem Fizika me zanima





Grafikon 1: Sprememba v odnosu do fizike



Ko sva svoje delo evalvirala tudi sama, sva ugotovila, da je

najino mnenje precej podobno tistemu, ki so ga izrazili dijaki.

Strinjala sva se, da je zelo dobra aktivnost, ki bi jo bilo treba še

nadgraditi in tudi kdaj izvesti. V skupini sva opazila nekaj

posameznikov, ki so se takega načina dela lotili zelo težko in jim

ni ustrezalo. To sva predvidevala že pred začetkom dela. Zato

sva načrtovala delo v parih in po izbiri samostojno. A za te

posameznike bo treba najti še kakšen način dela, da se odprejo in

pri pouku zaživijo tudi na ta način.



5 Zaključek

Po premisleku glede opravljenega dela in rezultatov

ugotavljava, da je opisana aktivnost primerna za trajno vključitev

v pouk obeh predmetov. Po pridobljeni izkušnji je bilo zaslediti,

da podpira izvedbo ciljev obeh predmetov, ki jih logično

povezuje in ustvarja nekakšen okvir, kjer eni omogočajo izvedbo

drugih in obratno. Fizika namreč ni samo stvar laboratorijev,

Slika 3: Prikaz merjenja temperature. ampak stvar realnega življenja, informatika pa se prepleta v vse

niti našega življenja s tehnologijo in razvojem novih načinov



pogovora. V obeh aktivnostih so dijaki izrazili veliko zanimanja Glede na izvedbo, ki je bila pripravljena, dokumentirana in za podobno delo še v prihodnje. analizirana še s strani strokovnjakov s področja RIN na FRI in Za konec smo izvedli še evalvacijo s pomočjo ankete in sprejemanja realnosti.

FMF ter drugih učiteljev projekta MINUT Napoj, bo ta aktivnost



ostala vključena v redno delo pri pouku še v naslednjih letih.

Glede na spremembe, ki se trenutno dogajajo na področju

posodabljanja učnih načrtov, dejavnost ustreza trendom na tem

področju. Oba učitelja si želiva slediti spremembam pedagoške

in didaktične prakse. Zato načrtujeva, da bo nastalo še več

podobnih aktivnosti.



Zahvala

Zahvala za izvedbo celotnega projekta gre v prvi meri

dijakom, ki so vztrajali in nama dovolili pri pouku

eksperimentirati z njihovim časom in znanjem. Hkrati bi se

zahvalila tudi vsem sodelavcem, ki so nama priskočili v pomoč

z nasveti, in vodstvu šole, ki je imelo dovolj razumevanja za

drugačnost dela, ki je bilo opravljeno. Zahvaljujeva se najinim

kolegom iz projekta MINUT NAPOJ, ki so s svojimi komentarji

že v osnovi v nama spodbujali razmišljanje o pouku na drugačen

način in seveda mentorjema, dr. Andreju Brodniku in mag. Matiji

Lokarju, ki sta bila zelo dobra podpora in vodnika na tej poti

odkrivanja.



Viri

[1] Brodnik, A., 2023: 1+1=4. Zbornik povzetkov konference STEAMColab

2023. https://digitalna.uni-lj.si/wp-

content/uploads/2023/09/STEAMCOLAB23_zbornik-%E2%80%93-

koncna-verzija_14_9.pdf.

[2] Lokar M., 2023: Projekt MINUT NAPOJ. http://naslokar.fmf.uni-

lj.si/FMF/MINUT-NAPOJ/MINUT-NAPOJ-predstavitev.html#domov.

[3] Devine, j. et. al. MakeCode and CODAL: Intuitive and efficient

embedded systems programming for education. Journal of Systems

Architecture, Volume 98, September 2019, Pages 468-483,

https://www.sciencedirect.com/science/article/pii/S1383762118306088

[4] S. Hodges, S. Sentance, J. Finney and T. Ball, "Physical Computing: A

Key Element of Modern Computer Science Education," v Computer, vol.

53, no. 4, str. 20-30, april 2020,

https://ieeexplore.ieee.org/document/9062372.

[5] Przybylla M., Romeike R., Key Competences with Physical Computing,

https://publishup.uni-potsdam.de/opus4-

ubp/frontdoor/deliver/index/docId/8290/file/cid07_S351-361.pdf.



Prepletanje kmetijske stroke, strokovne informatike in



statistike ter knjižničnih znanj



Interweaving Agricultural Expertise, Professional Informatics,



and Statistical and Library Knowledge



Darinka Bosnar

Šolski center Šentjur

C. na kmetijsko šolo 9, 3230 Šentjur

darinka.bosnar@sc-s.si



context in which it can be understood in relation to agricultural

POVZETEK production.

Prispevek predstavlja primer dobre prakse povezave Students are faced with a practical challenge, where they must

kmetijske stroke in računalništva. Z dodatkom knjižničnih znanj, identify the appropriate potato variety based on client

to je navajanja literature, pridemo do končnega izdelka requirements and educate the clients about the characteristics of

študentov. the variety and the necessary agronomic measures.

Namen predstavljenega primera je, da predavatelj med Finally, they present their product and ensure correct source

izvajanjem učne ure vključi čim več spletnih vsebin, s pomočjo citation.

katerih podkrepi navedena dejstva. Poleg tega je tudi študentom

lažje, saj lahko sami ponovijo in predelajo obravnavano snov, kar KEYWORDS

je še pomembneje v primeru študija na daljavo. potato, seed houses, hyperlink, SiStat database, online

S pomočjo spletne podatkovne baze Statističnega urada RS resources

študenti poiščejo ustrezne podatke in jih ovrednotijo v času in

prostoru. Na konkretnem primeru študentje razložijo pomen

predstavljenih statističnih podatkov ter v kakšnem kontekstu jih 1 UVOD

lahko razumemo glede kmetijske proizvodnje. Današnji čas nam v izobraževanju postavlja vedno več

Študenti so postavljeni pred praktični izziv, saj morajo na izzivov. Na razpolago imamo več učnih metod, več učnih

podane zahteve strank poiskati ustrezno sorto krompirja in sredstev, ki omogočajo raznolike in bolj pestro obarvane pristope

stranke poučiti o posameznih sortnih značilnostih ter k poučevanju. Prav gotovo je medpredmetno povezovanje tista

agrotehničnih ukrepih. dodana vrednost, ki omogoča preplet različnih predmetov in tako

Na koncu predstavijo svoj izdelek in poskrbijo za pravilno bolj obogateno podajanje učnih vsebin. Pri tem pridobi predvsem

navajanje virov. stroka, saj lahko obogatena s splošnimi predmeti ponudi nekaj

več. Hkrati pa se na ta način še bolj pokaže pomen

KLJUČNE BESEDE splošnoizobraževalnih predmetov in podpornih znanj, saj so

krompir, semenarske hiše, hiperpovezava, podatkovna baza oplemenitena s stroko veliko bolj koristna in dosežejo tisti pravi

SiStat, spletni viri namen. Saj poznamo tisto … kaj nam bo matematika, saj je ne

bomo rabili. Če pa ji je dodan praktični pomen, potem je to čisto



connecting agricultural expertise with computer science. By This article presents an example of good practice in V tem članku bo predstavljena praktična povezava stroke in računalništva. ABSTRACT nekaj drugega.

incorporating library skills, particularly literature citation,



students are able to produce a final product. 2 ŠOLSKI CENTER ŠENTJUR

The purpose of the presented example is to encourage the

lecturer to include as many online resources as possible during Je šola z dolgoletno tradicijo. Njen začetek sega v leto 1910.

the lesson, thereby reinforcing the facts presented. This approach Danes Šolski center Šentjur sestavljata dve enoti: Srednja

also benefits students, as they can independently review and strokovna in poklicna šola ter Višja strokovna šola. V sklopu

process the material, which is especially important in the context zavoda deluje tudi dijaški dom. Srednješolsko izobraževanje

of distance learning. obsega programe nižjega, srednjega poklicnega, srednjega

Using the online database of the Statistical Office of the strokovnega in poklicno tehniškega izobraževanja s področja

Republic of Slovenia, students find relevant data and evaluate it kmetijstva, živilstva in veterine. Višja strokovna šola ponuja štiri

in terms of time and space. In a specific example, students izobraževalne programe: Gostinstvo in turizem, Naravovarstvo,

explain the significance of the presented statistical data and the Upravljanje podeželja in krajine ter Živilstvo in prehrana. V

sklopu šole je še šolsko posestvo [1].

Permission to make digital or hard copies of part or all of this work for personal or

classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s). Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia D. Bosnar



3 ŠTUDIJSKI PROGRAM UPRAVLJANJE 3.2.1 Podatkovna baza SiStat

PODEŽELJA IN KRAJINE Na spodnji sliki 2 so prikazane naloge, ki so jih študenti dobili

v zvezi s statističnimi podatki.

Študenti imajo tekom izobraževanja v programu Upravljanje

podeželja in krajine tudi predmeta Gospodarjenje v poljedelstvu

in vrtnarstvu [2] ter Strokovno informatiko in statistične metode

vrednotenja [3].

Sama imam v omenjenem programu predavanja in vaje iz

»računalništva«, obenem pa sem tudi knjižničarka. Moja

osnovna stroka je kmetijska, zato jo še kar negujem, da mi ne gre

preveč v pozabo. Sem tudi praktične narave, zato vedno iščem

možnosti, kako se da uporabiti znanja in veščine splošnih

predmetov v praksi, v vsakdanjem življenju.

In tako sem prišla na idejo, da bi pri vajah iz statistike

obravnavali praktične primere, ki bi jim dodali še teoretične

vsebine. Želela sem namreč celo vsebino še nadgraditi in jo

postaviti v bolj strokovne okvire, da bo kot celota imela zadostno

težo. Za začetek so se mi zdele najbolj primerne poljščine. In

tako sem po dogovoru s predavateljico »poljedelstva« izbrala Slika 2: Prvi del praktične naloge – statistični podatki

krompir.

V ta namen smo uporabili podatkovno bazo Statističnega

3.1 Obravnava snovi s pomočjo hiperpovezav urada Slovenije, SiStat [4]. Izbrali smo področje Kmetijstvo,

Cela zadeva je ponovitev srednješolske snovi o gojenju gozdarstvo in ribištvo, znotraj tega pa Rastlinska pridelava →

krompirja z nadgradnjo vsebin in praktičnimi nalogami. V ta Pridelki in površina → Pridelava poljščin. Na sliki 3 je prikazan

namen sem naredila predstavitev v Microsoftovem PowerPointu zaslonski posnetek izbirnih polj v konkretnem primeru.

in vsebino opremila s spletnimi viri, torej hiperpovezavami Navodila na delovnem listu:

(rumeni simboli na sliki 1). Tako sem vključila tudi knjižnična Glede na vašo številko s povezave »Naloge« izberite ustrezne

znanja, saj je pomembno, da študenti tudi sami dalje raziskujejo kriterije v podatkovni bazi SiStat: in tako pridobivajo širša znanja. Prav tako je pomembno, da • Meritve

znajo pravilno navajati spletne vire. • Vrste posevkov na njivah

• Leto



Izpišite podatke, kopirajte tabelo in v ta delovni list na označeno



mesto prilepite le-to tabelo. Nato kliknite na tej spletni strani na

grafikon, ga kopirajte in tudi njega prilepite spodaj na označeno

mesto. Pomanjšajte ga tako, da bo ves viden. Na pripravljeno

mesto še prilepite vir. To je spletni naslov, na katerem ste

»narisali« grafikon. Spodaj napišite komentar grafikona.





Slika 1: Izsek študijskega gradiva



Obe s sodelavko sva bili prisotni pri vseh teh aktivnostih v

razredu, pri obeh predmetih, ki sta potekala v računalniški

učilnici. Najprej je kolegica izvedla predavanje o krompirju,

njegovi pridelavi in agrotehničnih ukrepih ter zakonodaji s tega

področja. V ta namen je uporabila e-predstavitev, ki sem jo

vnaprej pripravila. Nato so študenti vsak na svojem računalniku

snov ponovili, s tem da so uporabljali hiperpovezave. Slika 3: Zaslonska slika podatkovne baze SiStat in izbirnih

polj glede pridelave krompirja v Sloveniji



3.2 Praktična naloga Da smo dodali obravnavi celotne snovi še nekaj več, smo

Sledil je praktični del obravnave snovi. Vsak študent je dobil poiskali tudi statistične podatke, ki so na voljo v spletni bazi

delovni list in izvlekel številko, pod katero so se kasneje skrivale svetovne organizacije za prehrano in kmetijstvo FAOSTAT [5],

naloge. Prvi del naloge je zajemal statistične podatke, drugi pa kot je prikazano na sliki 4.

strokovni nasvet stranki glede izbora sorte krompirja. Tako smo zaokrožili strokovne vsebine glede statistike.

Interweaving Agricultural Expertise, Professional Informatics, and

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

Statistical and Library Knowledge





3.3 Rezultati obravnavane snovi in praktične

naloge

Naloga študentov je bila, da doma ob ogledu e-predstavitve

ponovijo učno snov ter s pomočjo hiperpovezav pregledajo

spletne vire, ki bolj podrobno opisujejo posamezne vsebine.

Na ta način je študente lažje usmeriti na ciljne vsebine, ki so

verodostojne. Prav tako si prihranimo čas pri izdelavi učnih

gradiv, saj lahko podamo samo bistvene podatke, na študentih pa

je, da sami preberejo in predelajo dodatne vsebine.

Študenti so bili zelo zadovoljni s takim načinom obravnave

učne snovi, saj so točno vedeli, kaj je bistvo in kaj je tista

podporna snov, ki temu bistvu daje neke širše dimenzije.

Poseben izziv jim je predstavljala praktična naloga. Tako so

se tudi lotili reševanja; resno in zavzeto. Ker je imel vsak

Slika 4: Zaslonska slika podatkovne baze FAOSTAT in drugačen primer, je bilo potrebno več samostojnosti, torej je bil

izbirnih polj glede pridelave krompirja po svetu vsak prisiljen razmišljati sam in tako tudi opraviti nalogo. Ob

enem pa so si lahko nudili tehnično pomoč. Iskanje sort na spletu

3.2.2 Strokovni nasvet stranki, kupcu v trgovini je nekaterim povzročilo kar nekaj preglavic, ob enem pa tudi

Drugi del praktične naloge je bil agronomsko obarvan. zadovoljstvo, ko so našli ustrezno sorto. Je pač potrebno narediti

Študenti so bili v vlogi prodajalca v kmetijski zadrugi. Vsak vse, da bo stranka zadovoljna in se še vrnila k nam v trgovino.

primer je imel drugačne zahteve, kar je razvidno iz slike 5, in so Na spodnji sliki 6 je prikazan en primer dela rešene naloge.

jih mogli študenti izpolniti. Lahko rečem, da so se študenti potrudili in podali dobre

odgovore. Kot prodajalci bi bili prepričljivi. Tudi sami so bili

zadovoljni s tovrstno simulacijo konkretnega primera prodaje in



svetovanja v kmetijstvu. In tako so ugotovili, da ni tako lahko

prodajati krompir, če imaš zahtevno in izbirčno stranko.

Za potrebe pedagoškega procesa so uporabljali elektronske

vire, ki so na razpolago. V prodajalni bi najbrž uporabili tiskan

katalog semenskega krompirja.





Slika 5: Drugi del praktične naloge – agrotehnično

svetovanje stranki



Navodila na delovnem listu:

Ste prodajalec. K vam v kmetijsko zadrugo pride stranka, ki želi

kupiti semenski krompir. Vpraša vas za nasvet glede semenskega

izbora. Poda vam nekaj informacij, na osnovi katerih:

• svetujete nakup določene sorte,

• seznanite stranko s priporočenimi tehnološkimi

navodili in agrotehniko pridelave.

Zdaj je vaša naloga, da na povezavi »Naloge« najdete,

kopirate in sem na določeno mesto prilepite strankine zahteve

glede sorte in površine, ki jo bo namenila sajenju.

Upoštevajoč strankine želje na spletu poiščite primerno sorto

slovenskih dobaviteljev semenskega krompirja.

Na določeno mesto napišite ime izbrane sorte, prilepite opis

te sorte ter sliko (pazite na dosledno navajanje spletnih virov).

Utemeljite svojo odločitev. Če je možno, jo podkrepite z rezultati

sortnih poskusov. Stranki ponudite tudi alternativo. Izračunajte,

koliko semenskega krompirja rabi.

Slika 6: Rešen primer dela delovnega lista



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia D. Bosnar



Edina slaba stvar pri vsem skupaj je to, da takšen pristop od

učitelja zahteva res ogromno dela. Vendar se da vse pripravljeno

gradivo uporabiti tudi naslednja leta, s tem da je treba preveriti

4 ZAKLJUČEK vse spletne vire, če so še primerni oz. jih osvežiti. No, to pa je že

Ob nastanku ideje nisem vedela, kakšen bo odziv študentov, čar vedno večje digitalizacije, saj so nam spletni podatki na voljo

izplen, sem pa pričakovala, da dober. In res je bil. Celo zelo praktično vedno in povsod.

dober. Študenti so doma zavzeto predelali učno snov in si tudi Prikazan primer govori o tem, da se da tudi v strokovne

pridno pogledali dodatne učne vsebine, ki so jih našli na predmete zelo uspešno in učinkovito vključiti razpoložljive

hiperpovezavah. Tako so natančno vedeli, kaj se od njih zahteva, spletne vire in na ta način naše mlade navajati, da sami iščejo

katere strokovne vsebine morajo poznati. ažurirane podatke in jih umestijo v ustrezen kontekst ter tako

Poleg tega so jim bile vaje iz statistike dosti bolj zanimive, saj koristno uporabijo.

so se nanašale na konkretne primere. Velika večina se jih je prvič

srečala s temi spletnimi podatkovnimi bazami. Tako so tudi VIRI IN LITERATURA

spoznali pomen statističnih podatkov ter jih še dalje sami [1] Šolski center Šentjur. (b. d.). https://sc-s.si/

raziskovali. Bili so navdušeni nad možnostmi, ki so na voljo za [2] Katalog znanj. Višja strokovna šola. Upravljanje podeželja in krajine.

iskanje rezultatov po različnih kriterijih. Te spletne statistične Gospodarjenje v poljedelstvu in vrtnarstvu. (2007). Center RS za poklicno

izobraževanje. https://cpi.si/wp-content/uploads/2020/08/gospodarjenje-v-

podatke bodo prav gotovo še iskali za kakšne seminarske naloge poljedelstvu-in-vrtnarstvu.pdf

in morda na koncu za diplomsko nalogo. [3] Katalog znanj. Višja strokovna šola. Upravljanje podeželja in krajine.

Ko smo preverili znanje, so si študenti res veliko zapomnili Strokovna informatika in statistične metode vrednotenja. (2007). Center RS za

in verjamem, da jim bo to znanje ostalo. informatika-in-statisticne-metode-vrednotenja.pdf poklicno izobraževanje. https://cpi.si/wp-content/uploads/2020/08/strokovna-



Na enak način bi lahko obdelali vse z vidika kmetijstva [4] Statistični urad Republike Slovenije. (b. d.). Podatkovna baza SISTAT.

pomembne kulturne rastline. Ne samo v poljedelstvu, tudi v https://pxweb.stat.si/SiStat/sl

zelenjadarstvu, sadjarstvu in vinogradništvu. [5] FAO, Food and Agriculture Organization of the United Nations. (b. d.). Food

and agriculture data FAOSTAT. https://www.fao.org/faostat/en/#data/QCL

Tudi pri živinoreji bi se dalo s takim načinom dela študentom

bolj kvalitetno približati učno snov.



Digitalne dimenzije ob Prešernovi klasiki: IKT kot orodje



za sodobno obravnavo Krsta pri Savici



Digital Dimensions of Prešeren's Classic: ICT as a Tool for



Contemporary Interpretation of the Krst pri Savici



Andreja Breznik

Srednja šola Slovenj Gradec in Muta

Muta, Slovenija

andreja.breznik@sc-sg.si



POVZETEK KEYWORDS

V prispevku je predstavljena obravnava umetnostnega Study of literary works, information and communication

besedila Franceta Prešerna Krst pri Savici s pomočjo technology, modern teaching methods, e-materials

informacijsko-komunikacijske tehnologije. Obravnava

književnega dela v izobraževalnem procesu je pomemben del

pouka književnosti, saj dijakom omogoča poglobljeno 1 PREGLED OBSTOJEČEGA E-GRADIVA

razumevanje literarnih del, razvoj kritičnega mišljenja in Za popestritev obravnave Prešernovega dela Krst pri Savici

izboljšanje njihovih jezikovnih ter interpretativnih veščin. smo pri pouku slovenščine uporabili informacijsko-

Profesorji se pri pouku književnosti soočamo z upadom komunikacijsko tehnologijo. V uvodni uri sem dijake motivirala

zanimanja za domače branje in analizo oz. interpretacijo z napovedjo, da bo pouk potekal malo drugače in da bodo lahko

literarnih del. Na izvirne načine se trudimo dijake motivirati in pri pouku uporabljali računalnike in telefone. To delo so dijaki

jih spodbuditi k branju s kritičnim razmišljanjem ter zmožnostjo že prebrali za domače branje, tako da so poznali avtorja, vsebino

aktualizacije. Klasična obravnava dela žal zaradi veliko in književne osebe. Prva zadolžitev je bila, da so v Google vpisali

suhoparnih podatkov vodi v dolgočasenje dijakov. Učitelji zato iskalni niz: Prešeren ‒ Krst pri Savici. Pregledati so morali nekaj

velikokrat iščemo zanimive načine za izvedbo pouka. Eden e-gradiv, tudi videoposnetke na Youtubu, ki so se jim je na prvi

izmed njih je zagotovo uporaba IKT-orodij (CapCut, Crossword pogled zdeli zanimivi, in nato podati svoje mnenje – ali se jim

Puzzle, Kahoot, umetna inteligenca). gradivo zdi zanimivo, poučno in uporabno. Skupaj smo nato

naredili nabor najboljših gradiv ‒ na ta način so dijaki že

KLJUČNE BESEDE natančneje spoznavali delo. Med njimi so našli tudi zelo

Obravnava literarnega dela, informacijsko-komunikacijska uporabne PP-predstavitve, videoposnetke in zvočne zapise.

tehnologija, sodobne metode pouka, e-gradiva Izpostavljam tri e-gradiva, ki smo jih z dijaki ocenili kot najbolj

zanimiva in poučna.



ABSTRACT • E-gradivo za slovenščino v gimnazijah in v poklicnih



The article presents an approach to teaching France Prešeren's šport RS in Evropski socialni sklad pri EU šolah. Projekt sta sofinancirala Ministrstvo za šolstvo in

literary work "Krst pri Savici" using information and

(https://gradiva.txt.si/slovenscina/slovenscina-za-

communication technology. The study of literary works in the

gimnazije-srednje-sole/2-letnik/knjizevnost/romantika-na-

educational process is a crucial part of literature classes, as it

slovenskem/012_krst_pri_savici/uvod-58/. Vsebuje tudi

enables students to gain a deeper understanding of literary texts,

zvočni zapis branja Uvoda.

develop critical thinking, and enhance their language and

• E-učbeniki: Uvod h Krstu pri Savici ‒ obravnava dela z

interpretative skills. Literature teachers face the challenge of



analyzing or interpreting them. We strive to motivate students in • Zvočni posnetek branja Krsta na Youtubovem kanalu creative ways, encouraging them to read with critical thinking Slovenska kultura and the ability to contextualize the material. Traditional (https://www.youtube.com/watch?v=H4p-lQHKutQ). declining interest among students in reading assigned texts and (https://eucbeniki.sio.si/slo8/2352/index.html). interaktivnimi vajami in rešitvami

approaches often result in boredom due to the abundance of dry

information. Therefore, teachers frequently seek engaging

methods for conducting lessons. One of them is certainly the use

of ICT tools (CapCut, Crossword Puzzle, Kahoot, artificial

intelligence).





2 UPORABA IKT





2.1 USTVARJANJE V APLIKACIJI CAPCUT



V naslednjih urah nismo imeli več možnosti izvajanja pouka

v računalniški učilnici, zato smo prešli na uporabo telefonov.

Večina dijakov ima na svojem telefonu aplikacijo CapCut, zato

sem dijake razdelila v skupine in jih spodbudila, da na izviren in

zanimiv, lahko tudi nenavaden način v tej aplikaciji predstavijo

vsebino ter osebe tega dela. Da bi jim malo olajšala oziroma

nadgradila njihovo delo, sem jim namignila, da lahko vsebinski

del predstavijo tudi s pomočjo neumetnostnih besedil: poročilo,

reportaža, intervju … To je razred, ki je že drugače zelo

ustvarjalen, zato jim je bila naloga zelo všeč. Pri delitvi v skupine

smo pazili na to, da je v vsaki skupini bil dijak, ki delo v tej

aplikaciji obvlada, ostali pa so prispevali svoje zamisli za

izvedbo glede na svoje zmožnosti. Sprva so v zvezke zapisali



osnutek načrta in cilje dela, nato so se lotili ustvarjanja.

CapCut je priljubljena brezplačna aplikacija za urejanje

videoposnetkov in fotografij. Dijaki so pri ustvarjanju na podlagi

fotografij uporabili naslednje funkcije:

• Obrezovanje in rezanje videoposnetkov, da so

odstranili neželene dele.

• Dodajanje prehodov: različni prehodi med prizori, ki

videoposnetke naredijo bolj tekoče.

• Učinki in filtri, ki omogočajo prilagajanje

videoposnetkov.

• Besedilo in nalepke: dodajanje besedila, naslovov in

nalepk za večjo ustvarjalnost.

• Glasba in zvoki: Uporabniki lahko dodajo glasbo,

zvoke ali svoje glasovne posnetke.

• Hitrostni učinki: Urejanje hitrosti videoposnetkov,

vključno z upočasnjevanjem ali pospeševanjem določenih delov.

• Animacije: Dodajanje različnih animacij, ki izboljšajo Slika 1: Reportaža, narejena v CapCutu

dinamiko videoposnetka.

Nastale so zanimive predstavitve: reportaža, poročilo,



domišljijska pripoved in intervju (slike 1, 2, 3 in 4.

Prva skupina je ustvarila reportažo – reporter se je javljal s

prizorišča boja, izpred Ajdovskega gradca, in poročal o boju med

Črtomirjevo in Valjhunovo vojsko. Fotografije so poiskali na

spletu, jih dodali v program in obdelali ter dodali besedilo in

animacije. Pri tem so ob vnosu besedila morali ves čas natančno

spremljati Prešernovo besedilo Krst pri Savici in ga prilagajati

izbrani besedilni vrsti.

Druga skupina je ustvarila poročilo – na podoben način kot

prva skupina reportažo, paziti so morali le, da je bilo vse zapisano

v pretekliku, imeli so malo težjo nalogo, saj je reportaža zajemala

le vsebino Uvoda, poročilo pa vsebino vseh treh delov pesnitve.

Tretja skupina se je odločila za domišljijsko pripoved z

naslovom Bogomila pripoveduje – v ospredje so postavili

Bogomilo in čez fotografije, ki prikazujejo potek dogajanja skozi

Bogomilino videnje, posneli pripoved dijakinje. Bogomila je

tako v prvoosebni pripovedi in izpovedi predstavila vsa dejstva,

predvsem pa izrazila svoja čustva in doživljanja ob bridkem

slovesu.

Četrta skupina je pripravila intervju z glavnim junakom –

Črtomirjem tik pred bojem. Črtomir je tako iz svoje perspektive

predstavil težke trenutke pred zadnjim bojem in kako je

nagovoril svoje brate – vojake. Slika 2: Poročilo, narejeno v CapCutu





Slika 5: UI-klepetalnik (AI-chatbot)



2.3 IZDELAVA KRIŽANKE IN KVIZA

Po končanem delu je sledila predstavitev ustvarjenega. Dijaki

so pozorno in navdušeno spremljali predstavitve sošolcev.

Zapletena vsebina, zgodovinsko ozadje in književne osebe so se

jim kar izrisali pred očmi in pravijo, da jim bodo ostali v spominu

dolgo časa prav zaradi teh nenavadnih, a zanimivih predstavitev.

Predlagali so, da z delom nadaljujemo. Zastavili smo si nove

naloge, da preverimo usvojeno znanje. Dijaki so se razdelili v

dve skupini.

Slika 3: Bogomila pripoveduje ‒ CapCut Prva skupina je izdelala križanko v spletnem programu

Crossword Puzzle (https://www.education.com/worksheet-

generator/reading/crossword-puzzle/) (slika 6). Sprva so določili



gesla križanke, nato pa njihovo razlago. Druga skupina pa je



ustvarila kviz v spletni platformi Kahoot (slika 7). Tako so se

pomerili v znanju. Ponovno so morali prebrati Krst pri Savici in

gradivo, da so lahko sestavili vprašanja za sošolce. Večkrat so

navdušeno prebirali posamezne verze in na ta način neprisiljeno

spoznavali podrobnosti dela. Prepričana sem, da ob enkratnem

branju v šoli ne bi tako natančno spoznali pesnitve, marsikateri

verz in podrobnost bi bila spregledana.





Slika 4: Črtomirjev intervju ‒ CapCut Slika 6: Izdelava križanke



2.2 UI-KLEPETALNIK (AI-CHATBOT)

V peti skupini pa sem dijakom predstavila orodje umetne

inteligence. Predlagala sem, da ustvarimo AI-chatbota, ki bo

odgovarjal na vprašanja o zgodovinskem ozadju, zgradbi in

sporočilu dela (slika 5). Pred tem so dijaki morali zbrati vse

podatke in jih naložiti v model umetne inteligence.





4 ZAKLJUČEK

Želja vsakega učitelja je, da svoje dijake motivira za delo, za

aktivno sodelovanje in sooblikovanje pouka, na takšen način se

namreč lažje povežejo s predstavljeno vsebino in usvojijo nova

kakovostna znanja. Nenehno iščemo različne načine dela, s

katerimi bi te cilje dosegli. Velika nagrada za učitelja je, da ob

delovni vnemi in velikem zanimanju za delo dijaki sami

predlagajo dodatne naloge in izzive ter se jih lotijo z

navdušenjem in iščejo izvirne rešitve. Po pregledu e-gradiv, ki

obravnavajo Krst pri Savici, smo IKT-orodja preizkusili tudi

sami. Dijaki so ustvarili digitalne predstavitve, kjer so izrazili

svoje razumevanje in interpretacijo dela. Uporaba IKT je bila

zagotovo nadgradnja in popestritev klasičnih obravnav literarnih

Slika 7: Kviz, izdelan v Kahootu del. Cilji so bili doseženi, dijaki motivirani za delo. Z

ustvarjalnostjo in lastno aktivnostjo so podrobneje spoznali



3 pomembno delo slovenskega romantika. POMEN UPORABE IKT PRI POUKU



Ko dijaki aktivno uporabljajo IKT, se izobraževalni proces LITERATURA IN VIRI

obogati, njihovo učenje se poglobi, pouk se popestri, kar prispeva [1] Aplikacija CapCut. Dostopno na naslovu https://www.capcut.com/ (20. 8. k bolj dinamičnemu, interaktivnemu učnemu okolju. Delo z IKT 2024).

prinaša veliko koristi: [2] Spletni program za izdelavo križank Crossword Puzzle. Dostopno na

naslovu https://www.education.com/worksheet-

Razvoj digitalnih kompetenc: dijaki z uporabo različnih orodij generator/reading/crossword-puzzle/ (19. 8. 2024).

pridobijo pomembne digitalne veščine, kar jih pripravi na [3] Spletna platforma Kahoot. Dostopno na naslovu https://kahoot.com/ (19.

učinkovito uporabo tehnologije v vsakdanjem življenju in na 8. 2024.

[4] Mojca Dolinar, Anita Poberžnik. Lidija Jerše.(2013). Vodenje in podpora

delovnem mestu. učencem pri pridobivanju digitalnih kompetenc. Dostopno na naslovu

Dostop do virov in vsebin: https://www.zrss.si/wp-content/uploads/2023/08/Vodenje-in-podpora- IKT omogoča enostaven dostop do

ucencem-pri-pridobivanju-digitalnih-kompetenc.pdf. (15. 8. 2024).

številnih izobraževalnih virov (e-knjige, izobraževalni [5] Viri vseh slik so lastni, zajem zaslonske slike.

videoposnetki …). To učencem omogoča, da poglobijo svoje

znanje in raziskujejo teme, ki jih posebej zanimajo.

Samostojnost in odgovornost: Aktivno delo z IKT spodbuja

dijake k samostojnosti, saj pogosto sami raziskujejo, iščejo

informacije in rešujejo naloge. To povečuje njihovo odgovornost

za lastno učenje, saj morajo sami načrtovati svoje delo in reševati

težave, ki se pojavijo med procesom.

Ustvarjalnost, izvirnost, inovativnost: S pomočjo IKT orodij

lahko dijaki ustvarjajo lastne projekte, ki so lahko bolj ustvarjalni

in inovativni kot navadne naloge. Lahko izdelajo videoposnetke

in digitalne predstavitve, kar spodbuja njihovo ustvarjalno

mišljenje.

Sodelovanje, interaktivnost in komunikacija: IKT omogoča

lažje sodelovanje med dijaki, izmenjujejo si ideje in podatke. To

spodbuja timsko delo in izboljšuje njihove komunikacijske

veščine. Z uporabo interaktivnih tabel, tablic, računalnikov in

pametnih telefonov lahko učitelji ustvarijo bolj interaktivno učno

okolje. Dijaki lahko neposredno sodelujejo pri reševanju nalog,

kvizih, glasovanjih ali diskusijah.

Kritično mišljenje in reševanje problemov: Aktivno delo z

IKT zahteva od dijakov, da kritično razmišljajo, ocenjujejo

informacije in reševanje problemov. Ko se soočajo z izzivi pri

uporabi tehnologije, morajo poiskati rešitve, kar krepi njihove

analitične sposobnosti.

Motivacijo in angažiranost: Uporaba tehnologije poveča

njihovo motivacijo za učenje. Interaktivne naloge, takojšnja

povratna informacija in možnost za ustvarjalno izražanje

prispevajo k aktivnemu sodelovanju pri pouku, saj takšni izzivi

lahko naredijo učenje bolj zabavno, privlačno in sproščeno.

Programiranje + fizika → Igra se začne



Programming + Physics → The Game Begins



Zdenka Candellari

OŠ Toneta Pavčka

Zdenka.candellari@ostpavcka.si



učencem pa motivacije. Da bi se pri tem “ne bodi ga treba” še

POVZETEK medpredmetno povezovali, je res znanstvena fantastika.

V prispevku je predstavljeno, kako preko računalniške igrice, V projektu NAPOJ MINUT pa najdemo možnost, kako

ki jo otroci sami sprogramirajo, utrjevati učno snov pri fiziki v mimogrede utrjevati osnove fizike, to je fizikalne količine in

OŠ. S tem smo naslavljali tako učne cilje s področja informatike enakomerno gibanje ob programiranju računalniške igrice v

kot tudi fizike. Opisana sta ideja, ki je pripeljala do izvedbe te Scratchu.

teme, in način priprave ustreznega gradiva. Predstavljeni sta tudi Projekt NAPOJ MINUT združuje (predvsem) učitelje

izvedba v razredu in ključne ugotovitve o tej izvedbi. V računalništva in informatike (RIN) in ostalih naravoslovnih

zaključku so navedeni načrti glede razvoja obravnave te teme in predmetov, torej MINUT - M(atematika) I(nformatika)

tudi uporaba tovrstnega pristopa pri obravnavanju drugih tem. N(aravoslovje) U(metnost) T(ehnologija) predmetov.

V okviru projekta se sodelujoči:



KLJUČNE BESEDE - Redno sestajajo na spletnih (ZOOM) sestankih in izmenjujejo izkušnje. Programiranje, Scratch, fizika, računalniška igrica, - Razvijajo gradiva, kjer se prepletata RIN in še en formativno spremljanje predmet.



ABSTRACT - Se srečujejo na delavnicah, kjer razvijajo učna gradiva.

- Sodelujejo na konferencah: NAK, STEAMColab,

The article presents how to consolidate learning material in DigiDakta, VIVID.

physics in elementary school through a computer game that - Obiskujejo kolege iz drugih držav, ki poučujejo na

children program themselves. With this, we addressed both podoben način.

learning objectives in the field of informatics and physics. The Na sliki 1 vidimo logotip projekta, ki je nastal prav med

idea that led to the implementation of this topic and the method razvojem teme, kjer so se prepletali umetnost, matematika in

of preparing the relevant material are described. The classroom računalništvo [2]. implementation is also presented and the key findings about this



implementation are presented. In the conclusion, plans are given

regarding the development of both the treatment of this topic, as

well as the use of this type of approach in the treatment of other

topics.



KEYWORDS

Programming, Scratch, physics, computer game, formative

monitoring



Slika 1: Logotip projekta NAPOJ MINUT



1 Medpredmetno povezovanje in projekt V sklopu projekta je nastalo že več kot 10 obravnav tem, NAPOJ MINUT določene pa so še v razvoju. Za več informacij o projektu se

Medpredmetno povezovanje je stara ideja, ki jo pogosto kar lahko obrnete na napoj@404.si.

precej umetno vlečemo skozi celoten učni načrt. Običajno smo

brez pravih idej, kako bi se stvari lotili, s katerim predmetom in 1.1 Ideja projekta

sodelavcem bi se povezovali, da bi stvar stekla gladko in Kot že rečeno je težje pripraviti učence do utrjevanja znanja

predstavila učencem naravo, tako kot je − torej celota, v kateri je kot do usvajanja novih znanj. Zastavili smo si vprašanje, kako

vse povezano. Najlepše steče kakšen dan dejavnosti, ko se istega utrjevati zaključen učni sklop pri fiziki, da bi bilo učencem

problema lotijo različni predmetni učitelji vsak s svojega zornega zanimivo in bi hkrati predstavljal nadgradnjo njihovega znanja.

kota, učenci pa krožijo po delavnicah in doživijo celoto preko Ob tem se pojavi še vprašanje, kaj učenca vleče naprej, kaj je

razdrobljenosti, ki jih povežejo v svojih glavah. njegova motivacija, kdo mu daje povratno informacijo, kaj mu

Še večji problem kot usvajanje novih znanj pa je v šoli daje zadovoljstvo [1]

utrjevanje znanja. Zanj učitelju pogosto zmanjkuje časa,

Glede na privlačnost računalniških igric [1] smo predvidevali,

da bi lahko ustvarjanje oziroma programiranje in potem

medsebojno preskušanje igric imelo željene učinke.

Kot trdijo avtorji v [3] ustvarjanje računalniške igrice učenca

motivira. S pomočjo programa, ki bo deloval, če ga bo učenec

pravilno sestavil, bo znanje učenca sprotno preverjeno in tudi

nagrajeno prav s svojim delovanjem.

Dodana vrednost tega povezovanja pa se bo pokazala tudi

učitelju fizike, saj je prav delovanje ali nedelovanje programa

učencu povratna informacija, ki jo potrebuje za uspešno

napredovanje. Na ta način bo RIN tudi opravil del formativnega

spremljanja [3], ki bi ga sicer moral opravljati učitelj. Z Slika 3: Del fizikalnih nalog

delovanjem programa bo učenčevo delo tudi ustrezno nagrajeno.

Tako se ob povezovanju dveh predmetov fizike ter Predviden potek delavnice po urah in razporeditev nalog je

računalništva in informatike učenec nauči programirati v podan na sliki 4.

Scratchu ter hkrati utrjuje znanje iz enakomernega gibanja.



2 O projektu

Projekt omogoča igrifikacijo [1] učenja osnovnih pojmov iz

enakomernega gibanja. Ko postane utrjevanje že dolgočasno,

vendar je še vedno potrebno, se lahko vprašamo, zakaj ne bi

razširili znanja fizike in ga povezali z računalništvom? Pri naši

nalogi gre za povezovanje osnovnih pojmov enakomernega

gibanja in osnovnih konceptov programiranja, združenih v

okolju Scratch.





Slika 4: Del priprave za izvedbo





Izvajalke dogodka smo se najprej pripravljale in povezovale



v živo na delavnici NAPOJ MINUT, potem pa preko spletnih

srečanj.



2.1 Operativni učni cilji

Pri sklopu enakomerno gibanje pri fiziki v 8.razredu želimo

učence pripeljati do prepoznave in nadalje avtomatske uporabe

simbolov ter enot za osnovne fizikalne količine. Pri usvajanju

enakomernega gibanja definiramo hitrost s pomočjo poti in časa

in izvedemo eksperimente, pri katerih merimo dve fizikalni

količini in tretjo izračunamo. V naslednjem koraku pa učenca

preko konkretnih besedilnih nalog izzovemo, da uporablja tako

znanje fizikalnih količin kot tudi formulo za izračun enakomerne



Slika 2: Delček programa hitrosti in jo po potrebi tudi obrne oziroma izrazi neznanko ter

tako izračuna manjkajočo fizikalno količino.

Ker sem kot učiteljica fizike prisotna na šoli celo šolsko leto, V računalniškem delu delavnice stremimo k osvojitvi

je bil dogovor, da bomo pri fiziki pri rednih urah usvojili osnovnih konceptov programiranja, kot so pogojni stavki, vejitve

operativne učne cilje iz fizike (glej razdelek 2.1), potem pa se in zanke, ter razvoju računalniškega oziroma algoritmičnega

nam za izvedbo delavnice pridruži študentka računalništva in mišljenja. Učenci nove koncepte urijo v okolju Scratch, ki pri

skupaj izvedeva še računalniški del projekta ter tako učenci programiranju uporablja bloke kode, s katerimi učenci

usvojijo tudi učne cilje iz računalništva in informatike (glej sestavljajo programe in jih dopolnjujejo z ustreznimi vrednostmi.

razdelek 2.1). Eden izmed, predvsem za začetnike kompleksnejših konceptov,

Del fizikalnih nalog iz enakomernega gibanja z rešitvami je s katerim se v delavnici srečujejo učenci, je sočasnost dogodkov

predstavljenih na sliki 3. oziroma sinhronizacija elementov, ki jih v okolju Scratch

dosežejo s t. i. pošiljanjem sporočil.



2.2 nekaj novega, da se bodo učili nekaj, kar bo povečevalo znanje Izvedba



oziroma radovedne učence. Na ta dogodek so se lahko prijavili Naše povezovanje smo izvedli v delavnici za nadarjene na obeh področjih tako fizike kot programiranja. Po izvedeni delavnici je v anketi glede fizike 40 % učencev ocenilo, da se je njihov odnos do fizike izboljšal. učenci 8. in 9. razreda do zapolnitve mest (24). Glede računalništva in informatike pa po izvedeni delavnici Delavnica je trajala 5 šolskih ur. Pred začetkom delavnice in ugotavljajo, da je računalništvo bolj zanimivo, manj težko in da po njenem zaključku so učenci izpolnili anketo, s katero smo jih bolj veseli, kot so predvidevali. Po izvedeni delavnici je na želeli meriti uspešnost izvedenega dogodka. trditev »rad programiram« 20 % več učencev kot pred izvedbo Po uvodni predstavitvi delavnice, ki jo je izvedla zunanja odgovorilo pritrdilno. izvajalka, so učenci spoznavali okolje Scratch in si pripravili Moje mnenje je, da je delavnica uspela. Operativni cilji so bili svoje ozadje. Hitrejši so imeli tudi nekaj časa, da so lahko že doseženi in učenci zadovoljni. Res pa je, da je bila delavnica samostojno raziskovali po okolju. precej zahtevna, saj so morali učenci najprej usvojiti kar nekaj Sledila je predstavitev že narejene igrice; najprej so jo sami računalniškega znanja, da so igrici po programu lahko sledili, jo preizkusili, potem pa so se spustili v program igrice in spoznavali, spreminjali in končno sestavili nekaj novega. Tudi njihova prosta kako delujejo posamezni bloki in kombinacija le-teh v Scratchu. izbira udeležbe na delavnici je gotovo botrovala motivaciji in Za spoznavanje okolja Scratch in učenje posameznih uspehu. Glede na področje utrjevanja fizike je cilj zagotovo računalniških konceptov, kot so pogojni stavki, vejitve in zanke, dosežen. je bil pripravljen prazen mini program (primer končnega mini

programa: https://scratch.mit.edu/projects/1029715695 in prazen



mini program: https://scratch.mit.edu/projects/1029727769). 4 Zaključek Učenci so najprej skupaj z izvajalko delavnice spoznali končni

program in ga nato v nadaljnjih urah poskušali postopoma V nadalje bi bilo smiselno izvesti delavnico ločeno v nivojih

vodeno poustvariti z dodatnimi razlagami posameznih konceptov. s tistimi, ki že poznajo Scratch, in tistimi, ki se z njim prvič

V drugi uri so spoznavali osnovno delovanje vejitev in srečajo. Za tiste, ki Scratch prvič uporabljajo, bi bilo potrebno

pogojnih stavkov (if-else stavkov), logike (Booleovo algebro) in nalogo olajšati, morda tako, da bi programirali igrico po delih.

zank, v tretji uri pa pošiljanje sporočil znotraj programa, Način izvedbe v okviru delavnice za nadarjene in radovedne

sinhronizacije elementov ter koncept »vprašanje in odgovor«. učence se je izkazal za zelo primeren in izvedljiv v šoli. Opisani

V tej točki so bili učenci opremljeni z vsem potrebnim način utrjevanja bi zagotovo dobro deloval pri mnogih

znanjem programiranja v Scratchu, ki ga bodo uporabili pri zaključenih enotah fizike, kemije in matematike.

reševanju fizikalne naloge in izdelavi igrice. V Scratchu so učenci uspešno izdelali igrico na podlagi

Učencem smo na začetku četrte ure delavnice razdelili fizikalne vsebine. Tako so pridobili znanje iz osnov

različne fizikalne naloge iz enakomernega gibanja in njihova programiranja. Usvojene spretnosti so prva stopnička na poti do

naloga je bila, da pripravijo oz. prilagodijo program tako, da bo znanja, ki je danes na trgu dela iskano in cenjeno.

deloval za njihovo nalogo. Za to nalogo je bil pripravljen prazen

oziroma deloma izpolnjen glavni program (prazen glavni ZAHVALA

program: https://scratch.mit.edu/projects/1029774314 in končni Zahvaljujem se organizatorjem projekta NAPOJ MINUT za

glavni program: https://scratch.mit.edu/projects/1020356210). primerne izzive in podporo pri nastajanju naših gradiv ter

V tem delu so učenci uporabljali svoje fizikalno znanje; spodbudo za njihovo izvedbo. Projekt sva pripravili z Mio Zalo

prepoznavanje fizikalnih količin in enot iz besedila, zapisovanje Smrečnik, študentko matematike in računalništva na UL PeF, pri

le teh s simboli v program, priklic formule in izražanje neznanke, konkretni izvedbi nama je na pomoč priskočila Anja Količ, prav

računanje rezultata in vnos le-tega v program, da bo v primeru tako študentka smeri računalništvo in matematika na UL PeF,

uspešno rešene igrice potrdil igralca s pohvalo in aplavzom. iskrena hvala obema.

Dogovor je bil, da so pravilni rezultati le tisti, ki vsebujejo simbol

fizikalne količine, število in enoto. VIRI

Za delo so bili zelo motivirani. V primeru zapleta so za pomoč [1] M. Milovič, „Kako psihologija oblikuje naše delo in odnose na delovnem



poklicali učiteljico. Vse narejene/popravljene naloge so naložili mestu,“ Psihologija dela, 6 December 2021. [Elektronski]. Available: https://psihologijadela.com/2021/12/06/igrifikacija-v-izobrazevanju/. v studio, kjer so postale dosegljive za vse sodelujoče in za javnost [Poskus dostopa 23 Avgust 2024].

(studio s končnimi izdelki: [2] Uroš Ocepek in Mojca Baloh, Digitalna umetnost in matematika,

https://scratch.mit.edu/studios/33589412/). https://novi.napoj.si/digitalna-umetnost-in-matematika/ in

https://ucilnica.acm.si/mod/resource/view.php?id=2164

Nadaljevali so z medsebojnim igranjem igric, ki so jih [3] Jože Rugelj, Špela Cerar, Andrej Brodnik, An architecture of an ICT based

sestavili sošolci in si dajali tudi povratne informacije. system for constructivist based learning in higher education, 15-17

september 2010, DOI: 10.13140/2.1.4494.8486.



3 Refleksija

Učenci so po delavnici izražali zadovoljstvo. Tisti, ki jim je

uspelo oddati delujočo igrico v studio, so bili ponosni. Takih je

bilo 17/23. Glede na njihove odgovore v anketi bi izpostavila

njihova pričakovanja in spremembo odnosa do obeh predmetov,

ki sta bila povezana v projektu. Pričakovanja so bila; da bo

zanimivo, da bo drugače, kot je sicer pouk, da se bodo naučili



Razvijanje digitalnih kompetenc na OŠ Prežihovega



Voranca Maribor



Developing Digital Competencies at Prežihovega Voranca



Elementary School Maribor



Natalija Carmona

OŠ Prežihovega Voranca Maribor

Maribor, Slovenija

natalija.carmona@ospvmb.si



POVZETEK KEYWORDS

V prispevku je predstavljeno delovanje na šoli OŠ Digital competences, Innovative pedagogy 5.0, modern

Prežihovega Voranca, ki se je v šolskem letu 2023/24 pridružila teaching concepts

v projektu Inovativna pedagogika 5.0. Projekt temelji na

razvijanju sodobnega pouka, ki je podprt z digitalnimi

tehnologijami ter je usmerjen v razvijanje digitalnih kompetenc 1 UVOD

in temeljnih vsebin RIN. Predstavljeni so tudi izzivi, s katerimi Na šoli smo učitelji in vodstvo šole začutili veliko potrebo po

se soočamo na šoli in kako jih premagujemo. razvijanju digitalnih kompetenc. V času šolanja na daljavo smo

Natančneje je predstavljen tudi konkretni primer učne ure, ki spoznali, da nas je večina učiteljev na osnovni ravni znanj

je bil izveden pri pouku zgodovine in je podrobneje razvijal digitalnih veščin in smo v času večje uporabe računalnikov za

kompetenco avtorske pravice in licence. Pri pripravi sodobne izvajanje kvalitetnega pouka potrebovali več znanja iz tega

učne ure moramo kot učitelji razmišljati celostno, vključiti področja. Vodstvo šole se je zato odločilo, da se v šolskem letu

znanja, spretnosti in veščine, ki so podprte z digitalnim. V pridruži projektu »Dvig digitalne kompetentnosti«, ki je potekal

ospredje si moramo postaviti dejstvo, da morajo učenci postati pod okriljem Zavoda za šolstvo Republike Slovenije.

aktivni člen učnega procesa in ne samo tisti, ki prejemajo znanja Namen projekta, ki je potekal med leti 2021 in 2023 je bil

učitelja. To so izzivi učiteljev, ki jih lahko rešujemo s skupnimi predvsem izboljšati kakovost in učinkovitost izobraževanja ter

močmi in počasi odpravljamo strahove, s katerimi se na šolah razvoj inovativnih učnih okolij. Cilj projekta je bil predvsem

srečujemo. spodbujanje strokovnih delavcev v vzgoji in izobraževanju v

razvoj prožnejših oblik učenja s pomočjo rabe digitalnih

KLJUČNE BESEDE tehnologij. [1]

Digitalne kompetence, Inovativna pedagogika 5.0, sodobni Z izvajanjem projekta smo bili strokovni delavci deležni

koncepti poučevanja različnih strokovnih izobraževanj, ki so bila glede na željo ter

različno raven znanj organizirana s strani projektnega tima. Na

ABSTRACT šoli je bilo zaznati večje zanimanje za sodobne koncepte

The paper presents the work at the Prežihov Voranca poučevanja, posledično je bilo opaziti tudi večjo motivacijo za

Elementary School, which joined the Innovative Pedagogy 5.0 učenje pri učencih. V šolskem letu 2023/2024 se je šola

project in the 2023/24 school year. The project is based on the priključila projektu Inovativna pedagogika 5.0.

development of modern teaching that is supported by digital V prispevku je predstavljeno, kako se kot vodja projekta

technologies and is focused on the development of digital Inovativna pedagogika 5.0 soočam z novimi izzivi, novimi

competences and basic content of RIN. The challenges we face pridobljenimi znanji in spretnostmi ter kako to prenašam med

at school and how we overcome them are also presented. strokovne delavce in učence.



history lesson that developed the competence of copyright and digitalnih tehnologij, ki jih učenci s pridom in spretno uporabljajo. Zavedamo se pozitivnih kot tudi negativnih posledic, license in more detail is also presented. When preparing a A concrete example of a lesson that was conducted in a Učitelji se tako kot družba soočamo s hitrim napredkom



modern lesson, we as teachers need to think holistically, ki nam jih napredek na tem področju prinaša. Kot izobraževalci smo se v prvi vrsti dolžni z izzivom napredka soočiti, saj smo incorporating knowledge, skills and skills that are supported by tisti, ki lahko našim učencem predstavimo učinkovito rabo digital. We need to focus on the fact that students need to become digitalnih tehnologij in jih učimo njihove varne rabe. an active member of the learning process and not just those who

receive the teacher's knowledge. These are the challenges of

teachers that we can solve together and slowly eliminate the fears

we face in schools.



razvoj temeljinh vsebin RIN pri učečih se. Razvojni timi in

učiteljski zbori vključenih šol pripravljajo inovativne učne

scenarije skozi katere se predstavljajo inovativni pristopi k

sodobnejšemu učnemu procesu, ki učeče se pripravlja na

učinkovito soočanje s sodobno družbo.



Slika 1: Logo projekta Inovativna pedagogika 5.0 – 2.3 Delovanje projekta na OŠ Prežihovega

Inovativna šola. Voranca Maribor

Na OŠ Prežihovega Voranca Maribor smo v šolskem letu

2 PREDSTAVITEV PROJEKTA 2023/2024 pričeli sodelovati v projektu Inovativna pedagogika

Inovativna pedagogika 5.0 je projekt, ki se oddvija v hitro 5.0. Najprej je bila oblikovana razvojna skupina v zavodu, ki

spreminjajočem svetu. Projekt odgovarja na potrebe sodobne vključuje ravnateljico, pomočnika ravnateljice, koordinatorico

družbe, ki od posameznika zahteva razvoj kompetenc RIN, koordinatorico ROID ter koordinatorico projekta. Naloga

ustvarjalnosti, kritičnega razmišljanja, reševanja problemov in razvojnega tima je celostno izvajati naloge znotraj zavoda, ki so

veščin, ki so povezane z razvojem in uporabo digitalne ključne za dosego namena projekta.

tehnologije. Hkrati projekt poudarja razvoj sodelovanja, V mesecu septembru 2023 sem se skupaj z ravnateljico in še

usklajevanja in učinkovitega komuniciranja v družbi. Vse to pa enim članom šolskega tima udeležila uvodnega srečanja v

pomeni, da se mora izobraževalni proces premakniti od projetku Inovativa pedagogika 5.0. Po navodilih iz sestanka sem

ustaljenih vzorcev ter omogočiti razvijanje digitalnih kompetenc, v Forms anketi izrazila svoj inters za sodelovanje v posamezni

temeljnih vsebin RIN, kritičnega mišljenja, kreativnosti in razvojni skupini. V prvem šolskem letu smo določili dva

ustvarjalnosti tako strokovnih delavcev kot tudi učencev. [2] inovativna oddelka, kjer so učitelji razvijali digitalne

V prvi fazi projekta so bili v vsakem vzgojno- kompetence. Glede na izbrana inovativna oddelka je razvojni tim

izobraževalnem zavodu poleg ravnatelja vključeni še vsaj 4 k sodelovanju povabil tudi strokovne delavce, ki so bili

strokovni delavci. V prvem letu je bilo tako vključenih več kot oblikovani v projektni tim.

200 vodstvenih in strokovnih delavcev. V drugem letu bodo Učitelje na šoli, ki so tudi člani projektne skupine, sem na

vključeni dodatni inovativni oddelki ter dodatnih vsaj 300 krajšem seznanitvenem sestanku seznanila s cilji projekta in

strokovnih delavcev. V letu 2026 se ne načrtuje vključevanja nalogami v projektu. Na šolski spletni stani sem poskrbela za

dodatnih strokovnih delavcev temveč dvig kakovosti dela že objavo projekta.

vključenih. [3] V mesecu oktobru sem sklicala srečanje razvojnega tima,

namen je bil seznanitev z nalogami projekta in razmislek o

2.1 prirpavi načrta delovanja na šoli. V šolski spletni učilnici sem Organizacija projekta

uredila posebno mapo, ki je namenjena projektu, vanjo sem

Organizacija projekta se deli na projektni svet in razvojne

dodala vse učitelje v projektni skupini, prav tako razvojni tim.

skupine. Projektni svet usmerja in potrjuje odločitve na nivoju

Moja naloga kot koordinatorica projekta na šoli je bila tudi

celotnega projekta. Sestavljen je iz vodij posameznih razvojnih

prevzem uredništva in objavljanje na Facebook strani projekta ter

skupin. Vodi in usmerja ga vodja projekta, dr. Andrej Flogie. [3]

seznanjanje projektnega tima z novostmi. Člane projektne

Razvojne skupine predstavljajo jedro razvojno

skupine sem usmerila tudi do spletne učilnice E-skupnost

raziskovalnega projekta. Vsako razvojno skupino sestavljano

učiteljev v projektu Inovativna pedagogika.

predstavnik razvojne institucije, predstavnik vodstva šole ter

Kot pomembno nalogo semi mela tudi pripraviti akcijski

inovativni strokovni delavci. Skupaj je v delo razvojnih skupin

načrt medpredmetnih povezovanj s poudarkom na razvoju

vključenih 140 strokovnjakov in znanstvenikov. Delo razvojnih

digitalnih kompetenc ter na podlagi tega načrtovanja pripraviti

skupin usmerja, vodi in preverja projektni svet. Razvojne

tudi okvirni načrt medsebojnih hospitacij.

skupine so sledeče: katalog digitalnih kompetenc, skupina za

V mesecu novembru sem organizirala delovno srečanje

vzorčne učne scenarije, skupina za digitalno didaktiko, skupina

projektne skupine, kjer sem članom razdelila gradivo za

za usposabljanje, skupina za izvedbeni kurikul, skupina za

seznanitev z digitalnimi komeptencami. Delo sem organizirala

sodobne tehnologije, učeče se skupnosti, skupina za promocijo,

tako, da so učitelji po že vnaprej dogovorjenih medpredmetnih

skupina za uvajanje temeljnih vsebin RIN, skupina za

povezovanjih debatirali o možnostih razvijanja digitalnih

implementacijo in skupina za evalvacijo. [3]

kompetenc ter iskali možnosti vključevanja digitalnih tehnologij.



2.2 Kot koordinatorica projekta na šoli sem se mesečno Namen projekta

udeleževala videokonferenčnih srečanj, kjer smo se seznanjali z

Namen projekta je razvoj didaktike za dvig digitalnih novostmi in aktivnostmi v projektu ter izpostavljali izzive, s

kompetenc in temeljnih vsebin RIN pri učencih z uvajanjem katerimi smo se srečevali na šolah. Videokonferenčna srečanja

sodobnih pristopov in uporabo sodobnih digitalnih orodij. S smo imeli tudi v razvojni skupini, v katero sem bila dodeljena po

takšnim pristopom je ključni namen projekta učeče se pripraviti predhodno opravljenem anketnem vprašalniku v Formsu.

na izzive, ki jih prinaša sodobna družba in svet. [3] V mesecu januarju 2024 smo pričeli na šoli izvajati kolegialne

Konzorcij projekta, ki ga vodi Zavod Antona Martina hospitacije, ki sem jo kot prva izvedla sama. Po vsaki hospitaciji

Slomška, sestavlja 6 razvojnih inštitucij, 10 mentorskih šol in 30 smo opravili tudi poglobljeno pedagoško diskusijo, saj so člani

razvojnih šol po celotni Sloveniji. [3] projektnega tima opazovali razvijanje izbrane digitalne

Naloga vključenih v projekt je z razvojem celostnega pristopa kompetence. Prednost takšnega načina opravljanja hospitacij je

doseči razvoj diditalnih kompetenc pri učečih se in učiteljev ter

predvsem v tem, da se lahko osredotočeno diskutura o prednostih pregledu učnega načrta in DigComp 2.2 sem se odločila, da se

takšnega načina poučevanja kot tudi negativnih vidikih, ki se bomo postavili pod kompetenco avtorske pravice in licence.

lahko pojavijo. Pri odpiranju vrat učilnic sem spoznala, da imajo Oblikovala sem tudi namene učenja, ki sem jih učencem

strokovni delavci možnost spoznati nove koncepte poučevanja, predstavila. Nameni so bili naslednji: seznanili se bomo s

vidijo motiviranost učečih se, hkrati pa izgubljajo strah pred področjem življenja in napredka obdobju prve in druge

novostmi. industrijske revolucije, razumeli bomo, kako je tehnološki

Velika novost, ki jo je prinesel projekt, je bilo tudi pisanje napredek vplival na življenje različnih družbenih slojev, sklepali

učnih scenarijev. Učitelji v projektnem timu so se soočali z bomo o posledicah, ki jih je posamezen izum ali odkritje prinesel

velikimi izzivi, zato sem organizirala srečanje, kjer smo s v življenje, znali bomo pridobljene informacije strniti v smiselno

skupnimi močmi oblikovali učne scenarije, debatirali o celoto, znali bomo selekcionirati spletne vire glede na CClicence,

prednostih in slabostih takšnega vodenja učnega procesa. seznanili se bomo z relevantnimi spletnimi viri, seznanili se

Na šoli smo pričeli tudi z izvajanjem dodatnih izobraževanj bomo s spletnim orodjem Oblak365 in v njem ustvarjati

za strokovne delavce, ki so bili organizirani v ovkuru Inovativne dokumente ter jih deliti z ostalimi.



bil predvsem v seznanjanju z novostmi v digitalni tehnologiji, 1.1.2 Preverjanje predznanja pedagogike 5.0 kot tudi znotraj šole. Namen theh izobraževanj je



prav tako pa smo s tem želeli zmanjšati strah strokovnih delavcev Z računalničarko sva uro pričeli tako, da sva preverili predznanje učencev o avtorskih pravicah. Najprej sva jim pred vključevanjem sodobnih digitalnih orodij v pouk. predvajali posnetek posseted FAQs on copyright for teachers -

Observatory (europa.eu) in po posnetku opravili razgovor z

3 učenci. Skozi pogovor sva učence pripeljali do ugotovitve, da RAZVIJANJE DIGITALNIH KOMPETENC



PRI POUKU ZGODOVINE bomo pri projektnem delu spoznavali avtorska dela in poudarili avtorske pravice.

3.1 Nato so učenci razmišljali o lastnih izdelkih, ki so jih Digitalne kompetence

izdelovali pri tehniki in tehnologiji. Vodila sem jih skozi

Digitalne spretnosti za delo in življenje so med

razgovor in z njihovimi odgovori ugotavljala, kakšno je njihovo

najpomembnejšimi prednostnimi nalogami evropske politike.

znanje o avtorskih pravicah. V lažji razmislek sem jim postavila

Cilj strategije EU za pridobivanje digitalnih spretnosti in z njo

naslednja vprašanja: kdo je avtor izdelka, ali lahko tvoj sošolec

povezanih pobud je okrepiti digitalne spretnosti in kompetence

izdelek fotografira in ga uporabi v svojem referatu, ali lahko na

za uspešnejšo digitalno preobrazbo. [4]

spletu vzamem sliko ali besedilo in ga uporabim v referatu?

V okviru DigComp 2.2 so kompetence razvrščene na pet

Ker so učenci delali tudi s spletnimi viri, sem jim prebrala en

področij, in sicer: informacijska in podatkovna pismenost,

primer rabe digitalnih vsebin. Učenci so imeli nalogo, da

komuniciranje in sodelovanje, ustvarjanje digitalnih vsebin,

ugotovijo, kaj je pri prebranem primeru bilo napačno. Hkrati pa

varnost in reševanje problemov. Prva tri področja vključujejo

so podajali predloge, kako bi lahko oseba iz zgodbe uporabila

kompetence, ki so povezane z dejavnostmi in opravili. Četrto in

spletni vir in ne kršila avtorskih pravic. Učenci so skozi razgovor

peto področje pa vključujeta kompetence, ki so prečne, kar



pomeni, da jih potrebujemo pri izvajanju najrazličnejših pregledali že dogovorjene načine navajanja uporabljenih virov. kritično ovrednotili navajanje uporabljenih virov. Skupaj smo

aktivnosti povezanih z uporabi digitalnih tehnologij. [4]

Naš pogovor se je nadaljeval tako, da so učenci morali s



3.2 svojimi besedami pojasniti naslednje pojme: avtorsko delo in Praktičen primer razvoja digitalne

komepence pri pouku zgodovine avtorske pravice. S pridobljenim znanjem so učenci morali

dopolniti svojo pisno projektno nalogo. Poudarjali smo predvsem

Kot učiteljica zgodovine in koordinatorica projekta to, da morajo pri projektni nalogi pisati z lastnimi besedami in

Inovativna pedagogika 5.0 sem se odločila, da bom pri pouku v obvezno ob uporabi drugih gradiv navesti avtorje.

inovativnih oddelkih postavila delo učencev pod tretje področje

kompetenčnega okvirja, torej ustvarjanje digitalnih vsebin in



podrobneje razvijala kompetenco 3.3 avtorske pravice in licence.

Z učenci v osmem razredu sem dve izbirni temi iz učnega

načrta izvajala v obliki projektnega dela. Učna sklopa sta bila

prve kapitalistične velesile in izumi, ki so spremenili življenje

ljudi in življenje ljudi 19. stoletja. Projektno delo je bilo

načrtovano medpredmetno, tako so učenci pri pouku zgodovine

spoznavali posamezne teme, medtem ko so pri predmetu tehnika

in tehnologija praktično ustvarjali izdelke. V preteklih letih so

učenci takšen način dela pozitivno sprejeli, saj so pridobljeno Slika 2: Zapis učenke o avtorskih pravicah in izdelkh.

znanje pri zgodovini lahko uporabili tudi pri praktičnem delu v

delavnici. V šolskem letu 2023/2024 sem se odločila, da bom Delo je potekalo tako, da sva učence v oddelku razdelili na

takšen način dela dopolnila še z razvijanjem digitalne polovico, saj nam računalniška učilnica ne dopušča, da bi v njej

kompetence. delal celoten oddelek (28 učencev). Tako je polovica učencev

delo opravila v računalniški učilnici, preotanek razreda pa je delo

1.1.1 Nameni učenja



Med načrtovanjem učnega procesa sem si zamislila namene prenosne računalnike. opravljal v učilnici zgodovine, kjer so imeli na razpolago

učenja, ki jih želim skozi učni proces pri učencih razvijati. Po

V naslednjem koraku so bili učenci seznanjeni s CClicencami.

V ta namen so dobili povezavo do spletnih strani, ker so lahko

spoznali posamezne licence. Njihova naloga je bila ponovno

takšna, da so morali vsako licenco posebej s pomočjo obeh

učiteljic preoblikovati s svojimi besedami. Da so lahko praktično

preizkusili uporabo licenc, so dobili nalogo, da na spletu poiščejo

slikovno gradivo, ki se nanaša na njihovo projektno nalogo.

Učiteljici sta učencem pokazali, kje lahko najdejo podatek o Slika 4: Zapisan razmislek učencev o avtorskih pravicah.

avtorstvu slikovnega gradiva ter kakšno CClicenco ima

dodeljeno. 1.1.4 Razmislek učiteljice ob koncu projektnega dela

Delo je potekalo brez večjih težav, vidno je bilo, da so učenci

v tem šolskem letu že razvijali digitalne kompetence in težav z



deljenjem dokumentov ter urejanje spletne različice Worda niso



imeli. Še vedno ostaja nekaj učencev, ki potrebujejo pri delu

pomoč in usmerjanje. Učencem s posebnimi potrebami smo delo

organizirali tako, da so bile prisotne tudi izvajalke DSP, kar je

bila velika prednost pri izvajanju aktivnosti.

Pred pričetkom izvajanja seznanjanja z novimi vsebinami

sem ob pomoči ustnega razgovora ugotovila, da se učenci

zavedajo, da je fizičen izdelek/slika/risba intelektualna lastnina

avtorja. Manj učencev se je zavedalo, da je intelektualna lastnina

tudi prisotna na spletu in da je potrebno uporabiti tiste, za katere

imamo dovoljenje avtorja.

Slika 3: Zapis učenke o licencah. Učenci so z razmislekom pokazali, da so razumeli pomen

upoštevanja avtorskih pravic. Prav tako so pri iskanju fotografij

Učenci so se prijavili v spletno učilnico in na računalnike uporabljali pridobljeno znanje, torej kje preveriti CC licenco.

prenesli pripravljen dokument. Dokument so dopolnili s svojimi Ključnega pomena je, da takšno delo poteka v šolskem

podatki in novimi spoznanji, ki so jih pridobili skozi razgovor. prostoru, saj je za nekatere učence to preveik zalogaj in

Učenci so dokument shranili v Oblak 365 in ga delili z učiteljico. potrebujejo vodneje in usmerjanje učitelja.

Tega koraka z učenci nisem preverjala, saj sem vedela, da so

takšen način že uporabili pri predmetu matematika in sem se

lahko osredotočila na pomoč tistim učencem, za katere sem 4 ZAKLJUČEK

vedela, da še imajo težave pri deljenju dokumnetov. Z vstopom v projekt Inovativna pedagogika 5.0 je naša šola

naredila pomemben korak k sodobnemu pouku, pouku ki podpira



potekalo 6 šolskih ur, prednost je ta, da imamo blok ure in so tako vlogo, zato moramo biti med prvimi, ki si upamo preizkusiti nekaj novega. lahko intenzivno delali. Vse podatke so povzemali v deljen Kot koordinatorica projekta na šoli se zavedam svoje vloge dokument in sproti navajali uporabljeno literaturo ter vire. na šoli, saj sem že v prvem letu naletela na neodobravanje Učenci so v referat vključili vsaj dva slikovna gradiva. Pri izbiri vnašanja sprememb, predvsem med učitelji. Moja naloga je bila slikovnih gradiv so morali paziti na upoštevanje CC licenc. velikokrat usmerjena v pomirjanje in seznanjanje, saj so se tako Učenci so ob zaključku pisanja referata zapisali tudi učbeniku, knjižnih virih in spletnih virih. Delo učencev je Učenci so svojo temo individualno raziskovali, iskali so v spodbuja uporabo digitalnih tehnologij. V procesu sprememb izobraževalnega sistema imamo učitelji veliko in pomembno 1.1.3 razvoj digitalnih kompetenc med učečimi se in učitelji ter Uporaba lastnega znanja in razsikovalno delo



samorefleksijo o lastnem delu in napredku, v zapis razmisleka so strokovni delavci pomirili in poskusili narediti kakšen korak naprej. V ta namen se je kot zelo dobro izkazalo izvajanje vključili tudi razmišljanje o pomembnosti navajanja virov ter hospitacisjkih ur, kjer smo lahko učitelji skupaj razmišljali o upoštevanju avtorskih pravic. izvedeni učni uri, se usmerjeno pogovarjali o prednostih in Delo obeh učiteljic je bilo namenjeno predvsem usmerjanju izzivih, ki smo jih učitleji pri pripravi ur podprtih z digitalno in pomoči posameznikom, ki so pomoč potrebovali. Ker so tehnologijo imeli. učenci njihov dokument delili z menoj, je bilo preprosto odpreti Naša ugotovitev je bila tudi ta, da moramo biti pri njihov dokoument in jim zapisati povratno informacijo. Povratna vključevanju digitalnih tehnologij zelo previdni, saj učenci informacija učiteljice je bila tako sprotna in pisna, tako so lahko velikokrat njihovo uporabo postavijo na prvo mesto, pozabijo pa učenci hitro popravili pomanjkljivosti. na ostala znanja, ki jih bi jih naj v izobraževalnem procesu Učenci so z lastnim avtorskim izdelkom, ki je nastal pri pridobili. Učitelj, ki vključuje razvoj digitalnih kompetenc in z predmetu tehnika in tehnologija, predstavili novo pridobljeno digitalnimi tehnologijami dopolnjuje svoj pouk, mora proces znanje o izumitelju in izumu. Pri sami predstavitvi zastaviti premišljeno. Digitalne tehnologije morajo predstavljati zgodovinskega dela so morali vključiti tudi lasten razmislek o orodje, s katerim bodo učenci lahko na zanimivejši in aktivnejši tem, zakaj je pomemno, da priznamo delo nekoga drugega in ga način pridobivali znanja. Vključevanje razvijanja digitalnih pravilno tudi navedemo. kompetenc pa bo učencem nudilo priložnost, da se aktivneje

vključijo v učni proces.



Večje odobravanje pri takšnem načinu poučevanja vidim pri

učečih se, ki se dnevno srečujejo z digitalno tehnologijo. Zato je

naloga učiteljev, da novosti na tem področju spoznavamo skupaj

z učenci ter jim dajemo možnosti koristne rabe digitalnih

tehnologij v varnem okolju.

Naša naloga je, da učencem omogočimo, da izobraževalni

proces zaključijo tehnološko pismeni, saj je tehnologija sestavni

del številnih poklicnih področij. Vključevanje tehnologije v učni

proces lahko pomaga učencem razviti potrebne spretnosti za

prilagajanje sodobnemu svetu, ki vse bolj zahteva digitalne

spretnosti, kot so programiranje, analitično razmišljanje in

reševanje kompleksnih problemov.



ZAHVALA



Zahvaljujem se vodstvu šole, ki podpira in spodbuja vse moje

ideje o vključevanju sodobnih konceptov poučevanja. Prav tako

se zahvaljujem sodelavcem, ki se trudijo in mi pomagajo pri

uresničevanju ciljev projekta.



LITERATURA

[1] Dvig digitalne kompetentnosti. (29. 3. 2022) DOI: Dvig digitalne

kompetentnosti - Zavod RS za šolstvo (zrss.si) [13. 8. 2024].

[2] Družbene spremembe. DOI: Vizija našega dela – Inovativna pedagogika

5.0 (inovativna-sola.si) [13. 8. 2024].

[3] Andrej Flogie (Ur.). 2024. INOVATIVNA pedagogika 5.0 :

transformacija poučevanja za 21. stoletje: celovit razvoj digitalnih

kompetenc ter uvajanje temeljnih znanj računalništva in nformatike.

Maribor: Zavod Antona Martina Slomška.

[4] DigComp 2.2: Okvir digitalnih kompetenc za državljane. Z novimi primeri

rabe znanja, spretnosti in stališč. 2023. DigComp-2-2-Okvir-digitalnih-

kompetenc.pdf (zrss.si). Ljubljana: Zavod Republike Slovenije za šolstvo.

[13. 8. 2024].





Povezanost psihoedukativnega pristopa in ravni




pismenosti o anksioznih motnjah



The Relationship between a Psychoeducational Approach and



Literacy Levels in Anxiety Disorders



Nataša Dernovšček Hafner, Lilijana Šprah

natasa.dernovscek-hafner@zrc-sazu.si

lilijana.sprah@zrc-sazu.si

Družbenomedicinski inštitut ZRC SAZU

Ljubljana, Slovenija



workshops and an online platform to improve knowledge about

POVZETEK mental health, especially about mental disorders. The study

Motnje razpoloženja, kot so anksiozne in depresivne motnje, examined the level of knowledge about anxiety disorders among

so v današnji družbi zelo pogoste in predstavljajo velik the participants of the online and on-site workshops, who were

javnozdravstveni problem. Kljub njihovi razširjenosti ostajajo asked to complete an anxiety literacy questionnaire before and

slabo prepoznane in neustrezno obravnavane. Koncept after the psychoeducational topic on anxiety disorders. The

pismenosti o duševnem zdravju (PDZ) je ključnega pomena za results show that the average level of anxiety competence

ozaveščanje javnosti o duševnih motnjah, njihovem zdravljenju increased by 17% (p < 0.05) after the workshop compared to the

in zmanjševanju stigme. Program OMRA, ki se je v Sloveniji pre-workshop level, confirming the effectiveness of

začel izvajati leta 2017, je zasnovan na konceptu PDZ in psychoeducation as a tool to improve understanding of mental

vključuje izobraževalne delavnice ter spletno platformo za disorders. The OMRA programme continues to face challenges

izboljšanje pismenosti o duševnem zdravju, zlasti o motnjah in its efforts to reduce the social and economic burden of mental

razpoloženja. V raziskavi smo preverjali raven pismenosti o disorders by improving access to information and promoting

anksioznih motnjah pri udeleženkah in udeležencih spletnih oz. timely help-seeking. Its long-term goal is to improve mental

terenskih delavnic, ki so bili pred in po izvedbi health care and reduce the

psihoedukacijskega tematskega sklopa o anksiznosti povabljeni,

da izpolnijo vprašalnik pismenosti o anksioznosti. KEY WORDS

Rezultati kažejo, da se je povprečna raven pismenosti o Psychoeducation, mental health literacy, mood disorders,

aksioznosti po delavnici povečala za 17 % (p < 0,05) v primerjavi anxiety disorders, OMRA programme

s predhodno ravnijo pismenosti, kar potrjuje učinkovitost

psihoedukacije kot orodja za izboljšanje razumevanja duševnih

motenj. Izzivi programa OMRA na področju prizadevanja za 1 Uvod

zmanjševanje družbenega in ekonomskega bremena duševnih

motenj z izboljšanjem dostopa do informacij in s spodbujanjem 1.1 Motnje razpoloženja - motnje moderne dobe

pravočasnega iskanja strokovne pomoči, ostajajo. Njegov Duševne motnje so v splošni populaciji zelo pogoste in

dolgoročni cilj je boljša skrbi za duševno zdravje in zmanjšanju verjetnost, da bo nekdo obolel za kakšno od njih, je med 40 in

stigme, povezane s težavami v duševnim zdravju. 50 % [1]. Med najbolj razširjene duševne motnje uvrščamo

motnje razpoloženja, med njimi anksiozne (tesnobne) in

KLJUČNE BESEDE depresivne motnje, kar predstavlja tudi resen javnozdravstveni

Psihoedukacija, pismenost o duševnem zdravju, motnje problem tako zaradi trpljenja posameznikov in njihovih družin

razpoloženja, anksiozne motnje, program OMRA kakor znatnih stroškov, ki so povezani z njihovim zdravljenjem

in posledicami za družbo, kot so absentizem, dolgotrajna



ABSTRACT bolniška odsotnost, prezgodnja upokojitev [2]. Tako so na primer

skupni letni stroški za anksiozne motnje leta 2012 v EU dosegli

Mood disorders such as anxiety and depression are very

74,4 milijarde evrov [3]. Anksiozne motnje so skupina motenj

common in today's society and represent a major public health

(generalizirana anksiozna motnja, panična motnja, agorafobija,

problem. Although they are widespread, they remain poorly

specifične fobije, socialna fobija, obsesivno-kompulzivna

recognised and inadequately treated. The concept of mental

motnja) za katere je značilna trajna ali občasna izjemno

health literacy (MHL) is crucial to raise public awareness of

poudarjena tesnoba, ki posameznika ovira pri funkcioniranju, mu

mental disorders and their treatment and to reduce stigma. The

povzroča težave v vsakodnevnem življenju in slabša kakovost

OMRA programme, launched in Slovenia in 2017, is based on

življenja. Badelow in Michaelis [4] ocenjujeta, da ima anksiozne

the concept of mental health literacy and includes educational

motnje v življenju do 33,7 % prebivalstva. Opozarjata tudi na



slabo prepoznavanje anksioznih motenj in posledično slabše duševno zdravje, ki so namenjene laičnim javnostim. Poleg tega,

zdravljenje oz. njihovo nezdravljenje. Ni dokazov, da bi se da omogoča učinkovito posredovanje informacij, ljudi

razširjenosti anksioznih motenj v zadnjih letih spreminjala. opolnomoči, omogoča izbiro, ohranja dostojanstvo, zmanjšuje

Anksiozne motnje imajo kronični potek, z leti pa se njihova (samo)stigmo, povečuje upanje za izboljšanje ter povečuje skrb

izraženost omili. Zelo pogosto imajo ljudje po več različnih za lastno telesno in duševno zdravje [16]. Psihoedukativni

anksioznih motenj hkrati ali imajo različne anksiozne motnje po pristop se lahko uporabi v različnih modalitetah ko so pisna

enkrat ali nekajkrat v življenju. Pogosto pa so anksiozne motnje gradiva, spletne vsebine, skupinske delavnice, lahko pa tudi v

pridružene drugim duševnim motnjam [5]. obliki individualnega svetovanja. Zadnji izsledki obširne

pregledne raziskave [18] razkrivajo, da je psihoedukacija

1.2 Koncept pismenosti o duševnem zdravju najučinkovitejši ukrep za spodbujanje kognitivnih veščin in

Kljub temu, da so anksiozne motnje zelo pogoste, so še vedno pismenosti na področju duševnega zdravja.

slabo prepoznane. Če na eni strani raziskave razkrivajo, da



obstaja veliko sivo polje neustrezne obravnave oseb, ki poiščejo 2 Namen raziskave pomoč pri strokovnih službah zaradi težav v duševnem zdravju,

pa ne gre zanemariti dejstva, da je tudi odločitev posameznikov, Leta 2017 smo v Sloveniji začeli izvajati nov promocijsko

da poiščejo strokovno pomoč, kompleksen problem, na katerega preventivni program na področju duševnega zdravja OMRA

vpliva vrsta dejavnikov. Pomemben faktor, ki vpliva na odločitev (akronim programa Z večjo pismenostjo o duševnem zdravju do

o iskanju pomoči, je prepoznava, torej ali oseba dojema svojo obvladovanja motenj razpoloženja) v obliki in vsebini, ki v

težavo kot duševno motnjo. Izsledki raziskav kažejo, da znaten slovenskem prostoru do takrat še ni bila vzpostavljena [19][20].

delež laične javnosti ne prepoznava duševnih motenj [6]. Program je bil utemeljen na konceptu PDZ z osrednjim ciljem

Duševne motnje se pogosto enači s psihosocialnimi pritiski ali dvigniti raven pismenosti o motnjah razpoloženja (depresija,

stresom, šibkostmi osebe, ne pa z zdravstveno težavo [7][8], kar anksiozne motnje, bipolarna motnja razpoloženja) v laični

se odraža pri iskanju strokovne pomoči, zanimanju za zdravljenje javnosti. V ta namen so bila razvita različna orodja in pristopi

in prepričanjih o učinkovitosti zdravljenja [9]. (www.omra.si). Ena od aktivnosti opismenjevanja, ki se še vedno

Eden učinkovitejših pristopov ozaveščanja in informiranja izvaja, poteka tudi v obliki delavnic o prepoznavanju in rešitvah

prebivalstva o značilnostih in posledicah motenj razpoloženja težav, povezanih s stresom, depresijo, anksioznimi in drugimi

izhaja iz koncepta pismenosti o duševnem zdravju (angl. Mental duševnimi motnjami, kjer uporabljamo psihoedukacijski pristop.

Helath Literacy, v nadaljevanju PDZ) [10][11][12]. Koncept Delavnice izvajamo tako na terenu kakor tudi v spletni obliki.

PDZ opredeljuje pismenost kot skupek kognitivnih in socialnih Namen predstavljene raziskave je bil preveriti raven

veščin posameznika, ki omogočajo vedenje, kako pridobiti in pismenosti o anksioznosti pri udeleženkah in udeležencih

vzdrževati dobro duševno zdravje, razumevanje duševnih motenj delavnic pred in po njihovem obisku ter ugotoviti ali na raven

in njihovega zdravljenja, zmanjševanje stigme, povezane z pismenosti o anksioznosti vplivajo izbrane sociodemografske

duševnimi motnjami, znanje, kdaj in kje poiskati pomoč, spremenljivke.

razvijanje kompetenc, namenjenih izboljšanju skrbi za duševno

zdravje, in sposobnost samopomoči [10]. V novejših študijah

smo priča temu, da se koncept PDZ vse bolj usmerja v pismenost 3 Metodološki pristop



o specifičnih duševnih motnjah. V ta namen se tudi hitro 3.1 Opis vzorca raziskave razvijajo specifična orodja, s katerimi se ocenjuje poznavanje in

razumevanje določenih duševnih motenj, npr. pismenosti o V raziskavo so bili vključeni udeleženci in udeleženke 58

anksioznosti [13], depresiji [14], shizofreniji [15]. terenskih in 35 spletnih delavnic Stres povzroča stiske, okrepite

svoje moči!, ki so se izvajale v obdobju od oktobra 2018 do

1.3 Psihoedukacija in duševno zdravje marca 2024. Od skupno 2152 udeležencev in udeleženk delavnic,

Psihoedukacija je posebna oblika izobraževanja, njen osnovni jih je bilo v raziskavo vključenih 1384. Iz analize podatkov so

namen je pomagati ljudem s težavami v duševnem zdravju, da si bili izključeni tisti, ki so oddali nepopolne oz. neveljavne

pridobijo širše znanje in vedenje o svoji motnji. Psihoedukacija vprašalnike oz. jih sploh niso izpolnjevali. Podrobnejše

pomaga, da se pacienti in njihovi svojci naučijo uporabljati sociodemografske značilnosti udeležencev in udeleženk, ki so

različne strategije, ki jim bodo pomagale pri soočenju z duševno sodelovali pri preverjanju učinka delavnic na raven pismenosti o

boleznijo [16]. anksioznosti, so predstavljene v tabeli 1.



Psihoedukacija se je najprej uveljavila v psihiatrični 3.2 Pripomočki in postopek pridobivanja klinični praksi. Z njeno pomočjo se pacientom z duševnimi

motnjami in njihovim svojcem na sistematski način podaja podatkov

informacije o simptomih, etiologiji, predstavi naravo in Udeleženci in udeleženke spletnih oz. terenskih delavnic so

predviden potek psihične motnje, možnosti zdravljenja in bili pred in po izvedbi psihoedukacijskega tematskega sklopa o

rehabilitacije in dostopnost zdravstvenih in ostalih podpornih anksiznosti povabljeni, da izpolnijo vprašalnik pismenosti o

vladnih in nevladnih organizacij, informira o delovanju zdravil z anksioznosti. Na terenski delavnici so prejeli natisnjen

njihovimi stranskimi učinki, predstavi pomen drugih vzporednih vprašalnik, medtem ko jim je bila na spletni spletni delavnici bila

terapevtskih pristopov ter vlogo družine pri poteku zdravljenja Prvi sklop vprašanj se je nanašal na osnovne sociodemografske

[17]. Psihoedukacija pa se je dobro uveljavila tudi kot oblika podatke (spol, starost, izobrazba, kraj prebivanja, stan). Trditve,

preventivno promocijskih aktivnosti na področju skrbi za dobro

ki so sledile pa so bile namenjene preverjanju poznavanja Tabela 1: Sociodemografske značilnosti udeležencev in

značilnosti tesnobnih motenj ter njihovem zdravljenju. udeleženk raziskave (N = 1384).

Za namene ocenjevanja pismenost o anksioznosti je bil



uporabljen že uveljavljen vprašalnik iz tega področja A-lit

(Anxiety Literacy) (Australian National University: National

Institute for Mental Health Research [20][21]. Vprašalnik ima

dobre psihometrične značilnosti; notranja konsistentnost- α =

0,76; test-retest zanesljivost- r = 0,83, p = 0,003. Vprašalnik je

bil za namene uporabe v delavnicah OMRA preveden v

slovenščino.



Vprašalnik sestavlja 22 trditev o anksioznosti. Udeleženci 4 Rezultati

in udeleženke so pri vsaki trditvi označili odgovor PRAVILNO,



če so menili, da trditev drži, ali NAPAČNO, če so presodili, da 4.1 Psihometrične značilnosti vprašalnika

navedeno ne drži. V vseh treh vprašalnikih smo uvedli tudi pismenosti o anksioznosti kategorijo odgovora NE VEM, v kolikor o trditvi niso imeli

mnenja oz. niso poznali odgovora. S tem pristopom smo se želeli Koeficient notranje konsistentnosti (α) vprašalnika A-lit je

izogniti prisilnemu izbiranju odgovorov in ugibanju v primerih, znašal 0,77. Ker so podatki vprašalnika statistično pomembno

ko udeleženci in udeleženke ne bi bili prepričani o svojem odstopali od normalne porazdelitve (Kolmogorov-Smirnov test;

odgovoru. Pravilni odgovor se je točkoval z eno točko, nepravilni p < 0,01), smo v nadaljevanju analize podatkov uporabili

odgovori in odgovori NE VEM pa so šteli nič točk. Najvišji neparametrične statistične teste.

skupni rezultat na vprašalniku je znašal 22 točk, najmanjši pa nič Analiza podatkov je pokazala, da je bil delež udeležencev

točk. in udeleženk, ki je po delavnicah podal pravilne odgovore na

vprašalniku pismenosti primerljiv saj razlike niso bile statistično



3.3 pomembne (Mann-Whitneyjev U-test: p < 0,05) kar kaže na to, Vsebina in potek psihoedukativne delavnice Vsebino in potek tematskega sklopa delavnice o anksioznosti o anksioznosti da je vsebinski del delavnic ustrezno zastavljen in da so učinki primerljivi, ne glede na modaliteto izvedbe delavnic. Zato smo v

nadaljevanju opravili analizo učinkov delavnic na vzorcu

smo zasnovali tudi s pomočjo intervjujev in fokusnih skupin, kjer udeleženk in udeležencev terenskih in spletnih delavnic skupaj

smo s sogovorniki preverjali, o katerih vsebinah bi se želeli (N =1384). izobraževati, in v kakšni obliki naj bi bile podane. Udeležba na

delavnicah je bila prostovoljna, prijava pa je potekala na spletni 4.2 Vpliv delavnic na raven pismenosti o

platformi programa ali preko lokalnih deležnikov, pridruženih anksioznosti partnerjev programa OMRA. Vsebinsko je bila delavnica o

anksioznosti oblikovana tako, da smo v dvournem izobraževanju Raven pismenosti o anksioznosti se je pri udeležencih in

naslovili ključne tematike, ki se povezujejo z anksioznimi udeleženkah delavnic ocenjevala s pomočjo vprašalnika A-lit

motnjami. Podrobneje smo osvetlili problematiko stresa, njegove pred izvajanjem psihoedukacije o anksioznosti in po njej.

značilnosti in povezanost z motnjami razpoloženja kot njihovem Analiza podatkov je pokazala, da se je povprečna raven

sprožilcu in vzdrževalcu. Zatem so bile predstavljene značilnosti pismenosti o aksioznosti pomembno povečala po delavnici v

anksioznosti, vzroki in pristopi zdravljenja. Naslovili smo tudi primerjavi z ravnijo pismenosti pred njo (p < 0,05). Delež ljudi,

problematiko (samo)stigme in dileme samorazkrivanja v ki so imeli boljši rezultat po delavnici se je povečal za 16,75 %

zasebnem in delovnem okolju. (slika 1). Povprečno so imeli ljudje po delavnici 3,7 več

odgovorov pravilnih kot pred delavnico. 75,5 % ljudi je imelo



3.4 boljši rezultat po delavnici; 10,2 % je imelo enak rezultat pred in Statistična obdelava podatkov

po delavnici; 10,2 % je imelo enak rezultat pred in po delavnici.

Za statistično obdelavo podatkov je bil uporabljen

Opravili smo tudi dodatno analizo postavk vprašalnika A-

računalniški programski paket IBM SPSS Statistics 22.

lit (tabela 2), kjer smo preverjali katere postavke so udeleženci

Preverjanje normalnosti distribucije rezultatov je potekalo s

in udeleženke delavnic v največjem deležu označili z napačnim

pomočjo Kolmogorov-Smirnovega testa. Za namene preverjanja



njih, za primerjavo učinkov spletnih in terenskih delavnic in za je izstopala postavka Antidepresivi so učinkovit način preverjanje učinkov sociodemografskih spremenljivk na raven zdravljenja anksiozne motnje, kjer se je po delavnici povečal pismenosti je uporabljen neparametrični Mann-Whitneyjev U- delež oseb za 54,8 %, ki je pravilno odgovoril na postavko. razlik v ravni pismenosti o anksioznosti pred delavnicami in po na istih postavkah podali pravilni odgovor, se je povečal. Najbolj odgovorom pred izvedbo delavnic. Delež oseb, ki so po delavnici

test. V post hoc analizi povezanosti med izbranimi

sociodemografskimi spremenljivkami in dosežki na vprašalniku

A-lit je bil uporabljen neparametrični Spearmanov koeficient

korelacije (ρ).



omeniti nekaj omejitev, ki bi lahko vplivale tudi na naravo

izsledkov. Najprej: uporabljeno orodje pismenosti o anksioznosti

ni standardizirano za uporabo v slovenskem okolju in zato je pri

določenih postavkah lahko prišlo tudi do neenotnega

razumevanja njihovih vsebin in posledično pristranskega

odgovarjanja udeležencev. Nadalje: vzorec na katerem je bila

opravljena raziskava vsebuje določene pristranosti kot je večje

število žensk (86,6 %), prevladujočo visokošolsko izobrazbo

(52,4 %) in visok delež (74,5 %) sodelujočih, ki je imelo status

zaposlenih oseb. Delavnice smo izvajali tako na terenu kot v

spletni obliki. Statistična analiza je sicer pokazala, da med njima

Slika 2: Primerjava povprečnih ravni pismenosti o ne prihaja do statistično pomembnih razlik na ravni pismenosti o

anksioznosti pred in po psihoedukacijski delavnici. anksioznosti. Vendar pa pri udeležencih in udeleženkah spletnih

delavnic ne moremo popolnoma izključiti »goljufanja«, kjer bi

4.3 vprašalnike namesto udeleženca ali udeleženke delavnic Vpliv sociodemografskih spremenljivk na

raven pismenosti o anksioznosti izpolnjeval nekdo drug, ali da so osebe bile priključene na

spletno delavnico, niso pa sledile njenim vsebinam. Spletno

Analiza podatkov je pokazala, da so nekatere izvajanje delavnic je pokazalo, da so se jih udeleževale

sociodemografske pomembno vplivale na raven pismenosti o predvsem osebe, ki so bile digitalno pismene, z dostopom do

anksioznosti. Starost se je izkazala kot pomemben dejavnik, ki spleta in pretežno mlajše od 50 let. vpliva na raven pismenosti. Post hoc analiza je pokazala, da je

povezava med starostjo in učinkom delavnic negativna (ρ = -

0,143, p < 0.001). Torej z višjo starostjo se učinek delavnice 5 Zaključek blago manjša. Tudi raven izobrazbe se je izkazala kot statistično

Duševne motnje predstavljajo veliko družbeno in ekonomsko

pomemben dejavnik, ki vpliva na raven pismenosti o anksioznih

breme in po ocenah strokovnjakov bo v prihodnje njihova

motnjah. Udeleženci s fakultetno izobrazbo so statistično

pojavnost pri odraslih v primerjavi z ostalimi kategorijami

pomembno dosegali boljše rezultate na vprašalniku A-lit v

bolezni še naraščala, prav tako pa se bodo povečevali tudi stroški,

primerjavi s tistimi s končano osnovno (p = 0,008), poklicno (p

povezani z njihovo obravnavo. Poročilo Health at a Glance:

= 0,011) ali srednjo šolo (p = 0,002) in tistimi z višjo šolo (p =

Europe 2018 (OECD/EU, [22] razkriva, da so gospodarski in

0,013). Status udeležencev in udeleženk psihoedukacijskih

družbeni stroški duševnih motenj ocenjeni na več kot 4 % BDP

delavnic je pomembno vplival na dosežke pismenosti o

ali več kot 600 milijard evrov (ocena za Slovenijo je znašala 1,6

anksioznosti saj so pokazale statistično pomembne razlike med

milijarde evrov oz. 4,13 % BDP). Anksiozne motnje se v EU

upokojenimi in ostalimi skupinami. Upokojeni ljudje so izkazali

pojavljajo pri približno 34 % populacije, kar jih uvršča med

nižjo raven pismenosti v primerjavi z zaposlenimi (p < 0,001) in

najpogostejše duševne in vedenjske motnje.

nezaposlenimi (p < 0,001) ter višjo v primerjavi s

Kljub težkemu bremenu duševnih motenj za posameznike,

dijaki/študenti(p = 0.009).

gospodarstvo in družbo, se mu je mogoče izogniti oz. s

Čeprav so v povprečju ženske dosegale boljši rezultat na

pravočasno prepoznavo in ukrepanjem, situacijo vsaj omiliti.

vprašalniku po delavnici, razlika med spoloma ni bila statistično

Številne evropske države imajo že vzpostavljene politike in

pomembna (p = 0.69). Tudi kraj prebivališča (mesto/podeželje)

programe za obravnavo duševnih motenj pri različnih starostih in

ni pomembno vplival na raven pismenosti o anksioznosti (p =

ciljnih skupinah, vendar je za ohranjanje in spodbujanje

0,246) kljub temu, da se izkazoval trend boljših rezultatov pri

duševnega zdravja mogoče storiti še veliko več. Duševne motnje

udeležencih in udeleženkah iz podeželja.

so od vseh kroničnih nenalezljivih bolezni tiste, o katerih imata

laična in strokovna javnost najbolj različna stališča in prepričanja.

Tabela 2: Predstavitev postavk vprašalnika A-lit, na katere Predsodki in napačne informacije so bolj pravilo kot izjema, kar

so udeleženci pred psihoedukacijskimi delavnicami negativno vpliva tudi na pogostost in načine iskanja pomoči. To

najpogosteje odgovorili napačno oz. z možnostjo ne vem. izpostavljajo tudi raziskave, ki poudarjajo, da približno polovica

ljudi, ki potrebuje zdravljenje zaradi duševnih motenj, ne prejme

nikakršne zdravstvene pomoči [23].



V Sloveniji smo v letu 2017 pod okriljem Ministrstva za

zdravje začeli izvajati inovativni program OMRA: Z večjo

pismenostjo o duševnem zdravju do obvladovanja motenj

razpoloženja. Program je zasnovan večdisciplinarno in vključuje

aktualne smernice na področju izvajanja promocije duševnega

zdravja in opismenjevanja o duševnih motnjah, ki so trenutno

najpogostejše – motnje razpoloženja. Cilj programa je povečati

dostopnost do različnih informacij, povezanih z zgodnjim



4.4 prepoznavanjem motenj razpoloženja in razpoložljivih oblikah Omejitve raziskave

(samo)pomoči s pomočjo spletne izobraževalne platforme in

Rezultati raziskave so se pokazali skladnost s primerljivimi preko izvajanja psihoedukativnih delavnic. Izvajanje delavnic, ki

mednarodnimi raziskavami s tega področja, vendar vseeno velja so zasnovane na konceptih psihoedukacije in pismenosti o



posameznih motnjah razpoloženja (anksioznost, depresija in [13] Adrian Furnham and Corinne Lousley. 2013. Mental Health Literacy and

bipolarna motnja razpoloženja), se je izkazalo za zelo učinkovito the Anxiety Disorders. Health 5, no. 3: 521–531.

[14] Kathleen Griffiths, Helen Christensen, Anthony F. Jorm, Karen Evans and

obliko dvigovanja obveščenosti o duševnem zdravju, saj se raven Christine Groves. 2004. Effect of Web-Based Depression Literacy and

pismenosti npr. o anksioznosti poveča pri udeležencih in Cognitive-Behavioural Therapy Interventions on Stigmatizing Attitudes

to Depression: Randomized Controlled Trial. The British Journal of

udeleženkah delavnic v povprečju za okrog 17 %. Psychiatry 185: 342–349.

Visoka raven pismenosti o duševnem zdravju je tako za [15] Liliana M Loureiro, Anthony F. Jorm, Rita A. Oliveira, Ana M. Mendes,

družbo kot posameznika ključnega pomena, saj bistveno José C. dos Santos, Maria A. Rodrigues, and Célia S. Sousa. 2015. Mental

Health Literacy about Schizophrenia: A Survey of Portuguese Youth.

pripomore k pravočasnemu iskanju ustrezne strokovne pomoči, Early Intervention in Psychiatry 9, no. 3: 234–241.

zniževanju stigme in preventivni naravnanosti populacije [24]. [16] Joseph Walsh. Psychoeducation in Mental Health. 2009. Oxford

University Press.

Raziskave poročajo, da je v okoljih z visoko stopnjo pismenosti [17] Nicholas A. Cummings and Janet L. Cummings. 2008. Psychoeducation



o duševnem zdravju skrb za dobro duševno zdravje boljša, več je in Conjunction with Psychotherapy Practice. In Practical Resources for the Mental Health Professional, Evidence-Based Adjunctive Treatments, iskanja strokovne pomoči, opažena pa je tudi povečana edited by William T. O'Donohue and Nicholas A. Cummings, 41–59.

implementacija učinkovitih preventivno promocijskih strategij, Academic Press.

kar na daljši rok zmanjšuje socialno in ekonomsko breme, ki ga [18] Gonzalo Salazar de Pablo, Anna De Micheli, Dominic H. Nieman,

Christoph U. Correll, Lars V. Kessing, Angelika Pfennig, Andreas

povzročajo duševne motnje pri posamezniku in v družbi Bechdolf, et al. 2020. Universal and Selective Interventions to Promote

[10][25][26]. Oblikovalci programa OMRA zato upamo, da bo Good Mental Health in Young People: Systematic Review and Meta-

Analysis. European Neuropsychopharmacology: The Journal of the

tudi Slovenija odločneje zakorakala v smeri povečane in European College of Neuropsychopharmacology 41: 28–39.

pravočasne skrbi za dobro duševno zdravje in da se bo poleg [19] Lilijana Šprah. 2019. Uvajanje inovativnega programa pismenosti o

izobraževalne platforme OMRA vzpostavilo še več preventivno motnjah razpoloženja v Sloveniji: rezultati in učinki programa. V Z več

znanja o motnjah razpoloženja do izhodov iz labirintov: izkušnje in

promocijskih programov, ki bodo z dokazi podprtimi razmisleki. (1st izd.). Lilijana Šprah, urednik. (1st izd.). Ljubljana:

intervencijami spodbujali višjo raven pismenosti o duševnem Založba ZRC, ZRC SAZU, Ljubljana, 51-73.

[20] Australian National University: National Institute for Mental Health

zdravju v različnih ciljnih skupinah. Research. 2014. The Anxiety Literacy Questionnaire (A-lit). Retrived

August 16, 2024 from https://nceph.anu.edu.au/research/tools-



AVTORSTVO resources/anxiety-literacy-questionnaire-lit.

[21] Amelia Gulliver, Kathleen M. Griffiths, Helen Christensen, Andrew

Obe avtorici sta enakovredno prispevali k pripravi članka. Mackinnon, Alison L. Calear, Alison Parsons, Kylie Bennet, Philip J.

Batterham, Rosanna Stanimirovic. 2012. Internet-Based Interventions to

Promote Mental Health Help-Seeking in Elite Athletes: An Exploratory

LITERATURA Randomized Controlled Trial. Journal of Medical Internet Research, 14:

69.

[1] Ronald C. Kessler, Matthias Angermeyer, James C. Anthony, Ron de [22] OECD/EU. Health at a Glance: Europe 2018: State of Health in the EU

Graaf, Koen Demyttenaere, et al. 2007. Lifetime Prevalence and Age-of- Cycle. 2018. OECD Publishing, Paris. Retrived August 18, 2024, from

Onset Distributions of Mental Disorders in the World Health https://www.oecd-ilibrary.org/docserver/health_glance_eur-2018-

Organization's World Mental Health Survey Initiative. World Psychiatry en.pdf?expires=1565290964&id=id&accname=guest&checksum=D82E0

6, no. 3, 168–176. 2BC84BBFE365E77757EE37019D6.

[2] Hans-Ulrich Wittchen, Frank Jacobi, Jürgen Rehm, Anders Gustavsson, [23] Tessa Roberts, Georgina Miguel Esponda, Dzmitry Krupchanka, Rahul

Mikael Svensson, Bengt Jönsson, et al. 2011. The Size and Burden of Shidhaye, Vikram Patel, Sujit Rathod. 2018. Factors Associated with

Mental Disorders and Other Disorders of the Brain in Europe 2010. Health Service Utilisation for Common Mental Disorders: A Systematic

European Neuropsychopharmacology 21, no. 9: 655–679. Review. BMC Psychiatry 18: 262.

[3] Jes Olesen, Anders Gustavsson, Mikael Svensson, Hans-Ulrich Wittchen, [24] Anthony F Jorm. 2000. Mental Health Literacy: Public Knowledge and

and Bengt Jönsson on behalf of the CDBE2010 study group in the Beliefs About Mental Disorders. British Journal of Psychiatry 177: 396-

European Brain Council. 2012. The Economic Cost of Brain Disorders in 401.

Europe: Economic Cost of Brain Disorders in Europe. European Journal [25] Amy J Morgan, Anna Ross, Nicola J Reavley. 2018. Systematic Review of Neurology 19, no. 1: 155–162. and Meta-Analysis of Mental Health First Aid Training: Effects on

[4] Borwin Bandelow and Sophie Michaelis. 2015. Epidemiology of Anxiety Knowledge, Stigma, and Helping Behaviour." PLoS One 13, no. 5):

Disorders in the 21st Century. Dialogues in Clinical Neuroscience 17, no. e0197102. DOI: 10.1371/journal.pone.0197102 3: 327–335. [26] Chiara Samele. 2016. Increasing Momentum in Prevention of Mental

[5] Norman Sartorius, T. B. Ustun, J. A. Costa e Silva, et al. 1993. An Illness and Mental Health Promotion Across Europe. BJPsych

International Study of Psychological Problems in Primary Care: International.

Preliminary Report from the World Health Organization Collaborative

Project on Psychological Problems in General Health Care. Archives of

General Psychiatry 50, no. 10: 819–824.

[6] Matthias C. Angermeyer and Susanne Dietrich. (2006) Public Beliefs

about and Attitudes towards People with Mental Illness: A Review of

Population Studies. Acta Psychiatrica Scandinavica 113: 163–179.

[7] Laura Altweck, Tara C. Marshall, Nadia Ferenczi and Katharina

Lefringhausen. 2015. Mental Health Literacy: A Cross-Cultural Approach

to Knowledge and Beliefs about Depression, Schizophrenia, and

Generalized Anxiety Disorder. Frontiers in Psychology 6: 1272.

[8] Herbert Matschinger and Matthias C. Angermeyer.1996. Lay Beliefs

about the Causes of Mental Disorders: A New Methodological Approach.

Social Psychiatry and Psychiatric Epidemiology 31: 309–315.

[9] Amy J. Morgan, Nicola J. Reavley and Anthony F. Jorm. 2014. Beliefs

about Mental Disorder Treatment and Prognosis: Comparison of Health

Professionals with the Australian Public." Australian & New Zealand

Journal of Psychiatry 48, no. 5: 442–451.

[10] Anthony F. Jorm. 2012. Mental Health Literacy: Empowering the

Community to Take Action for Better Mental Health. American

Psychologist 67, no. 3: 231–243.

[11] Stan Kutcher, Yifeng Wei and Christopher Coniglio. 2016. Mental Health

Literacy: Past, Present, and Future. Canadian Journal of Psychiatry 61, no.

3: 154–158.

[12] Don Nutbeam. 2000. Health Literacy as a Public Health Goal: A Challenge

for Contemporary Health Education and Communication Strategies into

the 21st Century. Health Promotion International 15, no. 3: 259–267.





Kaj lahko postorimo za svojo varnost na osebnem




računalniku?



What Can We Do for Our Security on a Personal Computer?



prof. dr. Saša Divjak

Univeza v Ljubljani, Fakulteta za računalništvo

in informatiko

sasa.divjak@fri.uni-lj.si



POVZETEK ABSTRACT

Cilj predavanja je udeležence opremiti z znanjem in The aim of the lecture is to equip participants with knowledge

praktičnimi nasveti za izboljšanje varnosti na njihovih osebnih and practical tips for improving security on their personal

računalnikih, s čimer bodo zmanjšali tveganje za izgubo computers, thereby reducing the risk of data loss and

podatkov in kibernetske napade. Predavanje se osredotoča na cyberattacks. The lecture focuses on the key steps individuals can

ključne korake, ki jih lahko posamezniki sprejmejo za zaščito take to protect their data and privacy while using a personal

svojih podatkov in zasebnosti pri uporabi osebnega računalnika. computer. It is based on personal experiences. The emphasis will

Predavanje temelji na osebnih izkušnjah. Poudarek bo na be on the following topics:

naslednjih temah: Basics of Cybersecurity: Understanding the importance of

Osnove kibernetske varnosti: Razumevanje pomena varnosti online security, identifying potential threats (e.g., viruses,

na spletu, prepoznavanje potencialnih groženj (npr. virusi, malware, phishing), and how these threats affect computers and

zlonamerna programska oprema, phishing) in kako te grožnje personal data.

vplivajo na računalnik in osebne podatke. Updates and Security Settings: The importance of regularly

Posodobitve in varnostne nastavitve: Pomembnost rednega updating the operating system and software, as updates often

posodabljanja operacijskega sistema in programske opreme, saj include fixes for security vulnerabilities. A review of the basic

posodobitve pogosto vključujejo popravke za varnostne security settings that should be enabled on a personal computer.

ranljivosti. Pregled osnovnih varnostnih nastavitev, ki jih je treba Using Strong Passwords: Practical advice for creating and

omogočiti na osebnem računalniku. managing strong, unique passwords, as well as using two-factor

Uporaba močnih gesel: Praktični nasveti za ustvarjanje in authentication for added protection.

upravljanje močnih, unikatnih gesel ter uporaba dvofaktorske Safe Use of the Internet and Applications: How to identify

avtentikacije za dodatno zaščito. malicious websites, avoid downloading unverified programs, and

Varna uporaba spleta in aplikacij: Kako prepoznati the importance of downloading software from official sources.

zlonamerne spletne strani, izogibanje prenosu nepreverjenih Data Backup: The reasons why regular data backups are

programov in pomembnost prenosa programske opreme iz important and various methods for backing up data, such as cloud

uradnih virov. services and external hard drives.

Varnostno kopiranje podatkov: Razlogi, zakaj je pomembno Protection against Social Engineering: Recognizing social

redno varnostno kopirati podatke ter različne metode za engineering techniques used by criminals to obtain sensitive

varnostno kopiranje, kot so oblačne storitve in zunanji trdi diski. information and how to protect oneself from them.

Zaščita pred socialnim inženiringom: Prepoznavanje tehnik

socialnega inženiringa, s katerimi kriminalci pridobijo občutljive

informacije, in kako se pred njimi zaščititi.



Digitalna pismenost v osnovnošolskem izobraževanju



Digital Literacy in Primary Education



Aleksandar Dončev Eva Jereb

Ministrstvo za vzgojo in izobraževanje, Ljubljana, Univerza v Mariboru, Fakulteta za organizacijske vede,

Slovenija Kranj, Slovenija

aleksandar.doncev@gov.si eva.jereb@um.si



digitalizacijo izobraževalnih procesov in nadgradnjo

POVZETEK izobraževalnega sistema [1].

Prispevek obravnava pomen digitalne pismenosti in veščin v Ta preobrazba je tesno povezana z razvojem tehnologij,

sodobnem izobraževalnem sistemu. V ospredju je potreba po digitalizacijo procesov in virov, ter zahtevo po stalnem

digitalni transformaciji osnovnošolskega izobraževanja, ki posodabljanju na najnovejše tehnološke rešitve [2].

vključuje uporabo sodobnih tehnologij in digitalnih virov v V tem kontekstu so se tudi osnovne šole odločile za digitalno

učnem procesu. Opisan je pomen izobraževanja učiteljev in preobrazbo, pri čemer so vključile tehnologije v svoje

učencev o varni in odgovorni rabi digitalnih orodij ter izobraževalne, administrativne in komunikacijske procese.

informiranja o kritičnem presojanju digitalnih medijev. Različne izobraževalne platforme in elektronske komunikacije,

Prispevek prav tako obravnava izzive in prednosti digitalne kot so e-pošta in sporočila, so postale nepogrešljiv del vzgojno-

pismenosti, ter opozarja na nevarnosti, kot so zloraba osebnih izobraževalnega procesa [3].

podatkov in spletno nadlegovanje. Izpostavljen je sistematičen S tem so učitelji in učenci pridobili dostop do sodobnih

pristop k izobraževanju o digitalni pismenosti, ki presega zgolj tehnologij in digitalnih virov, kar je spodbudilo šole, da

tehnične vidike in vključuje širši družbeni in kritični pogled na prenovijo svoje tradicionalne izobraževalne modele. Vodenje in

digitalne medije. zagotavljanje digitalne preobrazbe v osnovnošolskem

izobraževanju pa zahteva razvoj spretnosti tako učiteljev kot



KLJUČNE BESEDE učencev, da lahko učinkovito izkoristijo napredne tehnološke rešitve [4]. Digitalna pismenost, digitalne veščine, digitalna medijska

pismenost, digitalna bralna pismenost, osnovnošolsko



izobraževanje 2 DIGITALNA PISMENOST, VEŠČINE IN

KOMPETENCE

ABSTRACT

Nastanek in širitev socialnih omrežij v digitalni dobi je

The paper discusses the importance of digital literacy and

privedla do družbenih sprememb, ki imajo velik vpliv na

skills in the modern education system. In the foreground is the

področje osnovnošolskega izobraževanja. Programi za

need for digital transformation of primary school education,

izobraževanje učiteljev se soočajo s izzivom priprave bodočih

including modern technologies and digital resources in the

učiteljev, da lahko kritično interpretirajo digitalne medije.

learning process. The importance of educating teachers and

Učitelji morajo biti v tem delu odgovorni in kritični, če želijo iz

students about the safe and responsible use of digital tools and

otrok razviti državljane, ki so dobro obveščeni. Pri tem se v

informing them about the critical assessment of digital media is

ospredju postavlja dilema, ali so učitelji v skladu z razvojem

described. The paper also discusses the challenges and benefits

digitalnih medijev sposobni kritičnega mišljenja [5].

of digital literacy and warns of dangers such as misuse of

Uporabniki digitalne tehnologije morajo za učinkovito učenje

personal data and online harassment. A systematic approach to

s pomočjo tehnologije že zgodaj pridobiti znanje in veščine.

digital literacy education is highlighted, which goes beyond mere

Eden od izzivov je, da morajo uporabniki pridobivati nove in

technical aspects and includes a broader social and critical view

razvijati obstoječe veščine za delovanje in smiselno uporabo

of digital media.

tehnologije. Digitalna preobrazba osnovnošolskih učilnic



KEYWORDS presega zgolj uporabo orodij in programske opreme, vpliva na

znanje in veščine njenih uporabnikov, učiteljev in učencev ter

Digital literacy, digital skills, digital media literacy, digital didaktiko in procese izobraževalnih ustanov [6].

reading literacy, primary school education Primeri uporabe izobraževalne tehnologije se lahko najdejo v

digitalnih pedagogikah, tehnoloških modelih ter prilagodljivih,



1 odprtih in pametnih tehnologijah. Vsi deležniki osnovnošolskih UVOD

izobraževalnih ustanov morajo za učinkovito digitalno učenje

Tehnološki napredek je na podlagi digitalne preobrazbe tako skozi prizmo digitalne preobrazbe vseskozi slediti inovacijam na

v javnem kot tudi v zasebnem sektorju omogočil podlago za tem področju, predvsem pa morajo biti pozorni in pripravljeni,

da sledijo spremembam in implementirajo novosti, ki jih prinaša



digitalna preobrazba osnovnošolskega izobraževanja. Ker se naslavljati pomanjkanje digitalne pismenosti med otroci in

tehnologija nenehno razvija, nove spretnosti ne bodo ostale mladostniki in jih pripraviti na izzive sodobne digitalne družbe

statične, ampak se bodo nenehno spreminjale [7]. ter opolnomočiti za samostojno, odgovorno in kritično uporabo

digitalnih medijev [14].



3 DIGITALNA MEDIJSKA PISMENOST

Zagovorniki sodobnega izobraževanja, ki je vse bolj pod 4 DIGITALNA BRALNA PISMENOST

vplivom digitalne transformacije, poudarjajo pomen krepitve Nasloviti izzive, ki jih prinaša digitalna preobrazba

digitalne pismenosti tako otrok kot tudi učiteljev v slovenskega sistema osnovnošolskega izobraževanja, omogoča

osnovnošolskem izobraževanju. Tradicionalne metode se neprekinjen prehod od tradicionalnega k modernemu

pogosto osredotočajo zgolj na praktično uporabo digitalnih izobraževalnemu sistemu in s tem izboljšanje digitalne bralne

orodij, ne upoštevajo pa njihovega širšega vpliva. Vedno bolj se pismenosti.

priznava, da na digitalne tehnologije ne bi smeli gledati le kot na Tradicionalni sistem osnovnega šolskega izobraževanja, ki ga

tveganja ali priložnosti, ampak skozi celovito lečo, ki vključuje poznamo danes, se sooča z spremembami in reformami.

digitalno pismenost in spletno varnost. Namesto ozke Povečuje se povpraševanje po novem znanju in večji digitalni

osredotočenosti na digitalne prakse je bistveno razviti kritičen bralni pismenosti v modernem izobraževalnem sistemu. V

pristop k razumevanju vpliva digitalnih medijev v zadnjih letih se je s predstavitvijo digitalnih tehnologij v

osnovnošolskem izobraževanju. To vključuje posodobitev slovenskih osnovnih šolah znatno povečalo število digitalnih

obstoječih pedagoških okvirov in strategij za soočanje z novimi učnih gradiv, ki so na voljo učiteljem in učencem. Hkrati so se

izzivi, ki jih prinašata digitalizacija in razvoj družbenih medijev pojavili specifični izzivi pri njihovi uporabi, kot so razlike med

[5]. učitelji, pomanjkanje usposabljanja učiteljev za delo z

Razmerje med tehnologijami in produkcijo digitalnih medijev digitalnimi učnimi gradivi ter pomanjkanje ustrezne opreme in

ni novo, vendar postaja bolj zapleteno, saj sedanje in prihajajoče naprav v šolah.

tehnologije preusmerjajo naše medijske prakse stran od Z zagotavljanjem celovitega zajema okvirja ocenjevanja

množične produkcije k različnim oblikam komunikacije. Obstaja branja in izvajanjem skupinskega prilagodljivega načrta je

potreba po novi fazi izobraževanja o digitalni medijski nastala raziskava PIRLS 2021, ki predstavlja najobsežnejše in

pismenosti predvsem v razredih osnovnih šol. Čeprav se številne najbolj kompleksno mednarodno ocenjevanje digitalne bralne

šole ukvarjajo s konceptom "digitalne tehnologije" v različnih pismenosti učencev doslej, ki je vključevalo 18 nizov besedil in

stopnjah, je poudarek predvsem na tehničnih veščinah. To je nalog ePIRLS. PIRLS 2021 je edino mednarodno ocenjevanje

razumljivo, saj učitelji v osnovnih šolah običajno nimajo izobraževalnih dosežkov, ki je uspešno zbralo podatke med

formalnega izobraževanja ali usposabljanja na področju digitalne motnjami v šolanju učencev v času epidemije Covid-19.

medijske pismenosti [8]. Mednarodna zbirka podatkov PIRLS 2021 zagotavlja izjemno

V ospredju je miselnost, da bi morala medijska pismenost bogat in dragocen vir podatkov za raziskovanje, ki so jih

otrok presegati zgolj frazo z dobronamernim pomenom, ki povzročili vplivi Covid-19 na digitalizacijo poučevanja in učenje

zahteva medijsko izobraževanje kot sistematičen program branja. Raziskava se osredotoča na ocenjevanje bralne

poučevanja in učenja za vse otroke v osnovnih šolah. Vendar to pismenosti otrok ob koncu 4. razreda osnovne šole, kar pomeni,

ne izključuje potrebe po regulaciji saj potrebujemo oboje. da so udeleženci raziskave v povprečju stari približno 10 let. Ta

Digitalno medijsko izobraževanje ni namenjeno zgolj ocenjevanja se izvajajo v rednih petletnih intervalih. Junija 2021

posamezniku, ki se uči spopadati z izzivi povečane rabe IKT gre je bilo tako izvedeno merjenje bralne pismenosti na šolah. V

tudi za zahtevo po spremembah v dojemanju novodobnih sklopu raziskave se zbirajo podatki s pomočjo vprašalnikov, ki

medijev [9]. jih izpolnijo učenci, učitelji in ravnatelji. S temi vprašalniki se

V kontekstu usposabljanj in izobraževanj na področju pridobijo informacije o različnih dejavnikih, ki lahko vplivajo na

digitalne pismenosti je potrebno izpostaviti Erasmus+ projekt uspešnost pri branju. Rezultati nalog in podatki, pridobljenih iz

"Digitalna medijska pismenost" Združenja EPEKA, katerega vprašalnikov, se nato primerjajo in analizirajo. Tako se

namen je izboljšanje digitalne medijske pismenosti med otroci in proučujejo bralni dosežki učencev in povezave med temi dosežki

mladostniki. V okviru tega projekta je bila junija leta 2023 ter različnimi dejavniki, tako med državami kot znotraj Slovenije

izvedena diseminacijska aktivnost, kjer so otrokom in [11].

mladostnikom v lokalnem okolju predstavili izsledke projekta in Mednarodna raziskava PISA - Raziskava PISA (ang.

jih opremili z dodatnimi znanji na področju digitalne medijske Programme for International Student Assessment) je globalna

pismenosti. Cilj te aktivnosti je bil omogočiti pridobitev ključnih raziskava OECD, ki meri kompetence 15-letnikov v matematiki,

kompetenc za odgovorno in kritično uporabo digitalnih medijev. naravoslovju in bralni pismenosti, vključno z digitalno

Projekt se osredotoča na deset korakov za spodbujanje digitalne pismenostjo. Rezultati raziskave PISA v Sloveniji nudijo

pismenosti med otroci in mladostniki. Ti vključujejo vpogled v digitalne kompetence slovenskih otrok. PISA V

izobraževanje skozi delavnice, seminarje in predavanja, razvoj Sloveniji je 75 odstotkov učencev doseglo raven osnovnih

kritičnega mišljenja, dostop do kakovostnih virov in orodij, bralnih spretnosti (raven 2), kar je za 8 odstotnih točk manj kot

spodbujanje kreativnosti ter izmenjavo izkušenj. Poudarjajo tudi leta 2018. Najvišjo stopnjo bralnega znanja in spretnosti (ravni 5

varno uporabo digitalnih medijev, aktivno udejstvovanje pri in 6) doseže 4 odstotke slovenskih učencev in 7 članic OECD.

ustvarjanju digitalnih vsebin, razvoj komunikacijskih veščin, Opredelitev bralne pismenosti v raziskavi PISA je opredeljena

sodelovanje s starši in izobraževalnimi ustanovami ter sledenje kot: razumevanje, uporaba, razmišljanje o zapisanem besedilu in

trendom na področju digitalnih medijev. Namen projekta je angažiranost ob branju besedila, ki bralcu oziroma bralcu



omogoča doseganje zastavljenih ciljev, razvijanje lastnega ciljev. S tem pristopom lahko digitalno tehnologijo uporabimo

znanja in spretnosti ter potencialov ter aktivno sodelovanje v ne le kot orodje za zamenjavo tradicionalnih metod poučevanja,

družbi [12]. temveč kot sredstvo za inovacije in globlje učenje.

Ob prvem pregledu rezultatov raziskave PISA lahko opazimo,

da naši učenci dosegajo nadpovprečne rezultate v primerjavi z

vrstniki iz drugih držav OECD, kar kaže na učinkovitost našega 6 ZAKLJUČEK

izobraževalnega sistema. Vendar, za celovito razumevanje in Povečana uporaba digitalnih tehnologij v osnovnih šolah

smiselno interpretacijo teh rezultatov v kontekstu našega odpira nove priložnosti, vendar prinaša tudi izzive, kot so spletno

nacionalnega osnovnošolskega sistema je potrebno dodatno nadlegovanje, neprimerna spletna komunikacija, izsiljevanje,

premisliti o natančnih dejavnikih teh rezultatov [13]. kraja identitete in širjenje lažnih informacij. Deviantno vedenje

V raziskavi PISA je še posebej pomembno prepoznati vpliv učencev pri digitalizaciji izobraževanja se lahko kaže v

uporabe informacijsko-komunikacijske tehnologije (IKT) na odvisnosti od digitalnih naprav in iger, kar vpliva na akademski

digitalno pismenost otrok v osnovnošolskem izobraževanju. uspeh ter socialne in čustvene vidike. Pomembno je zagotoviti

Digitalna pismenost, ki vključuje zmožnost učinkovite uporabe ustrezno infrastrukturo in opremo, da se omogočijo enake

IKT orodij in virov, je ključna za uspešno vključevanje v možnosti za vse učence, hkrati pa preprečiti, da prevelika

sodobno družbo in izobraževanje. Vendar pa razlike v dostopu odvisnost od tehnologije negativno vpliva na socialne veščine in

do IKT in raznolikosti učnih priložnosti lahko prispevajo k kompetence otrok. Digitalne tehnologije lahko izboljšajo učni

digitalnemu razkoraku med učenci. Zato je pomembno, da se šole proces, če se z njimi modro ravna. Ključna je tudi ustrezna

in izobraževalne ustanove osredotočijo na enakopravno izobrazba učiteljev in učencev o varni in odgovorni uporabi IKT,

dostopnost do IKT orodij ter razvijanje digitalnih veščin pri vseh da bodo učitelji znali pravilno širiti znanje o teh temah in učenci

učencih. [14]. učinkovito uporabljati pridobljene veščine.

Internetna podjetja, ki omogočajo digitalne socialne



5 UPORABA DIGITALNE TEHNOLOGIJE V za njihovo vsebino, saj objavljajo in distribuirajo medijske platforme pri tem ne morejo biti popolnoma izvzeta odgovornosti

UČNEM PROCESU vsebine, ne glede na to, kdo jih ustvari. Potrebna je nova

Za uspešno vključevanje digitalnih tehnologij v regulacija, kot npr. za politične oglase, ki so na televiziji strogo

osnovnošolski izobraževalni sistem je potrebno, da tako učenci nadzorovani. Podobna pravila morajo veljati tudi za internet.

kot učitelji že v zgodnji fazi osvojijo ustrezno znanje in veščine. Druga težava je zasebnost, saj ljudje pogosto ne berejo pogojev,

Primarni namen uporabe digitalne tehnologije pri pouku je preden se strinjajo z njimi, kar omogoča zlorabo njihovih

izboljšanje kakovosti poučevanja in učenja, hkrati pa postopno podatkov. Otroci so na to posebej ranljivi saj so na spletu

razvijamo digitalne spretnosti, koncepte in njihovo uporabo v izpostavljeni škodljivim medijskim kampanjam. Pri izvajanju

različnih okoljih. To omogoča uspešno reševanje kompleksnih kampanj se morajo podjetja zavzemati za bolj sprejemljiv in

življenjskih problemov. uravnotežen pristop pri promociji izdelkov. Monopoli niso

Kot primer uporabe digitalnih tehnologij v učnem proces koristni za zaščito zasebnosti podatkov, zato bi morala vlada

lahko navedemo model SAMR, ki vključuje zamenjavo (S – ang. spodbujati konkurenco. Medijska in tehnološka podjetja bi

Substitution), obogatitev (A – ang. Augmentation), spremembo morala plačati več davkov, kar bi lahko financiralo digitalno

(M – ang. Modification) in redefinicijo (R – ang. Redefinition) medijsko izobraževanje v osnovnih šolah in skupnostih.

in je eden izmed pristopov k vključevanju digitalne tehnologije Za zagotavljanje kakovostnega, varnega in spodbudnega

v učni proces. Ta model predstavlja štiri različne načine, kako učnega okolja je nujno, da naslovimo ta vprašanja in poiščemo

vključiti tehnologijo v poučevanje in učenje, pri čemer se ravni ustrezne odgovore, ki bodo prispevali k izboljšanju digitalne

razlikujejo glede na učinkovitost uporabe digitalne tehnologije. pismenosti in splošnega učnega uspeha otrok v osnovnošolskem

Na najnižji ravni (zamenjava) digitalna tehnologija nadomesti izobraževanju.

tradicionalna orodja brez dodane vrednosti. Na primer, učenci

namesto pisanja v zvezke tipkajo na računalnike. Naslednja

raven (obogatitev) izboljša učinkovitost učnih aktivnosti, kjer

digitalni dokumenti omogočajo lažje urejanje in deljenje

informacij, kar naredi učne procese bolj učinkovite. Na višji

ravni (sprememba) tehnologija spreminja učne aktivnosti in

omogoča doseganje bolj kompleksnih in višjih učnih ciljev.

Uporaba simulacij in interaktivnih orodij omogoča globlje

razumevanje zapletenih konceptov. Najvišja raven (redefinicija)

vključuje učne aktivnosti, ki so možne le s pomočjo digitalne

tehnologije, kot so globalne kolaboracije med učenci iz različnih

držav preko spleta, kar omogoča učenje in izmenjavo idej na

načine, ki brez digitalne tehnologije ne bi bili mogoči [15].

Razvoj digitalnih kompetenc pri učencih in dijakih je

ključnega pomena za njihovo uspešno vključevanje v sodobno

digitalno družbo. Model SAMR ponuja strukturiran pristop k

vključevanju tehnologije v učni proces, ki omogoča postopno

nadgradnjo učnih aktivnosti in doseganje kompleksnejših učnih





LITERATURA


[1] Pihir, I., Tomičić-Pupek, K., & Furjan, M. T. (2019). Digital

transformation playground - literature review and framework of concepts.

Journal of Information and Organizational Sciences, 43(1), 33–48.

https://www.researchgate.net/publication/334367976_Digital_transforma

tion_playground_literature_review_and_framework_of_concepts

[2] Schallmo, D., Williams, C. A., & Boardman, L. (2017). Digital

transformation of business models-best practice, enablers, and roadmap.

International Journal of Innovation Management, 21(8), 1–17.

https://ideas.repec.org/a/wsi/ijimxx/v21y2017i08ns136391961740014x.h

tml

[3] Sjöberg, J., & Lilja, P. (2019). University teachers’ ambivalence about the

digital transformation of higher education. International Journal of

Learning, Teaching and Educational Research, 18(13), 133–149.

https://www.researchgate.net/publication/339420291_University_Teache

rs%27_Ambivalence_about_the_Digital_Transformation_of_Higher_Ed

ucation

[4] Bond, M., Marín, V. I., Dolch, C., Bedenlier, S., & Zawacki-Richter, O.

(2018). Digital transformation in German higher education: Student and

teacher perceptions and usage of digital media. International Journal of

Educational Technology in Higher Education, 15(1), 1–20.

https://docs.opendeved.net/lib/WU3RAAUR

[5] Castellví, J.; Díez-Bedmar, M.-C.; Santisteban, A (2020). Pre-Service

Teachers’ Critical Digital Literacy Skills and Attitudes to Address Social

Problems. 134. https://www.mdpi.com/2076-0760/9/8/134

[6] Salmieri, L. (2019). The rhetoric of digitalization in Italian educational

policies: Situating reception among digitally skilled teachers. Italian

Journal of Sociology of Education, 11(1), 162–183.

https://www.researchgate.net/publication/332304834_The_Rhetoric_of_

Digitalization_in_Italian_Educational_Policies_Situating_Reception_am

ong_Digitally_Skilled_Teachers

[7] NY UNESCO. 2017. Working group on education: Digital skills for life

and work. In Broadband commission for sustainable development.

https://unesdoc.unesco.org/ark:/48223/pf0000259013

[8] Levido, A. (2024). Media literacy and the concept of ‘technologies’ in

primary school classrooms: moving beyond technical skills. Learning,

Media and Technology, 1–13.

[9] Buckingham, D. (2019). The Media Education Manifesto. New Jersey:

Wiley, 230-239. https://cejc.ptks.pl/Volume-13-No-2-26-Special-Issue-

2020/BOOK-REVIEW-David-Buckingham-2019-The-Media-Education-

Manifesto-Cambridge-UK-Polity

[10] Združenje EPEKA. 5. 6. 2023. Digitalna medijska pismenost –

Diseminacijsko srečanje. https://epeka.si/digitalna-medijska-pismenost-

diseminacijsko-srecanje/

[11] Pedagoški inštitut. IEA PIRLS, (2021).

https://www.pei.si/wpcontent/uploads/2023/06/PIRLS21_NacionalnoPor

ocilo_Splet.pdf.

[12] Pedagoški inštitut. PISA 2022. (2022) https://www.pei.si/wp-

content/uploads/2023/12/Porocilo_PISA22_FINAL.pdf.

[13] Šterman Ivančič, K. (2019) PISA 2018: Nacionalno poročilo s primeri

nalog iz branja. Ljubljana: Pedagoški inštitut.

[14] Šterman Ivančič, K. in Štremfel, U. (2020) Assessment policy and practice

of Slovenia. V Harju-Luukkainen, H., Mcelvany, N., in Stang, J. (ur.).

Monitoring student achievement in the 21st century: European policy

perspectives and assessment strategies. Cham: Springer.

https://doi.org/10.1080/17439884.2024.2308609

[15] Bouchrika, I. (2023). How to Use SAMR Model in Designing Instruction.

Pridobljeno na https://research.com/ education/how-to-use-samr-model-

in-designing-instruction



Učenje španščine z jezikovno aplikacijo Duolingo



Learning Spanish with the Duolingo Language App



Ana Dvoraček

Srednja ekonomska šola Ljubljana

Ljubljana, Slovenija

ana.dvoracek@seslj.si



language and vice versa, sorting words into the correct word

POVZETEK order, thematic stories with comprehension questions, etc.). In

Dandanes ima praktično že vsak dijak pametni telefon, na the research, we examined the effect of Duolingo on students of

katerega lahko naloži številne aplikacije za učenje jezika in z Spanish. We found that students use Duolingo 1 time per day

njimi dopolni klasičen način poučevanja. Glede na upad (42%) and thus maintain a streak of consecutive days in the app,

motivacije dijakov za šolo in učenje na sploh, je modro, da that the vast majority of students (80%) find the app easy to use,

učitelji najdemo še druge načine poučevanja, ki so dijakom bližje. that it helps them learn Spanish (60%) and that they like learning

Mobilno učenje dijakom omogoča fleksibilen način samoučenja with it (53%), which is very positive and we can conclude that it

izven pouka, ki pa lahko postopoma poveča tudi motivacijo za is an effective tool for independent Spanish learning.

sam pouk. S številnimi vajami, ki so razdeljene na lekcije in

spretnosti, jezikovna aplikacija Duolingo uči besedišče in KEYWORDS

slovnico. Pri tem se ves čas uporabljata tako izvirni kot ciljni Duolingo, language app, learning foreign languages, self-

jezik. Vaje v aplikaciji vključujejo različne spretnosti, kot so study, Spanish

izgovorjava, prevajanje, slušno razumevanje, in vajami

(prepoznavanje besede s pomočjo slike, narek, prevajanje iz

izvirnega v ciljni jezik in obratno, razvrščanje besed v pravilen 1 UVOD

besedni red, tematske zgodbe z vprašanji za razumevanje, itd.). Dandanes ima praktično že vsak dijak pametni telefon, na

V raziskavi smo preverjali, kakšen učinek ima jezikovna katerega lahko naloži številne aplikacije za učenje jezika in z

aplikacija Duolingo na dijake španščine. Ugotovili smo, da dijaki njimi dopolni klasičen način poučevanja.

Duolingo uporabljajo 1-krat na dan (42 %) in s tem vzdržujejo Kot pojasni Ramírez Montoya [6], je mobilno jezikovno

niz zaporednih dni v aplikaciji, da je za veliko večino dijakov učenje naslednik e-učenja, ki je kakršnokoli učenje s podporo

(80 %) aplikacija enostavna za uporabo, jim pomaga pri učenju elektronskih orodij in virov. Mobilno učenje oz. m-učenje

španščine (60 %) in imajo radi učenje z njo (53 %), kar je zelo dijakom omogoča fleksibilen način samoučenja izven pouka, ki

pozitivno in lahko sklepamo, da gre za učinkovito orodje za pa lahko postopoma poveča tudi motivacijo za sam pouk. V

samostojno učenje španščine. nadaljevanju se bomo osredotočili na jezikovno aplikacijo

Duolingo, do katere lahko dijaki dostopajo z mobilnimi telefoni

KLJUČNE BESEDE ali po spletu.

Duolingo, jezikovna aplikacija, samoučenje, španščina,

učenje tujih jezikov

2 PREGLED LITERATURE



ABSTRACT Duolingo je jezikovna aplikacija, ki sta jo novembra 2011

Nowadays, practically every student has a smartphone, on ustvarila Luis Von Ahn in Severin Hacker. Gre za brezplačno

which they can download many language learning applications aplikacijo, ki ponuja tudi plačljivo nadgradnjo. Slednja

and use them to supplement the classical teaching method. Given uporabniku omogoči neomejeno ponavljanje brez izgubljanja

the decline in students' motivation for school and learning in življenj in draguljev, ki so neke vrste valuta v aplikaciji. S

general, it is wise for teachers to find other ways of teaching that številnimi vajami, ki so razdeljene na lekcije in spretnosti,

are closer to students. Mobile learning allows students a flexible Duolingo uči besedišče in slovnico. Pri tem se ves čas uporabljata

way of self-learning outside of classes, which can also gradually tako izvirni kot ciljni jezik. Nabor jezikov in kombinacije

increase motivation for the class itself. With many exercises izvirnih ter ciljnih jezikov so številne, npr. angleščina in

divided into lessons and skills, the Duolingo language app španščina, ruščina in kitajščina itd. Cilj aplikacije je, da

teaches vocabulary and grammar. Both the source and target uporabniki z vsemi rešenimi vajami in ponovitvami dosežejo

languages are used at all times. The exercises in the application nivo B2 v ciljnem jeziku. Ena glavnih značilnosti aplikacije je

include various skills such as pronunciation, translation, listening ponavljanje s pomočjo algoritmov, ki zaznajo, katere besede ali

comprehension, and exercises (recognizing a word with the help stavke mora uporabnik ponoviti glede na pretekle napake. Vaje

of a picture, dictation, translating from the original to the target v aplikaciji vključujejo različne spretnosti, kot so izgovorjava,

prevajanje, slušno razumevanje, in vajami (prepoznavanje niz dni reševanja vaj v aplikaciji, doseženo določeno število točk

besede s pomočjo slike, narek, prevajanje iz izvirnega v ciljni izkušenj XP, minimalno število napak pri reševanju vaj, dnevni

jezik in obratno, razvrščanje besed v pravilen besedni red, čas reševanja vaj, skupen izziv opravljanja vaj s prijateljem itd.

tematske zgodbe z vprašanji za razumevanje, itd.) [5]. 4. Profil uporabnika, ki vključuje ustvarjen avatar ali

Pomembno je tudi izpostaviti, da Duolingo uporabnikom poljubno fotografijo, uporabniško ime, datum pridružitve

hkrati nudi takojšnjo povratno informacijo po vsaki opravljeni aplikaciji, sledilce in osebe, ki jim sledi, statistiko in dosežke.

vaji. Po mnenju Garcíe je povratna informacija podana na več

načinov: kot pohvala, kot komentar na opravljeno aktivnost, s

točkami izkušenj XP in z dragulji [2]. S slednjim Duolingo

uporabnike usmerja k samoučenju. Gre za »učenje, pri katerem

konceptualizacijo, oblikovanje, vodenje in vrednotenje učnega

projekta usmerja uporabnik sam« [1].

Kot razloži Munday, je aplikacija razdeljena na različna

področja [5]:

1. Drevo, ki vključuje vsebinska področja z vajami, ki

spreminjajo barve, začenši s sivo (ki označuje, da se z vajo še ni

začelo), nadaljuje se z različnimi barvami (ki označujejo začetek

opravljanja vaje) in konča z zlato (ki označuje, da je uporabnik

usvojil vajo in opravil vse ponovitve), kot prikazuje Slika 1. Pri

vsakem vsebinskem področju Duolingo nudi tudi kratek pregled Slika 2: Dosežki sledilcev in reakcije slovnice in besedišča. Po vsaki opravljeni vaji je uporabnik

nagrajen s točkami izkušenj XP, dragulji in pohvalo s strani Kot lahko opazimo, je Duolingo zelo moderna jezikovna

aplikacije. aplikacija, ki je podobna igri, čeprav je večina vaj zelo

tradicionalnih, osnovanih na prevodu, nareku in izgovorjavi [5].



Kot zatrjuje Krashen [4], vaje v Duolingo vključujejo zavestno



učenje, ki pa po njegovem mnenju ne vodi do jezikovne

kompetentnosti, saj gre za tradicionalni način poučevanja, ki ga

je v 60. letih 20. stoletja nadomestila komunikativna metoda.

Pomembno se je zavedati, da z aplikacijo ne želimo nadomestiti

poučevanja v razredu, temveč ga popestriti in dijake usmeriti k

samoučenju. Hall, Graham in Cook [3] celo poudarjajo, da je

prevajanje v številnih kontekstih naravno in učinkovito sredstvo

učenja jezika, razvija pomembno spretnost, odgovarja

učenčevim potrebam in ščiti učenčevo jezikovno ter kulturno

identiteto.

Ravno z raziskavo na Srednji ekonomski šoli Ljubljana bomo

preverili, kaj dijakom ustreza in kaj ne.



3 REZULTATI

Cilj raziskave je presoditi, ali se lahko jezikovna aplikacija

Duolingo uporablja kot učinkovito orodje za samostojno učenje

španščine in za dopolnitev pouka španščine. Dijaki lahko

Slika 1: Drevo z vsebinskim področjem aktivnosti in aplikacijo uporabljajo kadarkoli in kjerkoli pod pogojem, da

vremena imajo dostop do interneta, si sami prilagajajo težavnost glede na

vaje, ki jih rešujejo, in utrjujejo že pridobljeno znanje španščine.

2. Sodelovanje z drugimi uporabniki je omogočeno na

Raziskava temelji na deskriptivni in kavzalno-

podoben način kot na drugih socialnih omrežjih. Uporabniki



se na ta način spodbujajo, kot prikazuje Slika 2. V aplikaciji so anketiranja, kot inštrument pa anketni vprašalnik. Podatke smo tudi lestvice najboljših po vsem svetu v različnih ligah, ki v pridobili preko spletne ankete 1KA na začetku pouka španščine. uporabnikih vzbujajo tekmovalnost in željo po zmagi. Izhodišče raziskave je vzorec 173 dijakov srednje strokovne lahko sledijo drug drugemu, reagirajo na opravljene dosežke in raziskovanja. Za zbiranje podatkov smo uporabili tehniko neeksperimentalni metodi empiričnega pedagoškega



Uporabniki so nanje uvrščeni po točkah izkušenj XP, ki jih šole, in sicer Srednje ekonomske šole Ljubljana. V vzorcu je bilo pridobijo z opravljanjem vaj. Vsako nedeljo ob 19. uri po času anketiranih 114 žensk in 59 moških. Povprečna starost je bila 16 ZDA se lestvice osvežijo. let (1. letnik). Prvi del vprašalnika zajema splošna podatka o 3. Dosežki uporabnikov, ki se beležijo sproti in so zanje dijakih: spol in letnik, ki ga na srednji strokovni šoli obiskujejo. uporabniki nagrajeni z različnimi značkami, dvojnimi točkami Sledijo sklopi vprašanj, ki se nanašajo na namen raziskave. izkušenj XP, možnostjo zamrznitve zaporednega niza dni Sklopi so razdeljeni glede na pogostost rabe Duolinga, na reševanja vaj ali celo z nekajdnevno brezplačno nadgradnjo motivacijsko plat aplikacije in na rabo aplikacije tudi v prihodnje. aplikacije, ki je sicer plačljiva. Kot dosežki se smatrajo zaporedni

Vprašanja so zaprtega tipa. Anketni vprašalnik obsega 5 vprašanj.

Podatki so prikazani grafično tabelarično.

S Slike 3 je razvidno, da dijaki Duolingo uporabljajo 1-krat

na dan (42 %) in s tem vzdržujejo niz zaporednih dni v aplikaciji.

Nekateri so še bolj motivirani in aplikacijo uporabljajo večkrat

na dan (16 %). Drugi aplikacijo uporabljajo le redko ali nikoli

(35 %).





Slika 5: Vprašanje glede uporabe aplikacije Duolingo v

bodoče



4 ZAKLJUČEK

V raziskavi smo preverjali, kakšen učinek ima jezikovna

aplikacija Duolingo na dijake španščine. Najprej je Duolingo

začela uporabljati profesorica, nato so si aplikacijo pri pouku na



Slika 3: Pogostost uporabe aplikacije Duolingo svoje mobilne telefone naložili še dijaki. Ugotovili smo, da dijaki Duolingo uporabljajo 1-krat na dan (42 %) in s tem vzdržujejo



Na Sliki 4 so vidni odgovori dijakov na dane trditve o niz zaporednih dni v aplikaciji, da je za veliko večino dijakov

Duolingu. Za veliko večino dijakov (80 %) je aplikacija (80 %) aplikacija enostavna za uporabo, jim pomaga pri učenju

enostavna za uporabo, jim pomaga pri učenju španščine (60 %) španščine (60 %) in imajo radi učenje z njo (53 %), kar je zelo

in imajo radi učenje z njo (53 %), kar je zelo pozitivno in lahko pozitivno in lahko sklepamo, da gre za učinkovito orodje za

sklepamo, da gre za učinkovito orodje za samostojno učenje samostojno učenje španščine. Čeprav iz raziskave ni tako zelo

španščine. Pri bolj specifičnih trditvah so dijaki večinoma razvidno, smo povečano motivacijo dijakov za pouk španščine

neopredeljeni. Odkar uporabljajo aplikacijo, ne vedo, če so bolj opazili iz številnih komentarjev na hodniku in pred samim

motivirani za pouk španščine (47 %) ali če pouku lažje sledijo poukom španščine. Bili so predvsem na temo zaporednih dni in

(42 %). Kljub vsemu 35 % dijakov pa zaradi uporabe Duolinga novih dosežkov v aplikaciji. Tisti, ki aplikacijo redno uporabljajo,

lažje sledi pouku. Neopredeljeni tudi, ko profesorica spremlja so opazili tudi vključenost profesorice, in sicer, da redno

njihov napredek (46 %), čeprav smo menili, da bo to pozitivno spremlja njihove dosežke z danimi emodžiji v aplikaciji.

vplivalo na njihovo motiviranost in to v 39 % primerih tudi drži. Kljub vsemu ima Duolingo tudi določene pomanjkljivosti.

Domnevamo, da so neopredeljeni, ker se počutijo nadzirane in Vrsta vaj je omejena, ponavljanje določenih besednih zvez

ne dodatno motivirane za učenje španščine. Kar se tiče razloga (pre)pogosto, novih besed ter besednih zvez ni veliko in po

za uporabo Duolinga, prav tako niso enotni. Nekateri ga določenem času lahko uporaba aplikacije postane monotona,

uporabljajo le zaradi opravljenega izziva (37 %), kar je eden dolgočasna. Hkrati lahko postane tudi naporno opravljati vaje

izmed delov alternativnega ustnega ocenjevanja pri pouku vsak dan, da se vzdržuje niz zaporednih dni, vendar je to za

španščine, drugi imajo tudi druge razloge (30 %). dijake izziv, ki jim bo koristil tudi v življenju na sploh, saj se s

tem učijo vztrajnosti.

Predlagamo, da učitelji tujih jezikov premislijo, če bi tudi oni



aplikacijo uporabili za dodatno motivacijo za pouk tujega jezika



in morda tudi kot del ocene v primeru določenega števila

zaporednih dni.



ZAHVALA

Zahvalila bi se ravnateljici za podporo in dijakom Srednje

ekonomske šole Ljubljana za aktivno uporabo aplikacije

Duolingo ter sodelovanje v raziskavi.



LITERATURA IN VIRI

[1] Stephen Brookfield, 2009. Self- Directed Learning. In R. Maclean & D.

Slika 4: Trditve glede uporabe aplikacije Duolingo Wilson (Eds.), International Handbook of Education for the Changing

World of Work, (2615-2627). Springer, Rotterdam.

Na Sliki 5 so vidni odgovori dijakov glede uporabe aplikacije DOI: https://doi.org/10.1007/978-1-4020-5281-1_172

[2] Ignacio Garcia, 2013. Learning a Language for Free While Translating the

Duolingo v bodoče, pri čemer je več kot večina dijakov (56 %) Web. Does Duolingo Work? International Journal of English Linguistics,

odgovorila pritrdilno in 35 % dijakov z morda, kar ponovno Vol. 3, No. 1. 19-25. DOI: https://doi.org/10.5539/ijel.v3n1p19

[3] Hall G, Cook G. Own-language use in language teaching and learning.

potrdi dejstvo, da je Duolingo lahko učinkovito dopolnilno Language Teaching, 45 (3), 271-308.

orodje za pouk tujega jezika in za samostojno učenje. DOI: https://doi.org/10.1017/S0261444812000067



[4] Stephen Krashen, 2014. Does Duolingo “trump” university-level language

learning? International Journal of Foreign Language Teaching, 9 (1), 13-

15.

[5] Pilar Munday, 2015. The case for using DUOLINGO as part of the

language classroom experience. RIED. Revista Iberoamericana de

Educación a Distancia. 19 (1), 83-101. DOI:

https://doi.org/10.5944/ried.19.1.14581

[6] María S. Ramírez Montoya, 2009. Recursos tecnológicos para el

aprendizaje móvil (mlearning) y su relación con los ambientes de

educación a distancia: implementaciones e investigaciones. RIED. Revista

iberoamericana de educación a distancia, 12 (2), 57-82.





Vloga umetne inteligence pri ustvarjanju zgodb v




animiranih filmih: »Vpliv na kognitivne procese in



izobraževalni sistem«



Role of Artificial Intelligence in Animated Storytelling:



»Impact on Cognitive Processes and the Education System«



mag. Mojca Guštin

Srednja medijska in grafična šola Ljubljana

Pokopališka ulica 33, Ljubljana, Slovenija

mojca.gustin@smgs.si



pripisujejo človeške lastnosti in namere neživim objektom na

POVZETEK podlagi njihovih gibanj in interakcij;

Pripovedovanje zgodb je od nekdaj ključno orodje za prenos  Ugotoviti, kako lahko tehnologija, podprta z UI orodji,

idej, kultur in znanja med ljudmi. S prihodom umetne inteligence bistveno izboljša, predvsem pa pohitri, proces ustvarjanja in

(UI) se je način pripovedovanja predvsem pa ustvarjanja zgodb kreiranja novih zgodb ter izboljša učni proces, saj omogoča

bistveno spremenil, saj so se pojavile inovativne metode za bolj učinkovito in motivirano učenje dijakov;

ustvarjanje in doživljanje zgodb. V Dreamfarm studiu v članku  prikazati, kako formativno spremljanje in sumativno

»How Artificial Intelligence will change animation forever« [1] ocenjevanje znanja omogočata celostno izobraževanje, kjer

raziskujejo, kako lahko z UI izboljšamo scenarije, in sicer z dijaki ne le pridobijo teoretično znanje, temveč tudi razvijajo

analizo uspešnih vzorcev v prejšnjih produkcijah in praktične spretnosti in kompetence;

spremljanjem povratnih informacij gledalcev. Z uporabo umetne  predstaviti primer dobre prakse z uporabo različnih orodij za

inteligence lahko izboljšate svojo zgodbo s prilagoditvijo ustvarjanje zgodb in generiranje slik ter kako so uporabljena

vsakega dialoga. Algoritmi lahko izboljšajo pisanje scenarijev za orodja vplivala na dijakov ustvarjalni proces in učne rezultate.

animacijo z analizo vzorcev uspeha v prejšnjih produkcijah in V prispevku bomo predstavili implementacijo in integracijo

dominantnih povratnih informacij gledalcev. Poleg scenarija se UI orodij predvsem za ustvarjanje zgodb (Simpleshow in

lahko posvetujete z UI, ki vam predlaga ustvarjalne ideje ali ChatGPT), UI orodje za izdelavo vizualne pripovedi zgodbe v

prilagodi obstoječe. Seveda pa UI na tem področju še potrebuje obliki zgodborisa (Boords) ter UI orodja za generiranje slik (Kitl

veliko dela in človeški ustvarjalec je še vedno zelo pomemben AI Image Generator in Dalle). Ta orodja so bila vključena v učni

člen v verigi celotnega procesa ustvarjanja animiranih filmov. proces z namenom izboljšanja ustvarjalnih zmožnosti dijakov in

Cilj tega članka je raziskati in predstaviti, kako UI spreminja premagovanja omejitev, ki jih postavljajo obstoječa programska

področje pripovedovanja ter ustvarjanja zgodb v animaciji, zlasti orodja. Ustvarjalne zmožnosti znotraj te implementacije, v

v kontekstu animiranega filma. Poleg tega bomo preučili, kako nasprotju z omejitvami tradicionalnih tehnik in metod

dojemanje zgodb vpliva na kognitivne procese dijakov in kako poučevanja, omogočajo dijakom, da preizkusijo različne pristope

lahko tehnologija, podprta z UI orodji, izboljša proces k ustvarjanju zgodb in vizualnemu pripovedovanju zgodbe. S

ustvarjanja in kreiranja novih zgodb ter izboljša učni proces, saj kombinacijo teh orodij in individualnemu pristopu dela v

omogoča bolj učinkovito in angažirano učenje dijakov. Naše skupinah smo dijakom omogočili bolj dinamično in interaktivno

raziskave temeljijo na prejšnjih ugotovitvah Fritza Heiderja in učno okolje. Za formativno spremljanje, samovrednotenje in

Marianne Simmelove [2], Davida Eagleman [3] ter različnih vrstniško vrednotenje znanja smo uporabili različna orodja

sodobnih raziskavah o pomembnosti pripovedovanja zgodb za (1KA, Padlet in spletne učilnice). Pričakuje se, da bodo zaključki

človeški um avtorjev, kot so Dwayne Harapnuik [4], Rachel iz tega prispevka vodili k boljšemu razumevanju ustvarjalnega

Gillett [5] in Rahul Prakash [6]. Temeljno vodilo izvedbe je bilo procesa v kontekstu integracije UI orodij v izobraževalni proces.

poiskati ustrezne učne cilje, učne aktivnosti in načine To vključuje povezovanje tehnoloških inovacij in ustvarjalnega

ocenjevanja, v okviru katerih bi lahko vsak dijak ob koncu dela, kar omogoča dijakom, da razvijejo tako teoretične kot

zaključnega letnika sam ali v skupini izdelal svoj animirani film praktične spretnosti pri izdelavi medijskih vsebin, v našem

s poudarkom na sporočilnosti oz. pomenu zgodbe, pri katerem bi primeru animacije.

si pomagal s sodobnimi orodji, ki podpirajo uporabo UI. Namen

prispevka je bil: KLJUČNE BESEDE

 pojasniti, kako dojemanje ljudi vpliva na proces ustvarjanja Animirani film, formativno ocenjevanje, kognitivni procesi,

zgodb, pri čemer se osredotočamo na način, kako ljudje pripovedovanje zgodb, umetna inteligenca



ABSTRACT software tools. The creative possibilities within this

Storytelling has always been a crucial tool for transmitting implementation, in contrast to the limitations of traditional

ideas, cultures, and knowledge among people. With the rise of techniques and teaching methods, allow students to experiment

artificial intelligence (AI), the way stories are told, and specially with different approaches to storytelling and visual storytelling.

created, has fundamentally changed, introducing innovative By combining these tools and an individualized approach to

methods for story creation and experience. The article "How group work, we enabled students to engage in a more dynamic

Artificial Intelligence Will Change Animation Forever" by and interactive learning environment. For formative monitoring,

Dreamfarm Studio explores how AI can enhance scripts by self-assessment, and peer assessment of knowledge, we used

analyzing successful patterns in previous productions and various tools (1KA, Padlet and online classrooms). The

tracking viewer feedback. Using AI, you can refine your story by conclusions from this article are expected to lead to a better

making small adjustments to each dialogue. Algorithms can understanding of the creative process in the context of integrating

improve animation scriptwriting by analyzing success patterns in AI tools into the educational process. This includes connecting

past productions and dominant viewer feedback. Beyond technological innovations with creative work, allowing students

scripting, AI can be consulted to suggest creative ideas or adjust to develop both theoretical and practical skills in creating media

existing ones. However, AI in this field still requires significant content, specifically in our case, animation.

development, and the human creator remains a vital part of the

entire animation production process. KEYWORDS

The goal of this article is to explore and present how AI is Animated Film, Formative Assessment, Cognitive Processes,

transforming the field of storytelling and story creation in Storytelling, Artificial Intelligence

animation, particularly within the context of animated films.

Additionally, we will examine how the perception of stories

impacts the cognitive processes of students and how technology, 1 UVOD

supported by AI tools, can enhance and accelerate the process of Sodobni izobraževalni sistemi se soočajo z izzivom

creating new stories, improving the educational process by vključevanja novih tehnologij v pedagoške prakse. UI ponuja

enabling more efficient and engaged student learning. Our obetavne priložnosti za izboljšanje izobraževanja z uporabo

research is based on previous findings by Fritz Heider and animacije. Ta pristop omogoča bolj interaktivne in angažirane

Marianne Simmel [2], David Eagleman [3], and various učne izkušnje, kar je še posebej pomembno za današnje dijake,

contemporary studies on the importance of storytelling for the ki so navajeni vizualnih in digitalnih vsebin. Poleg tega razvoj

human mind by authors such as Dwayne Harapnuik [4], Rachel orodij s podporo UI omogoča formativno spremljanje znanja, kar

Gillett [5], and Rahul Prakash [6]. učiteljem pomaga pri prilagajanju učnih načrtov in metod

The fundamental guiding principle of the implementation was poučevanja potrebam posameznih dijakov. V tem prispevku

to identify appropriate learning objectives, activities, and raziskujemo preplet UI in ustvarjanja zgodb v kontekstu

assessment methods through which each student could create medijskih vsebin, predvsem v animiranem filmu. S poudarkom

their own animated film by the end of their final year, na percepciji dijakov glede drugačnega načina poučevanja,

emphasizing the message and meaning of the story, aided by uporabi UI orodij za generiranje slik in strukturiranju zgodb ter

modern tools that support AI use. The purpose of the article was formativnem spremljanju znanja želimo prikazati potencial in

to: izzive, ki jih prinaša integracija UI v izobraževanje. Raziskava

 Explain how human perception influences the storytelling vključuje analizo anketnih podatkov dijakov, ki so si ogledali

process, focusing on how people attribute human video o gibanju geometrijskih oblik in ustvarili zgodbo ter

characteristics and intentions to inanimate objects based on izdelali kratki animirani film z uporabo UI orodij. Poleg tega

their movements and interactions; bomo predstavili orodja, ki nam omogočajo formativno

 Determine how AI-supported technology can significantly spremljanje znanja, ki je ključnega pomena za prilagajanje učnih

improve and speed up the process of creating new stories, procesov in izboljšanje učnih izidov.

thereby improving the educational process by facilitating more

efficient and motivated student learning;

 2 METODA DELA Demonstrate how formative assessment and summative

evaluation of knowledge allow for education, where students Raziskava je vključevala dijake, ki so si ogledali animirani

not only gain theoretical knowledge but also develop practical film z geometrijskimi oblikami, ne da bi vedeli, kdo je film

skills and competencies; naredil ali kaj naj bi pomenil. Po ogledu animiranega filma je 64

 Present a case of good practice in using various tools for story dijakov izpolnilo anketni vprašalnik, ki je meril njihove

creation and image generation, and how these tools influenced interpretacije in čustvene odzive na gibanje oblik. Anketna

the students' creative process and learning outcomes. vprašanja so bila zasnovana za merjenje percepcije, razumevanja

In this article, we will present the implementation and in čustvenega odziva dijakov na animirane vsebine. Poleg ankete

integration of AI tools primarily for story creation (Simpleshow so dijaki izdelali tudi končni izdelek, kratki animirani film. Na

and ChatGPT), an AI tool for creating visual storytelling in the koncu so izrazili svoje dojemanje, mnenje in razmišljanje s

form of storyboards (Boords), and AI tools for image generation pomočjo zastavljenih vprašanj ter orodjem Padlet, ki nam je

(Kitl AI Image Generator and Dalle). These tools were integrated omogočal takojšnjo diskusijo in medsebojno interakcijo.

into the learning process to improve the students' creative

capabilities and overcome the limitations posed by existing

2.1 nezavedni procesi vplivajo na naše odločitve in obnašanje [3]. Kognitivni procesi v socialni percepciji in

nagnjenost k sklepanju Predstavitev teh študij dijakom je ključna, ker jim omogoča, da

razumejo, kako močno zgodbe vplivajo na naše možgane in zakaj

Nedavne znanstvene raziskave [7] kažejo, kako zgodbe

so tako učinkovite pri komuniciranju kompleksnih idej. Pri

spreminjajo naš odnos, prepričanja in vedenje. Fritz H. in



Marianne S. [2] sta izvedla klasično študijo, ki je preučevala, te ugotovitve za: pisanju in ustvarjanju animiranih filmov lahko dijaki uporabijo

kako ljudje pripisujejo človeške lastnosti, namere in čustva  razvijanje zgodb, ki so čustveno nabite in povezujejo preprostim geometrijskim oblikam, kot so trikotnika in krog, na

gledalce na globlji ravni, s čimer izboljšajo angažiranost in

podlagi njihovega gibanja in interakcij. S pomočjo animiranega

pomnjenje vsebine;

filma, v katerem so trikotniki in krogi komunicirali med seboj  uporabo kognitivnih procesov za oblikovanje likov in zgodb, skozi gibanje, sta odkrila, da ljudje te oblike pogosto

ki odražajo človeške lastnosti in namere, kar omogoča

interpretirajo kot namerne akterje, ki sodelujejo v kompleksnih

gledalcem, da se povežejo z liki in razumejo njihove motive;

socialnih interakcijah, kot so preganjanje, izogibanje, boj in

sodelovanje. Ta raziskava je pokazala, kako močno je človeško  za učinkovito sporočanje idej in vrednot, kar poveča vpliv



nagnjenje k antropomorfizaciji in kako pomembni so kognitivni animiranega filma na gledalce in spodbuja kritično razmišljanje. procesi v socialni percepciji. V povezavi s to študijo smo izvedli Integracija teh študij v izobraževalni proces omogoča raziskavo med dijaki, ki so si ogledali animirani film z dijakom, da bolje razumejo pomembnost pripovedovanja zgodb geometrijskimi oblikami, ne da bi vedeli, kdo je film naredil ali in kako lahko uporabijo te tehnike za izboljšanje svojega dela pri kaj naj bi pomenil. Raziskavo smo izvedli, ker nas je zanimalo, ustvarjanju animiranih filmov. S tem se ne le izboljša kakovost kako današnje generacije, ob tako močnem vplivu digitalizacije, njihovih projektov, temveč tudi njihovo splošno razumevanje dojemajo te vsebine. Slika 1 nam prikazuje študijo animiranega kognitivnih procesov in sposobnost učinkovitega komuniciranja filma Fritza H. in Marianne S. [2] ter dijake med gledanjem skozi vizualne in narativne medije. animiranega filma.



2.2 Vloga umetne inteligence pri



pripovedovanju zgodb v animiranih filmih



Dijaki so za pisanje scenarija sprva uporabili orodje

ChatGPT. Kljub temu da jim je ChatGPT pomagal oblikovati

zgodbo, rezultati niso bili povsem zadovoljivi glede kreativnosti

in izvirnosti. Zato so se dijaki odločili, da raje uporabijo svojo

avtorsko zgodbo, ki so jo nato dodatno razvili in strukturirali z



Slika 1: Prikaz študije anim. filma Fritza H. in Marianne S. ključne točke zgodbe v koherentno strukturo, vključno z [2] ter dijake med gledanjem anim. filma. uporabo UI orodja Simpleshow, ki jim je pomagal organizirati

uvodom, zapletom, vrhuncem, razpletom in zaključkom. Ko je

bil scenarij dokončan, so dijaki prešli na ustvarjanje zgodborisa.

Človeški um deluje kot procesor zgodb in ne kot logični

Slika 2 nam prikazuje uporabo UI orodja Boords, za risanje

procesor. To pomeni, da so zgodbe ključen način, kako ljudje

zgodborisov. Dijaki so ustvarili okvirne risbe za vsak prizor iz

dojemajo svet, si zapomnijo informacije in komunicirajo med

scenarija, pri čemer jim je UI orodje Boords omogočil, da

seboj [4]. Zgodbe nam pomagajo ustvarjati smiselne povezave

vizualizirajo dogajanje in določijo ključne točke zgodbe.

med novimi in starimi izkušnjami, kar je temeljno za proces

Zgodboris je vključeval pomembne vizualne elemente, kot so

učenja. V izobraževalnih okoljih je zato priporočljivo uporabljati

postavitev likov, premiki kamere in zaporedje dogodkov.

pristop, ki omogoča avtentične učne priložnosti, kjer imajo

učenci izbiro, lastništvo in glas v procesu učenja. Dijaki lahko z



uporabo zgodb celo izboljšajo svoje učenje in razumevanje

kompleksnih konceptov. Naši možgani hrepenijo po zgodbah,

ker nam pomagajo ustvarjati čustvene povezave in bolje razumeti

ter si zapomniti informacije [5]. Med branjem zgodb so naši

možgani bolj aktivni, saj se aktivirajo enaki deli možganov, kot

bi se ob dejanskem doživljanju izkušnje. Zgodbe pomagajo tudi

pri izboljšanju spomina in povezanosti s prejeto vsebino, kar je

ključnega pomena za učinkovito trženje in komunikacijo.

Študija, objavljena na spletnem dnevniku Prakasha, R. [6] Slika 2: Izdelana vizualna pripoved zgodbe z uporabo UI

poudarja, da je pripovedovanje zgodb ključno pri sprejemanju orodij.

odločitev. Ljudje uporabljajo zgodbe za oblikovanje in krepitev

svojih kognitivnih procesov, saj zgodbe pomagajo pri Animacija je orodje, s katerim animatorji poskušajo popeljati

razumevanju kompleksnih situacij in omogočajo globlje gledalca skozi umetno ustvarjen svet na tak način, da se gledalec

čustvene povezave. Zgodbe so močno orodje, ki vpliva na naše vživi vanj. Za prvi prizor so dijaki želeli uporabiti futuristični

zaznavanje in odločanje, saj nam pomagajo interpretirati svet svet. Z uporabo UI orodja Kitl AI Image Generator so ustvarili

okoli nas in sprejemati premišljene odločitve. Pri ustvarjanju realistično in estetsko privlačno sliko, ki ustreza viziji njihove

zgodb je pomembno, da poznamo dejstva, kako naši možgani zgodbe. Dijaki so vnesli ključne elemente vpisa v Kitl AI Image

obdelujejo zvoke in slike, kako zaznavamo resničnost ter kako Generator, ki je na podlagi teh podatkov zgeneriral več možnih

slik. Dijaki so izbrali najbolj primerne slike, jih uredili v orodjih in učenje. Glavni namen je zagotoviti sprotne povratne

za urejanje slik, jih združili in uporabili v animaciji. Slika 3 nam informacije, ki pomagajo pri prilagoditvi poučevanja in učenja

prikazuje primer umetno ustvarjenega sveta v tehniki 2D [8, 9]. Sumativno ocenjevanje je proces ocenjevanja učenčevega

računalniške animacije. znanja in spretnosti na koncu učnega obdobja ali enote, da bi se

ocenila stopnja doseganja učnih ciljev. Glavni namen



sumativnega ocenjevanja je ovrednotiti končno znanje in

dosežke učencev ter jih primerjati s standardi ali cilji [10].



Rezultati ankete so pokazali, da dijaki še vedno interpretirajo

gibanje oblik na podoben način kot v študiji Fritza H. in

Marianne S., kar potrjuje trajnost teh kognitivnih procesov. Iz

raziskave dojemanja in interpretacije gibanja oblik, kjer je

sodelovalo 64 dijakov, so se pokazali zanimivi rezultati. Večina

dijakov (75 %) je menila, da video prikazuje zgodbo o lovu in

pobegu, medtem ko jih je 13 % videlo preprosto gibanje oblik, 8

% pa je menilo, da gre za nekakšno igro. Pri vprašanju o zgodbi,

ki so jo videli v gibanju oblik, jih je 81 % zaznalo prepir in



Slika 3: Umetno ustvarjen svet v tehniki 2D računalniške animacije. preganjanje, 11 % ljubezensko zgodbo, 8 % pa naključno gibanje

brez pomena. Glede čustev, ki so jih zaznali pri gibanjih



pomočjo Adobe programskega okolja, ki prav tako vključuje Ostale elemente zgodbe so dijaki izrisali in animirali s posameznih oblik, je 73 % dijakov zaznalo strah in jezo, 22 % nevtralnost, 5 % pa veselje in srečo. Kar 95 % dijakov je odnos med večjo in manjšo trikotno obliko opisalo kot konflikt, podporo UI. Adobe orodja, kot so Adobe Illustrator, Adobe After medtem ko jih je 3 % menilo, da ni nobenega posebnega odnosa, Effects in Adobe Photoshop, so jim omogočila, da ustvarijo 2 % pa je zaznalo sodelovanje. Krožna oblika v zgodbi je za 47 profesionalno izgledajoče animacije in vizualne učinke. UI % dijakov predstavljala žrtev, za 30 % heroja, za 23 % pa funkcionalnosti v teh orodjih so jim pomagale optimizirati nevtralno figuro. Pri vprašanju, kako bi spremenili video za delovni proces in izboljšati kakovost končnega izdelka. Slika 4 ustvarjanje drugačne zgodbe, bi 69 % dijakov spremenilo gibanje nam prikazuje posamezne prizore iz končnega animiranega filma oblik, 19 % bi spremenilo ozadje, 13 % pa bi dodalo več oblik. Kibo. Najbolj presenetljivo pri gibanju oblik je bilo za 69 % dijakov

način interakcije oblik, za 20 % hitrost gibanja, za 11 % pa barva

oblik. Glede pripisovanja čustev ali namenov neživim

predmetom je 45 % dijakov odgovorilo, da to počnejo občasno,

34 % pogosto, 19 % pa nikoli. Slika 5 nam prikazuje rezultate

ankete v obliki toplotne karte. Močnejša kot je rdeča barva, več

dijakov je izbralo to trditev. Na sliki se lepo vidi, kako so

elementi »dramatičnosti« v gibanju poganjali domišljijo zgodb.





Slika 4: Prikaz posameznega prizora iz končnega



animiranega filma Kibo.



2.3 Formativno in sumativno spremljanje in

ocenjevanje znanja

Spremljanje, preverjanje in ocenjevanje znanja je zaradi

kompleksnosti tisti del izobraževalnega procesa, ki zahteva

premišljeno načrtovanje in izvedbo. S stalnim spremljanjem

razvoja znanja posameznega učenca zbiramo podatke o vrzeli v

znanju zato, da najdemo vzroke za motnje učenja, medtem ko

lahko ugotovitve vključimo v naslednjo fazo načrtovanja, učenja



na srednješolskem nivoju pri pouku odprtega kurikula Animacije. Slika 5: Prikaz rezultata ankete: Raziskovanje dojemanja in poučevanja. V prispevku bomo predstavili, kako to izvajamo

in interpretacije gibanja oblik.

Uporabili bomo tako formativno kot tudi sumativni način



spremljanja in ocenjevanja znanja. Ta podobnost med zgodovinsko študijo in našimi rezultati Formativno ocenjevanje je proces zbiranja podatkov o poudarja trajnost in univerzalnost kognitivnih vzorcev v učenčevem učenju med učnim procesom, da bi učitelji in učenci interpretaciji animiranih vsebin. Naše ugotovitve so v skladu s lahko prepoznali močne in šibke točke ter izboljšali poučevanje sodobnimi raziskavami o pomembnosti pripovedovanja zgodb za

človeški um [4]. Zgodbe imajo ključno vlogo pri našem dijak in učitelj skupaj ugotovila, da sta dosegla želeni cilj oz.

dojemanju sveta in pri učenju. standard znanja glede na predhodno določene kriterije in

Rachel [5] poudarja, da naši možgani hrepenijo po zgodbah, navodila, zato sta lahko končala proces učenja z vrednotenjem.

ker nam pomagajo bolje razumeti in si zapomniti informacije. Dijaki so delali samostojno ali v skupinah in so se samovrednotili

Poleg tega pripovedovanje zgodb vpliva na naše odločitve in ter medvrstniško vrednotili (uporaba orodij Padlet, osebna

krepi kognitivne procese, kar še dodatno potrjuje pomen naših demonstracija svojih rešitev, kritičnega prijateljevanja, diskusija

ugotovitev v izobraževalnem kontekstu. ...). Preverjanje vmesnih faz izdelave je bilo usmerjeno v to, da

V kontekstu, kako UI spreminja področje animacije s dijaki ustrezno pojasnijo, kako rešitve danih izzivov delujejo,

poudarkom na ustvarjanju zgodb in uporabi UI orodij za ocenjevanje pa se izvaja na podlagi odobritve načeloma ustreznih

generiranje scenarijev, slik in ostalih elementov, učni cilji na rešitev. Učitelj je zagotavljal sprotne formativne povratne

formativni ravni obsegajo predvsem razumevanje kognitivnih informacije o tem, kako izboljšati pristope v reševanju izzivov.

procesov v socialni percepciji, spoznavanje različnih tehnik Projektna naloga je bila predstavljena na zaključnem seminarju z

animacije, razumevanje razvoja animacije kot umetniške in diaprojekcijo. Menimo, da če lahko dijak dokaže dvig učnih

pripovedne oblike ter razumevanja, kako UI vpliva na proces dosežkov in kakovost znanja s tem, da izdela kratki animirani

ustvarjanja animacije. Na formativni ravni se učni cilji film z uporabo UI orodij, ki ima močno sporočilnost in je

osredotočajo na razvoj spretnosti in kompetenc skozi sposoben samovrednotenja oz. medvrstniškega vrednotenja, je

kontinuirano ocenjevanje in povratne informacije. Cilj je nalogo uspešno opravil. Tak pristop omogoča celostno

pomagati dijakom pri izboljšanju njihovega učenja in izobraževanje, kjer dijaki ne le pridobijo teoretično znanje,

prilagajanju učnih pristopov. V kontekstu animiranega filma in temveč tudi razvijajo praktične spretnosti in kompetence, kot jih

uporabe UI v izobraževanju so dijaki aktivno sodelovali pri poudarjajo v članku omenjene študije [4, 5, 6].

razvijanju svoje sposobnosti ustvarjanja in pisanja zgodb z

vključevanjem vizualne pripovedi zgodbe. Dijaki so razvijali Slika 7 nam prikazuje uporabo aplikacije Padlet, kjer so dijaki

spretnosti ustvarjanja animacij, pri čemer so uporabljali različne ob zaključku in po predstavitvi projekta podali svoje mnenje.

tehnike in orodja, ki so podpirala UI. Z uporabo kritične analize Dijaki so s pomočjo QR kode zajeli vprašanja ter nanje anonimno

in refleksije so se dijaki naučili analizirati in kritično odgovarjali. Odgovarjali so na pet vprašanj:

samovrednotiti ter tudi vrstniško vrednotiti animirane vsebine, ki 1. Kaj že vem o zgradbi zgodbe, pisanju scenarijev, kakšne so

so jih ustvarili sami ali njihovi sošolci. Skozi proces so dijaki moje predstave o izdelavi anim. filma s pomočjo UI?

reflektirali svoj učni proces in sprejemali ukrepe za izboljšanje 2. Zakaj so snov, veščine, vrednote zame pomembne? Kako jih

svojih spretnosti in znanj. Učitelj je ves čas formativno spremljal bom lahko uporabil/-a pri drugih predmetih, v bodočem

napredek posameznega dijaka oz. skupine skozi posamezne poklicu?

procese izdelave (povratne informacije na osnutke nalog, 3. Koliko časa potrebujemo za izdelavo zgodbe, scenarija in

diskusije v razredu, samoevalvacija in vrstniška evalvacija) ter koliko za izdelavo anim. filma?

jim posredoval redne povratne informacije, ki so dijakom 4. Kako bom dokazal/-a, da sem cilj dosegel/-a?

pomagale razumeti, kaj so naredili dobro in kje so potrebne 5. Kaj sem se naučil/-a? Kako učinkovit/-a sem bil/-a?

dopolnitve.





Slika 6: Prikaz diskusije v razredu.



Slika 7: Prikaz analize po zaključku izdelanega projekta.

Integracija formativnih učnih ciljev ter sumativnega

ocenjevanja v izobraževalnem procesu omogočata celostno Opaziti je bilo, da je uporaba aplikacije Padlet spodbudila

izobraževanje, kjer dijaki ne le pridobijo teoretično znanje, učence k bolj kritičnemu in iskrenemu premisleku o zgradbi

temveč tudi razvijajo praktične spretnosti in kompetence. zgodbe, pisanju scenarijev in predstavi o izdelavi animiranega

Uporaba UI v tem kontekstu omogoča bolj inovativne in filma s pomočjo UI. Učenci so bili zmožni anonimno izraziti

dinamične učne metode, ki spodbujajo kreativnost, kritično svoje dojemanje, kar je povečalo stopnjo odprtosti in iskrenosti

mišljenje in aktivno učenje. Formativno ocenjevanje se torej v povratnih informacijah. Učenci so tudi dobro sprejeli možnost,

osredotoča na zbiranje sprotnih informacij od dijakov med učnim da svoje mnenje delijo brez pritiska javnega izpostavljanja, kar

procesom, medtem ko sumativno ocenjevanje analizira končne je izboljšalo kakovost njihovih odgovorov. Kot pozitivno bi

rezultate in dosežke dijakov po zaključku učne enote. Postopek izpostavili, da je orodje omogočilo takojšnje razmišljanje in

ocenjevanja se je začel v fazi sumativnega povzemanja, ko sta interakcijo, kot izziv pa bi se lahko štelo omejevanje globine

razprave zaradi narave kratkih odzivov na platformi. Za prvo



vprašanje lahko podamo, da so dijaki pokazali široko informacije, ki so jim pomagale izboljšati njihove pristope in

razumevanje zgradbe zgodb, pisanja scenarijev in uporabe UI pri rešitve, kar je ključno za uspešno formativno spremljanje in

izdelavi animiranih filmov. Navdušeni so nad potencialom UI, ki ocenjevanje znanja. Učenje torej postane vidno, ko učitelj pri

olajša delo, pospeši procese in generira nove ideje, vendar se sebi razjasni namene poučevanja, predvsem pa z učenci razjasni

zavedajo njene kompleksnosti in omejitev. Razumejo, da mora namene učenja. Jasni nameni učenja izboljšajo kakovost učenja

biti zgodba smiselno zgrajena in kreativna, scenarij pa vključuje in učnih dosežkov učencev. Formativno spremljanje in

opise dogajanja in dialoge, pri čemer UI pomaga avtomatizirati vrednotenje je pokazalo, da izboljša strategijo učenja in

proces, vendar človeški prispevek ostaja ključnega pomena. poučevanja, poveča kakovost znanja, predvsem pa vpliva na

Menijo, da UI dopolnjuje človeško kreativnost, saj je za končno večjo odgovornost dijakov za lastno učenje in znanje. Sklepamo

izdelavo čustveno bogate zgodbe nujno vključevanje človeške lahko, da je uporaba UI orodij v izobraževalnem procesu

perspektive. Pri drugem vprašanju, zakaj so snov, veščine, učinkovita metoda za izboljšanje učnih izidov, saj omogoča bolj

vrednote za njih pomembne in kako jih bodo lahko uporabili, so dinamične in interaktivne učne izkušnje. Tak pristop ne le

dijaki prepoznali pomembnost snovi, veščin in vrednot, saj te izboljša kakovost izobraževanja, temveč tudi pripravi dijake na

prispevajo k njihovemu celovitemu razvoju in uspehu v bodočem prihodnje izzive v svetu, kjer bo umetna inteligenca igrala vse

poklicu. Znanje, pridobljeno pri tem projektu, jim omogoča bolj pomembno vlogo.

nenehno izboljševanje, kritično mišljenje in ustvarjalnost.

Poudarjajo, da so pridobljene veščine uporabne pri drugih

predmetih ter pri reševanju realnih izzivov v poklicnem LITERATURA IN VIRI

življenju, medtem ko vrednote usmerjajo njihove etične [1] Dreamfarm Studios. 2023. Future of Animation: How Artificial



odločitve in vedenje. Ugotovili so, da jim umetnost in kreativnost Intelligence Will Change Animation Forever. Pridobljeno na: https://dreamfarmstudios.com/blog/how-ai-change-animation/ dodajata smisel življenja ter omogočata, da v vsakdanje življenje [2] Heider, F., & Simmel, M. 1944. An Experimental Study of Apparent

vnašajo delčke umetnosti. Koliko časa potrebujejo za izdelavo, Behavior. University of Illinois Press The American Journal of

so dijaki odgovorili, da je čas, potreben za izdelavo zgodbe, Psychology

Vol. 57, No. 2 (Apr., 1944), pp. 243-259 (17 pages). Dostopno na:

scenarija in animiranega filma, odvisen od zapletenosti projekta http://cs.uky.edu/~sgware/reading/papers/heider1944experimental.pdf

in kakovosti, ki jo želijo doseči. Pisanje zgodbe lahko traja od [3] Eagleman, D. 2015. The Brain: The Story of You, Canongate Books

[4] Harapnuik, D. 2018. The Human Mind is a Story Processor, Not a Logic

nekaj dni do tednov, medtem ko scenarij zahteva več tednov do Processor. Dostopno na: https://www.harapnuik.org/?p=7445

mesecev. Izdelava animiranega filma je najdaljša faza, ki lahko [5] Gillett, R. 2014. Why Our Brains Crave Storytelling In Marketing.

traja od nekaj mesecev do več let, odvisno od dolžine in Pridobljeno: https://www.fastcompany.com/3031419/why-our-brains-

crave-storytelling-in-marketing

zahtevnosti animacije. Glede doseganja cilja so dijaki menili, da [6] Prakash, R. 2023. Decision making and storytelling. Dostopno na:

ga bodo dosegli, ko bodo zadovoljni s končnim izdelkom https://www.rahulprakash.me/blog-2/decisionmaking

[7] Mohamed, A. 2021. Storytelling-A Scientific Analysis in Life/Philosophy.

animiranega filma. Uspeh bodo merili s ponosom na dosežene Dostopno na: https://countercurrents.org/2021/05/storytelling-a-



rezultate in prepoznavanjem lastnega napredka. Na vprašanje, scientific-analysis/ [8] Carnegie Mellon University. 2024. What is the difference between kaj so se naučili in kako so bili učinkoviti, so odgovorili, da so formative and summative assessment? Dostopno na:

se naučili celoten proces izdelave animiranega filma, pridobili https://www.cmu.edu/teaching/assessment/basics/formative-

izkušnje timskega dela, razporeditve nalog in praktičnih tehnik summative.html

[9] Brodnik, V. 2024. Formativno spremljanje in vrednotenje znanja in

animacije. učenja. Zavod RS za šolstvo. Gradivo za izobraževanje dijakov na daljavo

s portala Jazon na povezavi: https://jazon.splet.arnes.si/formativno-

spremljanje-in-vrednotenje-znanja-in-ucenja/



3 [10] University of San Diego. 2024. Formative vs. Summative Assessment: ZAKLJUČEK

What’s the Difference? USD Pro Ed. Dostopno na:

https://pce.sandiego.edu/formative-vs-summative-assessment/

V raziskavi smo preučili vpliv UI na ustvarjanje zgodb v

animiranih filmih in njegov pomen za izobraževalni sistem.

Ugotovili smo, da lahko uporaba UI orodij, kot so Simpleshow,

Boords in Kitl AI Image Generator, bistveno pohitrijo in

izboljšajo proces ustvarjanja zgodb in animiranih vsebin, kar

omogoča bolj učinkovito in angažirano učenje dijakov. Naša

raziskava je pokazala, da so dijaki, ki so sodelovali v procesu,

razvili sposobnosti za ustvarjanje kompleksnih in čustveno

nabitih zgodb, kar je povečalo njihovo angažiranost in

sposobnost pomnjenja vsebine. Analiza rezultatov ankete je

potrdila ugotovitve prejšnjih študij Fritza H. in Marianne S. [2],

saj so dijaki interpretirali gibanje geometrijskih oblik trikotnikov

in krogov kot zgodbe z elementi lova, zasledovanja in konfliktov.

Študije Eaglemana [3], Harapnuika [4], Gillett [5] in Prakasha

[6] poudarjajo pomembnost pripovedovanja zgodb za človeški

um, saj zgodbe pomagajo bolje razumeti, si zapomniti

informacije ter vplivajo na sprejemanje odločitev. Integracija

formativnih učnih ciljev v izobraževalni proces omogoča

celostno izobraževanje, kjer dijaki ne le pridobijo teoretično

znanje, temveč tudi razvijajo praktične spretnosti in kompetence.

Skozi celoten proces so dijaki prejeli sprotne povratne



Matematika, logika in Mathematica



Mathematics, Logics and Mathematica



Izidor Hafner

Upokojeni nezaslužni docent

Univerze v Ljubljani

izi.hafner@gmail.com



Tako je eno od maturitetnih vprašanj 'kdaj sta dve množici

POVZETEK enaki?' V logiki bi se vprašali 'kako ugotovimo, da gre za isto

Sodobna (matematična) logika izvira iz matematike, v množico?' Še nekaj primerov definicij. Diferenčni količnik

nasprotni smeri pa se še vedno v matematiki uporabljajo oznake, funkcije f v x pri spremembi h = (f(x+h)-f(x))/h. Odvod funkcije

ki niso povsem v skladu z logičnimi pravili. Na primer, nariši f v točki x: f'(x)=lim h->0 (f(x+h)-f(x))/h. Desna stran je odvisna

graf funkcije f(x), namesto nariši graf funkcije f. Gre za definicije, of f in x (to sta prosti spremenljivki), ni pa odvisna od h (vezana

kjer ni dovolj poudarka na vlogi parametrov v izrazih. spremenljivka). Pozor, definirano je f'(x), ne pa tudi (f(x))'.

Programski jeziki morajo upoštevati logiko, zato prihaja do Zadnje je očitno prepuščeno dijakom. Gre za pravilno definicijo.

določenih težav pri razumevanju matematike s strani tistih, ki jim Oznaka za nedoločeni integral je ena od najslabših matematičnih

matematika ne leži. oznak. Nedoločeni integral funkcije f je vsaka takšna funkcija F,

da je F'=f (to je pravilna definicija). V simbolizmu F  NI(f) 



KLJUČNE BESEDE F'=f ( pomeni je). (Ne pa F = NI(f)  F'=f . ) Ni pa prav

Definicija, modul, postopek, funkcija, spremenljivka, naslednje: 2xdx=x2+C (leva stran je neodvina of C, desna pa je.

parameter 2xdx=x2+C bi morali zapisati 2xdx={x2+C; xR}. Postavlja se

vprašanje, ali je naslednje sklepanje pravilno: 2xdx=x2+1,



ABSTRACT  2 2 2 2 2 2 x d x = x , x +1= x , 0=1. Iz x  2 x d x in x+12xdx ne moremo

Modern (mathematical) logic originates from mathematics, sklepati 2 2 x +1= x. Iz aM in bM ne sledi a=b; iz a=c in b=c pa

but in the opposite direction, notations are still used in sledi a=b. Enostavna rešitev bi bila da definirano pojma

mathematics that do not fully conform to logical rules. For nedoločen in nedoločeni integral. Nedoločen integral funkcije f

example, plot the graph of the function f(x) instead of plot the je vsaka takšna funkcija F, da je F'=f. Nedoločeni integral pa je

graph of the function f. These are definitions where there is not dogovorjeno izbran nedoločen integral (po izrazu najbolj

enough emphasis on the role of parameters in the expressions. enostaven). Torej, nedoločen integral funkcije sinx je {-cosx+C;

Programming languages must use logic, so there are certain C R}, nedoločeni integral funkcije sinx je -cosx (po dogovoru).

difficulties in understanding mathematics by those who do not Oznako  bi uporabljali za nedoločeni integral (torej sinxdx=-

like mathematics. cosx). Na vprašanje, ali je to edina funkcija, pa je odgovor ne.

Vse funkcije -cosx+C imajo odvod sinx. Pri večini nalog v



KEYWORDS gimnaziji potrebujemo (Newton-Leibnizovo pravilo) le eno



Definition, module, procedure, function, variable, parameter primitivno funkcijo (nedoločen integral). Če potrebujemo vse primitivne funkcije pa dodamo C. Problem nedoločenega

intagrala je v tem, da je to inverzna relacija k odvajanju, vendar

ni enolična. Podobno je z inverzi trigonometričnih funkcij. Tole

1 je pravilna definicija: y = arctg(x)  (y[- x = tg(y)). Matematika

Zelo slaba je tudi tale opredelitev: funkcija f:A->B, je predpis, ki

Matematična praksa je, da se porabi čim manj znakov, zato

2 vsakemu elementu množice A priredi natanko določen element

takile izrazi: |x|, x , sinx, n!, an. To nekoliko zamegljuje pravilo

množice B. Slabost te definicije je izraz 'predpis'. Tu je rešitev

definicije. Včasih so bile funkcije (realne spremenljivke x)

teorija množic, kjer je funkcija opredeljena kot enolična relacija,

matematični izrazi, v katerih nastopa samo spremenljivka x.

relacija pa je neka množica urejenih parov. Tu lahko definiramo

Naloga se je lahko glasila, da narišemo funkcijo sinx in se ne

tudi pojem inverzne relacije in pojem omejitve relacije na neko

razlikuje med sinx (vrednost funkcije pri x) in funkcijo sin.

Naloga, dan je polinom p(x)=x +x3+ax2+2bx+6. Za a=0 in b=3

4 množico.



poiščite realne ničle polinoma. Določite a in b tako, da bo



polinom p deljiv s q(x)=x2+bx+2 (pri definicijah za p in q 2 Logika parametri niso navedeni in izgleda, da leva stran ni odvisna od

parametrov, desna pa je). Splošna kvadratna funkcija je podana Matematična logika je logika, v kateri lahko aksiomatsko

z enakostjo f( 2 izpeljemo celotno matematiko. Tu nastopata v bistvu dve x )=a x +b x +c, kjer so a, b in c dana števila in ni a=0.

V matematiki se znak '=' bere 'je enako' v logiki pa 'je isto kot'. možnosti. Prva je teorija množic, to je teorija, ki jo formuliramo

v logiki prvega reda z enakostjo (identiteto). Aksiom, da skozi



dve različni točki poteka natanko določena premica, bi v logiki

izrazili brez 'različni', saj če imamo dve točki, gotovo nista

istovetni. Vsi matematični objekti so množice in je na primer

1=sin slovnično pravilen, a neresničen stavek. Druga možnost je

teorija tipov, v kateri so na primer funkcije nekaj drugega kot

števila. Tu 1=sin sploh ni stavek (števila spadajo v drug tip kot

funkcije). Oglejmo si sedaj pravila definicij v logiki.

Najenostavnejši primer je eksplicitna definicija novega

funkcijskega znaka. Na primer tanx=sinx/cosx. Ali pa f(x)=2x+3,

ki pa velja le znotraj posamezne naloge. Splošno kvadratno

funkcijo lahko uvedemo z definicijo K[a,b,c](x)=ax2+bx+c.

Pravilo je naslednje. Definicija uvaja nov znak 'K', ki mu v

oklepaju (ali oklepajih) sledijo vse spremenljivke in parametri

(vsak samo enkrat), ki nastopajo prosto na desni strani. Zgornjo

nalogo s polinomi bi po naše formulirali takole: dan je polinom

(ali dvoparametrična družina polinomov) p[a,

b]( 4 3 2 x )= x + x +a x+2bx+6. Poiščite realne ničle polinoma p[0, 3].

Določite a in b tako, da bo polinom p deljiv s q[b]( 2 Fig 1: Graf y=sin(x) in x=sin(y) x )= x +b x +2.

Oglejmo si še nekaj pravilnih definicij relacij. Identiteta funkcij:



f=g  xR(f(x)=g(x)), a|b  c(b=ac) (a deli b). Za relacijo

'večkratnik' nimamo posebnega znaka, lahko pa bi uporabili |-1,

če smo prej definirali pojem inverzne relacije: xR-1y  yRx.

V matematiki pogosto nastopajo definicije kot so: Lik A je

paralelogram  A je štirikotnik, ki ima dva para vzporednih

stranic. Lik A je deltoid Fig 2: Graf y=sin(x) in y=arcsin(x)  A je štirikotnik, ki ima dva para



ima en par vzporednih stranic. To so definicije z navedbo Če zapišemo Integrate[Sin[x],x] bomo dobili -Cos[x] in ne - Cos[x]+C, kar je v skladu s predlogom, da je nedoločeni integral enakih sosednih stranic. Lik A je trapez  A je štirikotnik, ki



štirikotnik) in vrste s specifično razliko (differentia specifica). izbran nedoločen integral. Če pa rešujemo diferencialno enačbo najbližjega roda (genus proximum, v zgornjih primerih je to



krožnica spremenljivke x, določena z izrazom c1-Cos[x] . Če definiramo  (  S  r>0)(L={A; d(A,S)=r}), L je krog  f= Function[{x}, c1-Cos[x]] in pišemo f[x], dobimo c1-Cos[x]. ( Definicije z ekvivalenco so pogoste v matemtiki. Primeri: L je y'=sin(x) dobimo Function[{x}, c1-Cos[x]] (beremo: funkcija



A v B Sr>0)(L={A; d(A,S) ≤r}), funkcija f je injektivna funkcija iz Tu je še ostanek tradicionalizma, saj bi pričakovali rešitev  (  a,b  A)(a ≠ b  f(a) ≠ f(b)). Pojem praštevila diferencialne enačbe v obliki Function[{x, c1}, c1-Cos[x]], torej definiramo lahko takole: število n je praštevilo  tudi kot funkcijo parametra. Seveda pa lahko samo definiramo (  m  N)(m|n  m=1  m=n). g[x_,C_]:= C-Cos[x].



3 REFERENCES Wolfram Mathematica [1] Definicije v logiki prvega reda http://www.logika.si/revija/revije-

Če bi v definiciji funkcije napisali 'algoritem' na mesto splet/Logika-2017-2018_4.pdf

'predpis', bi bili že pri programiranju. Temu bi lahko rekli tudi

konstruktivna matematika. Mathematica je programski sistem

(jezik), ki je zelo blizu običajni matematični praksi. Splošno

linearno funkcijo dobimo z stavkom L[k_,n_][x_]:=kx+n. Brez

tistih črtic pri spremenljivkah, je to običajna definicija. Če bomo

vpisali to definicijo, potem pa vnesli L[1,2][3], bo odgovor 5.

Zgornja definicija praštevila ni konstruktivna, saj bi lahko začeli

preverjati deljivost števil n+1, n+2, … z n. Konstruktivna je



definicija P(n)  (mm=1m=n).

Če hkrati narišemo y=sin(x) in x=sin(y), dobimo fig 1, če

narišemo hkrati y=sin(x) in y=arcsin(x) pa fig 2.





Reklama za branje knjig




Promotion of Reading Books



Polona Jelen

šolska knjižnica

Osnovna šola Griže

Griže, Slovenija

polona.jelen@guest.arnes.si



students in a team-oriented learning environment with the little

POVZETEK help of modern technology (Actionbound application).

Živimo v dobi pospešene digitalizacije in komuniciranja

preko mobilnih naprav in družbenih omrežij. Največji izziv v tem KEYWORDS

času predstavlja za učitelje in skrbnike, kako otroke odvrniti od Reading literacy, promotion of reading, Actionbound

prekomernega gledanja v ekrane in jih preusmeriti v branje knjig. application, team work, cross-curricular learning, stumulating

Že res, da tudi preko spleta veliko beremo, vendar se moramo learning environment, formative assesment.

zavedati, da je to branje samo površinsko in ne daje globokega

ter kritičnega razmišljanja o prebranem. Če želimo vzgojiti

mladega otroka v kritičnega in samostojnega intelektualca, je 1 MOTIVACIJA ZA BRANJE

potrebno spodbujati interes za branje kakovostne litereature, ki Prepoznavanje različnih elementov bralne motivacije je prvi

bo bogatila besedni zaklad ter kritično razmišljanje. Sodobna korak k razumevanju in spodbujanju zavzetosti za vseživljenjsko

tehnologija nam naj bo v podporo pri delu in učenju, ne pa glavno branje (Bucik in Pečjak, 2004) [1]. Pri otrocih poznamo različne

orodje za zapolnitev prostega časa. V prispevku smo povzeli motivacijske dejavnike za branje, ki jih delimo na zunanje in

aktivnost, ki smo jo izvajali v šoli v sklopu knjižnično- notranje. Pri zunanjem motivacijskem dejavniku otrok bere

informacijskih znanj pri starejših učencih v osnovni šoli in z njo zaradi posledic, ki jih branje prinese, na primer pohvala drugih,

poskušali učence notranje motivirati za branje, jih spomniti, nagrada, ocena ipd., in ne, ker bi ga to veselilo. Pri notranji

zakaj je branje tako pomembna aktivnost v dobi šolanja, ter motivaciji za branje pa želimo, da otrok dojame branje kot

hkrati na zabaven način s pomočjo sodobne tehnologije pozitivno aktivnost, ki prinaša užitek, sprostitev in tudi znanje.

(Actionbound aplikacija) povezati učence v timskem spobudnem To motivacijo je pri otrocih najtežje doseči in jo vzdrževati. Zato

učnem okolju. je pomembno, da v času digitalizacija poudarjamo pomen branja

knjig zaradi pozitivnih posledic na otrokov bodoči intelektualni

KLJUČNE BESEDE razvoj in izobrazbo. Branje mora biti povezano s pozitivnimi

Bralna pismenost, promocija branja, aplikacija Actionbound, bralnimi izkušnjami in pozitivnimi čustvi. Bucikova in

timsko delo, medpredmetna povezava, spodbudno učno okolje, Pečjakova (2004) ugotavljata, da morajo pri bralnih spodbudah

formativno spremljanje sodelovati vsi pedagoški delavci z več branja v šoli, z večjo

možnostjo izbire bralnih gradiv, z večjim upoštevanjem bralnih

ABSTRACT interesov učencev, s sistematičnim razvijanjem bralnih

We live in an era of accelerated digitalization and sposobnosti in zavedanjem, da je branje process, ki ga je

communication via mobile devices and social networks. The potrebno spodbujati na različne načine in v vseh obdobjih

biggest challenge at this time for teachers and parents is how to življenja, zlasti pa v osnovni šoli. Zelo pomembno je redno

distract children from excessive using of mobile devices and obiskovanje knjižnic in knjigarn, pogovor o prebranem ter zgled

redirecting them to reading books. The truth is that we read a lot in spobude staršev vse od zgodnjega otroštva dalje.



online, but this kind of reading is shallow, not deep and without 1.1 ZAKAJ MORAMO BRATI? critical reflection on what is read. If we want to raise a young

child to be a critical and independent intellectual, it is necessary To vprašanje si pogosto zastavljajo učenci, ko se učijo črke,

to encourage interest in reading quality literature that will enrich povezovati v besede, besede pa v povedi. Prav tako se to

student’s vocabulary and critical thinking. Modern technology sprašujejo učenci zadnje triade osnovne šole. Vsako šolsko leto

should support and not lead our work and study. In the paper we pripravim učno uro na to temo in z razlinimi oblikami ter

summarized the activity that we carried out at during lesson of aktivnostmi pripeljem na koncu učence do spoznanja, da brez

library and information skills (KIZ) for older student in primary branja ni znanja. Učence sem v skupinah (po 3 ali 4) pozvala, da

school and we tried to motivate students to read (more), remind na plakat s pomočjo možganske nevihte zapišejo čimveč

them why reading is such an important activity, especially during razlogov, zakaj je branje koristno. Njihovi skupni odgovori so

primary education and at the same time a fun way to connect bili znani vsem: širjenje besednega zaklada, učenje novih

spoznanj, razvijanje empatije, razvijanje domišljije, izboljšanje digitalni časovni trak dogodkov ali ogled zanimivih krajev z

spomina, kritično razmišljanje, telovadba za možgane, uporabo koordinat GPS ter vnaprej postavljenih kod in skrivnosti.

razgledanost, sprostitev in zabava. Ugotovili so veliko razlogov Odlično orodje za vaje prebijanja ledu, zgodovinska ali

za branje, a za mnoge učence še vedno ne dovolj za začetek arheološka najdišča, ali preprosto za predstavitev vizije za

branja knjig. prihodnost. (povz. po https://en.actionbound.com/, 2. 8. 2024)

[4].

Učenci so jo uporabili za snemanje in montiranje kratkih

2 SPODBUDNO UČNO OKOLJE filmčkov kot reklama za branje knjig ter na koncu za

Številne raziskave kažejo na velik pomen, ki ga ima varno in medvrstniško evalvacijo.

spodbudno učno okolje na kakovost in uspešnost vzgoje in



izobraževanja.

K dobri vključenosti učencev in njihovi učni zavzetosti

odločilno prispevajo: zanimiv pouk, ki ga učenci doživljajo kot

smiselnega in vrednega zavzemanja, učni proces, v katerega so

učinkovito vključeni, dober odnos oz. stik učitelja z učenci in

učinkovito vodenje razreda. (Rutar Ilc, 2021) [2].

Spodbudno in varno učno okolje je prav tako ena od smernic

na šoli, h kateri stremimo in si zanjo prizadevamo. Učenci so bili

tekom dveh ur zelo motivirani za delo, dobili so jasna navodila

in spodbude s strani knjižničarke.



3 Slika 1: Aplikacija Actionbound (naslovna stran) MEDPREDMETNO POVEZOVANJE

Medpredmetno povezovanje je didaktični pristop, kjer

povezujemo vsebine in spretnosti različnih predmetnih področij. 5 FORMATIVNO SPREMLJANJE

Ima veliko pozitivnih učinkov, kot npr. spodbujanje aktivne S formativnim spremljanjem učitelj omogoča učencem

vloge učencev v učnem procesu, ki pridobivajo trajnejše znanje, soustvarjanje učnega procesa (namenov učenja, kriterijev

uporabo v novih situacijah. Učenci so za delo in učenje bolj uspešnosti, osebnih ciljev, zastavljanje vprašanj,

motivirani. Učitelji tako prihranimo na času, saj se vsebine v samovrednotenje, vrstniško vrednotenje itd.), kar vodi k

učnih načrtih zelo povezujejo. V učni uri sem povezala aktivnejši vlogi učencev ter h kakovostnemu in trjnejšemu

slovenščino v 8. razredu (reklama) in KIZ (motiviranje za branje). znanju. Učenci vrednotijo svoje delo in so deležni kakovostnih

Učenci so spoznali cilje iz UN za SLJ 3. VIO: tvorijo besedilo povratnih informacij tako s strani učitelja kot sošolcev z

določene vrste (oglasno besedilo). namenom izboljševanja učenja (ugotavljanja v kolikšni meri so

Standardi znanja/pričakovani dosežki: dosegli zastavljene cilje ter izboljševanja dosežkov

- tvorijo besedilnovrstno ustrezna, smiselna in (http://www.zrss.si/strokovne-resitve/formativno-spremljanje-2/,

razumljiva besedila, skladna z jezikovno in slogovno normo 2. 8. 2024) [5].

knjižne zvrsti, Na šoli že več let sledimo načelom formativne pedagogike ter

- vrednotijo zanimivost, verodostojnost, razumljivost, jo vnašamo v svoje pedgoške ure. Elemente formativnega

zgradbeno ustreznost in jezikovno ustreznost in pravilnost spremljanja pogosto vpeljujem v svojo prakso, saj prinašajo

besedila, ustreznost nebesednih spremljevalcev govorjenja oz. pozitivne učinke na učenje in poučevanje. Tokrat sem uporabila

pisanja ter utemeljijo svoje mnenje in predlagajo izboljšave. metodo možganske nevihte za iskanje razlogov za branje (namen

Cilji učne ure KIZ-a v 3. VIO: učenja) ter skupno oblikovanje kriterijev uspešnosti o dobri

Poznajo knjižnico kot prostor predstavljanja ustvarjalnosti. reklami, načrtovanje aktivnega dela - snemanja filma (scenarij)

Berejo različne leposlovne vsebine za osebno rast in ter končno medvrstniško evalvacijo posnetih reklam. Prav

razvedrilo. povratna informacija je bistven element učenja.

Motivirajo sošolce za branje knjig. S konstruktivno, pozitivno in prilagojeno povratno

(https://www.gov.si/assets/ministrstva/MVI/Dokumenti/Osnovn informacijo podpiramo vsakega učenca na njegovi učni poti ter

a-sola/Ucni-nacrti/obvezni/UN_slovenscina.pdf, 20. 8. 2024) mu omogočamo razvijanje samozavesti, samostojnosti in

[3]. notranje motivacije za nadaljnje učenje (www.formativno.si, 5.

8. 2024) [6]. Prav tako uporabna je povratna informacija učencev

učitelju, saj dobi učitelj potrditev ali mnenje učencev o učni uri.

4 APLIKACIJA ACTIONBOUND

Actionbound je aplikacija za igranje digitalno interaktivnih

lovov na zbiralce, ki učenca vodijo na pot odkrivanja (slika 1). 6 POTEK UČNE URE

Uporablja se lahko na pametnih telefonih ali tablicah. Temu Učence sem na začetku dveh učnih ur naključno razdelila v

večpredstavnostnemu lovu pravimo 'meje'. Program dobesedno skupine (po 3 ali 4) in jih pozvala naj zapišejo razloge, zakaj je

širi našo resničnost tako, da izboljša interakcijo ljudi v resničnem koristno brati. Sledilo je poročanje vsake skupine in razgovor o

življenju med uporabo njihovih pametnih telefonov in tablic. slišanem. Skupaj smo ugotavljali, kaj bo namen teh učnih ur.

Ustvarite lahko svojo igro pobega, ki temelji na aplikaciji, Napovedala sem, da bodo morali posneti reklamo za branje knjig.

Ker je reklama del učne snovi pri slovenščini v 8. razredu, sva z [3] https://www.gov.si/assets/ministrstva/MVI/Dokumenti/Osnovna-

učiteljico slovenščine dorekli, da bomo to snov medpredmetno sola/Ucni-nacrti/obvezni/UN_slovenscina.pdf

[4] https://en.actionbound.com/

povezali z uro knjižnične vzgoje. Sledil je pogovor o lastnostih [5] http://www.zrss.si/strokovne-resitve/formativno-spremljanje-2

dobre reklame in hkrati skupno oblikovanje kriterijev uspešnosti [6] https://formativno.si/

za njihov izdelek – video reklama za branje knjig. Z učenci smo

dorekli, da mora biti reklama privlačna na pogled, imeti mora

slogan, po katerem si reklamo zapomnimo, govorci morajo pred

kamero govoriti jasno, razločno in glasno, vsebinsko mora

vsebovati vsaj tri razloge za branje knjig ter vsaj en interaktivni

element (glasbo, napis, zvočni ali video efekt).

Naslednjo učno uro je sledilo načrtovanje snemanja,

porazdelitev vlog v skupini, pisanje scenarija ter snemanje po

različnih kotičkih šole. Pri snemanju z mobilnimi telefoni so

učenci spoznali apliakcijo Actionbound in jo hitro usvojili. Ko

so bili zadovoljni s svojim izdelkom, so ga naložili v skupen

prostor na aplikaciji. Sledil je zabaven zaključek ure – ogled vseh

video reklam (slika 2). Učence sem usmerila na kriterije

uspešnosti, ki smo jih predhodno oblikovali, saj so morali podati

ustno povratno informacijo pri vsakem video izdelku glede na

kriterije. Učenci so se zelo zabavali ob gledanju video reklam in

hkrati večkratno obnovili, zakaj je branje res pomembno. Na

koncu učne ure so prejeli izstopni listek, na katerem so izrazili,

kako so se počutili med urama ter kaj jim je bilo najbolj všeč.





Slika 2: Skupina osmošolcev pri snemanju reklame.



7 ZAKLJUČEK

Učni uri z naslovom Reklama za branje knjig sta dokaz, da

lahko učinkovito združimo več metod in oblik dela v skop dveh

učnih ur, in sicer: skupinsko delo, formativno pedagogiko,

spodbudno učno okolje, medpredmetno povezanost ter IKT

orodja. Učenci so bili nad skopom ur navdušeni, saj so v izhodnih

listkih, kjer sem želela izvedeti o počutju in pozitivnih vidikih

ure, največkrat izjavili, da jim je bila ura všeč, ker so bili aktivni,

ker so sodelovali v skupini s pomočjo telefonov in aplikacije, ker

so bili motivirani, da izdelajo najboljšo video reklamo za branje

knjig. Hkrati smo večkrat spomnili drug drugega na prednosti

branja knjig, na katere učenci velikokrat pozabijo. Ker opažam

same pozitivne lastnosti in veliko učinkovitost, bom tak način

dela in tovrstno aktivnost izvajala vsako šolsko leto.



VIRI

[1] Nataša Bucik in Sonja Pečjak, 2004. Učenčev izbor – ključ do motivacije

za branje?. Otrok in knjiga, letnik 31, številka 60 URN:NBN:SI:DOC-

9GSBQ2BJ from http://www.dlib.si.

[2] https://www.zrss.si/wp-content/uploads/2021/03/2019-05-10-kako-do-

spodbudnega-in-varnega-ucnega-okolja.pdf



Uporaba video kamere pri poučevanju športne vzgoje



Using Video Cameras in Teaching Physical Education



Anže Jurgele

Šolski center Kranj

Kranj, Slovenija

anze.jurgele@sckr.si



POVZETEK

1 UVOD

Informacijsko-komunikacijska tehnologija (IKT) je v

današnjem hitro spreminjajočem se svetu postala ključni element V učnem načrtu predmeta športne vzgoje je le-ta opredeljena

vsakdanjega življenja. Prispeva k sodobni družbi in omogoča kot »nenehen proces bogatenja znanja, razvijanja sposobnosti in

posameznikom bolj udobno življenje ter olajšuje številne napore. lastnosti ter pomembno sredstvo za oblikovanje osebnosti in

Z razvojem interneta in napredkom IKT se spreminja tudi način odnosov med posamezniki«. Med športno vzgojo poučujemo

učenja in poučevanja v srednjih šolah. Učitelji vse pogosteje otroke o pomenu vključevanja športnih aktivnosti v njihovo

integrirajo IKT v pouk, medtem ko dijaki izkoriščajo tehnologijo vsakdanje življenje, kar je ključno za ohranjanje zdravega

za svoje učenje. Sodobna tehnologija tako dopolnjuje in življenjskega sloga. S tem jim omogočamo, da skrbijo za svoje

izboljšuje izobraževalni proces. Z uporabo IKT lahko učitelji dobro počutje, zdravje, vitalnost ter ohranjajo pozitiven pogled

pouk naredijo bolj interaktiven in učinkovit, spodbujajo na življenje [1].

sodelovanje dijakov ter dodatno povečujejo motivacijo za učenje.

V prispevku je predstavljena uporaba video kamere kot Svetovna zdravstvena organizacija – World Health

pripomočka za poučevanje športnih vsebin pri predmetu športna Organization (WHO) opredeljuje gibalno aktivnost kot vsakršno

vzgoja, kar še dodatno obogati in raznoliko popestri izkušnjo gibanje telesa, ki ga povzročajo skeletne mišice in ki za to

dijakov. zahteva porabo energije. Gibalna neaktivnost (pomanjkanje

gibalne dejavnosti) je četrti vodilni dejavnik tveganja za



KLJUČNE BESEDE svetovno umrljivost – 6 % smrtnih ţrtev na svetovni ravni [2].

WHO že nekaj let opozarja na zaskrbljujoče razmere, saj večina

IKT, športna vzgoja, video kamera, video analiza

mladostnikov med 11. in 17. letom starosti premalo telovadi in



ABSTRACT predstavlja resno tveganje za njihovo zdravje, pa tudi za razvoj preveč časa preživi z zasloni mobilnih telefonov. To stanje

Information and Communication Technology (ICT) has možganov in socialnih veščin, opozarjajo strokovnjaki WHO.

become a key element of everyday life in today's rapidly WHO priporoča, da bi se otroci vsaj eno uro na dan gibali, vendar

changing world. It contributes to modern society, enabling je neaktivnost med dekleti še posebej pogosta. individuals to lead more comfortable lives and facilitating

numerous efforts. With the development of the internet and Podobne težave se pojavljajo tudi v Sloveniji. Leta 2016 je

advancements in ICT, the way teaching and learning are bilo kar 80 odstotkov mladostnikov premalo telesno aktivnih, kar

conducted in secondary schools is also evolving. Teachers je le za drobec manj kot leta 2001. Športna vzgoja se izkaže kot

increasingly integrate ICT into their lessons, while students eden od ključnih učnih predmetov, saj je gibanje uporabljeno kot

utilize technology for their learning. Thus, modern technology sredstvo za pridobivanje znanja. Tudi tisti učenci, ki imajo sicer

complements and enhances the educational process. Through the težave pri učenju akademskih vsebin, lahko izkažejo izjemne

use of ICT, teachers can make lessons more interactive and sposobnosti v športni vzgoji. A športna vzgoja ni le gibanje,

effective, promote student collaboration, and further increase zabava in sprostitev; je predmet, ki sledi določenim ciljem v

motivation for learning. The article explores the use of video razvoju gibalnih sposobnosti in pridobivanju znanja o pravilni

cameras as a tool for teaching sports topics in physical education, izvedbi gibov [3].

enriching and diversifying students' learning experiences. Po Škofu (2010) je prav skrb za razvoj telesne

pripravljenosti/zmogljivosti in zdravja mladih oziroma vzgoja za

KEYWORDS zdrav življenjski slog otrok in mladine osrednji cilj športne

ICT, Phyisical Education, video camera, video analysis vzgoje v današnjem času. [4].

Športni način življenja ni le niz vrednot; je celoten sistem

prepričanj, ki se razvijajo skozi naše življenje in postanejo naš

edinstven osebni pristop k življenju. Postopoma oblikujemo

svoje življenjske vzorce, ki izražajo naše individualne potrebe in

izhajajo iz našega notranjega jaza. Ta notranja usmerjenost nas

vodi k vzpostavitvi discipline in nam daje univerzalno



življenjsko energijo, ki odpira nova obzorja in dimenzije našega - izboljšanje pogoje učenja in poučevanja [12].

bivanja [5]. Uvedba IKT v pouk ima naslednje pomembne vplive na

izobraževanje:

- razvijanje spretnosti učencev, ki so potrebne za

2 KAJ JE INFROMACIJSKO- življenje in delo v 21. stoletju;

KOMUNIKACIJSKA TEHNOLOGIJA - učitelje spodbuja k spremembi načina učenja v razredu

(IKT)? z interaktivnimi in dinamičnimi viri;

Informacijska-komunikacijska tehnologija (IKT) je skupen - za učence zagotavlja bogatejšo izkušnjo učenja in več

izraz različnih računalniških, informacijskih in komunikacijskih motivacije [10].



naprav, ki so postale naš vsakdanji spremljevalec [6]. 2.2 UPORABA IKT PRI PREDMETU

Novljan IKT opredeljuje kot informacijsko pismenost, katero

dosegamo z določenimi komunikacijskimi mediji (računalniški ŠPORTNA VZGOJA

programi, medmrežje, videoposnetki in njihove analize itd.). Učenje in poučevanje s pomočjo IKT, svetovnega spleta in

Opredeljuje jo kot sposobnost opredelitve informacijske potrebe, virtualnih okolij, je vsekakor tudi pri pouku športne vzgoje

pridobivanja, vrednotenja in uporabe informacij iz različnih dobrodošlo in potrebno, saj ponuja številne priložnosti za

virov [7]. kvalitetnejše usvajanje zastavljenih ciljev ter doseganje čim

Informacijsko-komunikacijske tehnologije (IKT) so širok boljših rezultatov učenja. Uporaba IKT nadgradi in bogati

pojem, nadpomenka, ki vključuje vse vrste komunikacijskih klasično poučevanje in naj ne bo predmet učenja ali orodje za

naprav ali aplikacij. Zajemajo radio, televizijo, mobilno učenje, temveč medij, ki pomaga pri pridobivanju znanja.

telefonijo, računalnik, medmrežje, programsko in strojno Največkrat se pri športni vzgoji uporablja video kamera ali

opremo, satelitske sisteme, pomnilniške enote in mnogo drugih fotoaparat, razna CD in DVD gradiva, svetovni splet, merilnike

sodobnih tehnologij. Omenjene tehnologije nudijo številne srčnega utripa, štoparice in prenosni računalnik [13].

storitve, med njimi so videokonference in učenje na daljavo. Tudi Jurak s sodelavci (2007), navaja več primerov možnosti

te storitve opredeljujemo kot IKT. Na področju izobraževanja uporabe IKT pri pouku športne vzgoje:

nam IKT nudi podporo pri učenju in poučevanju, predstavlja pa - pri načrtovanju športno-vzgojnega procesa,

tudi razvoj poučevanja in okolja, v katerem poučujemo ali se - pri spremljanju in vrednotenju telesnih značilnosti in

učimo [8]. gibalnih sposobnosti učencev,

Neke enotne definicije, ki bi opredeljevala IKT ni. V literaturi - pri vzdržljivostni vadbi,

je mogoče zaslediti več različnih opredelitev IKT. - pri posredovanju, utrjevanju in ovrednotenju gibalnih

znanj,

2.1 IKT V IZOBRAŽEVANJU - pri posredovanju, utrjevanju in ovrednotenju športnih

Danes si je težko predstavljati izobraževanje brez uporabe teoretičnih znanj,

tehnologije. Informacijsko-komunikacijska tehnologija (IKT) se - pri objavi dosežkov učencev na šolskih športnih

vse bolj uveljavlja v vseh vidikih šolskega delovanja, kar vpliva tekmovanjih, obvestilih učencem in staršem,

ne le na izobraževalni proces, ampak tudi na splošno delovanje - pri posredovanju informacij učencem in staršem [7].

in razvoj šole kot celote.

Lahko ga uporabljamo kot sredstvo za poučevanje v različnih Z izkoriščanjem aktivnih učnih metod in vseh potencialov, ki

predmetih ali pa kot osrednji predmet, ki se ga učijo ločeno. [9]. jih ponuja IKT tehnologija, učitelji spodbujajo učence in dijake

Začetek vključevanja IKT v izobraževanje se je prvič pojavil v k sodelovanju v športnih dejavnostih tako v šoli kot v

prvi polovici sedemdesetih let prejšnjega stoletja, ko so šole po vsakdanjem življenju. Pri tem upoštevajo posameznikove

svetu začele v poučevanje vključevati računalniško tehnologijo. sposobnosti ter razpoložljivost, primernost in dostopnost opreme.

V nekaj letih so bile šole že sposobne povezati računalnike v 2.2.1 ATLETIKA

mreže, nato pa so v devetdesetih letih združile tako informacijsko Atletiko uvrščamo v skupino monostrukturnih športov, za

kot komunikacijsko tehnologijo, ki je tako postala stalni katere je značilna standardna struktura cikličnega ali acikličnega



naredijo bolj Nove tehnologije podpirajo izobraževanje, ga nadgrajujejo in lastnim telesom ali predmetom, ki ga mečemo. Sestavljajo jo fleksibilnega. Omogočajo učinkovitejšo številne atletske discipline, ki jih delimo v tri osnovne skupine: izobraževalni dejavnik v mnogih državah [10]. gibanja, pri čemer je osnovni cilj premagovanje prostora z

komunikacijo učitelja z učenci, ki pred uporabo IKT ni bila teki, skoki, meti [14].

mogoča [11]. Atletika kot ena izmed temeljnih športnih panog, ki vpliva na

S pomočjo IKT lahko učitelji učinkoviteje pripravijo, celovit otrokov razvoj, je sestavni del tako šolske športne vzgoje

načrtujejo in izvajajo pouk, pri čemer enostavno posredujejo tako kot interesnih dejavnosti, ki se odvijajo v prostem času v

praktične kot teoretične vsebine. S tem omogočajo dostop do organiziranih združenjih, kot so športna društva in klubi. ustreznejših informacij tako sebi kot tudi učencem.

Cilji uvajanja sodobne IKT v izobraževanje se lahko razdelijo 2.2.2 UČENJE POSKOKOV ATLETSKE ABECEDE S

v tri sklope: POMOČJO VIDEO POSNETKA

- pridobivanje določenih znanj in spretnosti vezanih na Pri poučevanju poskokov atletske abecede (nizki skiping,

sodobne tehnološke procese pri poznejšem vključevanju v delo; visoki skiping, hopsanje, striženje …) si lahko pomagamo z

- zagotavljanje ključnih informacij o IKT, njenem video kamero. Posnetek je zajemal:

delovanju in posledicah vsem učencem; - posnetek demonstracije pravilne izvedbe/gibanja in

ali ponovili oz. utrdili posamezne gibalne naloge tudi v času, ko

- opis pravilne končne izvedbe posameznega poskoka s niso imeli pouka športne vzgoje (npr. priprava na ocenjevanje,

poudarkom na najpomembnejših delih gibalne naloge (slikovna učenje v primeru različnih izostankov ipd.). To pa je bil tudi

in zvočna ponazoritev). glavni namen priprave teh posnetkov.





3 ZAKLJUČEK

V športni vzgoji lahko integracija IKT prinese številne koristi

tako učiteljem kot dijakom. Kljub temu pa je ključno, da jo

uporabljamo premišljeno, ne prekomerno, ter le takrat, ko je to

smiselno. IKT ne sme nadomestiti praktičnega dela pouka,

temveč naj ga dopolnjuje. Namenjena naj bo predvsem kot

pomoč učiteljem pri pripravah in izvedbi pouka ter kot orodje, ki

izboljšuje učne rezultate, kadar je to primerno.



Slika 1: Postavitev in gibanje rok pri visokem skipingu LITERATURA IN VIRI





[1] [1] Kovač, M., Markun Puhan, N., Lorenci, B., Novak, L., Planinšec, J.,

Hrastar, I., … Muha, V. (2011). Učni načrt , osnovna šola, športna vzgoja.

Ljubljana: Ministrstvo za šolstvo in šport: Zavod RS za šolstvo.

[2] [2] WHO. (2012). Physical activity. Pridobljeno s

http://www.who.int/dietphysicalactivity/pa/en/index.html.

[3]

[4] [3] Body and Mind. Pridobljeno (15. 1. 2024), s

http://publications.becta.org.uk/display.cfm?resID=25919.

[5]

[6] [4] Škof, B. (2010). Spravimo se v gibanje- za zdravje in srečo gre: kako

do boljše telesne zmogljivosti slovenske mladine?. Ljubljana: Fakulteta za

šport, Inštitut za šport.

[7] [5] Bizjak, M. (1999). Šport od mladosti k zrelosti. Ljubljana: Karantanija.

Slika 2: Postavitev zamašne noge in dela stopal pri hopsanju [8] [6] Zuljan, D. (2014). Tehnološka pismenost v obdobju zgodnjega učenja.

Koper: Univerzitetna založba Annales ludus.

[9] [7] Jurak, G., Kovač, M., Starc, G. in Strel, J. (2007). Informacijsko

komunikacijske tehnologije pri pouku športne vzgoje. Ljubljana: Univerza



v Ljubljani, Fakulteta za šport.

[10] [8] Kumar, R. (2007). Convergence of ICT and education. V Proceedings

of world academy of science, engineering and technology. Word Academy

of Science, Enginering & Technology. 30, 556 - 559. Pridobljeno iz:

http://www.waset.org/journals/waset/.

[11] [9] Tišler, T., Černilec, B., Vehovec, M., Korošec, D., Brezovar, D. in

Pungartnik, S. (2006). Vodenje za spodbujanje informacijsko-

komunikacijske tehnologije na šolah. Ljubljana: Evropski socialni sklad.

[12] [10] Brečko, B. N., Vehovar, V. (2008). Informacijsko-

komunikacijska tehnologija pri poučevanju in učenju v slovenskih šolah.

Ljubljana: Pedagoški inštitut.

[13] [11] Dawes, L. (2001). What stops teachers using new

technology? V: Leask, M. (2001). Issues in teaching using ICT. London:

Slika 3: Prikaz dela pete pri »zametavanju« Routledge.

[14] [12] Gerlič, I. (1998). Informacijska družba in sodobna šola,

Nekateri profesorji športne vzgoje so pripravili video vzročno-posledična razmerja. Novo Mesto: Pedagoška obzorja, letnik 13,

št. 3/4.

posnetke gibalnih nalog nekaterih športnih panog (atletika, [15] [13] Markun Puhan, N. (2010). Športna vzgoja in IKT.

gimnastika …) (Slika 1, 2 in 3). Vso gradivo je bilo objavljeno v Pridobljeno s http://skupnost.sio.si/course/view.php?id=5529.

[16] [14] Čoh, M. (2002). Atletika. Ljubljana: Fakulteta za šport,

razrednih skupinah MS Teams, do katerih imajo dostop tudi Inštitut za šport.A

dijaki. Na ta način smo vsem omogočili, da so se lahko naučili



Učenje preko ustvarjanja: uporaba IKT v projektu Knjiga



podajanka



Learning through Creation: Using ICT in the Pass the Story –



the collaborative book Project



Urša Kirn

Osnovna šola Bežigrad

Ljubljana, Slovenija

ursa.kirn@guest.arnes.si



everything created in connection with the book – illustrations,

POVZETEK chapters, opinions, corrections. Working through online

Sodoben pouk zahteva nove pristope in uporabo modernih classrooms makes creation much easier, as all information is

tehnologij. Čeprav se morda področje slovenščine kot maternega collected in one place and accessible to everyone in the group

jezika zdi tradicionalno, je možnosti uporabe IKT veliko in od whenever they need it.

učitelja je odvisno, kaj in kdaj ter kako bo tehnologijo pri pouku By following this method of work in a different, more modern

uporabil. way, we achieve many goals outlined in the curriculum.

Ena od možnosti je tudi ustvarjanje skupne knjige podajanke Additionally, we also achieve other objectives that extend



Dogovorijo se o žanru, ustvarijo književne osebe, določijo such as cooperation, assistance, support, problem-solving, discovery, and the development of diverse talents by integrating dogajalni kraj in čas. Vsak učenec napiše svoje poglavje knjige, – projekt za devetošolce. Cilj je napisati svojo knjigo. beyond Slovene language skills and contribute to life skills –

ga ilustrira in sodeluje pri nastajanju skupnega izdelka. V veliko various knowledge domains.

pomoč nam je sodobna IKT, saj tekom projekta uporabljamo

spletne učilnice, kjer uporabniki oddajajo vse, kar je nastalo v KEYWORDS

povezavi s knjigo – ilustracije, poglavja, mnenja, popravke. Delo Online classrooms, ICT, Slovene, pass the Story – the

preko spletnih učilnic ustvarjanje zelo olajša, saj so vse collaborative book project

informacije zbrane na enem mestu in dostopne vsem v skupini,

kadarkoli jih potrebujejo.

S to metodo dela na drugačen, sodobnejši način sledimo 1 UVOD

mnogim ciljem v učnem načrtu in jih tudi dosežemo. Dosežemo S sodelovanjem v današnjem tehnološko podprtem in

pa tudi druge cilje, ki niso povezani le s slovenščino, ampak informacijsko bogatem okolju državljani sooblikujemo digitalno

življenjem – veščine sodelovanja, pomoči, podpore, znajdenja, okolje, ki je ključno za demokratične procese in prakse ter v

povezanosti, odkrivanja in razvijanja raznovrstnih talentov ter katerem je mogoč medkulturni dialog. Digitalizacija tako skoraj

povezovanja različnih znanj. ni več izbira, temveč je pričakovanje, ki postaja nujnost. [1]

Digitalna kompetenca sodi med ključne kompetence v hitro

KLJUČNE BESEDE spreminjajoči se družbi. [2] Zato je povsem logično, da se je tudi

Spletne učilnice, IKT, slovenščina, knjiga podajanka sistem vzgoje in izobraževanja z razvojem informacijsko-

komunikacijske tehnologije (IKT) začel spreminjati. Ne le sistem,

ABSTRACT tudi učenci zahtevajo nove pristope in sodobne metode ter

aktivno vlogo v procesu učenja.

Modern education requires new approaches and the use of

Pouk slovenščine v duhu sodobnega pouka in priporočil

modern technologies. Although the field of Slovene as a mother

Zavoda RS za šolstvo poteka tudi v spletni učilnici, ki je

tongue may seem traditional, there are many possibilities for

opremljena z e-didaktičnimi gradivi in pripomočki ter orodji, ki

using ICT, and it depends on the teacher what, when, and how

omogočajo iskanje podatkov in informacij, dostop do

technology will be used in teaching.

brezplačnih učnih gradiv, avtorskih programov, spletnih

One option is also creating a collaborative book – Pass the

slovarjev, komuniciranje z uporabo internetnih storitev,

Story, a project for ninth graders. The goal is to write their own

sodelovanje v mrežnih projektih, forumih in različnih spletiščih

book. They agree on the genre, create literary characters,

za učenje slovenskega jezika kot materinščine. [3]

determine the setting and time. Each student writes their own

Učni načrti in katalogi znanj ne zapišejo, kdaj in v kolikšni





chapter of the book, illustrates it, and participates in creating a


meri naj učitelji in učenci pri pouku uporabljajo digitalno

common product. Modern ICT is very helpful to us because

tehnologijo, zato lahko učitelj sam izbira, kako jo bo vključil v

during the project we use online classrooms where they submit



Permission to make digital or hard copies of part or all of this work for personal or

classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s). Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia U. Kirn



pouk glede na interese učencev, znanje, ki ga ima o uporabi ima morda z velikim veseljem spoznava računalniška omrežja in

tehnologije in glede na tehnologijo, ki jo ima na voljo on in programe ter se izkaže kot oblikovalec končnega izdelka.

njegovi učenci. Z uvajanjem novih oblik in načinov dela z Nadarjeni na literarnem področju pa so glavni motor projekta, ki

uporabo sodobne informacijsko-komunikacijske tehnologije ženejo nalogo proti cilju, spodbujajo in pomagajo drugim.

učitelji krepijo digitalno pismenost učencev, a mora raba Cilj je tudi medpredmetno povezovanje, predvsem z likovno

tehnologije potekati skladno s cilji pouka in ne sme stopiti v umetnostjo in računalniškimi predmeti ter programom za

ospredje na način, da bi bila pomembnejša od zmožnosti in ciljev, nadarjene učence.

zapisanih v UN in KZ za slovenščino. [4]

V prispevku je predstavljen vsakoletni razredni projekt 2.2 Prvi koraki v projekt



devetošolcev pri pouku slovenščine, ki smo ga poimenovali Prvi korak je izbira žanra knjige. Najpriljubljenejši izbiri sta Knjiga podajanka. To je sodobna didaktična metoda, s pomočjo mladinski ali kriminalni roman. Izbor je logičen, saj je katere dosegamo več ciljev iz učnega načrta. Del učnega načrta mladostnikom prvi žanr najbližji, drugi pa je med najbolj pa je tudi (s)poznavanje IKT vsebin, kritična raba računalnika za izposojanim gradivom v splošnih knjižnicah na oddelkih za delo z informacijami, komunikacijo in sodelovanje v medmrežju. odrasle. [6] Če izbira ni soglasna, učenci glasujejo o predlogih in Natančneje, pri realizaciji projekta uporabljamo spletne učilnice, tako demokratično zmaga predlog žanra z največ glasovi. kjer nastaja vse, kar je povezano s knjigo. Ko je žanr dogovorjen, se lotimo dogajalnega kraja in

dogajalnega časa. Tudi kraj in čas izberemo skozi debato – kaj je

primerno žanru, kaj je lažje vplesti v zgodbo, kaj zgodbo podpre.

2 KNJIGA PODAJANKA

Po tem so na vrsti književne osebe. Skozi debato ugotovimo,

Vsako leto je devetošolcem predstavljena možnost kakšne književne osebe so primerne za izbrani žanr njihove

sodelovanja v večmesečnem oddelčnem oz. skupinskem (če na knjige. Vsak učenec nato doma ustvari eno književno osebo, jo

šoli poteka pouk v manjših učnih skupinah) projektu, ilustrira, poimenuje, opiše in označi. Na dogovorjeni dan v šolo

poimenovanem Knjiga podajanka. Projekt poteka vsako leto in prinesejo svoje izdelke (Slika 1).

ni se še zgodilo, da se učenci projekta ne bi lotili ter ga tudi

dokončali.



Projekt je skupinski, sodelujejo vsi v oddelku oz. skupini,

vsak po svojih zmožnostih, željah, sposobnostih, talentih. Cilj je

skupen izdelek – napisati čisto svojo knjigo in jo izdati tudi v

fizični obliki. Vsak učenec napiše svoje poglavje, ga (lahko)

ilustrira, podnaslovi. Za dosego cilja pa morajo poleg želje po

sodelovanju, raziskovanju in širjenju obzorij imeti aktiven še

AAI-račun. Učenci račune večinoma že imajo, saj spletno

učilnico pri pouku uporabljamo za različne dejavnosti.



2.1 Cilji projekta



Ciljev je več. V prvi vrsti sledimo ciljem iz učnega načrta –

splošnim, to je razmišljujoče in kritično sprejemanje ter

vrednotenje raznovrstnih umetnostnih besedil, razvijanje

pripravljenosti za pisanje, ubesedovanje svojega znanja, misli.

Učenci so sposobni tvoriti učinkovita slovnično-pravopisno

pravilna besedila. Skozi ta projekt pa dosegajo tudi specifične Slika 1: Ilustracija, oznaka in opis književne osebe

cilje, kot so določanje glavne in stranske književne osebe,

dogajalnega prostora in časa, teme besedila, pisanje zgodbe z Skupaj jih pregledamo, pokomentiramo, katera bi bila

različnih perspektiv, razvijajo nov domišljijski svet v različnih primerna za glavno književno osebo, katere osebe bodo stranske.

žanrih. V učnem načrtu za slovenščino je navedena tudi uporaba Učenci podajo povratno informacijo, če je pri kateri od oznak oz.

IKT in vključevanje kompetenc, povezanih s tem. Urijo se v opisov potreben popravek in kakšen (starost, poklic, značajska

tvorjenju besedil v računalniških programih, uporabi okolja značilnost).

spletnih učilnic ter elektronskih sporočil in uporabi elektronske

pošte. Učenci razvijajo kritično rabo računalnika za delo z 2.3 Spletna učilnica kot uporaben pripomoček

informacijami, komunikacijo in sodelovanje v medmrežju. [5] Isto uro pregledamo skupno spletno učilnico in razdelek, kjer

Morda pa so še pomembnejši cilji, ki jih dosegamo že med bo knjiga podajanka nastajala. Predstavimo delo v spletni

ustvarjanjem: krepitev medvrstniškega sodelovanja, učilnici – mapo, kjer bodo objavljene ilustracije z oznako in

povezovanje oddelčne skupnosti, pomoč med vrstniki, učenci opisom književnih oseb, mapo že napisanih poglavij, mapo z

kritično razmišljajo, ves čas krožijo povratne informacije, v ilustracijami poglavij, dokument, v katerem je zapisan izbrani

ospredju je ustvarjanje in kreativno pisanje. Nenazadnje poteka žanr, kraj in čas dogajanja knjige ipd. Nato mentorica vse izdelke

skozi proces tudi odkrivanje novih zanimanj in talentov, z optičnim bralnikom preslikam v elektronsko obliko in jih

razvijanje različnih področij nadarjenosti. Vsakič se v skupini objavim v spletni učilnici v razdelku Knjiga podajanka v mapi

izpostavi kakšen posameznik, ki si prej ni upal/zmogel pokazati, Književne osebe. (Slika 2)

da ga likovno področje zelo zanima in je pri tem nadarjen. Ali pa

Učenje preko ustvarjanja: uporaba IKT v projektu Knjiga podajanka Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



Slika 2: Znajdenje v spletni učilnici v razdelku Knjiga Slika 3: Ilustracija poglavja s pomočjo programa Slikar

podajanka

Nekatere generacije se odločijo, da bodo svoja poglavja tudi



Dogovorimo se, da bo imel vsak avtor časa za pisanje podnaslovile. Z lansko generacijo učencev smo preizkusili tudi

poglavja sedem dni. Zadnji rok za oddajo svojega poglavja v novejše orodje v spletnih učilnicah, in sicer možnost pripisovanja

spletno učilnico je večer pred uro slovenščine, med katero bomo v PDF-dokument. (Slika 4) Nekateri učenci so jo z veseljem

poglavje vsak teden prebrali in ga pokomentirali. Največkrat se uporabili, nekateri pa so svoje predloge raje posredovali v živo v

učenci že kar sami javijo in določijo, kdaj bi pisali svoje poglavje učilnici.

– po navadi se odločajo glede svojih drugih obveznosti.

Mentorica nato v spletno učilnico vnesem poimensko in

terminsko označene oddaje nalog, kamor lahko avtorji v

dogovorjenem času oddajo svoje poglavje.

Če se pri oddaji poglavja v spletni učilnici pojavijo težave, se

lahko učenci obrnejo na učiteljico ali si med seboj pomagajo z

nasveti, lahko tudi oddajo v imenu drugega učenca. Lahko pa

svoje poglavje pošljejo tudi preko elektronske pošte svoji

učiteljici.



2.4 Globoko v procesu fizičnega in virtualnega

ustvarjanja



Ko posamezni učenec napiše svoje poglavje, ga prikažemo Slika 4: Prikaz uporabe orodja pripisovanja v PDF-

preko projektorja na tablo in ga glasno preberemo. Zatem učenci dokument

z mentoričino pomočjo debatirajo o zapisanem: kaj je bilo dobro,

katere spremembe in kakšen nadaljnji razvoj zgodbe predlagajo.

Avtor poglavja spremlja debate in vnese morebitne popravke.

Pisec naslednjega poglavja pozorno posluša, aktivno sodeluje v Učenci prav tako sami oblikujejo naslovnico. Tudi to nalogo

pogovoru, saj so tako njegovi naslednji koraki veliko lažji. S prevzamejo tisti, ki jim je likovno ustvarjanje v užitek in tako

skupnimi idejami oblikujejo smer, v katero se bo njihova zgodba pripomorejo k najboljši verziji knjige podajanke. (Slika 5)

odvila. S povratnimi medvrstniškimi in mentoričinimi

informacijami se učenci učijo, urijo v jezikovnih spretnostih,



kreativnem pisanju, znajdenju v spletnem okolju in delom v



programih za urejanje besedil. Na tej točki tudi kritično presojajo,



utemeljujejo svoje ideje, kreativno razmišljajo. Če zgodba zaide



na stransko pot, jo morajo s skupnimi močmi pripeljati nazaj do



glavne teme. Paziti morajo tudi, da smiselno vključujejo

književne osebe, da dogajanja vestno zaključujejo ter natančno

sledijo zgodbi.

Poglavje lahko napišejo tudi v paru, če morda čutijo, da bi

potrebovali pomoč vrstnika.

Knjigo oz. poglavja tudi ilustrirajo. To nalogo po navadi

prevzamejo tisti učenci in učenke, ki radi likovno ustvarjajo.

Nekateri tudi s pomočjo risalnih tablic in računalniških

programov. (Slika 3).



Slika 5: Ilustracija naslovnice knjige podajanke



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia U. Kirn



2.5 Večmesečni projekt se bliža koncu Njihova lastna knjiga na koncu vedno požanje veliko



zaključi, je na vrsti izbor naslova. Tudi naslov se izbira s predlogi zdelo nedosegljivo. avtorjev, nato pa se glasuje za najljubšega. Izbranega se doriše Na praktičen način lahko dokažejo veliko usvojenega znanja oz. vstavi v naslovnico. ob koncu osnovne šole – tako s področja rabe jezika kot Ko se pisateljski in ilustratorski del ustvarjanja skupne knjige dokaz, da zmorejo nekaj, kar se jim je morda pred nekaj meseci zadovoljstva, ponosa in tudi nekaj solz sreče, saj imajo v rokah



zadnjih korakov zapišejo svoje misli o prehojeni poti. Svoje misli Učenci so ob zaključevanju projekta naprošeni, da kot enega književnosti. Učenci pa poleg tega dobijo tudi lep spomin na zaključek osnovne šole ter na skupni rezultat sodelovanja in so lahko tako kot svoja poglavja oddajali v spletni učilnici, lahko podpore. Ponosni so na ‘izid’svoje knjige, ki so jo snovali in pa so mi jih poslali po elektronski pošti. ustvarjali skupaj, od ideje do končnega izdelka. Pa sploh ni bilo Sledile so zadnje naloge – ponovno branje, popravki in “za oceno”! dopolnitve, lektoriranje, vstavljanje ilustracij, računalniško

oblikovanje. Teh nalog se loteva vsaka generacija drugače –



nekatere večino uspejo narediti same, druge potrebujejo več 4 ZAKLJUČEK mentoričine pomoči in usmerjanja. Predvsem jezikovni pregled,

včasih tudi računalniško oblikovanje in priprava na tisk, je Vsako leto s povratnimi informacijami učencev ustvarjalcev

največkrat naloga mentorice in manjše skupine učencev, ki knjige projekt še izboljšujem in nadgrajujem. Začeli smo ga

izkažejo zanimanja za določena področja. namreč fizično, s poglavji, natisnjenemi na listih, spravljenimi v

Ko je jezikovni pregled opravljen, računalniško oblikovanje kartonski mapi, ki je kot kuričkova torbica potovala od učenca

pri koncu, sledi še zadnji splošni pregled, ali je vse narejeno, kot do učenca. IKT je ta del posodobil in olajšal. Nova tehnološka

so bile želje in vizije. Če je vse potrjeno, je na vrsti tiskanje. Po orodja ne le prinašajo novosti v šolski prostor, temveč tudi

zaključku tiskanja pa se vsak izvod še ročno spne, da nastane pospešujejo prenos informacij, povečujejo zanimanje učencev za

prava knjiga v formatu A5. Vsak učenec namreč zadnji šolski snov oz. šolski projekt. Kot velika prednost se je pokazal tudi

dan ob zaključnem spričevalu prejme tudi svoj izvod knjige hiter in hkraten dostop do virov informacij ter pregled nad

podajanke. potekom projekta za celotno skupino ustvarjalcev.

V projekt se mi ne le zdi smiselno, ampak dandanes že celo

nujno vključevati IKT, četudi je področje slovenščine morda

malo netipično, a se z idejami, sodelovanjem, razvijanjem in

3 POVRATNE INFORMACIJE odprtostjo za spremembe da marsikaj izvesti. Ključni



pomena – tako o vsebini poglavij kot tudi znajdenja v spletnih dostopnost znanja in spretnosti. učinicah ter oblikovanja dokumentov z urejevalniki besedil ter Povratne informacije učencev, kolegov, staršev kažejo, da je uporabe digitalnih pripomočkov za ilustracije. Informacije so ta metoda zanimiva, sodobna in aktivna. Doseženi so cilji, kot so Povratne informacije skozi ves projekt so bile ključnega omogočimo inovativno in spodbudno učno okolje ter večjo preoblikovalci pedagoške prakse smo učitelji, ki učencem lahko



krožile tako med učenci, tudi v popoldanskem času med razvijanje in ohranjanje radovednosti, zanimanja za književnost, neformalnimi srečanji, in v šolskem okolju, kot tudi med učenci branje in literarno ustvarjanje, urjenje veščin uporabe in mentorico ter obratno. Vsaka generacija dobi kakšno novo računalnika in z njim povezanih orodij, znajdenje v okolju zamisel, poda drugačen predlog (npr. dva različna konca zgodbe, spletnih učilnic. Učenci so pri pouku v aktivni in ustvarjalni vzporedna zgodba ene od književnih oseb …), tako se tudi sama vlogi, kar jim sodoben pouk z drugačnimi metodami, a enakimi marsičesa vsako leto naučim. cilji, mora nuditi. To pa je nekaj misli učencev – avtorjev različnih generacij o Spregledati pa se ne sme dejstva, kako so učenci ob uspelem procesu pisanja knjige podajanke: skupinskem projektu in doseženem, včasih celo preseženem cilju



predebatiramo, kako bi zgodbo peljali naprej in da v tem projektu medsebojne vezi in na novo odkrili svoje talente in zanimanja. res sodelujemo vsi. Zdi se mi, da do sedaj vse napreduje odlično Motivirala ji ni ocena, ki za ta projekt sploh ni predvidena, ampak in verjamem, da bo tako še naprej.” sodelovanje in želja prispevati svoj košček v mozaiku “Knjiga podajanka mi je zelo všeč. Super mi je, da začetku zdel nedosegljiv, v procesu nastajanja knjige so poglobili zadovoljni in srečni. Ne le da so dosegli cilj, ki se jim je na

“Projekt mi je bil zelo všeč in mislim, da nam gre zelo dobro.



Všeč mi je, da imamo vsi možnost sodelovati in skupaj debatirati povratno infomacijo o svojem delu in se dokazati samemu sebi, oddelčnega oz. skupinskega projekta, prejeti (pozitivno)

o napisanem. Všeč mi je tudi to, da se učimo tudi uporabe

vrstnikom, staršem in mentorici.

spletnih učilnic in drugih uporabnih stvari za življenje. Ničesar

ne bi spremenila.”



“Ko sem začel pisati svoje poglavje, sem imel mešane LITERATURA IN VIRI

občutke. Sploh nisem znal začeti. Razmišljal sem o vseh možnih

kombinacijah, a začeti je bilo res težko. S pomočjo učiteljice in [1] Digitalna Slovenija 2030 – Krovna strategija digitalne preobrazbe

sošolcev mi je steklo in sem med pisanjem nepričakovano celo Slovenije do leta 2030. Ljubljana, marec 2023.

https://www.gov.si/assets/ministrstva/MDP/Dokumenti/DSI2030-potrjena-na-

užival. Vesel sem, da nas je učiteljica popeljala skozi ta izziv, saj Vladi-RS_marec-2023.pdf (18. 8. 2024)

smo se veliko naučili in zdaj imamo svojo knjigo! Koliko [2] https://skupnost.sio.si/pluginfile.php/432495/mod_resource/content/0

učencev se lahko pohvali s tem?” /Gradivo/Osem_kljucnih_kompetenc.pdf (18. 8. 2024)

[3] https://www.zrss.si/pdf/smernice-ikt-slo.pdf (18. 8. 2024)

[4] https://www.zrss.si/pdf/Digitalna_tehnologija_pouk_razvija.pdf (18.

8. 2024)



Učenje preko ustvarjanja: uporaba IKT v projektu Knjiga podajanka Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



[5] Učni načrt (posodobljena izdaja). 2018. Program osnovna šola,

Slovenščina. Ljubljana, Ministrstvo za šolstvo in šport, Zavod RS za šolstvo.

[6] https://plus.cobiss.net/most-read-

web/si/sl#contentTypes=52,53,54,55,56,57,58&libAcronym&libType&materialT

ypes=7&periodFrom=202407&periodTo=202407&pubType=1&publishYear&tar

getGroups=5 (18. 8. 2024)



Spopadanje z izzivi pri razvijanju digitalnih veščin pri



učencih z učnimi težavami v redni osnovni šoli



Addressing the Challenges of Developing Digital Skills among



Students with Learning Difficulties in Regular Primary School



Barbara Kokec

OŠ Prežihovega Voranca Maribor

Maribor, Slovenija

barbara.kokec @ospvmb.si



POVZETEK KEYWORDS

V prispevku so predstavljeni nekateri izzivi, s katerimi se Digital skills, help, students with learning difficulties, team

soočamo v redni osnovi šoli ob hitrem digitalnem razvoju. work

Sledenje trendom razvoja digitalnih veščin je pri poučevanju

učencev v šoli izrednega pomena, saj so oni tisti, ki bodo v

prihodnosti te veščine potrebovali za kvalitetno delo in življenje. 1 UVOD

V prispevku je poudarek na premagovanju ovir pri učencih z Živimo v času hitrih sprememb, ki jih prinaša pospešena

učnimi težavami, ki imajo posledično pogosto težave tudi pri informatizacija in digitalizacija. Pandemija koronavirusa je

vedno hitrejšem načrtnem razvijanju digitalnih veščin pri pouku. digitalni prehod in preobrazbo samo še močno pospešila. Delo in

Ti učenci potrebujejo posebno podporo ter drugačen način dela, učenje na daljavo sta postala naša stalnica. [1]

ki zahteva združevanje različnih profilov učiteljev, ki poučujejo Vpliv razvoja digitalne tehnologije seže na vsa področja

na šoli. Tako smo se združili učitelji predmetnega pouka z učitelji življenja posameznika in družbe, zato kot odgovor na

dodatne strokovne pomoči, to so pedagog, inkluzivni pedagog, spremenjen način življenja družbe sega vpliv digitalne

specialni in rehabilitacijski pedagog ter socialni pedagog, ki s preobrazbe tudi na področje izobraževanja. [2]

svojimi specifičnimi znanji pristopimo k tem učencem s svojimi Tudi na naši šoli se v zadnjem času zelo intenzivno

cilji in načinom dela in tako pripomoremo k najbolj optimalnemu ukvarjamo z razvijanjem digitalnih veščin pri učencih in učiteljih.

razvoju, tudi razvoju digitalnih veščin, vseh učencev. Posebno pozornost smo zaradi nuje temu namenili v času šolanja

na daljavo, kasneje pa smo svoja znanja, spretnosti in veščine

KLJUČNE BESEDE samo nadgrajevali in si s pridobljenim znanjem na veliko

Digitalne veščine, pomoč, učenci z učnimi težavami, timsko področjih močno optimalizirali učno delo in proces. Učitelji so

delo digitalne vsebine vnašali v vsakodnevno delo, pouk in učenje. Na

šoli smo posodobili računalniško opremo, dobili smo nove

ABSTRACT računalnike, tablice, inovativno učilnico, pametne table,... in tako

The paper presents some of the challenges we face in a regular učiteljem omogočili poučevanje s pomočjo digitalne tehnologije,

elementary school in the face of rapid digital development. če so le želeli. Sčasoma so se vsi učitelji začeli posluževati teh

Following the trends in the development of digital skills is very orodij, vsak na svojem področju, glede na svoja zanimanja,

important to teach students at school, as they are the ones who interes in po svojih zmožnostih.

will need these skills for quality work and life in the future. The Kot specialna pedagoginja sem budno spremljala ves ta

paper emphasizes overcoming obstacles for students with napredek in se seveda tudi sama izpopolnjevala na tem področju.



teaching, which requires bringing together the different profiles posebnimi potrebami in učenci z učnimi težavami posebni na vseh ostalih šolskih področjih, tako je potrebno upoštevati of teachers who teach at the school. In this way, we have teamed njihove učne posebnosti tudi na področju razvijanja in uporabe up with teachers of additional professional assistance, i.e. digitalnih veščin. Pogosto je delo v razredu kljub odlični pedagogues, inclusive pedagogues, special and rehabilitation učiteljevi pripravi potekalo prehitro, večino dela niso mogli pedagogues and social pedagogues, who approach these students opraviti samostojno in so imeli težave že z osnovnimi veščinami with their specific skills with their goals and different way of developing digital skills in the classroom at an increasingly rapid digitalnih veščin pri učiteljih in preizkušala pridobljena znanja tudi pri učencih. Ugotovila sem, da tako kot so učenci s pace. These pupils need special support and a different way of learning difficulties, who consequently often have difficulties in Sodelovala sem v vseh šolskih projektih, ki so spodbujala razvoj

teaching, thus contributing to the most optimal development, kot so odpiranje, shranjevanje dokumenta, urejanje besedila in

including the development of digital skills, of all students. podobno. Podobno so seveda opazili tudi njihovi učitelji in zato

smo se odločili za skupno strategijo dela, ki bo tudi učencem z

učnimi težavami omogočila razvoj digitalnih veščin optimalno v meji, ki jo je zmogel učitelj, ki jih je poučeval. Uporabljene

glede na njihove sposobnosti. aplikacije so bile večinoma namenjene utrjevanju znanja in od

učencev niso zahtevale zahtevnejših digitalnih spretnosti. Zelo

dobro so se zavedali nevarnosti, ki jim pretijo na spletu,

2 POMOČ UČENCEM Z UČNIMI teoretično so bili tudi zelo dobro poučeni o primernem obnašanju

TEŽAVAMI PRI RAZVIJANJU na spletu.

DIGITALNIH VEŠČIN V ČASU POUKA

Digitalne kompetence spadajo med ključne kompetence, ki so 2.1 Dejavnosti v okvira projekta Inovativna

potrebne za vseživljenjsko učenje. “Digitalna kompetenca pedagogika 5.0

vključuje samozavestno, kritično in odgovorno uporabo Zapletlo se je, ko smo se vključili v projekt Inovativna

digitalnih tehnologij in njihovo vključevanje pri učenju, delu in pedagogika 5.0 in so učitelji začeli sistematično razvijati

družbenem udejstvovanju: Vključuje informacijsko in digitalne kompete istočasno s poučevanjem svojih učnih vsebin.

podatkovno pismenost, ustvarjanje digitalnih vsebin (tudi Leta 2013 je bil izdan prvi trefernčni okvir DigComp, v

programiranje), varnost (tudi digitalno blagostanje in katerem so bile digitalne kompetence opredeljene kot

kompetence v zvezi s kibernetsko varnostjo), vprašanja kombinacija 21 kompetenc, razvrščenih v pet glavnih področij

intelektualne lastnine, reševanje problemov in kritično (slika 1). [3]



mišljenje.“ (Priporočilo Sveta o ključnih kompetencah za

vseživljenjsko učenje, 22. maj 2018, ST 9009 2018 INIT) [3]

To potrjuje tudi definicija digitalne pismenosti Organizacije

Združenih narodov za izobraževanje, znanost in kulturo –

UNESCO: »Digitalna pismenost je sposobnost varnega in

ustreznega dostopa, upravljanja, razumevanja, vključevanja,

komuniciranja, vrednotenja in ustvarjanja informacij z

digitalnimi tehnologijami za potrebe zaposlovanja, dostojnega

delovnega mesta in podjetništva. Vključuje kompetence, ki jih na

različne načine imenujemo računalniška pismenost, IKT

pismenost, informacijska pismenost in medijska pismenost.« [1]

Na področju življenja posameznika se poleg pasivne ali

aktivne splošne uporabe spleta opravlja tudi vse več storitev s

pomočjo digitalne tehnologije, od naročanja živil do oddajanja

različnih vlog v digitalnih storitvah države. Vsi ti vplivi digitalne

preobrazbe narekujejo vodstvenim ali strokovnim delavcem, da

v vzgojno izobraževalnih zavodih razvijajo znanja, spretnosti in

stališča, vezana na uporabo digitalne tehnologije, najprej pri sebi, Slika 1: Konceptualni referenčni model okvira DigComp

da bodo opolnomočeni lahko vnašali potrebne prvine v svoj [3]

vzgojno-izobraževalni proces. [2]

Na naši šoli smo v preteklih letih sodelovali v različnih Učitelji so si iz nabora izbrali kompetenco, ki so jo želeli

projektih, kjer smo pridobili znanja in spretnosti, ki smo jih razvijati, pripravili dejavnosti in jih z učenci tudi udejanjili.

potem prenašali tudi na učence. Najprej smo bili izbrani v Učenci so morali preko pridobivanja učne snovi načrtno

dvoletni projekt DDK (dvig digitalne kompetentnosti). Namen preko vnaprej načrtovanih aktivnosti razvijati tudi digitalno

tega projekta je bil izobraževati učitelje, da bodo poučevali tako, kompetenco, kar pa je bilo bistveno težje in je sodoben koncept

da bodo pri svojem pouku spodbujali razvoj digitalnih dela, ki vključuje razvoj in uporabo digitalnih kompetenc,

kompetenc. predstavljal veliko težavo pri pridobivanju znanja pri

V tem projektu se je nekaj naših učiteljev izobraževalo ter posameznem predmetu. Učitelji so ugotavljali, da sami v

svoje znanje nato preneslo na ostali kolektiv. V teh prepolnih razredih ne zmorejo kvalitetno izvajati pouka, kljub

izobraževanjih smo se naučili uporabljati nova uporabna spletna odlični pripravi in uporabi napredne tehnologije, ki bi jim naj pri

orodja, delati z orodji v Microsoft Office 365, izdelali smo svoj tem pomagala.

izobraževalni video, svoj izdelek s pomočjo gradnika H5P ter V prvem letu je bil v projekt vključen en osmi razred. V tem

poudarili pomen varnega shranjevanja v oblak, ki ga imamo na razredu je 7 učencev s posebnimi vzgojno-izobraževalnimi

voljo kot učitelji. Predstavili so nam nevarnosti uporabe spleta potrebami, poleg tega so še 4 učenci s splošnimi učnimi težavami.

ter kako se jim izogniti. Ob različnih priložnostih smo te veščine Že v preteklosti smo občasno prakticirali, da smo se učitelji

pokazali in predajali učencem, ki so s svojimi napravami imeli dodatne strokovne pomoči kot dodaten učitelj vključevali v pouk

možnost preizkusiti predstavljena orodja. V tej fazi so se učenci pri posameznih predmetih, največkrat pri matematiki, zgodovini

z učnimi težavami pravzaprav dobro znašli. Učitelji so po in slovenščini. Tokrat smo se za to odločili sistematično ravno

korakih razlagali potek dela, kar je za njih najboljši način učenja. zaradi uvajanja digitalnih vsebin v pouk.

Bili so motivirani, ker so lahko uporabljali tehnološko napravo, Prvo tako sodelovanje je potekalo z učiteljico zgodovine, ki

ki jim je znana in radi delajo z njo. Povečini niso imeli težav s je tudi vodja tega projekta in je takšen način dela uvajala že pred

tem, da so si aplikacijo naložili in jo potem po navodilih tudi sodelovanjem v tem projektu. Z učiteljico sva skrbno načrtovali

uporabljali. Pri tem so nekateri potrebovali nekaj pomoči, vendar ure, pri katerih so se poleg učnih vsebin načrtno razvijale tudi

digitalne kompetence. Težava, ki sva jo najprej zaznali in se ji



posvetili, je bila izredno slaba bralna pimenost učencev z učnimi programom, ponovno smo se morali naučiti tudi shraniti svoje

težavami. Tudi digitalne vsebine, ki jih je pripravljala učiteljica delo, da lahko učenci naslednjič nadaljujejo s svojim delom in si

zgodovine, namreč zahtevajo sposobnost branja, razumevanja z njim pomagajo pri kasnejšem učenju. Zaradi preobsežnosti

prebranega in tudi bralno učne strategije, ki so temelj snovi sva z učiteljem omejila naloge na tiste, ki so ponujale

kateregakoli učenja. Učenci so bili pri delu zelo počasni, površni, minimalni in nekatere temeljni standard znanja. Ob tem so

niso znali poiskati informacij, brez pomoči učitelja niso razumeli potrebovali tudi dodatno razlago in pomoč pri ponazoritvah s

digitalno predstavljenih vsebin, niso zmogli dati povratne programom. Takšen način dela je učencem omogočil sledenje

informacije o pridobljenem znanju ali podati povratno učni snovi. Sicer niso bili tako samostojni kot ostali, vendar je

informacijo drugim učencem o kvaliteti opravljene naloge. Poleg program kljub temu dosegel svoj namen, saj so si učenci lažje

tega so imeli težave z iskanjem dokumentov v spletni učilnici, predstavljali učno snov, uporabna vrednost Pitagorovega izreka

odpiranjem dokumentov, varnim shranjevanjem in še kaj. Na je namreč zanje precej abstraktna, program pa jim je omogočil

primer, v enem sklopu učnih ur so se učenci preko jasnejšo sliko, lahko so si prenesli izsek, ki je prikazoval uporabo

raziskovalnega dela (naslov so si glede na interese izbrali sami) Pitagorovega izreka in tako bolje razumeli učno snov in

učili o avtorskih pravicah. Šlo je tudi za medpredmetno posledično lažje naredili zahtevane naloge s tega področja.

povezovanje, zato so na isto temo izdelali še izdelek pri tehniki.

Že samo raziskovalno delo je za te učence velik zalogaj, ko pa so

morali paziti še na avtorske pravice, jih ustrezno označiti ter 3 ZAKLJUČEK

povezati svoje znanje z drugim predmetom, se je izkazalo, da bo Zaradi digitalne revolucije in ne dovolj hitrega prilagajanja

potreben drugačen pristop. izobraževanja in usposabljanja novim razmeram, smo ponovno v

Ugotovili sva, da je za njihovo učenje najbolje, da delajo situaciji, ko tehnologija prehiteva izobraževanje in tako povzroča,

ločeno od ostalih. Tako sva ustvarili manjše skupine, ki so da ljudje nismo ustrezno pripravljeni na življenje in delo z novo

večinoma delale samostojno in skupino učno manj uspešnih in tehnologijo ter težko sledimo takemu hitremu tehnološkemu

motiviranih učencev, ki so delali skupaj z mano, da se je razvoju.[1]

učiteljica lahko posvetila ostalim skupinam. Pri tem sva vsebine V šoli se je to izkazalo na več nivojih. Najprej so novi dosežki

prilagodili njihovim sposobnostim, dostopali so do enostavnejših predstavljali velik izziv nekaterim, predvsem starejšim učiteljem,

besedil, naloge so bile prilagojene, da so lahko samostojno prišli ki jim je nova tehnologija na začetku znatno upočasnila njihovo

do rešitev, pri tem pa so imeli ves čas na voljo učitelja, če so delo. Imeli smo kar nekaj nezaupljivih staršev, ki jih je digitalno

naleteli na težavo. Delali so v skupini, tako so si lahko tudi znanje njihovih otrok preraslo in jih več niso zmogli nadzirati ali

pomagali, en drugemu posredovali svoje razumevanje navodil ali jim pomagati. Nenazadnje so se s težavami na tem področju

učne snovi in tako urili tudi prečne veščine. Osredotočili smo se soočali tudi učenci z učnimi težavami, za katere je pouk

le na najpogostejše oznake za avtorske pravice, ki sem jih naennkrat potekal preveč kompleksno, prehitro in niso zmogli

poenstavljeno na konkretnih primerih razložila in tako so lahko slediti učnim zahtevam, kar je v njih povzročilo dodaten stres, ki

samostojno delali in naredili svoj izdelek, ki je bil res samo jim je onemogočal doseganje rezultatov v skladu z njihovimi

njihov. sposobnostmi. Tako kot vsakega izziva smo se tudi tega na naši

Naslednji primer takšnega sodelovanja je bil pri matematiki. šoli lotili sistematično, z veliko podpore s strani vodstva in bolje

Pri tem predmetu sta se učitelja odločila za pridobivanje nove usposobljenih učiteljev ter z željo slediti trendom in naše učence

učne snovi Pitagorov izrek preko spletnega orodja GEOGEBRA. kvalitetno pospremiti v novi svet digitalne tehnologije. Z

GeoGebra Classic je programska oprema za dinamično različnimi seminarji smo opolnomočili tudi starše, ki jih pri nas

matematiko za vse stopnje izobraževanja, ki združuje geometrijo, jemljemo kot pomembne partnerje v učnem procesu. Menim, da

algebro, funkcije, preglednice, grafe, statistiko in račun v enem smo na dobri poti uspeha tudi pri razvijanju digitalnih veščin pri

paketu, ki je enostaven za uporabo.[4] učencih z zmanjšanimi sposobnostmi slediti tako hitremu

Učenci so delali v spletni različici programa. Učitelj jim je razvoju. S tem, ko stopimo skupaj, združimo znanja iz različnih

najprej program predstavil, pokazal nekaj možnosti in osnovnih področij, opazujemo učence in njihovo delo, sproti prilagajamo

ukazov programa ter jih napotil na navodila za delo v spletni cilje in naloge, v naših učencih ustvarimo zaupanje, motivacijo

učilnici, kjer so sistematično raziskovali Pitagorov izrek, in željo po novih znanjih in veščinah.

njegovo uporabnost v različnih likih in telesih ter si s pomočjo

programa vse prikazali in naredili zapiske v digitalni obliki, ki so VIRI IN LITERATURA

jih nato shranili v oblak. Podobno kot pri prejšnjem primeru, so [1] [1] Lešnik Štefotič, V. (2021). Digitalne kompetence stanje in prihodnost. se učenci z učnimi težavami ponovno izgubili pri samsotojnem V Mednarodno inovativno poslovanje, letnik 13, številka 1. (str. 79-87).

delu. Pri samostojni uporabi programa so že pozabili osnove, ki http://www.dlib.si

[2] Dolinar, M., Poberžnik, A., Jerše, L. (2023). Vodenje in podpora učencem

jim jih je na začetku pokazala učitelj, tudi si niso znali konkretno pri pridobivanju digitalnih kompetenc, 1. del: Strokovna izhodišča in

prikazati učne snovi, kar jim je omogočal program in čemur je priporočila. (str. 8). Zavod RS za šolstvo.

bil seveda tudi namenjen. Učenci so potrebovali stalno asistenco Vodenje-in-podpora-ucencem-pri-pridobivanju-digitalnih-kompetenc.pdf

(zrss.si)

učitelja in tudi dodatno razlago njim preveč abstraktne učne [3] Vuorikari, R., Kluzer, S., Punie Y. (2023). DigComp 2.2: Okvir digitalnih

snovi. Kljub delu v manjših učnih skupinah, je bilo za enega kompetenc za državljane. Z novimi primeri rabe znanja, spretnosti in

stališč. (str. 3, 4). Zavod RS za šolstvo.

učitelja preveč dela v razredu, da bi lahko kvalitetno izpeljal digcomp_2_2_okvir_digitalnih_kompetenc.pdf (zrss.si)

pouk. Tako smo se ponovno združili, oblikovali manjšo skupino, [4] Geogebra: Introduction – GeoGebra

ki je potrebovala pomoč in za katere sem bila odgovorna jaz kot

specialna pedagoginja. Delo sem v sodelovanju z učiteljem

znatno upočasnila, večkrat ponovila osnovno ravnanje s



Knjižnično informacijska znanja pri pouku Slovenščine



v 7. razredu



Library Information Knowledge in Slovenian Lessons



in the 7th Grade



Martina Kokelj

OŠ Antona Martina Slomška Vrhnika

Vrhnika

martina.kokelj@osams.si



Poleg tega šolska knjižnica spodbuja sodelovanje med učenci,

POVZETEK učitelji in knjižničarjem, kar krepi skupno učenje. Knjižničarji

Na Osnovni šoli Antona Martina Slomška Vrhnika si igrajo pomembno vlogo pri usmerjanju učencev k ustreznemu

prizadevamo učencem približati učno snov skozi samostojno gradivu, organiziranju kulturnih dogodkov, kot so bralne značke,

raziskovanje in medpredmetno povezovanje. V prispevku je literarni večeri in razstave, ter pri pomoči pri raziskovalnih

predstavljeno sodelovanje med šolsko knjižnico in predmetom nalogah. Šolska knjižnica je zato neprecenljiv vir, ki bogati

slovenščina, kjer učenci spoznavajo osnove iskanja po Slovarju izobraževanje in prispeva k osebnemu in akademskemu razvoju

slovenskega knjižnega jezika in Slovarju slovenskih frazemov. učencev.

Šolska knjižnica je pomembno informacijsko središče vsake



KLJUČNE BESEDE osnovne šole in ima pomembno vlogo, saj šolski knjižničar vzgaja mladega učenca v samostojnega uporabnika knjižnice. Šolska knjižnica, medpredmetno povezovanje, slovarji Naloga šolskega knjižničarja je, da prilagodi vsebine znanj



ABSTRACT učenčevi starosti primerno, pri čemer pričakujemo, da bodo

učenci ob zaključku osnovne šole samostojni uporabniki.

At Anton Martin Slomšek Primary School in Vrhnika, we Leta 1996 je bila ustanovljena kurikularna komisija za šolske

strive to bring the curriculum closer to students through knjižnice, ki je prenovila dotedanji program knjižne in knjižnične

independent research and interdisciplinary integration. This vzgoje. Program Knjižnična informacijska znanja za osnovno

article presents the collaboration between the school library and šolo je bil sprejet na Strokovnem svetu Republike za splošno

the Slovenian language course, where students learn the basics izobraževanje leta 1999 (Posodobitve pouka v gimnazijski praksi,

of using the Dictionary of Standard Slovenian Language and the 2010). Prednost se daje predvsem povezovanju znanja. Pri tem

Dictionary of Slovenian Idioms. je pomembno sodelovanje učiteljev, da medpredmetno povežejo

učno snov.

KEYWORDS

School library, cross-curricular conections, dictionaries

2 MEDPREDMETNO SODELOVANJE S

SLOVENŠČINO



1 Največ medpredmetnih sodelovanj v naši knjižnici izvajamo UVOD

s predmetom slovenščina. Med standardi znanja, ki jih predpisuje

Šolska knjižnica ima ključno vlogo v izobraževalnem procesu, Učni načrt za slovenščino (2018) za sedmi razred osnovne šole,

saj je prostor, kjer se spodbuja bralna kultura, raziskovanje in je tudi uporaba Slovarja slovenskega knjižnega jezika in Slovarja

vseživljenjsko učenje. Pomen šolske knjižnice presega zgolj slovenskih frazemov. Učenci se pri tem učijo razvijati

izposojo knjig; je središče, ki podpira učne procese, razvija poimenovalno zmožnost, kar pomeni, da morajo znati razložiti

informacijsko pismenost in omogoča dostop do širokega spektra frazeme v besedilu in jih ustrezno uporabiti v svojih pisnih

znanja in virov. izdelkih.

Knjižnica učencem nudi priložnost, da raziskujejo svet Učni načrt predvideva tudi uporabo slovarjev v knjižni in

literature, se poglobijo v različne teme ter razvijajo kritično elektronski obliki. Pred leti smo učence učili iskanja v knjižnih

mišljenje. Poleg tega je knjižnica prostor, kjer lahko učenci izdajah slovarjev, danes pa se osredotočamo predvsem na iskanje

pridobijo spretnosti za iskanje, vrednotenje in uporabo po spletnih različicah, občasno pa se preizkusimo tudi v iskanju

informacij, kar je ključno v sodobni družbi, preplavljeni z po knjižnih izdajah.

informacijami. Z dobro urejeno zbirko knjižničnega gradiva, ki

vključuje tako tiskane kot digitalne vire, knjižnica podpira učni 2.1 Slovar slovenskega knjižnega jezika

načrt ter pomaga učiteljem pri izvajanju pouka. Učni načrt za sedmi razred osnovne šole predvideva usvajanje

znanja uporabe Slovarja slovenskega knjižnega jezika, pri čemer

se povežemo s šolsko knjižnico. Za to učno snov namenimo tri besedilih (slika 4). Tudi naši učenci se v tem preizkusijo. Da

šolske ure, nato pa učenci svoje znanje utrjujejo še pri urah lahko frazem uporabijo v povedi, ga morajo razumeti.

slovenščine z učiteljico.

Prvo uro knjižničarka učencem predstavi zgodovino nastanka

Slovarja slovenskega knjižnega jezika in njegovo zgradbo, kar

učenci zabeležijo v zvezke. Drugo uro razdelimo učence v

skupine po največ štiri, vsaka skupina pa prejme knjižno izdajo

slovarja, tablico in delovni list z nalogami (slika 1). Rešitve nalog,

ki vključujejo celotne povedi, zapišejo v zvezek. Prvo nalogo

rešijo s pomočjo knjižne izdaje slovarja, drugo pa s pomočjo

spletne različice. Učencem reševanje nalog ne predstavlja večjih

težav, vendar večina raje uporablja spletno obliko slovarja, saj

jim iskanje po abecednem vrstnem redu v knjižni izdaji povzroča

več izzivov.





Slika 2: Delovni list (Slovar slovenskih frazemov)





Slika 1: Delovni list (Slovar slovenskega knjižnega jezika)





2.2 Slovar slovenskih frazemov

Učni načrt za slovenščino (2018) predpisuje prepoznavanje

frazemov v besedilu, ki ga morajo učenci uporabiti v svojem

besedilu.

Slika 3: Ilustracija frazema igrati se z ognjem

V šolski knjižnici imamo na voljo samo en Slovar slovenskih

frazemov in ne more skupina največ štirih učencev istočasno



delati s knjižno obliko slovarja, zato si ga izmenjujejo, ko

skupina opravi nalogo na delovnem listu (slika 2).

Na začetku učencem razložimo, kaj sploh so frazemi,

razložimo razliko med reklom in rečenico in predstavimo način

iskanja frazemov, ki se razlikuje od iskanja po Slovarju

slovenskega knjižnega jezika. V iskalni niz Slovarja slovenskih

frazemov ne moreš samo vpisati celotnega frazema, saj ne dobiš

nobenega zadetka. V iskalnik moraš vpisati samostalnik v

osnovni obliki (imenovalnik ednine), ki se pojavi v frazemu.

Iskalnih ti ponudi vse frazeme, ki se pojavijo na temo določenega

samostalnika in moraš poiskati pravega.

Učenci imajo z iskanjem po Slovarju slovenskih frazemov

veliko več težav, saj včasih niso pozorni, da se razlaga ne nanaša

na iskani frazem in prepišejo napačno razlago.



postane tudi zabavna. Učenci se preizkusijo v risanju frazemov Slika 4: Primer besedila, v katerem učenci uporabijo frazeme Slovenščina pa ni samo iskanje po slovarju, ampak lahko

in uporabi le-teh v svojih besedilih. Med ustvarjanjem se

zabavajo in hkrati veliko naučijo. Vsak učenec izbere svoj

frazem, ki ga ilustrira (slika 3). 3 ZAKLJUČEK

Učni načrt za slovenščino (2018) ima zapisan standard znanja,

Učenci se najbolj naučijo, ko so aktivni in sami raziskujejo.

da znajo učenci razumejo in uporabljajo frazeme v svojih

Na začetku učne ure smo jim pokazali, kako iskati po slovarjih,



nato pa so se sami lotili raziskovanja. Priznali so, da je iskanje

po Slovarju slovenskega knjižnega jezika precej lažje kot po

Slovarju slovenskih frazemov, saj jim nepoznavanje frazemov

predstavlja dodatno oviro. Danes poznajo zelo malo frazemov, a

hkrati priznavajo, da je raziskovanje le-teh zelo zabavno in da so

pri tem pridobili veliko novega znanja.

Medpredmetno sodelovanje med učitelji je ključnega pomena

za kakovosten pouk, saj omogoča celostno obravnavo učnih

vsebin in povezuje različna področja znanja. Ko učitelji različnih

predmetov sodelujejo, lahko ustvarijo smiselne povezave med

svojimi predmeti, kar učencem pomaga bolje razumeti in

uporabiti pridobljeno znanje v različnih kontekstih. Tako učenje

postane bolj praktično, saj se učenci zavedajo, kako so različna

znanja med seboj povezana. Poleg tega medpredmetno

sodelovanje spodbuja inovativne pristope k poučevanju in

omogoča učiteljem, da izmenjujejo ideje ter izkušnje, kar bogati

njihovo pedagoško delo in prispeva k večji motivaciji učencev.

Medpredmetno sodelovanje med knjižničarjem in učiteljem

slovenščine je izjemno dragoceno, saj omogoča poglobljeno in

raznoliko obravnavo literarnih in jezikovnih vsebin. Knjižničar s

svojim strokovnim znanjem o knjižničnem gradivu,

raziskovalnih metodah in informacijskih virih pomembno

dopolnjuje pouk slovenščine. Sodelovanje omogoča, da učenci

razvijajo bralne in informacijske veščine, hkrati pa poglabljajo

svoje razumevanje književnosti in jezika.

Knjižničar lahko učencem predstavi širok spekter literature,

pomaga pri iskanju dodatnega gradiva za bralno značko ali

raziskovalne naloge ter jih nauči učinkovite uporabe knjižničnih

virov, kot so slovarji, enciklopedije in digitalne zbirke. Učitelj

slovenščine pa lahko v sodelovanju s knjižničarjem oblikuje

naloge, ki spodbujajo kritično branje in razmišljanje, ter

organizira projekte, kjer učenci raziskujejo določeno temo s

pomočjo knjižničnega gradiva.

Takšno sodelovanje spodbuja medpredmetno učenje, ki

povezuje jezikovne spretnosti z veščinami informacijske

pismenosti, ter učencem pomaga razvijati celostne kompetence,

ki jih bodo potrebovali v življenju.



LITERATURA

[1] Posodobitve pouka v gimnazijski praksi. Knjižnično informacijsko znanje.

(2010). Ljubljana: Zavod RS za šolstvo.

[2] Učni načrt, Slovenščina. 2018. Ljubljana: Ministrstvo za izobraževanje,

znanost in šport, Zavod Republike Slovenije za šolstvo. Dostopno na

naslovu: https://www.gov.si/teme/programi-in-ucni-nacrti-v-osnovni-

soli/ (Pridobljeno 16. 8. 2024).



Matematika – vir motivacije za programiranje



Mathematics - A Source of Motivation for Programming



Gregor Kopinč

Gimnazija Brežice

Brežice, Slovenia

gregor.kopinc@gib.si



• sodelovanje na konferencah kot so NAK,

POVZETEK STEAMColab, DigiDakta, VIVID

V prispevku je predstavljena izvedba projekta Programiranje • obiski kolegov iz drugih držav, ki poučujejo na

v matematiki v sklopu večjega projekta NAPOJ MINUT. Projekt podoben način

je bil izveden v 1. letniku gimnazijskega programa. Namen V sklopu projekta so nastala gradiva že za več kot 10

projekta je bil med dijaki ozavestiti potrebo po programiranju in medpredmetnih tem, določene pa so še v razvoju. V projekt se

jim pokazati, kako s pomočjo le-tega rešujejo probleme z lahko vključijo vsi zainteresirani učitelji. Za več informacij o

različnih področij. V projektu so dijaki s pomočjo programskega projektu se lahko obrnete na: napoj@404.si.

jezika Python reševali različne matematične probleme.



2 Povezovanje predmetov matematika in

KLJUČNE BESEDE

informatika

Matematika, programiranje, gimnazija

Do ideje sodelovanja je prišlo iz potrebe pokazati, kako s



ABSTRACT programiranjem rešiti določene matematične naloge in prikazati programiranje kot pomembno orodje za reševanje problemov. The article presents the implementation of the project Projekt NAPOJ MINUT je ponudil pravo okolje in motivacijo za »Programming in Mathematics« as part of a larger project called izvedbo medpredmetne povezave. NAPOJ MINUT. The project was carried out in the first year of



the high school program. The aim of the project was to raise 2.1 Ideja projekta awareness among students about the need for programming and

to show them how it can be used to solve problems in various Dijaki naj bi skozi določene matematične naloge začutili

fields. In the project, students used the Python programming potrebo, da jih lahko rešijo s programiranjem. V okviru projekta

language to solve different mathematical problems. bi reševali take matematične naloge, za katere bi drugače na

klasični način porabili veliko časa. Na običajen način bi torej



KEYWORDS nalogo lahko rešili, če je število podatkov majhno, pri večjem številu podatkov pa se pojavi potreba po orodju, s katerim bi Mathematics, programming, high school pohitrili računanje. Take naloge bi lahko rešili s programiranjem.

Obenem pa je reševanje matematičnih nalog s programiranjem



1 odlična motivacija za učenje programiranja. Dijaki bi Uvod programirali v programskem jeziku Python v okolju Thonny. Projekt NAPOJ MINUT je nastal na pobudo učiteljev, ki so

pri svojem delu čutili potrebo po kvalitetni obliki

medpredmetnega sodelovanja. Pri tem sodelovanju naj bi 3 O projektu predmeti ne služili zgolj kot orodje ali vir podatkov, ampak naj

Na projektu sva sodelovala učiteljica informatike in učitelj

bi učeči se usvajali temeljna znanja iz vseh v sodelovanje

matematike. V sklopu priprav na izvedbo, sva določila rok, do

vključenih predmetov. Projekt NAPOJ MINUT združuje

katerega bodo dijaki določenega 1. letnika gimnazije pri pouku

(predvsem) učitelje Računalništva in informatike (RIN) in

informatike obdelali programiranje v programskem jeziku

ostalih naravoslovnih predmetov, torej MINUT - M(atematika)

Python [2], pri pouku matematike pa poglavja učnega načrta [1],

I(nformatika) N(aravoslovje) U(metnost) T(ehnologija)

iz katerih bodo matematični problemi: naravna in cela števila,

predmete. Aktivnosti v sklopu projekta so predvsem:

deljivost naravnih števil, osnove logike in teorije množic,

• redno sestajanje na spletnih (ZOOM) sestankih in

osnovni pojmi statistike.

izmenjavanje izkušenj

Pred izvedbo projekta smo pri urah matematike naredili

• razvoj gradiva, kjer se prepletata RIN in še en predmet

precej nalog, ki smo jih na običajen način (na papir) reševali zelo

• delavnice v živo, kjer se razvijajo gradiva

dolgo. Prišli smo do ideje, da bi reševanje lahko pohitrili, če bi

• učitelji dobijo nove ideje za bolj učinkovito

sam postopek izvedli programsko. A v ta namen smo potrebovali

načrtovanje in izvedbo medpredmetnega sodelovanja

ustrezno znanje programiranja. Zato smo se povezali s

predmetom Informatika.



Matematične naloge za ta projekt je bilo treba skrbno izbrati. spremenljivke, ki jih bodo uporabili v svojem programu in

Učitelj matematike običajno predlaga take naloge, ki jih dijaki opredelijo tip spremenljivk, zapišejo program za rešitev

načeloma znajo rešiti, vendar je pri izbiri nalog potrebno problema v Pythonu in ga preizkusijo, analizirajo in evalvirajo

upoštevati tudi, kako zahtevno je napisati program, ki tako zapisane rešitve.

nalogo reši, saj mora programiranje ustrezati nivoju znanja, ki ga MAT: uporabljajo računske operacije v množici naravnih in

ima dijak. Tako nalogo Število 2024 razcepi na prafaktorje zna celih števil, uporabljajo osnovne kriterije za deljivost, poznajo in

rešiti vsak gimnazijec, napisati program, ki bo to naredil, pa je uporabljajo lastnosti relacije deljivosti, izračunajo največji

precej bolj zahtevno. Zato je za ustrezno izbiro nalog, ki bi jih skupni delitelj in najmanjši skupni večkratnik dveh ali več

vključili v projekt, potrebno vedeti, kaj vse pri programiranju naravnih števil, uporabljajo osnovni izrek o deljenju celih števil,

obravnavajo dijaki in kakšno je njihovo dejansko znanje le-tega. uporabljajo Evklidov algoritem za iskanje največjega skupnega

Najbolje je, če učitelj sam predela snov programiranja iz učnega delitelja, uporabljajo zvezo Dv = ab, zapišejo sestavljeno izjavo

načrta informatike in tako dobi jasen vpogled v pričakovano s simboli, računajo z množicami, uporabljajo formulo za moč

znanje dijakov. Učitelj tako pride do novih oz. drugačnih idej za unije dveh množic, izračunajo aritmetično sredino podatkov.

pripravo ustreznih matematičnih nalog, kot jih je morda imel na

začetku. Način razmišljanja pri pripravi medpredmetne povezave 3.2 Izvedba

pa je dokončno šel v pravo smer prav s projektom NAPOJ Izvedba medpredmetne povezave je potekala v računalniški

MINUT, kjer sva učitelja pridobila ogromno novih uvidov za učilnici v obsegu štirih šolskih ur. Vsaka dvojica je prejela svoj

pripravo najinega sodelovanja. listek z nalogami, ki so jih lahko reševali dve uri. Učitelja sva

V projektu sodelujoči učitelj matematike je tudi sam predelal spodbujala dijake, da si osnutek programa zapišejo na list. Za

in usvojil vso snov programiranja [2], ki jo obravnavajo dijaki v vsako podnalogo so izdelali svoj program. Uporaba orodij

1. letniku gimnazijskega programa. Za to imajo po aktualnem umetne inteligence (Microsoft Copilot, ChatGPT in drugi) ni bila

učnem načrtu informatike dijaki predvidenih približno deset dovoljena. Pri delu so lahko uporabljali e-učbenik [2] in zvezek.

šolskih ur, kar je premalo za pridobitev poglobljenega znanja Učitelja sva ves čas spremljala delo dijakov ter jih po potrebi

programiranja. Zato je bilo potrebno še dodatno prilagoditi usmerjala in jim pomagala. Vse naloge so na koncu reševanja

naloge. Predpostavljali smo, da bodo dijaki, ki so zelo uspešni oddali. Učitelja sva oddane rešene naloge pregledala. Naslednji

pri matematiki, tudi uspešno programirali, dijaki, ki pa imajo pri dve uri so dvojice predstavljale svoje rešitve nalog pred vsemi

matematiki dostikrat težave, pa bodo le-te imeli tudi pri dijaki. Naloge in rešitve smo projicirali na iTabli. Tako so vsi

programiranju. dijaki lahko videli vse naloge vseh dvojic in njihove izdelke.

Matematične naloge je pripravil tako, da so zajemale čim več Nastale programe smo tudi preizkušali in ugotovili, da vedno ne

predelane snovi, s tem pa so pri reševanju tudi čim bolj delujejo. Učitelja sva opozorila na napake, ki smo jih skupaj z

obnavljali svoje znanje matematike. V nalogah se pojavijo nasveti ostalih dijakov odpravili. Razpravljali smo tudi, kako

večkratniki, delitelji, relacija deljivosti, osnovni izrek o deljenju nalogo rešiti še na druge načine. Dijaki so pri delu bili zelo

celih števil, največji skupni delitelj in najmanjši skupni motivirani in zavzeti. Znanje programiranja se je pri

večkratnik, Evklidov algoritem, praštevila, številski sestavi, moč marsikaterem dijaku izboljšalo.

unije dveh množic in aritmetična sredina. Skozi reševanje pa je Na koncu smo opravili analizo projekta.

potrebno uporabljati znanje s področja algoritmov in V nadaljevanju je navedenih nekaj primerov nalog, ki so jih

programiranja in sicer predvsem s podpodročij spremenljivk, dijaki reševali. nadzora in modularnosti [3].

Dijaki so naloge reševali v dvojicah. Sestavo dvojic nisva Primer 1

prepustila dijakom samim. Ustrezno smo jih sestavili glede na a) Izpišite prvih 100 naravnih števil in izračunajte njihovo

njihovo znanje matematike in programiranja. Vsaka vsoto.

matematična naloga je bila sestavljena iz več podnalog – od b) Izpišite n naravnih števil in izračunajte njihovo vsoto.

konkretnega primera do posplošenega primera oz. od lažjega c) Izpišite kvadrate prvih n naravnih števil in izračunajte

primera k težjemu primeru. Da bi bile naloge po težavnosti čim njihovo vsoto.

bolj enakovredne, so bili v nekaterih nalogah podani s d) Izpišite m-te potence prvih n naravnih števil za naravno

psevdokodo zapisani algoritmi, z uporabo katerih so lahko dijaki število m (1

ampak glede na znanje, ki ga dijaki v dvojicah imajo. Primer 2

Za reševanje nalog so imeli dijaki na voljo dve šolski uri. a) Zapišite program, ki ugotovi ali je vneseno naravno

Nato so naloge oddali. Dve uri pa smo porabili še za predstavitev število praštevilo.

in analizo izdelkov. Rešenih nalog učitelja nisva ocenjevala. b) Zapišite funkcijo, ki ugotovi ali je vneseno naravno

Ključna želja je bila pokazati, kako je lahko matematika vir število praštevilo.

motivacije za programiranje. Pri dijakih sva želela ohraniti oz. c) Izpišite vsa praštevila med 1 in 100.

vzpostaviti pozitiven odnos do programiranja. d) Izpišite vsa praštevila med 1 in n.



3.1 Operativni učni cilji Primer 3

Naštejmo, kakšni so bili operativni učni cilji, ki sva jih a) V tabelo zapišite vse naravne delitelje števila 2024 in

zasledovala pri izvedbi pri obeh predmetih. jih preštejte.

RIN: problem raziščejo in analizirajo načine reševanja, b) V tabelo zapišite vse naravne delitelje števila n in jih

izdelajo algoritem za dani matematični problem, določijo preštejte.

c) Zapišite program, ki ugotovi ali je vneseno naravno Vse izbrane naloge, kot tudi ustrezno anonimizirane rešitve

število n popolno. Naravno število n je popolno, če je enako vsoti dijakov najdete na povezavah:

vseh svojih pravih deliteljev (deliteljev števila n, ki so manjši od • Listki z nalogami

n). Primer: 6 =1+2+3. Števila 1, 2, 3 so pravi delitelji števila 6. https://gimnazijabrezice-

d) Zapišite program, ki ugotovi ali sta vneseni naravni my.sharepoint.com/:w:/g/personal/teah_gib_si/Ef7_WXIRVDV

števili m in n prijateljski. Različni naravni števili a in b sta CpZW6TBSle5UBIS7finpl3xefOrp5ApZuvA?rtime=JqDkF_X

prijateljski, če je število a enako vsoti vseh pravih deliteljev B3Eg&wdLOR=cC4A3DDDE-0CBD-4684-8862-

števila b in število b je enako vsoti vseh pravih deliteljev števila 6FF83DBD83D9

a. Primer: 220 in 284 sta prijateljski števili. • Izdelki dijakov

https://drive.google.com/drive/folders/1BSMLkIOOmZv6N

Primer 4 k9pPF1pzOo1jZRZpHEZ

a) S pomočjo dane psevdokode zapišite funkcijo, ki

izračuna največji skupni delitelj dveh naravnih števil.

4 Refleksija

funkcija EvklidovAlgoritem(a, b): Ugotovili smo, da je bilo za reševanje nalog premalo časa.

dokler b ≠ 0: Mnogim dijakom je zmanjkalo časa za reševanje vseh podnalog.

ostanek = a % b Za reševanje bi bilo potrebno namesto dveh, nameniti tri šolske

a = b ure. Tako bi dijaki lahko rešili vse naloge, programe dobro

b = ostanek preizkusili, jih sproti popravili in v program zapisali komentarje.

vrni a Prav tako je bilo premalo časa za pregled in analizo izdelkov, kjer

b) Zapišite funkciji, ki za naravni števili a in b izračunata bi namesto dveh ur potrebovali tri šolske ure časa. Analiza

D(a, b) in v(a, b). izdelkov pred vsemi dijaki se nam je zdela zelo pomembna in

c) Izračunajte, koliko naravnih števil od vključno 1 do dragocena. S tem se dijaki veliko naučijo in so zelo aktivni.

vključno n je deljivih z 8 ali z 12. Dijaki so povedali, da delo v dvojicah spodbuja njihovo

d) Izračunajte, koliko naravnih števil od vključno 1 do aktivno sodelovanje. Začutili so potrebo, da bi probleme lahko

vključno n je deljivih z a ali z b, kjer sta a in b naravni števili. rešili s programiranjem. Menijo, da so matematični problemi

odlični za iskanje programskih rešitev in utrjevanje

V Slika 1 – Slika 4 je nekaj primerov izdelkov dvojic. programiranja. Prišli smo do sklepa, da s programiranjem lahko

bistveno hitreje rešujemo določene probleme. Večina jih je



menila, da bi za izvedbo celotne aktivnosti potrebovali več časa.

Menijo, da se je njihovo znanje poglobilo. Programiranje je

mnogim postalo sedaj bolj všeč, ga pa še vedno obravnavajo kot

zahtevno. Sedaj bolj razumejo, zakaj je programiranje



Slika 1: Primer 1 a) pomembno in potrebno na ostalih področjih. Dijaki želijo, da bi občasno izvajali medpredmetne povezave informatike tudi z

ostalimi predmeti.



Učitelja meniva, da bi za izvedbo potrebovala več ur npr. v

obliki projektnega dne. Izjemno pomembna se nama zdita tudi

predstavitev in analiza izdelkov. Tako se dijaki veliko naučijo. V

glavnem so se dijaki lotili programiranja kar na računalniku brez

Slika 2: Primer 1 b) predhodne priprave oz. skice programa na papirju. Ravno tako

sva želela, da bi dijaki programe opremili s komentarji, kar v

večini niso naredili.



Ocenjujeva, da so bili cilji doseženi. Matematika je res

odličen vir motivacije za programiranje. Potrdila se je tudi

domneva, da dijaki, ki so uspešni pri matematiki, tudi uspešno

programirajo in obratno. Učitelj matematike sem prepričan in

motiviran, da bomo tudi v bodoče še pri ostalih poglavjih učnega

Slika 3: Primer 3 a) načrta kdaj reševali naloge s programiranjem. S pripravo nalog

bo sedaj bistveno lažje in hitreje. S tem, ko sem učitelj

matematike obnovil znanje programiranja, bomo z dijaki lahko



pri urah matematike tovrstne naloge reševali tudi brez pomoči

učitelja informatike oz. z manj njegove pomoči.

Meniva, da bi bilo potrebno znotraj učnega načrta informatike

nameniti več ur programiranju, znotraj učnega načrta matematike

pa več ur nameniti tovrstnim medpredmetnim povezavam.

Učitelj matematike v gimnazijskem programu po aktualnem

Slika 4: Primer 3 b) učnem načrtu praktično nima manevrskega prostora, kjer bi

lahko izvajal tovrstne medpredmetne povezave. Učitelji

matematike smo zainteresirani za medpredmetne povezave,

vendar je to ob zelo obsežnem učnem načrtu in samo štirih urah [2] G. Anželj, A. Brodnik in drugi, „Programiranje in algoritmi,“ 2015.

matematike na teden zelo težko izvedljivo. Morda bi bilo [Elektronski]. Available: https://lusy.fri.uni-

lj.si/ucbenik/book/1201/index.html. [Dostop 20. avgust 2024].

smiselno razmisliti o povečanju števila ur informatike na teden, [3] A. Brodnik in drugi, „Okvir računalništva in informatike od vrtca do

kamor bi potem lahko vključili tudi medpredmetne povezave. srednje šole,“ januar 2022. [Elektronski]. Available:

https://redmine.lusy.fri.uni-

Na slikah 5 in 6 je nekaj rezultatov ankete za dijake po lj.si/attachments/download/3060/Porocilo_RINOS_10_1_22.pdf. [Dostop

končani izvedbi projekta. 20. avgust 2024].





Slika 5: Zadovoljstvo dijakov z opravljeno aktivnostjo





Slika 6: Ugotovitve po opravljeni aktivnosti



5 Zaključek

Učitelja sva skozi projekt NAPOJ MINUT ogromno pridobila

in bova tudi v bodoče sodelovala na tem projektu. Izvajanje

medpredmetnih povezav bi želela vključiti še v ostale oddelke,

tudi v višje letnike gimnazije, kjer je predmet Informatika samo

še izbirni predmet. Želiva, da bi vsi dijaki vsako leto vsaj malo

obnovili znanje programiranja, se morda zato v večjem številu na

maturi odločili za izbirni predmet Informatika, predvsem pa

želiva, da bi videli potrebo po znanju programiranja in njegovo

uporabnost na vseh področjih. Prav tako priporočava projekt

NAPOJ MINUT vsakemu učitelju – pridite po napoj za ideje, ki

so naravnane v pravo smer in prinašajo dobre rezultate.



Zahvala

Zahvaljujem se izjemnima mentorjema projekta NAPOJ

MINUT dr. Andreju Brodniku in mag. Matiju Lokarju za vso

posredovano znanje in odlične nasvete. Ravno tako se

zahvaljujem svoji sodelavki učiteljici informatike Tei Habinc, ki

me je spodbudila k sodelovanju pri projektu in mi je bila v veliko

pomoč.



VIRI

[1] A. Žakelj in drugi, „Gimnazija: matematika - učni načrt,“ 14 februar 2008.

[Elektronski]. Available: https://dun.zrss.augmentech.si/#/. [Dostop 20.

avgust 2024]..





S projektnim delom in umetno inteligenco pri pouku




tehnike



With Project Work and Artificial Intelligence in Technology



Lessons



Veronika Koščak

profesorica matematike in tehnike

Osnovna šola Janka Modra, Dol pri Ljubljani

Videm 17, Dol pri Ljubljani, Slovenija

veronika.koscak@guest.arnes.si



students to work actively, to cooperate, to think and give impulse

POVZETEK ideas and to experiential learning, to condition cooperative

Večina ljudi, ki uči, se ne zaveda, da je biti učinkovit učitelj relationships between teacher in student. This way of acquiring

morda najtežje delo v naši družbi. Učinkovit učitelj je tisti, ki knowledge and above all manual skills has a great advantage.

uspe prepričati vse učence, da v šoli kakovostno delajo. To The school is striving for the quality knowledge for all

pomeni, da učenci izkoristijo svoje sposobnosti, ne pa da v šoli students. Learning is not only theacquisition of knowledge, skills

preganjajo dolgčas, kot jih na žalost večina počne. Projektno in habits, but is also a progressive change of an individual based

učno delo sodi med tiste postopke učenja, ki temeljijo na on one's own activities, with the existing knowledge being linked

spodbujanju učencev k aktivnemu delu, sodelovanju, to the new one. This way of learning also depends on the

razmišljanju, impulzivnosti idej in zamisli, izkustvenem učenju expectations of emotional and social perception and the

in nenazadnje tudi pogojujejo kooperativne odnose med individual's views. The lesson is based on learning with

učiteljem in učencem. Tovrstni način osvajanja znanja in discovery, constructivist learning, interactive learning, project

predvsem ročnih spretnosti ima veliko prednosti. mode of work, cross-curricular integration in current learning.

V današnjem času si šola prizadeva za kakovostno znanje Students are active during the course of the project task. With

vseh učencev. Učenje ni le pridobivanje znanj, spretnosti in questions on worksheet and with the help of artificial intelligence,

navad, ampak je učenje progresivno spreminjanje posameznika they independently discover the substance of the rectangular

na osnovi lastne aktivnosti, pri čemer se obstoječe znanje projection, they search for possible solutions and formulate

povezuje z novim. Takšno učenje je odvisno tudi od pričakovanj suggestions to solved problems. And finally they achieve results

emocionalnega in socialnega doživljanja ter posameznikovih and get new knowledge.

stališč. Pouk temelji na učenju z odkrivanjem, problemskemu

učenju, konstruktivističnem učenju, interaktivnem učenju, KEYWORDS

projektnemu načinu dela, medpredmetnim povezovanjem in Project work, Rectangular Projection, artificial intelligence in

tekočim učenjem. Učenci so med potekom projektne naloge technology

aktivni izvajalci, s pomočjo delovnega lista in zastavljenih

vprašanj s pomočjo umetne intelligence samostojno odkrivajo

snov pravokotne projekcije, sami iščejo možne rešitve in 1 VSEBINE TEHNIKE V OSNOVNI ŠOLI

oblikujejo predloge o načinu reševanja problemov ter na koncu Predmet tehnika in tehnologija učencem predstavlja načine,

ugotavljajo rezultate svojega dela, predvsem pa si veliko sredstva in organizacijske oblike spreminjanja narave ter učinke

zapomnijo. nanjo. Opredeljujejo ga štiri področja, ki se pri pouku prepletajo

in jih učenci spoznavajo (predvsem) s svojo dejavnostjo:

KLJUČNE BESEDE  tehnična sredstva,

Projektno učno delo, pravokotna projekcija, umetna  tehnologija,

inteligenca pri tehniki  organizacija dela in

 ekonomika.

ABSTRACT Pri pouku spoznavajo, kako se naravne zakonitosti

Not everyone who is teaching and educating are aware that uporabljajo v tehniki in tehnologiji. Odkrivajo in spoznavajo

being an effective teacher may be one of the hardest work in our preproste tehnične in tehnološke probleme ter z uporabo orodij

society. An effective teacher is the one who succeed to achieve in sredstev iščejo načine za njihovo reševanje. Tako ustvarjalno

that all students in school work on highest quality. That means povezujejo naravoslovna in tehnična znanja s prakso. Tehnika in

that students take advantage of their abilities. Project work is one tehnologija je predmet, ki simbolno raven udejanja v resničnosti.

of those learning processes that are based on encouraging



Ob izdelovanju predmetov imajo učenci možnost, da

razvijajo svoje sposobnosti za iskanje in oblikovanje novih

rešitev ter za odločanje zanje. Tehnika in tehnologija velja v

osnovni šoli za predmet, ki pri učencih najbolj celovito spodbuja

razvoj psihomotoričnih sposobnosti in oblikovanje socialnih

vrednot. Pri delu v skupini razvijajo sposobnosti sodelovanja in

vodenja. Ob uporabi orodij, strojev, naprav in instrumentov

razvijajo delovne spretnosti ter navade in sposobnosti za

praktično ustvarjanje. Pri praktičnem delu spoznavajo nevarnosti

in varnostne ukrepe ter razvijajo spretnosti in navade pri uporabi

zaščitnih sredstev. Oblikujejo tudi sposobnosti za ustno, pisno in

grafično sporazumevanje. Seznanijo se z značilnostmi

posameznih vrst del na tem področju. Vsebina in organizacija

delovnega procesa omogoča celovit razvoj učenčeve osebnosti.

Tehnika in tehnologija jim odkriva ter razvija interese za

poklicno usmeritev. Učenci si oblikujejo zavest o tem, kako Slika 1: Učni cilji snovi o pravokotni projekciji pri

uporaba tehnika in tehnologija spreminja svet, v katerem živijo. predmetu tehnika in tehnologija v 7. razredu OŠ

Spoznavajo, da se tehnološkega razvoja ne da ustaviti ali



zaobrniti. Ob tem se v njih oblikuje spoznanje o blagodejnem

vplivu tehnike in tehnologije na človekovo udobje ter hkratnem

škodljivem vplivu na okolje, kar spodbuja zavest in voljo, kako

ga preprečiti, če pa je že, kako ga zmanjšati in odpraviti. Pomaga

jim razviti prilagodljivost, da lahko ustvarjalno živeli v svetu

nenehnih sprememb.

Tehnična vzgoja je vsebinsko in organizacijsko povezana z

življenjsko prakso, zato se njena sestava in vsebina oblikuje

podobno, kot je oblikovan proizvodni proces. Učenec tako

spoznava tehniko in tehnologijo po delih. Do konca osnovne šole Slika 2: Opisnik za preverjanje in ocenjevanje znanja iz

pa zgradi celoten sistem. pravokotne projekcije



1.1 Pravokotna projekcija v 7. razredu 1.2 Izobraževanje učencev za informacijsko

Vsak učitelj je pri svojem delu avtonomen in od vsakega družbo

posameznika je odvisno koliko učnih ur bo namenil določeni Nov, sodobni, inovativni tip izobraževanja, ki pri učencih

temi. Ure so predpisane z učnim načrtom, a vsekakor je še nekaj zahteva aktivno (sodelujoče) učenje, pri učitelju pa izvajanje

nerealiziranih ur, ki jih vsak posameznik razporedi po svoji njegove kvalitativne komunikacije, lahko v pedagoški praksi

presoji oziroma jih planira v začetku šolskega leta z letnim uresničuje le nov tip učitelja, za katerega lahko rečemo, da je

delovnim načrtom predmeta. Učenci v 7. razredu morajo pri resnični (akcijski) raziskovalec. Torej raziskovalec, ki zna

tehniki in tehnologiji razumeti namen načrtovanja predmeta, poučevati učinke lastnega pedagoškega dela, zato, da jih

znati prebrati načrt in ga opisati, znati poimenovati posamezne kakovostno izboljšuje, dograjuje ter tako širi prostor

rise predmeta in določiti mere, razumeti pojem projekcija, inventivnosti, ne le svoji, temveč vseh sodelujočih pri razvijanju

utemeljiti risanje predmetov v pravokotni projekciji, razložiti pedagoškega procesa.

nastanek projekcij na vseh treh projekcijskih ravninah, znati Učitelj in učenec dobivata množico informacij iz okolja in

razložiti zvrat ravnin in izdelati prostorski kot iz papirja, ki jim različnih občil. Spremembe v našem in svetovnem družbenem

je v pomoč pri uvodnih urah pravokotne projekcije (Slika 1). prostoru so silovite. Šola mora biti elastična in se mora na takšne

Tudi ocenjevanje znanja, veščin in navad pri vzgojno - spremembe odzivati pravočasno, kajti le tako bodo učenci

izobraževalnem delu nimamo natančnih instrumentov. Učitelji se spoznali pomen izobraževanja, ne le uresničevanje svojih

na osnovi pedagoško - psihološkega znanja približujejo čim bolj intelektualnih zmožnosti, temveč tudi za razvoj svoje celotne

objektivni oceni. Pri tem uporabljajo tudi razne oblike testov - še osebnosti.

posebej je učiteljem v pomoč taksonomija vzgojno - Najpomembnejše je, da učenec dobiva sodobno informacijo.

izobraževalnih ciljev. Torej, če cilji pouka niso natančno Ta mu mora biti posredovana na zanimiv način, tako, da ga že

formulirani, se ta pomanjkljivost pokaže pri preverjanju in sam didaktični način vpeljuje v stalno, nadaljnje individualno

ocenjevanju znanja. iskanje in v objektivno kritično razmišljanje o pomembnih

To področje v pedagoški praksi je najmanj razdelno in se informacijah. Za bogatitev učenčeve osebnosti in za bogatitev

ravno tu pojavijo številne nepravilnosti. njegovega znanja učitelj lahko uporablja številne didaktične

Preverjanje in ocenjevanje znanja pa služi učitelji v prvi vrsti pristope kot so projektno učno delo, socialne igre, mikropouk,

za ocenjevanje svojega dela in za ocenjevanje posameznih izkustveno učenje, diskusijske ure, didaktične delavnice in tudi

učencev. raziskovalne naloge.

Pedagoška dokimologija se torej ukvarja s preverjanjem in

ocenjevanjem celotnega poteka vzgojno - izobraževalnega dela

(Slika 2).





1.2.1 Projektno učno delo

Za projektno učno delo je značilno, da presega okvire pouka,

saj se ne omejuje niti vsebinsko niti organizacijsko, pa tudi ne

časovno in prostorsko na pogoje, v katerih je organiziran šolski

pouk. Projektnega učnega dela pa tudi ni mogoče uvrstiti med

učne metode (Novak, 1990, 21).

Ravno zaradi značilnosti, ki jih ima projektno učno delo, pa

ga lahko uvrstimo med didaktične sisteme. Projektno učno delo

namreč združuje faze učiteljeve razlage in vodenja učnega

procesa, ravno tako pa tudi samostojno delo učencev. Zato je

mogoče učiteljev in učenčev odnos prikazati preko didaktičnega

trikotnika V. Poljak (Poljak, 1970, 153).

Pri projektnem učnem delu učitelj učence le vodi postopno

skozi učni proces v smeri uresničevanja ciljev, jih spodbuja in

usmerja, učenci pa se samostojno učijo ob posredni učiteljevi

pomoči in sami prevzamejo načrtovanje projekta. Učenci se tako

veliko več naučijo, ker sami pridejo do določenih znanj, spoznanj

in ciljev.

Med projekte konstruktivnega tipa aktivnosti, ki so usmerjene

h konstrukciji izdelka spada izdelava nekega predmeta, to je



lahko stojalo za prtičke, obešalniki različnih oblik, šolsko glasilo, Slika 3: Odgovori chatgpt na zastavljena vprašanja seminarsko delo ipd. Lahko gre tudi za načrtovanje in izvajanje

določene akcije kot na primer postavljanje zaključne razstave

tehničnih izdelkov, priprava izdelkov na tekmovanje mladih

tehnikov ipd.

Pri projektu usvajanja in ovrednotenja gre za ustrezno



spoznanje in ovrednotenje nekega pojava, metode, razstave...

Projekt problemskega tipa je usmerjen k reševanju nekega

problema, kamor uvrščamo raziskovalne naloge. Projekt učenja

pa je sestavljen iz aktivnosti, s katerimi učenci usvajajo

spretnosti, sposobnosti ali znanja. Vse te dejavnosti srečujemo v

naših šolah in sicer največkrat v obliki projektov. Zato spada

projektno učno delo med odprte učne sisteme oziroma procese,

ki potekajo prek določenih vsebin, organizacije, učnih metod

oziroma tehnik in nas pripeljejo do cilja. Cilje pa postavljamo

učitelji skupaj z učenci.



1.3 Umetna inteligenca pri tehniki

Umetna inteligenca se izkazuje kot zmožnost stroja, da

prikaže človeške lastnosti oziroma lastnosti, ki jih uporablja

človek kot so učenje, mišljenje, načrtovanje in kreativnost. V

šolskem letu 2023/ 2024 smo bili učitelji vključeni v projekt

Digitrajni učitelj, preko katerega smo se izobraževali na različnih

področjih in tako tudi v praksi preizkušali novo usvojeno znanje.

Tako sem tudi sama preizkusila program Chatgpt in izvedla tudi

učno uro v 7. razredu z učenci pri pouku tehnike in tehnologije. Slika 4: Odgovori Chatgpt na zastavljena vprašanja

Poudariti moram, da je učitelj pred izvedbo učne ure dolžan

preveriti program, ki ga bo uporabljal, da mora program poznati,

da učence napelje in jih usmerja. Uro sem izvedla po predelani

in utrjeni snovi poglavja o pravokotni projekciji, kar se je

izkazalo za zelo dobro, saj bi v primeru izvedene uvodne ure v

pravokotno projekcijo učenci prejeli napačne informacije od

Chatgpt, kar bi pomenilo, da bi se snov tudi napačno naučili

(Slika3, Slika 4, Slika 5).



[4] https://www.europarl.europa.eu/topics/sl/article/20200827STO85804/kaj

-je-umetna-inteligenca-in-kako-se-uporablja-v-praksi (21. 8. 2024)



Slika 5: Odgovori Chatgpt na zastavljena vprašanja



Umetna inteligenca omogoča tehničnim sistemom, da

zaznavajo okolje, obdelajo, kar zaznajo, in rešijo problem, pri

čemer ravnajo v skladu z določenim ciljem. Računalnik sprejema

podatke, ki so predhodno pripravljeni, ali pa jih zbere sam s

senzorji, denimo kamero, jih obdela in se odzove.

Sistemi, ki delujejo na podlagi umetne inteligence, lahko na

podlagi analize učinkov svojih predhodnih dejanj do določene

mere samostojno prilagajajo svoje vedenje [4].



2 ZAKLJUČEK

Prvič sem izpeljevala učno uro z učenci pri ponovitvi z

vklučitvijo chatGPT. Še posebej je učence motiviralo njihovo

znanje, ki so ga izkazovali, kaj so se naučili in so preverjali

znanje umetne inteligence. Zadovoljni so bili, ko so spoznali, da

vedo več, da znajo bolje podajati kratke odgovore na zastavljena

vprašanja. Zadovoljni so bili s prikazom, ko se prostorski kot

razgrne v ravnino in so vedeli, kje so narisani vsi trije pogledi,

ChatGPT pa tega ni vedel. Težava se je pojavila takoj, ko vsi

učenci v razredu niso bili prepričani, katere dimenzije so vidne

na določenem pogledu, zato so slepo verjeli ChatGPT-ju, kar ni

bilo prav. Vedno je potrebno preveriti vsako dejstvo, predvsem ,

ko gre za specifična znanja s posameznih področji.

Menim, da je v umetni inteligenci tudi prihodnost v vzgojno-

izobraževalnih zavodih, a je za posamezna, specifična področja

kot je na primer tehnika in tehnologija, še potrebno počakati, da

umetna inteligenca pridobi vsa potrebna znanja. Do takrat pa

bodo učenci usvajali učne cilje in pridobivali znanje preko

projektnega dela in drugih metod, ter sredstev, ki jih bomo izbrali

učitelji.



REFERENCES

[1] Bezjak, J., 1997, Projektno delo pri pouku tehnike, kot uspešna oblika in

metoda sodobne inženirske pedagogike. Ljubljana: Zavod RS za šolstvo

in šport.

[2] Novak, H., 1990, Projektno učno delo: drugačna pot do znanja. Ljubljana,

DZS.

[3] Poljak, V., 1984, Didaktične inovacije i pedagoška reforma škole. Zagreb:

Školske novine.



Knjižnično informacijsko znanje in IKT – papir in oblak v



šolski knjižnici Gimnazije Ptuj



Library Information Skills and ICT – Paper Resources



and a Cloud in the School Library of Gimnazija Ptuj



Jelka Kosi

Gimnazija Ptuj

Ptuj, Slovenija

jelka.kosi@guest.arnes.si



literacies and the importance of reading on different media, both

POVZETEK on paper and with the help of information and communication

Prispevek opisuje uporabo informacijsko komunikacijske technology, as well as their advantages and disadvantages.

tehnologije pri izvajanju knjižnično informacijskega znanja ter KEYWORDS

drugih storitev in dejavnosti, ki potekajo v gimnazijski šolski School libraries, library information skills, ICT, web tools,

knjižnici. Predstavljena je uporaba spletne strani šolske knjižnice, web applications, motivation ki deluje na platformi Weebly, uporaba Microsoft Teamsov in

Microsoft Formsov, sodelovalnega okolja Padlet, QR kod, video

vodičev in nekaterih drugih spletnih orodij in aplikacij. 1 UVOD Omenjena orodja uporabljamo pri individualnem in skupinskem

Slovenija je ena redkih evropskih držav, ki ima enotno, z

delu ob poučevanju knjižnično informacijskega znanja, pa tudi

zakonom urejeno šolsko knjižničarstvo, ki se uspešno vključuje

skozi didaktiko poučevanja s knjižnico v timskih,

v knjižnično vertikalo od vrtca, osnovne in srednje šole, preko

medpredmetnih in kroskurikularih povezavah ter projektnem

knjižnic fakultet in specialnih knjižnic do Narodne in

delu. Opisana so tudi prizadevanja pri ozaveščanju naših

univerzitetne knjižnice. Šolske knjižnice (ŠK) v Sloveniji so

uporabnikov o pomenu povezovanja znanja, ki ga skozi šolanje

pomemben del vzgojno-izobraževalnega sistema ter s tem

na gimnaziji pridobivajo pri posameznih predmetih in

vsakega zavoda oziroma šole. Svojim uporabnikom nudijo

dejavnostih, tudi s pomočjo šolske knjižnice in knjižnično

dostop do različnih tiskanih in elektronskih virov, jih vodijo pri

informacijskih znanj. Posebej poudarjamo tudi pomen različnih

učenju, raziskovanju in razvijanju bralnih navad ter tako

pismenosti in pomen branja na različnih nosilcih, tako na papirju

sodelujejo pri informacijskem opismenjevanju, spodbujanju

kot s pomočjo informacijsko komunikacijske tehnologije ter

kakovostnega branja in pridobivanju veščin vseživljenjskega

njune prednosti in pomanjkljivosti.

učenja. V nadaljevanju bomo predstavili nekaj primerov

KLJUČNE BESEDE vključevanje informacijsko komunikacijske tehnologije (IKT) v

Šolske knjižnice, knjižnično informacijsko znanje, IKT, doseganje ciljev šolske knjižnice kot so se razvili na Gimnaziji

spletna orodja, spletne aplikacije, motivacija Ptuj.



ABSTRACT

The paper describes the use of information and 2 KIZ IN IKT V ŠOLSKI KNJIŽNICI

communication technology in the implementation of library and GIMNAZIJE PTUJ information knowledge, as well as other services and activities

Delovanje možganov in posledično učenje je pogojeno s

taking place in a high school library. The use of the school library

kognitivnimi, psihološkimi in biološkimi danostmi ter neločljivo

website, which operating on the Weebly platform, the use of

prepleteno s kognicijo in emocijami ter socialnimi interakcijami.

Microsoft Teams and Microsoft Forms, the Padlet collaborative

V delovanju šolske knjižnice Gimnazije Ptuj se vključenost

environment, QR codes, video guides and some other online

IKT odraža na dva načina. S polnopravnim članstvom v sistemu

tools and applications are presented. We use the mention tools in

COBISS, ki ji med drugim daje možnost aktivnega sodelovanja

individual and group work while teaching library information

v sistemu vzajemne katalogizacije omogoča prevzemanje

knowledge, as well as trough didactics of teaching with the

bibliografskih zapisov, prisotnost v aplikaciji COBISS+ in

library in team, cross-curricular and inter-subject connections

aplikaciji mCOBISS, dostop do baze podatkov COBIB.SI ter e-

and project work.

virov, izobraževanje in druge storitve. Ob omenjenem pa se

Efforts to raise awareness among our users about the

vključenost IKT odraža tudi v drugih dejavnostih ŠK, ki bodo

importance of connecting the knowledge they acquire through

opisane v nadaljevanju.

high school education in individual subjects and activities, also

Šolski knjižničar kot strokovnjak za upravljanje z

with the help of the school library and library information skills,

informacijami lahko v šolskem prostoru smiselno in uspešno

are also described. We emphasize the importance of different

deluje samo takrat, ko pri svojem delu lahko združuje in udejanja 3 ZASNOVA IN STRUKTURA SPLETNE

znanja in spretnosti bibliotekarskega in pedagoškega poklica. S STRANI ŠOLSKE KNJIŽNICE svojim strokovnim delom slovenski šolski knjižničarji

Eden od primerov uporabe IKT v ŠK je spletna stran, ki smo

vsakodnevno prispevamo k osmišljanju izobraževanja katerega

jo zasnovali na Gimnaziji Ptuj. Njen namen, razvoj in



Vrednost pridobljenega znanja, tudi znanja iz KIZ, je odvisna od uporabnost smo načrtovali že dalj časa. Zbirali smo gradivo in namen ni le prenos informacij, ampak predvsem prenos pomena.

vsebine za objavo, določili smo njeno strukturo. Nismo pa uspeli



učenja pri katerem ima šolska knjižnica nedvomno veliko vlogo. vključenosti v učni proces, predvsem v KIZ in v povezavi z njimi. Ta del je še čakal na uresničitev, Sodobni načini poučevanja in delovanja pa danes vključujejo V času izolacije, ko je potekal pouk na daljavo in smo se vanj tudi uporabo IKT. načinov poučevanja in pridobivanja spretnosti vseživljenjskega še zasnovati njene interaktivnosti in načina neposredne njegove uporabe. Ta misel na nek način povzema vlogo sodobnih



IKT pri pouku KIZ in delovanju šolske knjižnice na Gimnaziji V nadaljevanju bo predstavljenih nekaj primerov uporabe vključile tudi šolske knjižnice, se je naloga sprva zdela skoraj neizvedljiva. Generalno napotilo v tistem času je bilo, naj bomo pri aktivnostih realni, naj bodo dejavnosti zaporedne in ne po Ptuj. Šolsko knjižnico determinira njen prostor delovanja in njeni nepotrebnem preveč obremenjujoče in podobno. Najti je bilo uporabniki (Slika 1, 2, 3). potrebno rešitev za osmislitev vloge šolske knjižnice v do tedaj

neznanem stanju popolne izolacije.



Skoraj edina rešitev je bila misel na uresničitev ideje o

digitalni knjižnici kot vstopni točki »virtualnega posnetka«

dejanske knjižnice Gimnazije Ptuj. V samo nekaj dneh je

»Digitalna knjižnica GPT« (DK GPT) bila postavljena. GPT v

imenu spletne strani šolske knjižnice Gimnazije Ptuj predstavlja

njen akronim v vzajemni bazi COBISS. Spletna stran Digitalna

knjižnica GPT je bila oblikovana v času učenja na daljavo,

spomladi leta 2020. Zapis na naslovni strani DK GPT je

opredeljeval, da je namenjena ptujskim gimnazijcem za pouku

na daljavo, vsebovala pa je tudi vsebine za prosti čas. Objavljene

vsebine s predmetnega področja KIZ so se prepletale z vsebinami,

uporabnimi tudi za ostala predmetna področja. Zapisano je bilo

Slika 1: Prosti pristop v šolski knjižnici še, da smo dijakom Gimnazije Ptuj na voljo za dodatna pojasnila

k objavljenim vsebinam ter pomoč pri uporabi Cobissa in drugih



knjižničnih portalov, pri iskanju gradiva in ostalih raziskovalnih

postopkih, ustvarjanju referata, seminarske naloge ter v pomoč

pri iskanju vsebin za bogatenje šolskega znanja ali prostega časa.

Povezava do DK GPT je bila dostopna na spletni strani

Gimnazije Ptuj, v oddelku šolska knjižnica, kjer so tudi sicer

objavljene vsebine v povezavi s šolsko knjižnic. ter na šolski

Facebook in Instangram strani. DK GPT je delovala ves čas

izolacije in tudi še nekaj časa po vrnitvi k običajnemu načinu

izvajanja pouka v šoli.

Danes je na spletnem naslovu http://digitalnagpt.weebly.com/

dostopna naslovna stran s povezavo do spletne strani Gimnazije

Slika 2: Prostori šolske knjižnice Gimnazije Ptuj Ptuj, povabilom za uporabnike – vprašaj knjižničarja, napotila

varnost na spletu, netetika in spoštovanje avtorskih pravic (Slika

4). Ostale vsebine pa so dostopne preko arhiva spletne strani DK



GPT za pouk KIZ, druge dejavnosti in projekte v katerih sodeluje

šolska knjižnica Gimnazije Ptuj (Slika 5, 6, 7, 8, 9, 10,11).

DK GPT je bila zamišljena tudi kot vstopna točka šolske

knjižnice za komuniciranje z uporabniki, izobraževanje in

motiviranje za branje ter promocijo šolske knjižnice. Želeli smo,

da bi DK GPT bila zasnovana tako, da bi čim bolj odsevala

običajno delovanje šolske knjižnice na Gimnaziji Ptuj, torej ko

deluje v živo. Premišljeno smo v zasnovo vključili tudi teorijo

barv in spletno stran zasnovali s prevladujočo oranžno barvo, ki

po teoriji vzbuja prijetno, ustvarjalno, zabavno in poživljajoče

vzdušje. Je barva ognja in sonca. Združuje pozitivne lastnosti,



Slika 3: Utrinek iz šolske knjižnice Gimnazije Ptuj posreduje prešernost, prijateljstvo, vitalnost, veselje, igrivost, kot so veselje do življenja in vedrino. Stimulira duševnost in

čustvenost sigurnost in toplino. Je prijetna barva, ki pomaga

premagovati utrujenost. Pozitivno vpliva na črnogledost,

depresijo in pomanjkanje motivacije. S tem smo skušali

vzdrževati pozitivno naravnanost in optimizem naših

uporabnikov ter vabiti k branju in ustvarjalnosti. Tudi dijaški

ustvarjalnosti in dejavnostim knjižničnega kluba DOTIK

KNJIGE, ki že vrsto let deluje v sklopu šolske knjižnice, smo

namenili poseben zavihek v DK GPT.

Spletna stran je bila zasnovana z mislijo na učenje s

tehnologijo, samoučenje in vseživljenjsko učenje ter

razpolaganjem s časom na način, da uporabnik ponujene vsebine

in hipertekste uporablja v zanj najprimernejšem času, vse skupaj

pa omogoča preobrazbo učenja. Ker so bile vsebine za pouk v

DK GPT dostopne ves čas, so lahko uporabniki do njih dostopali Slika 5: Motivacijski nagovor z naslovne strani DK GPT

v zanje najprimernejšem času.

Za prosti čas so bila ponujena različna prostodostopna

gradiva, ki so jih ponujali muzeji, galerije, založbe in druge

ustanove brezplačno.

.





Slika 6: Naslovnica zavihka KIZ in prostodostopni viri v



DK GPT





Slika 4: Naslovna stran spletne strana Digitalne knjižnice



GPT



Vzpostavljena DK GPT je bila bogata z vsebinami za pouk

KIZ in v povezavi s KIZ, za pouk ostalih predmetnih področij in

prosti čas, a največji izziv je bila sama izvedba in dostop do Slika 7: Zavihek KIZ v DK GPT uporabnikov – dijakov in sodelavce. Dogovorjen komunikacijski

kanal je sicer bila komunikacija v e-Asistentu, a ta ni zagotavljala



vseh potreb za izvedbo pouka. Rešitev se je nakazala zelo kmalu,

ko smo bili strokovni delavci Gimnazije Ptuj deležni

izobraževanj kjer smo usvajali strukturo in uporabnost Microsoft

Teamsov (MST) pri pouku.

Za delovanje ŠK na daljavo in uporabo DK GPT smo v MST

oblikovali skupine za vse oddelke Gimnazije Ptuj. S tem je bila

omogočena izvedba pouka, tudi pouka KIZ na daljavo. Dijaki so

k pouku in k uporabi DK GPT bili pozvani preko dodeljenih

nalog v MST. Za izvedbo timskih povezav s KIZ smo v MST s

kolegi zasnovali skupine povezanih predmetov za posamezni Slika 8: Zavihek prostodostopni spletni portali pri KIZ

oddelek. K pouku so dijaki bili pozvani preko sestanka v



koledarju v MST, v dodeljenih nalogah so bili napotki kako naj

uporabijo gradivo iz DK GPT. Občasno smo za hitrejšo in lažjo

komunikacijo med poukom uporabili še sodelovalno orodje

Padlet.



Slika 9: Zavihek prostodostopne spletne povezave za pouk



[2] Kurikul. Knjižnično informacijsko znanje [Elektronski vir] : gimnazija :

splošna, klasična, strokovna gimnazija. (2008). Ljubljana : Ministrstvo za

šolstvo in šport : Zavod RS za šolstvo.

[3] O naravi učenja [Elektronski vir] : uporaba raziskav za navdih prakse.

(2013). Ljubljana : Zavod Republike Slovenije za šolstvo. Dostopno na

naslovu http://www.zrss.si/pdf/o-naravi-ucenja.pdf (15. 7. 2023).

[4] Šolska knjižnica kot središče branja in stičišče medpredmetnega

povezovanja [Elektronski vir]. (2022). Ljubljana: Zavod RS za šolstvo.

Dostopno na naslovu

http://www.zrss.si/pdf/Solska_knjiznica_kot_sredisce_branja_in_sticisce

_medpredmetnega_povezovanja.pdf (22. 7. 2023).

[5] Vilar, Polona, Zabukovec, Vlasta. (2023). Model kakovostne šolske

knjižnice: zasnova in uresničevanje v praksi. V Pogled na šolo 21. stoletja

Slika 10: Zavihek KIZ in prostodostopni viri v DK GPT v duhu kompetenc in pismenosti (str. 14–15). Zavod RS za

šolstvo.Elektronski vir]. Ljubljana: Zavod RS za šolstvo. Dostopno na

naslovu https://www.zrss.si/pdf/Pogled_na_solo_21_stoletja.pdf (17 7.

2023).



[6] Weel, Adrijan van der. (2015). Spreminjanje naše besedilne zavesti. Na

poti k digitalnemu redu znanja. Ljubljana: Cankarjeva založba.



Slika 11: Zavihek KIZ in prostodostopni viri v DK GPT





Slika 12: Padlet



Opisane izkušnje in nova spoznanja uporabe IKT so nam

odprla nova pedagoška in bibliotekarska obzorja.



4 ZAKLJUČEK

V prispevku je predstavljena uporaba IKT tehnologije pri

pouku KIZ in drugih storitev šolske knjižnice Gimnazije Ptuj.

Njena uporaba se je neizmerno pospešila v času izolacije, ko je

uporaba IKT bila edina možnost za izvajanje pouka, delovanje

šolske knjižnice ter komunikacijo. Dobro prakso ohranjamo še

danes za obogatitev izvajanja pouka KIZ, obveščanje,

sodelovanje v projektih, razstavni dejavnosti, promociji

knjižnice. Vse zapisano predstavlja pomembne priložnosti za

šolske knjižnice, da ostanejo relevantne in koristne za šolski

sistem ter spodbujajo branje, učenje in poučevanje v sodobnem

digitalnem okolju. S pomočjo informacijsko-komunikacijske

tehnologije šolske knjižnice vzdržujejo svojo osrednjo vlogo kot

središče za dostop do znanja, informacij in učenja v šolskem

okolju. Velik izziv in priložnost za krepitev vloge šolskih

knjižnic so orodja kot je ChatGPT in spremembe v učenju in

poučevanju, ki ga tovrstna orodja omogočajo.



LITERATURA IN VIRI

[1] Carr, Nicholas. (2011). Plitvine: kako internet spreminja naš način

razmišljanja, branja in pomnjenja. Ljubljana, Cankarjeva založba



Digitalne kompetence učitelja pri delu z učenci s



posebnimi potrebami



Digital Competences of Teachers when Working with Pupils



with Special Needs



Klaudija Kovač

Osnovna šola Glazija

Oblakova ulica 15

Celje, Slovenia

klaudija.kovac@osglazija-celje.si



POVZETEK motnje, motnje pozornosti in koncentracije). Učenci, vključeni v

V osnovno šolo s prilagojenim programom z nižjim isti oddelek, so si med seboj torej lahko zelo različni, zato se

izobrazbenim standardom so vključeni učenci z lažjo motnjo v mora učitelj pri poučevanju posluževati različnih metod,

duševnem razvoju, ki jo pogosto spremljajo še druge motnje, pristopov, da bi učencem učinkovito podal učne vsebine. Pri tem

ovire oz. primanjkljaji na področju učenja ali čustvovanja. seveda mora upoštevati tudi individualne prilagoditve za

Učitelj se pri poučevanju pogosto srečuje s težavami na področju posameznega učenca. Z dobro usvojenimi digitalnimi

pomanjkanja učnega gradiva za prilagojeni program z nižjim kompetencami si lahko učitelj, ki poučuje učence s posebnimi

izobrazbenim standardom, poleg tega mora upoštevati tudi potrebami, učinkovito pomaga pri pripravi učnih gradiv, ki jih je

individualne pristope in prilagoditve za posameznega učenca. v fizični ali elektronski obliki zelo malo.

Zaradi navedenega so dobro razvite digitalne kompetence lahko “Digitalne kompetence sestavljajo znanja, spretnosti in

učitelju v veliko pomoč pri pripravi in izvedbi pouka. stališča. Posamezniki potrebujejo znanja o digitalni tehnologiji

za sporazumevanje, ustvarjalnost in inovativnost ter

KLJUČNE BESEDE razumevanje kritičnega vrednotenja priložnosti in izzivov, ki jih

Učenci s posebnimi potrebami, prilagojeni program, digitalne prinaša uporaba digitalne tehnologije.” (Dolinar, Poberžnik,

kompetence Jerše, 2023, str. 17). Pri poučevanju otrok s posebnimi potrebami

nenehno iščem poti in načine, kako jim čimbolj učinkovito



ABSTRACT približati učne vsebine. Pri tem se pojavi dodatni izziv, saj

pogosto moram isti delovni list različnim učencem drugače

A primary school with an adapted programme with a lower

prilagoditi. Oblikovanje delovnih listov je lažje in učinkovitejše

standard of education includes pupils with a mild intellectual



obstacles or deficits in the field of learning or emotions. When zahteva premišljen in kritičen, a radoveden, odprt in napreden teaching, the teacher often faces problems in the area of lack of odnos do njihovega razvoja. Prav tako zahteva etičen, varen in teaching materials for an adapted programme with a lower odgovoren pristop k uporabi teh orodij.” (Ključne kompetence za disability, which is often accompanied by other disorders, upoštevati dejstvo, da “uporaba digitalnih tehnologij in vsebin z uporabo digitalne tehnologije. Ob tem je seveda potrebno



educational standard, and he must also take into account vseživljenjsko učenje, 2015). individual approaches and adaptations for each student. For this

reason, well-developed digital competences can be of great help

to teachers in the preparation and implementation of lessons. 2. USTVARJANJE DIGITALNIH VSEBIN ZA

UČENCE S POSEBNIMI POTREBAMI

KEYWORDS

Pupils with special needs, adapted programme, digital 2.1 Uporaba spletnega orodja Canva

competences Pri izdelavi delovnih listov za posamezni šolski predmet

pogosto uporabljam spletno orodje za ustvarjanje digitalnih



1. vsebin Canva, kjer lahko isti delovni list hitreje prilagodim (npr. UVOD

male/velike tiskanke, večje črke, poudarjeno navodilo ipd.).

Prilagojeni program z nižjim izobrazbenim standardom Canva je spletno orodje za grafično oblikovanje. Učitelju delo

obiskujejo učenci z lažjo motnjo v duševnem razvoju, ki se zelo olajša z že obstoječimi predlogami. “Predloge močno

pogosto pojavlja v kombinaciji z drugimi motnjami, ovirami oz. olajšajo in pohitrijo proces izdelave, kar pomeni prihranek na

težavami (avtizem, Downov sindrom, vedenjske in čustvene času, hkrati pa tudi podporo manj veščim v oblikovanju. Ko

izberemo določeno predlogo, ki nam ustreza, imamo možnost

posamezne elemente oblikovati po svoje.” (Canva- orodje za 2.2 Uporaba spletnega orodja Padlet



oblikovanje, 2016). “Inovativna učna okolja so okolja, ki vključujejo uporabo Delovni list Bralno razumevanje (slika 1) sem tako novih tehnologij, pristope in metode za izboljšanje procesa prilagodila za tiste učence, ki še slabše berejo male tiskanke in učenja. Ta okolja spodbujajo interakcijo, sodelovanje in aktivno sem namesto povedi uporabila le posamezne besede oz. besedne vključevanje učencev ter omogočajo prilagodljivost in zveze (slika 2) ter za tiste učence, ki še vedno berejo z velikimi prilagajanje individualnim potrebam učencev, posledično pa tiskanimi črkami (slika 3). pripomorejo k izboljšanju kakovosti izobraževanja.” (Kreuh,

Pesek, Košutnik, Čuk, Možina Podbršček, 2023, str. 10). Eno



izmed takih učnih okolij je po mojem mnenju tudi programska

oprema Padlet, ki jo uporabljam predvsem pri pripravi učnih ur

pri predmetu spoznavanje okolja ter za predstavitev vsebine

dnevov dejavnosti.



3. RABA DIGITALNIH TEHNOLOGIJ PRI

AKTIVNIH METODAH IN OBLIKAH

DELA TEKOM POUKA

Učenci s posebnimi potrebami potrebujejo za učenje več časa

ter različne pristope in metode dela. Učitelj se mora pogosteje

posluževati metode izkustvenega učenja, saj je zapomnitev

znanja pri učencih na ta način hitrejša in dolgotrajnejša. Podoben

učinek ima tudi aktivno sodelovanje učencev pri pouku z

Slika 1: Delovni list Bralno razumevanje- povedi z malimi uporabo digitalnih tehnologij. Kot navajajo Dolinarjeva,

tiskanimi črkami Poberžnikova in Jeršetova (2023, str. 31) “aktivne oblike in

metode učenja spodbujajo samostojnost učenja, razvoj kritičnega



mišljenja, komunikacijske spretnosti, reševanje problemov in

timsko delo.”

Rabe digitalnih tehnologij pri učencih se poslužujem

predvsem pri utrjevanju učnih vsebin, kjer učenci pogosto delajo

v dvojicah ali v manjših skupinah. Na ta način rešujejo spletne

kvize, ki jih ustvarim v orodju za spletne ankete 1KA ali po

navodilih v vizualnem programskem jeziku Scratch oblikujejo

nalogo in na ta način utrdijo obravnavano učno snov. Z

radovednostjo se poslužujejo tudi uporabe interaktivne bele table,

kar jih pri reševanju nalog dodatno motivira.



4. ZAKLJUČEK

Veliko prednost uporabe digitalnih znanj učitelja pri pripravi

in izvedbi pouka vidim v vizualizaciji oziroma večpredstavnosti.

Slika 2: Delovni list Bralno razumevanje- besede z malimi Učencem lahko na ta način bolj nazorno prikažem učne vsebine

tiskanimi črkami in jim omogočim lažje razumevanje zahtevnejših učnih vsebin.

Pri učencih s posebnimi potrebami, ki imajo lahko večje težave



z razumevanjem abstraktnih pojavov, so vizualne predstavitve še

toliko bolj zaželene in učinkovito izboljšajo dolgoročno

pomnenje učnih vsebin. Učenci s posebnimi potrebami se z rabo

digitalnih tehnologij tekom pouka tudi lažje povežejo z vrstniki

in z njimi bolje komunicirajo. Kot navaja Digitalna akademija

(2022) “digitalne kompetence predstavljajo eno od osmih

ključnih kompetenc in se nanašajo na samozavestno in kritično

uporabo digitalnih tehnologij za pridobivanje in izmenjavo

informacij, komuniciranje in sodelovanje, ustvarjanje digitalnih

vsebin, varno uporabo tehnologij informacijske družbe in

reševanje problemov.”

Usvojene digitalne kompetence so lahko učitelju pri pripravi

na pouk in izvedbi le-tega v veliko pomoč, ravno tako učencem

vzbudijo zanimanje za določene učne vsebine in popestrijo pouk.

Slika 3: Delovni list Bralno razumevanje- besede z velikimi So učinkovit pripomoček, s katerim lahko učitelj pri učencih

tiskanimi črkami poleg učnih vsebin uspešno razvija in krepi tudi ostale



pomembne kompetence kot so učenje učenja, komunikacija,

socialne kompetence, kulturno izražanje itd.

Spremembe na področju digitalnih tehnologij že zdavnaj več

ne trkajo na vrata šolstva, temveč je razvijajoča se digitalna

tehnologija v vedno večji meri prisotna v skorajda vseh razredih.

Še vedno pa je učitelj tisti, ki ima v rokah veslo ter zmore

učinkovito veslati med viharnimi spremembami in stalnim

okvirjem. Le na ta način bo moč vzgojiti digitalno odgovorne in

spoštljive člane naše družbe. Učence s posebnimi potrebami pa

učitelj na ta način opolnomoči in jim omogoči lažjo vključitev v

digitalno družbo.



REFERENCE

[1] Dolinar M., Poberžnik A., Jerše L. (2023). Vodenje in podpora učencem

pri pridobivanju digitalnih kompetenc. Ljubljana: Zavod RS za šolstvo.

[2] Kreuh N., Pesek I., Košutnik R., Čuk A., Možina Podbršček I. Projekt

Dvig digitalne kompetentnosti se predstavi. Ljubljana: Zavod RS za

šolstvo.

[3] Inkubator 4.0. (2016). Canva – orodje za oblikovanje. Ljubljana. Socialna

akademija. Pridobljeno s https://inkubator40.si/canva/.

[4] Digitalna akademija. (2022).Kaj so digitalne kompetence in zakaj so

pomembne? Pridobljeno s https://akademija.digitas.si/.

[5] Movit.si (2015). Ključne kompetence za vseživljenjsko učenje.

Pridobljeno s https://www.movit.si/.





Preverjanje predznanja z digitalno tehnologijo




Knowledge Assessment with Digital Technology



Ksenija Koželj Erika Smrekar

Ekonomska šola Ljubljana Ekonomska šola Ljubljana

Ljubljana, Slovenija Ljubljana, Slovenija

ksenija.kozelj@presernova.si erika.smrekar@presernova.si



creating their own projects, they also developed their algorithmic

POVZETEK skills, which are crucial for problem-solving in computer science

V učnem procesu učenci pogosto ne dobijo priložnosti, da bi and other natural sciences.

učne cilje več predmetov združili v eni aktivnosti. Da bi to vrzel By incorporating the Makey Makey educational kit, the

premagali, smo pri urah informatike uvedli inovativni pristop, ki students created physical interfaces that they connected with

povezuje matematiko z informatiko. Dijakom smo omogočili, da programming tools, allowing them to transfer theoretical

so preverjali svoje matematično predznanje in hkrati usvajali knowledge into practice. This way, they not only tested their

temeljna računalniška znanja. S programiranjem z delčki v mathematical knowledge but also improved their abilities to

Scratchu in z uporabo izobraževalnega kompleta Makey Makey create interactive solutions. This interdisciplinary approach

so ustvarjali interaktivne plakate, ki so jim nudili kreativno in helped students develop a wide range of skills essential for

praktično preverjanje že usvojenih matematičnih konceptov. successful functioning in the modern digital environment and

Programiranje v Scratchu jim je omogočilo spoznavanje in prepared them for more complex challenges in both mathematics

utrjevanje osnov računalniškega mišljenja in programerskih and computer science.

konceptov, kot so spremenljivke, pogojni stavki in zanke. Skozi

ustvarjanje lastnih projektov pa so hkrati razvijali tudi svoje KEYWORDS

algoritmične spretnosti, ki so ključne za reševanje problemov v Computational thinking, algorithmic thinking, physical

informatiki in ostalih naravoslovnih predmetih. computing

Z vključitvijo izobraževalnega kompleta Makey Makey so

dijaki ustvarili fizične vmesnike, ki so jih povezali z orodji za

programiranje, kar jim je omogočilo, da so teoretična znanja 1 MEDPREDMETNO POVEZOVANJE IN

prenesli v prakso. Tako so ne le preverili svoje matematično PROJEKT NAPOJ MINUT

znanje, ampak tudi izboljšali svoje sposobnosti za ustvarjanje Vse bolj postaja jasno, da je medpredmetno sodelovanje v

interaktivnih rešitev. Ta interdisciplinarni pristop je dijakom izobraževanju ključnega pomena za celosten razvoj učencev.

pomagal razvijati širok spekter veščin, ki so ključne za uspešno Projekt NAPOJ MINUT izhaja prav iz te potrebe po povezovanju

delovanje v sodobnem digitalnem okolju, ter jih pripravil na znanja različnih predmetnih področij, z namenom, da učencem

kompleksnejše izzive na obeh področjih, tako v matematiki kot omogoči bogato in praktično učno izkušnjo. Osnovna ideja

informatiki. projekta je ustvariti okolje, kjer se znanje iz področja

računalništva prepleta z znanjem matematike, naravoslovja,

KLJUČNE BESEDE umetnosti ali tehnologije, s poudarkom na enakovrednosti obeh

Računalniško razmišljanje, algoritmično razmišljanje, fizično predmetov. Na ta način se spodbuja aktivnost med učitelji

računalništvo različnih predmetov, kar vodi v ustvarjanje interdisciplinarnih

vsebin, ki učencem omogočajo celovito učno izkušnjo. Učenci se

ABSTRACT tako soočajo z iskanjem rešitev avtentičnih problemov, do

In the learning process, students do not often get the katerih jih pripelje znanje pridobljeno na različnih področjih.

opportunity to combine the learning objectives of multiple V okviru projekta MINUT NAPOJ je tako razvita vrsta

subjects into one activity. To bridge this gap, we introduced an projektov, ki poudarjajo tako obliko medpredmetnega

innovative approach in computer science classes that connects sodelovanja. Enega od teh projektov predstavljamo tudi v tem

mathematics with computer science. We enabled students to test prispevku.

their mathematical knowledge while simultaneously acquiring Dijaki 1. letnikov so ob vstopu v srednjo šolo običajno brez

fundamental computer skills. By programming with blocks in predhodnega znanja programiranja, zato so se morali pred

Scratch and using the Makey Makey educational kit, they created izvedbo projekta spoznati z okoljem Scratch in osnovnimi

interactive posters that provided a creative and practical way to programerskimi koncepti, kot so spremenljivke, zanke in pogojni

review previously acquired mathematical concepts. stavki.

Programming in Scratch allowed them to learn and reinforce Ker je poznavanje osnovnih obrazcev za izračun obsega in

the basics of computational thinking and programming concepts ploščine likov temeljno za uspešno nadgrajevanje znanja iz

such as variables, conditional statements, and loops. Through geometrije, sva v ta projekt vključili ravno te matematične cilje.

Hkrati pa so to tudi učni cilji, ki naj bi jih dijaki usvojili že v Scratch omogoča povezavo z več izobraževalnimi kompleti, med

osnovni šoli. drugimi tudi z Makey Makey.



1.1 Orodje in okolje

Pri iskanju ideje za projekt smo izbrali dijake 1. letnikov, saj

smo iz področja informatike vključili učne cilje iz programiranja,

ki jih z dijaki predelujemo v 1. letniku programa ekonomska

gimnazija. Odločili smo se, da bodo dijaki izdelovali interaktivne

plakate s programiranja z delčki v Scratchu in z uporabo

izobraževalnega orodja Makey Makey. Tako so dijaki med

preverjanjem predznanja iz matematike usvajali in utrjevali

temeljna znanja iz računalništva.

Izobraževalni komplet Makey Makey omogoča pretvorbo

vsakodnevnih predmetov z uporabo aligatorskih sponk in

prevodnih materialov v vnosni sistem (glej sliko 1). Z njegovo Slika 2: Prevodniki, ki jih lahko uporabimo pri izdelavi

pomočjo običajen plakat postane interaktiven.

Komplet ima na sprednji strani 6 vhodnih priključkov Za lažjo predstavo bomo opisali primer izdelave preprostega

(puščice gor, dol, levo, desno, presledek in klik), na spodnjem projekta, v katerem bo maček hodil levo in desno. Smer hoje bo

delu se nahaja 6 priključkov za ozemljitev. Z dotikom ustreznih določena glede na pritisk ustrezne tipke na plakatu. Na začetku

delov plakata tako simuliramo pritisk na omenjene tipke ter klik je potrebno poskrbeti za vklop razširitve v orodju Scratch. Ko to

z miško. Na zadnji strani je dodanih še 12 priključkov, ki storimo, se pojavita dva dodatna ukaza (Ko je pritisnjena tipka in

nadomeščajo pritiske na črke W, A, S, D, F in G, levi in desni Ko so pritisnjene tipke). Na plakatu označimo del na katerega bo

miškin klik ter premikanje miške gor, dol, levo in desno. uporabnik pritiskal. Pri tem lahko uporabimo železne razcepke,

ki jih povežemo s krokodiljim priključkom. Lahko pa uporabimo



tudi bakren trak ali prevodno barvo, s katero povežemo del

plakata in komplet Makey Makey. V našem primeru bomo na

plakatu puščico desno povezali s krokodiljim priključkom in ga

povezali na puščico na kompletu Makey (glej sliko 3).

Sprogramirati je potrebno še premikanje mačka. V Scratchu za

premikanje uporabiko ukaz pojdi 10 korakov. Da se bo maček ob

pritisku na tipko levo obrnil v pravo smer, pa uporabimo še ukaz

obrni se v smer. Na kompletu Makey Makey je potrebno dodati

še kabel za ozemljitev (v našem primeru je to kabel zelene barve).

Uporabnik mora med pritiskanjem na tipki na plakati držati tudi

kabel ozemljitve.





Slika 1: Sprednja in zadnja stran kompleta Makey Makey





Makey Makey na računalnik povežemo z USB kablom, za



priključke pa uporabimo krokodilje kable in povezovalne žice.



Na plakatu je za prevodnost možno uporabiti bakren

samolepilni trak, prevodno barvo, grafit (sled svinčnika) ali

železne razcepke (glej sliko 2). Tako se ob dotiku ustreznega dela

plakata preko navedenih prevodnikov simulira pritisk določenih

tipk (puščice, preslednica, črke,…).

Programsko orodje Scratch je brezplačno programsko orodje

namenjeno programiranju z delčki, kar omogoča uporabo tudi pri



mlajših učencih. Dostopno je preko spletne strani scratch.mit.edu. Slika 3: Plakat s povezovalnimi kabli in programska koda v Na spletni strani lahko najdemo tudi vodiče v slovenskem jeziku. Scratchu

To je primer enostavne uporabe kompleta Makey Makey in

orodja Scratch. Ob nakupu kompleta Makey Makey so kabli in

žice že priloženi, priporočamo še nakup železnih razcepkov in po

potrebi prevodnega traku in prevodne barve (glej sliko 2).



1.2 Projekt

Da bi zapolnili vrzel med povezovanjem znanja iz informatike in

matematike smo zasnovali projekt, ki vključuje uporabo

kompleta Makey Makey. Z njegovo uporabo dijaki usvajajo

programerske koncepte. Ta projekt je bil zasnovan v okviru

projekta MINUT NAPOJ.

Dijaki posameznega letnika so pri informatiki razdeljeni v dve

skupini po 16 dijakov. Pouk informatike se izvaja enkrat Slika 4: Interaktivni plakat

tedensko v dveh zaporednih šolskih urah. Dijaki so v mesecu

decembru in januarju usvojili osnovne programerske koncepte s Povratne informacije dijakov so bile pozitivne. Poudarili so, da

pomočjo izdelovanja različnih kratkih interaktivnih projektov. je bil projekt zabaven in poučen ter da so se naučili veliko novega.

Nato so pričeli z načrtovanjem interaktivnega plakata. Pripravili Z željo, da bi lahko še večkrat pokazali svojo ustvarjalnost, si

so skico plakata v katero so vključili vsebino, ki so jo kasneje želijo še več takšnih projektih v prihodnosti. Projekti jim

obravnavali pri matematiki. Dodali so tudi skico fizičnih omogočajo uporabo teoretičnega znanja na ustvarjalen način in

povezav s kabli. Vsak dijak je imel na voljo izobraževalni utrjevanje že pridobljenega znanja. Pri raziskovanju ideje o

komplet Makey Makey, bakreni trak, prevodno barvo in plakatu so razvijali tudi veščine raziskovanja.

razcepke. V Scratchu so morali z ustreznimi ukaznimi bloki

sestavi kodo, ki je ob pritisku na določene dele plakata izvedla

ustrezne akcije, kot so predvajanje zvoka, animacije, menjava 3 ZAKLJUČEK

ozadja ipd. Projekt z uporabo orodja Scratch in Makey Makey je dijakom

omogočil, da na kreativen način povežejo znanje iz različnih



predmetov in ustvarijo nekaj edinstvenega. Takšni projekti so



odličen način za spodbujanje inovativnosti in praktičnega učenja

ter za pripravo dijakov na prihodnje izzive v svetu tehnologije,

kjer se bodo vsakodnevno srečevali z izzivi. Verjamemo, da taki

projekti spodbujajo nujno potrebne veščine za reševanje

problemov in hkrati dijakom omogočajo uvid v uporabnost

povezovanja znanja. Za izvedbo takega projekta je pomembno,

da so učitelji željni usvajanja novih znanj in medsebojnega

sodelovanja s kolegi. Na ta način tudi učitelji krepijo veščine

sodelovalnega dela in komunikacijske veščine. Pomembna je

tudi podpora s strani vodstva šole, ki poskrbi za programsko

opremo in ostale pogoje kot so: prostor, čas,...

V prihodnje bo v projektu MINUT NAPOJ zagotovo razvitih še

Slika 3: Zadnja stran interaktivnega plakata več zanimivih gradiv, ki bodo na voljo tudi ostalim učiteljem, ki

bi radi nadgradili in popestrili ure svojega pouka.



2 REFLEKSIJA IN POVRATNA na naslovu napoj@404.si. Vabimo vse, da se projektu pridružijo. Vse informacije dobite

INFORMACIJA

Izvedba projekta je potekala v skladu s pričakovanji. Dijaki so VIRI

med izdelovanjem interaktivnega plakata utrdili znanje [1] Spletno orodje Scratch. (b.d.). (pridobljeno dne 28. 8. 2024) Pridobljeno s

programerskih konceptov, spoznali uporabo izobraževalnega https://scratch.mit.edu

[2] Izobraževalni komplet Makey Makey. (b.d.). (pridobljeno dne 27. 8. 2024)

kompleta Makey in Makey in hkrati preverili predznanje iz Pridobljeno s https://makeymakey.com

geometrije. Poleg tega so izboljšali veščine raziskovalnega dela [3] Projekt MINUT NAPOJ. (b.d.). (pridobljeno dne 27. 8. 2024) Pridobljeno

in s https://novi.napoj.si/aktivnosti-2/projekti/minut-ii medpredmetnega povezovanja. Med izdelovanjem

interaktivnega plakata naleteli tudi na nekaj težav, ki so jih

uspešno premostili. Zagotoviti so morali tudi pravilnost fizičnih

povezav med kompletom Makey Makey in plakatom. V

programskem orodju Scratch so morali povezave uskladiti z

ustreznimi ukazi, ki so se izvedli ob pritisku na ustrezen del

plakata, ki smo ga označili z železnim razcepkom. Z vsem

navedenim so zagotovili nemoteno delovanje programa. Da ne bi

uporabnik nenamerno odklopil povezave, so morali poskrbeti

tudi za to, da so bile vse fizične povezave skrite in varno pritrjene.





Izdelava interaktivnih tematskih listov s spletnim orodjem




Canva



Creating Interactive Themed Sheets with Canva



Marjeta Kuhar

OŠ Komenda Moste

Glavarjeva cesta 37

Komenda, Slovenija

kuhar.marjeta@oskomenda-moste.si



effective knowledge consolidation and the development of skills

POVZETEK necessary for working with various applications.

Danes si dela z informacijsko-komunikacijsko tehnologijo v

razredu praktično ne moremo več predstavljati, saj nas le-ta KEYWORDS

spremlja na skoraj vsakem koraku. Prav je, da učitelji gremo ˝v Canva, interactive themed sheets, student motivation

korak s časom˝ in posodobimo načine poučevanja.

Pri večletnem poučevanju desetletnikov opažam, da za delo

potrebujejo vedno več zunanje motivacije in da za klasičen pouk 1 UVOD

učenci niso več motivirani, saj precej raje rešujejo interaktivne Danes ljudje uporabljamo informacijsko-komunikacijske

naloge na računalnikih in tabličnih računalnikih. tehnologije (IKT) že praktično na vsakem koraku in na vseh

V prispevku je predstavljeno, kako lahko spletno orodje področjih. Zato je seveda pomembno in hkrati neizogibno IKT

Canva uporabimo za delo v razredu. S tem orodjem sem uvesti v čim večji meri že v osnovnošolsko izobraževanje. V

pripravila dva vsebinsko različna interaktivna tematska lista, ki današnjem času je razvoj tehnologije zelo hiter, zato je nujno, da

sta v prispevku natančneje predstavljena. Učenci zelo radi otroke izobrazimo in vzgajamo tako, da bodo ta hiter razvoj

utrjujejo in pridobivajo znanje na tak način. obvladali. Če bomo učence že v osnovni šoli navajali na uporabo

Za tako pripravo na pouk je potrebno nekoliko več časa, različne IKT, se bodo le-ti kot odrasli lažje prilagajali na vse

vendar je trud v razredu poplačan, saj lahko učitelj opazuje, s novosti. [4]

kakšno vnemo učenci rešujejo naloge, s tem utrjujejo znanje in Že nekaj let poučujem petošolce in vsako leto znova opažam,

hkrati pridobivajo spretnosti in veščine za delo z različnimi da se pri klasičnem delu z učbeniki in delovnimi zvezki

aplikacijami. dolgočasijo in da neradi posegajo po njih. Čeprav so ta gradiva

veliko privlačnejša kot pred nekaj desetletji, jih ne privlačijo tako

KLJUČNE BESEDE kot vsa virtualna okolja, ki so zanimiva za oči in kar kličejo k

Canva, interaktivni tematski list, motivacija usmerjeni pozornosti. Zelo radi se poslužujejo dela na

računalnikih in tabličnih računalnikih in pri tem sploh nimajo

ABSTRACT občutka, da se učijo, ampak tako delo doživljajo bolj kot zabavo.

Today, it is almost impossible to imagine classroom work V kolikor učitelj zna pripraviti ustrezna IKT gradiva oz.

without information and communication technology, as it has interaktivne naloge, učenci z reševanjem le-teh utrjujejo znanje

become integral to nearly every aspect of education. It is essential in hkrati pridobivajo spretnosti in veščine za delo na računalnikih

for teachers to "keep up with the times" and modernize their in drugih ˝pametnih˝ napravah.

teaching methods accordingly.



Having taught ten-year-olds for many years, I have noticed 2 INFORMACIJSKO-KOMUNIKACIJSKA

that they increasingly require external motivation to engage in TEHNOLOGIJA V ŠOLI their work. Traditional classroom methods seem less effective,

as students now prefer interactive tasks on computers and tablets. Poklic učitelja je zahteven. Zahteva nenehno izobraževanje

This article demonstrates how Canva, an online tool, can be učiteljev in njihovo prilagajanje na hitre sprememb. Učenci se

effectively used in the classroom. I have created two spreminjajo, skupaj s svojimi potrebami in interesi. Učitelji smo

substantively different interactive thematic sheets using Canva, pri svojem delu sicer avtonomni, saj je naša omejitev le učni

which are discussed in detail in this paper. Students enjoy načrt in njegovi cilji. Naša naloga je, da iščemo metode in oblike

consolidating and acquiring knowledge through these methods. dela, s katerimi bodo učenci prišli do znanja in osvojili cilje

Although this type of class preparation requires additional učnega načrta.

time, the effort is worthwhile. Teachers can observe the Pri pouku se trudim vključevati različne metode in oblike dela,

enthusiasm with which students approach tasks, leading to more kar mi uporaba IKT zagotovo omogoča. Dnevno uporabljam že

izdelano interaktivno gradivo založb, včasih le preprosto Power

Point predstavitev, pogosto za učence pripravim zanimive Prav tako lahko oblikujemo skupino, npr. sodelavcev,

interaktivne kvize ali pa s spletnimi obrazci preverim njihovo učencev in z njimi delimo svoje izdelke. Deljene izdelke lahko

znanje. Kot učiteljica sem vedno v dilemi, v kolikšni meri je potem vsak poljubno uporabi, spreminja, preoblikuje, dodaja.

smiselno učence posedati za računalnike. Svoje končane projekte lahko prenesemo oz. shranimo v več

Kot o pomenu IKT pravi Rebernak, je uporaba le-te smiselna, vrstah datotek: pdf, jpg, png, mp4, gif, svg … Lahko pa

ko z njo dosežemo boljše rezultate pri učenju. [2] enostavno kopiramo spletno povezavo do izdelka in ga

Sama zelo rada posegam po tehnoloških novostih, zato se z predvajamo kar preko spleta.

veseljem izobražujem na tem področju. Prav tako rada Je orodje, ki zahteva registracijo, za katero potrebujete

raziskujem in preizkušam različna spletna orodja, s katerimi Facebook, Google ali MS račun.

učencem pouk naredim bolj nazoren, zanimiv in hkrati bolj Do nje lahko dostopate preko spletnega mesta canva.com,

motivacijsko spodbuden. Delo na področju tehnologije mi je v preko mobilne aplikacije ali aplikacije, ki jo naložite na

izziv, ki se ga vedno rada lotim in zato mi ni težko posvetiti računalnik. Deluje tako na Androidu kot iOS.

popoldnevov raziskovanju in preizkušanju novih IKT orodij. Brezplačna je le osnovna verzija, ki zadovolji le manj

Sama sem mnenja, da moramo učitelji iti ˝v korak s časom˝ zahtevne uporabnike. Neprofitne organizacije pa lahko zaprosijo

in mislim, da je nujno, da ves čas nadgrajujemo veščine za za brezplačno pro različico. Tako lahko tudi učitelji dobimo

učinkovito in predvsem varno rabo sodobne tehnologije. Učenci izobraževalno brezplačno različico, ki pa omogoča zares

se s pametnimi napravami srečajo že zelo zgodaj, vendar opažam, ogromno različnih predlog, clipartov, ozadij, pisav … Učitelji

da dela z njimi niso vešči, razen igranja igric. Rada jim pokažem, morajo le skrbnikom poslati dokazilo o delu v šoli (npr. potrdilo

na kakšen način te naprave še lahko uporabljajo in ob tem o diplomiranju, plačilno listo).

pridobivajo še nova znanja in veščine za delo z različnimi Zelo rada v Canvi izdelam tudi predstavitev za roditeljske

aplikacijami. sestanke. Pri sami predstavitvi lahko do njega dostopam tako, da

V prispevku bom predstavila, kako sem popestrila uro pouka ga izvozim tudi kot power pointovo predstavitev, v kolikor pa ne

z uporabo IKT. V Canvi sem izdelala dva interaktivna tematska želim zasedati prostora na računalniku, pa lahko do njega

lista za učence. Pri prvem sem oblikovala interaktivno igralno dostopam preko spletne povezave. Take predstavitve so izredno

polje, preko katerega so učenci ob koncu leta ponovili snov nazorne, barvite in privlačne na oko. Vsak drsnica je lahko

celega šolskega leta. Drugi interaktivni učni list je nastal ob unikatna.

slovenskem kulturnem prazniku.

2.2 Interaktivni učni list – utrjevanje snovi

2.1 Canva Predhodno sem pripravila igralno polje in ga v Canvi

Canva je brezplačno spletno orodje za grafično oblikovanje, spremenila v interaktivni list. Vsako polje sem opremila s sličico,

ki ima na enem mestu zbrano vse, od predlog, clipartov, pisav, ki namiguje, kateri tip naloge se skriva na posameznem polju. Za

slik, ilustracij… Je eden od najbolj enostavnih in praktičnih vsakim poljem učence čaka interaktivna naloga, do katere

programov za grafično oblikovanje; v njej lahko ustvarjate dostopajo preko hiperpovezave, s klikom na sliko.

praktično vse, od predstavitev, plakatov, miselnih vzorcev, učnih

listov letakov, v njej lahko tudi snemate in montirate

videoposnetke … Canva olajša delo in z njo prihranite čas. [3]

Sama pogosto izdelam razne plakate, ki so zelo estetski in jih

učenci radi pogledajo. Samo izdelovanje plakatov je v Canvi

izredno preprosto in nezamudno.





Slika 2. Igralno polje (VIR: lasten, zajem zaslonske slike)





Na plošči imajo napisana navodila za igranje, ki si jih prej

preberejo. Običajno jim navodil ne razlagam, saj želim, da se jih

Slika 1. Plakat o deževniku (VIR: lasten, zajem zaslonske naučijo natančno in z razumevanjem prebrati.

slike) Lahko se igrajo v paru ali pa se igra vsak sam. Igro učenci

igrajo podobno kot namizne igre, le da tukaj zavrtijo virtualno

V Canvi lahko uporabimo že vnaprej pripravljene predloge

igralno kocko in se pomaknejo za ustrezno število mest po

ali projekte oblikujemo po meri.

igralnem polju. Na vsakem polju jih čaka različna naloga iz

različnega področja. Vključila sem kvize, naloge razvrščanja,

povezovanja … Večji del nalog sem pripravila sama, nekatere

med njimi pa sem našla na spletu. Nekatera polja so namenjena Učenci so bili bistveno bolj motivirani za ogled posnetkov,

počivanju, nekatera pa gibalni nalogi, kot so npr. počepi, kot v primeru, da bi jim povezave do posnetkov le pripela v

poskoki ... Mislim, da je prav, da učencem med dejavnostmi dokument. Še dodatno motivacijo jim je predstavljal podatek, da

ponudimo tudi gibanje, saj le-to pripomore k večji zbranosti, jih ob koncu ogleda čaka še zabaven kviz, ki ga bodo uspešno

koncentraciji in boljšemu pomnjenju. rešili le, če bodo posnetke pozorno poslušali. Reševanje takih

Zanimivo je bilo opazovati učence, kako so bili veseli, če so kvizov so že poznali in so vedeli, da bodo svoje točke oz.

prišli na polje z gibalno nalogo. Ko je prvi učenec prišel na to uvrstitev lahko primerjali s sošolčevimi, saj se med reševanjem

polje, me je malo začudeno pogledal in vprašal, če mora opraviti prikazuje lestvica najboljših. Kviz sem zanje pripravila s

tudi to nalogo . programu Quizizz, ki je za uporabo brezplačen.

Naslednji interaktivni list je nastal za utrjevanje snovi ob Za bolj zabaven zaključek pa sem poskrbela še z virtualno

koncu šolskega leta, saj so na njem zbrane naloge iz vseh klaviaturo, preko katere so se učenci lahko preizkusili v igranju

predmetov, ki jih poučujem. Tako se preizkusijo iz glasbene Prešernove Zdravljice. Notni zapis Premrlove skladbe sem

umetnosti, matematike, slovenščine, naravoslovja in družbe. namreč opremila s črkami in učenci so himno igrali tako, da so

Nekatere spletne naloge sem jim pripravila v orodjih, kjer je sledili črkam in ustrezno pritiskali na tipke klaviature.

povezava do njih časovno omejena. Učenci do njih lahko

dostopajo le nekaj dni. Da ne pozabim ponovno aktivirati naloge,

mi Canva omogoča, da si ob učni list napišem opomnik, ki me

opozori, na katerih poljih moram posodobiti povezavo do naloge.





Slika 5. Virtualna klaviatura (VIR: lasten, zajem zaslonske

slike)



Slika 3. Primer opomnika (VIR: lasten, zajem zaslonske

slike)



2.3 Interaktivni učni list – slovenski kulturni

praznik

Drugi interaktivni tematski list je nastal pri pouku na daljavo,

in sicer za spoznavanje Prešernovega življenja. Učence je na prvi

strani čakalo navodilo, naj poiščejo vse fige, ki jih je Prešeren

skril po Ljubljani. Ko jo najdejo, morajo nanjo pritisniti ter

pogledati in poslušati posnetek, ki se skriva za njo.





Slika 6. Pomoč za igranje na klaviaturi (VIR: lasten, zajem



zaslonske slike)





Ta naloga se jim je zdela najbolj zabavna, saj so klavir lahko

igrali tudi tisti, ki sicer nanj ne znajo igrati. Nekaj učencev je

imelo pri tej nalogi težave, saj niso znali istočasno odpreti dveh

Slika 4. Skrite fige v Ljubljani (VIR: lasten, zajem zavihkov.

zaslonske slike)



Tak način dela je bolj primeren za domače delo oz. je 5 ZAKLJUČEK



potrebno delo s slušalkami, saj posnetki vključujejo zvok, pa tudi Interaktivni učni list z igralnim poljem so letos reševali v igranje na klavir v razredu ni mogoč brez uporabe slušalk. zadnjem tednu pouka. Običajno v tem obdobju le s težavo sledijo

pouku in rešujejo naloge v delovnem zvezku. Letos so z veseljem



3 ˝vrteli˝ kocko in pritiskali na polja z nalogami ter jih z veliko DELO V RAČUNALNIŠKI UČILNICI vnemo reševali. Z zanimanjem so gledali na sosednje Učencem sem povezavo do interaktivnega lista pripela v MS računalnike in opazovali, kakšne naloge rešujejo sošolci. Teams, kamor so vključeni vsi. MS Teams je brezplačno Prav veselje jih je bilo opazovati, kako so bili motivirani za dostopen vsem učencem, ki jim šola priskrbi AAI račun. To je delo in tudi koncentracija jim ni tako hitro padla. Prav nihče se račun, s katerim lahko vsak brezplačno dostopa do orodja Oblak ni pritoževal nad delom. Po prvi uri dela so se začudili, ko je 365 in si prav tako brezplačno, namesti Office 365. Učencem že zvonilo, saj se jim je zdelo, da je ura hitro minila. septembra v MS Teams ustvarim razredno spletno učilnico, ki jo Mislim, da je prav, da učitelji učencem ponudimo delo z potem uporabljamo celotno šolsko leto. Učenci so bili zato že interaktivnim gradivom, moramo pa najti pravo ravnovesje med vešči dela v njej. uporabo IKT in ostalih metod poučevanja. Za reševanje nalog so imeli učenci na voljo dve šolski uri.



Naloge smo reševali v šolski računalniški učilnici. Posamezne ZAHVALA



samostojno reševanje, spet drugi za delo v parih. Nekaj parov je naloge, ki jim jih pripravim v interaktivni obliki. Prav tako sem vesela, ker opazijo moj trud ter pohvalijo zanimivost nalog. Tako predčasno prišlo do cilja in so pričeli spet od začetka, saj je bilo je moj trud poplačan. bolj malo verjetno, da bodo ponovno prišli na ista polja kot v pri nekaterih pa nekaj manj časa. Nekateri so se odločili za V prvi vrsti bi se zahvalila učencem, ki z veseljem rešujejo naloge so različno dolge in učenci so pri nekaterih porabili več,



prvem poskusu. Zahvala gre tudi sodelavkam, ki vedno kritično ovrednotijo moje delo ter ga tudi z veseljem uporabijo v svojih razredih ter

naši knjižničarki Andreji za lektoriranje prispevka.

4 PREDNOSTI IN SLABOSTI

Pred pripravo vsakega interaktivnega gradiva razmislim o LITERATURA



prednostih in slabostih, ki jih bo tako delo prineslo, saj je prav, [1] Kocjančič Kuhar, V. 2016. Sodelovalno učenje pri utrjevanju matematike.

Murska Sobota. Dostopno na naslovu: https://www.zrss.si/kupm2016/wp-

da učencem ponudimo tako metodo poučevanje, od katere bodo content/uploads/sodelovalno-ucenje.pdf (1. 8. 2024)

odnesli največ znanja. [2] Rebernak, B. 2008. Pomen IKT in e-gradiv pri pouku v sodobni šoli,

Vzgoja in izobraževanje XXXIX (5): 91–93. Dostopno na naslovu (1. 8.

Prednosti takih interaktivnih učnih listov je zelo veliko. 2024)

Učenci so bistveno bolj motivirani za delo, prav tako se lažje [3] Šefman M. 2024. Postopek pridobitve brezplačne Canve Pro za neprofitne

skoncentrirajo na delo in je njihova koncentracija daljša kot sicer. organizacije: primer Mestne knjižnice Ljubljana. Dostopno na naslovu:

https://knjiznicarske-novice.si/postopek-pridobitve-brezplacne-canve-

Ni jim težko celo šolsko uro reševati interaktivne naloge in celo pro-za-neprofitne-organizacije-primer-mestne-knjiznice-ljubljana/ (12. 7.

prosijo, če jih lahko rešujejo še doma. Take naloge na razredni 2024)

[4] Tišler, T., Černilec B., Vehovec M., Korošec, D., Brezovar, D. in

stopnji radi pokažejo tudi staršem. Pungartnik, S. 2006. Vodenje za spodbujanje informacijsko-



informacije o svojem delu oz. znanju, saj jim programi, s Učenci pri takem delu običajno dobijo takojšne povratne komunikacijske tehnologije na šolah. Ljubljana. Dostopno na naslovu: https://www.solazaravnatelje.si/ISBN/961-6637-04-5.pdf (22. 7. 2024)

katerimi sem izdelala interaktivne naloge, dajo takojšnje

informacije o pravilnosti rešitev. Prav se mi zdi, da učenci dobijo

takojšnje informacije o svojem znanju oz. rešitvah, saj se le tako

zavedajo pomanjkljivosti svojega znanja in napake lahko

popravijo.

Če načrtujem delo v paru oz. skupinsko delo, pri učencih na

tak način spodbujam sodelovalno učenje. To je vzgojno-

izobraževalna strategija, pri kateri delajo udeleženci v majhnih

skupinah. Poudarek je na delo v heterogenih skupinah.

Skupinsko delo temelji na njihovi medsebojni soodvisnosti in

odgovornosti. Pomembno je, da vsak član skupine sodeluje in da

vsak doda svoj prispevek k skupni nalogi. Pri takem učenju se

sodelujoči neposredno učijo tudi socialnih vešči. [1]

Pri takem delu nam ni potrebno tiskati učnih gradiv in s tem

delujemo tudi ekološko, saj ne trošimo papirja.

Slabost takega načrtovanja pouka je zagotovo možnost, da

tehnologija zataji, zato je prav, da naredimo rezervni plan.

Pri nekaterih vrstah interaktivnih nalog učitelj ne more

preveriti, kako uspešni so bili učenci pri reševanju.

Prav tako je priprava na tak pouk za učitelja zahtevnejša in

učitelj porabi več časa za pripravo na pouk.



Do rešitve avtentičnih problemov z matematiko in



informatiko



To the Solution of Authentic Problems with Mathematics and



Informatics



Nastja Lasič Andreja Kramar

Gimnazija Šentvid Gimnazija Šentvid

Prušnikova 98, Ljubljana, Slovenija Prušnikova 98, Ljubljana, Slovenija

nastja.lasic@sentvid.org andreja.kramar@sentvid.org



povezovanju. Pri matematiki so našteti tudi konkretni predlogi, a

POVZETEK je izvedba neobvezna in prepuščena učitelju. Posledično dijaki

V članku je predstavljena medpredmetna povezava bolj ali manj uspešno pridobivajo znanja in veščine ločeno po

matematike in informatike z namenom, da dijaki ob reševanju učnih predmetih in jih slabo prenašajo iz ene učne situacije v

avtentičnih nalog prenesejo znanja iz enega učnega področja na drugo. Ravno prenosljivost znanja pa je ključna pot k

drugega in obratno. Zanimalo nas je tudi, če taka oblika dela poglobljenemu razumevanju.

poveča zanimanje za računalništvo oziroma matematiko. Naloge Avtorji članka smo želeli z načrtovano aktivnostjo omogočiti,

so bile namenjene dijakom prvega letnika gimnazije in sodijo v da dijaki v avtentični situaciji prenesejo znanje iz matematike v

poglavje o naravnih številih. Dijaki so opazovali vzorce, informatiko in obratno. Naloge, ki smo jih reševali, je sicer

utemeljevali svoje ugotovitve, zapisovali korake algoritmov in možno reševati le pri matematiki oz. le pri informatiki, a bi bilo

programirali z delčki. Pred in po aktivnosti so izpolnili anketo, ki škoda izpustiti priložnost za povezovanje. Družba se digitalizira

meri zastavljene cilje. V članku so predstavljeni tudi rezultati na vseh področjih in pričakovati je, da bo računalniško mišljenje

ankete. postalo del vsakega področja znanosti in večjega dela bodočih

poklicev [3]. Zato je spajanje računalniškega mišljenja z



KLJUČNE BESEDE matematiko nujno, hkrati pa je matematika na srednješolskem nivoju bazen vsebin za učenje informatike. Žal pogrešamo več Vzorci, računalniško razmišljanje, psevdo koda, algoritem, primerov gradiv, kjer bi bilo to spajanje računalniškega in programiranje z delčki matematičnega mišljenja izvedeno na smiseln način. Projekt



ABSTRACT izvedbo tovrstne povezave. NAPOJ MINUT [5] pa je ponudil pravo okolje in motivacijo za

The article presents the interdisciplinary connection between Zanimalo nas je tudi, če tak način dela pri dijakih poveča

mathematics and computer science, with the aim of encouraging zanimanje za matematiko oziroma informatiko, posledično pa

students to transfer knowledge from one subject area to the other tudi za predmete MINUT - M(atematika) I(nformatika)

and vice versa while solving authentic tasks. We were also N(aravoslovje) U(metnost) T(ehnologija). interested in whether this form of work increases interest in

computer science or mathematics. The tasks were intended for

first-year high school students and are related to the topic of 2 OPIS PROBLEMA IN NAČRTOVANA

natural numbers. The students observed patterns, justified their AKTIVNOST findings, recorded algorithm steps, and programmed using

Za dijake smo pripravili nalogo iz teme večkratniki in delitelji

blocks. Before and after the activity, they completed a survey

z dvema razširitvama, ki odpirata številna dodatna vprašanja za

designed to measure the set goals. The article also presents the

dijaka.

survey results.

Osnovna naloga: Hana na poti do šole preskakuje ploščice.



KEYWORDS Vsakič jih preskoči enako število. Pot je podana v obliki spirale

(Slika1). Zanima nas, na katere je stopila in koliko ploščic ostane

Patterns, computational thinking, pseudo code, algorithem, ob koncu poti. Zato bi radi sestavili program, ki pobarva ploščice,

block-based programming na katere stopi Hana. Radi bi, da dijak ob nalogi razmišlja tudi o

naslednjih vprašanjih: koliko je vseh ploščic, kako zapišemo



1 zaporedne večkratnike, kako zapišemo števila med sosednjima UVOD

večkratnikoma, kaj predstavlja zadnje število,

Slovenski izobraževalni sistem temelji na pridobivanju znanj Druga naloga je izpeljanka prve. Hana gre večkrat po poti do

in veščin po posameznih področjih in v naprej definiranem šole in preskakuje ploščice. Na posamezni poti vedno preskakuje

obsegu. Prenovljeni učni načrti [1] vključujejo tudi didaktična enako ploščic, a število teh se med posameznimi potmi lahko

priporočila o uporabi učne tehnologije in medpredmetnem

razlikuje. Dijak naj pri barvanju posamezne poti uporabi različne 3.1 Izvedba v razredu

barve in pojasni značilnost ploščic, na katere Hana stopi večkrat. Avtorici članka sva aktivnosti izvedli dvakrat, vsakokrat s

približno 14 dijaki prvega letnika gimnazije. Večino časa sva bili



v učilnici prisotni obe in s številnimi vprašanji in namigi

usmerjali dijake, ki so delali v parih. Eno uro smo bili v običajni

učilnici, ostale ure pa v računalnici.

Dijaki so naloge reševali zavzeto, a so bile razlike v

uspešnosti kar velike. Prve težave so nastopile pri besednem

opisu, kako se Hana premika in kako se barvajo ploščice.

Nekateri so mislili, da Hana preskakuje k ploščic, drugi pa, da bo

stopila na vsako k-to ploščico. Vmesni del besednega opisa

(psevdo kode), kjer se nekaj ponavlja, so dijaki opisali dobro,

težje pa je bilo zapisati začetni in zaključni del poti. Približno

polovica dijakov je bila v programiranju z delčki uspešna, našli

so različne rešitve, ki smo jih prikazali na tabli in pokomentirali.

V nadaljevanje sta dve od rešitev (Slika 2 in Slika 3).

Slika 1: Pot od doma do šole, naloga v okolju Pišek





Tretja naloga je izpeljana iz druge: Pot do šole je tokrat

prikazana z ravnim trakom, na ploščicah pa je na začetku oznaka

0. Če Hana stopi na ploščico, se 0 spremeni v 1 in obratno,

oznaka na ploščici 1 se spremeni v 0. Dijak mora ugotoviti,

kakšni vzorčki nastanejo, če gre Hana večkrat do šole in pri tem

preskakuje vsakokrat drugačno število ploščic. Zanima nas tudi,

katera oznaka je ob koncu zapisana na k-ti ploščici.



2.1 Predznanje

Za uspešno reševanje morajo dijaki poznati pojme večkratnik,

delitelj in ostanek pri deljenju, ter zvezo med njimi. Poznajo naj

tudi osnove iz programiranja: zaporedje ukazov, spremenljivka,

vejitev, zanka. Prav tako je dobrodošlo, če se že prej seznanijo s

programiranjem v okolju Pišek, saj sta tam v ta namen bili

ustvarjeni dve nalogi [4].



2.2 Načrt za izvedbo

Aktivnost je zasnovana za izvedbo v treh etapah, skupaj

okvirno 5 šolskih ur. Prvi dve uri dijakom predstavimo delo, jih Slika 2: Ena izmed rešitev v Pišku

seznanimo z osnovno nalogo, ki jo rešijo na učnem listu in



zapišejo ugotovitve. Nato z besedami zapišejo postopek barvanja

ploščic (zapis psevdo kode) in se potem lotijo iste naloge v Pišku.

Na koncu svojo rešitev tudi testirajo. Delajo lahko samostojno ali

v parih.

Naslednja ura je namenjena reševanju druge naloge v Pišku,

ko gre Hana dvakrat do šole. Dijak s pridobljenimi izkušnjami

sestavi programček za barvanje ploščic in ga testira.

Zadnji dve uri sta namenjeni prikazu različnih rešitev in

pogovoru, potem pa sledi reševanje tretje naloge. Dijakom

pojasnimo, da je pot do šole tokrat predstavljena drugače in da

bomo namesto barv uporabljali ničle in enke. Dijaki dobijo učni

list z navodili in vprašanji. Na koncu dijaki poročajo o

ugotovitvah in jih argumentirajo. Celotno aktivnost na koncu

povzame še učitelj.



3 REZULTATI

V tem poglavju bomo na kratko predstavili, kako je potekala

Slika 3: Še ena rešitev v Pišku

izvedba v razredu in primerjali odgovore anket, ki smo ju izvedli

pred in po izvedeni aktivnosti z vsemi dijaki.

Z vsako skupino smo naslednjič reševali drugo nalogo in Ugotovili smo, da se je zanimanje za oba predmeta povečalo,

dijaki so bili precej bolj samostojni, več jih je prišlo na cilj. čeprav se jim oba še vedno zdita težka. Eno od vprašanj v anketi

Tretji del smo izvedli čez en teden skupaj z obema skupinama je spraševalo o odnosu do MINUT predmetov. V zaključni anketi

in imeli ob začetku kar nekaj težav s pojasnjevanjem pravila o je 12% dijakov zapisalo, da je njihov odnos do teh predmetov

zapisu z 0 in 1. Dijaki so dobili učni list s potjo dolžine 35, ki jo bolj pozitiven kot pred to aktivnostjo, kar je obetajoč rezultat.

Hana prehodi štirikrat in vsakokrat preskakuje drugačno število Vprašali smo jih tudi, če si v prihodnje še želijo takega

ploščic. Dijaki so ob vsakem potovanju zapisali spremembe na medpredmetnega povezovanja. V tabeli 3 so njihovi odgovori v

ploščicah in prišli do končnega zaporedja (slika 4). Mnogi so odstotkih, na sliki 5 pa še grafična primerjava istih odgovorov.

imeli težave, zaradi nezbranosti so se precej motili in na koncu

seveda zaporedja niso mogli komentirati.

Tabela 3: Ali si takih aktivnosti (še) želiš?





Prej (%) Potem (%)

Da, čim večkrat. 4 13

Da, občasno. 64 58

Raje ne. 32 21

Slika 4: Štirikrat na poti do šole, izdelek dijaka Nikakor ne. 0 8



Polovica dijakov je zapisala pravilno zaporedje, le peščica pa

je znala tudi napovedati in pojasniti, katero število se pojavi na

k-tem mestu. Nihče od dijakov se te naloge ni lotil z Da, čim večkrat.

računalnikom, morda zaradi premalo časa. Da, občasno.

Raje ne.

3.2 Rezultati ankete

Nikakor ne.

Ob začetku in koncu smo dijake anketirali in jih spraševali po

zanimivosti predmeta (matematika, informatika), reševali pa so 0 20 40 60 80

tudi naloge, ki preverjajo razumevanje navodil in zapisov za

premikanje od začetne do končne točke v mreži in iskanje ter prej potem

štetje poti v tej mreži. V začetni anketi je sodelovalo 29 dijakov,



naloge rešili pravilno. Slika 5: Primerjava odgovorov začetne in končne ankete v končni pa 26. V tabeli 1 so zabeleženi odstotki dijakov, ki so



Iz odgovorov lahko sklepamo, da se je zanimanje za tako

Tabela 1: Uspešnost dijakov pri nalogah v anketi povezovanje razslojilo. Povečalo se je število dijakov, ki si to še

želijo, prav tako pa se je povečalo število dijakov, ki tega v

Tip naloge prihodnje nikakor ne želijo. Izvedena medpredmetna povezava Prej (%) Potem (%)

Preštevanje poti jim je nekoliko odprla oči in so ugotovili, kaj jih veseli. 67 58

Zapis poti Zaključna misel enega od dijakov: »Med predmeti MINUT 38 57

imam rada matematiko in naravoslovje, kar se ni spremenilo.

Ugotovili smo, da so dijaki močno napredovali v Čeprav mi programiranje ni najbolj pri srcu, se je moje mnenje

razumevanju zapisov poti. Domnevamo, da je prišlo do rahlo izboljšalo. Če se taki predmeti povezujejo z mojimi

učinkovitega povezovanja (prenosa) znanj iz matematike in najljubšimi predmeti, jih imam malo raje in zanje pokažem večje

informatike, saj je bila naša aktivnost povezana s premikanjem v zanimanje. Izkušnja je bila pozitivna.«

mreži. V tabeli 2 je primerjava odgovorov na vprašanja o



matematiki (MAT) in informatiki (RAČ) kot učnih predmetih. 4 ZAKLJUČEK Naveden je odstotek dijakov, ki se strinjajo ali delno strinjajo z

zapisano trditvijo. Ostali dve možnosti sta bili ‘se ne strinjam’ in Reševanje avtentičnih nalog s povezovanjem dveh ali več

‘se ne znam odločiti’. učnih področij dijakom omogoča, da prepletejo znanja in jih

poglobijo. S tem se brišejo meje med učnimi predmeti, in dijaki

lahko iščejo rešitve na širšem območju, to omogoči globlji

Tabela 2: Strinjanje dijakov s trditvami v anketi vpogled. Tudi delo v parih se je izkazalo za dobrodošlo. Dijaki

so bili manj prestrašeni in bolj svobodno iskali rešitve, se

Trditev Prej (%) Potem (%) dopolnjevali, prepričevali, argumentirali in skupaj prišli do

RAČ je težko. 93 96 boljših rešitev. Avtorici članka meniva, da se pri aktivnosti niso

RAČ me veseli. 46 52 dolgočasili, ankete pa so potrdile, da se je okrepilo zanimanje za

Rad bi več RAČ. 25 32 predmete MINUT.

MAT je težka. 89 72

MAT me veseli. 74 80 Ta članek je nastal v okviru projekta MINUT NAPOJ I [5].

Rad bi več MAT. 37 52





REFERENCE


[1] Novi učni načrti: https://dun.zrss.augmentech.si/#/

[2] Weintrop, D., Beheshti, E., Horn, M. et al. Defining Computational

Thinking for Mathematics and Science Classrooms. J Sci Educ Technol

25, 127–147 (2016). https://doi.org/10.1007/s10956-015-9581-5

[3] Portal Pisek (pisek.acm.si) Naloga Hana hodi v šolo

(https://pisek.acm.si/contents/4907-319805995281415931-

146393949319149838-680361037148430359-1872704963165371564/ )

[4] Skupnost učiteljev Napoj https://napoj.si/ Projekt MINUT NAPOJ I



Uporaba digitalnih kompetenc v vzgojno izobraževalnem



zavodu



Use of Digital Competencies in Educational Institutions



Stanislava Letonja

OŠ Dušana Flisa Hoče, Hoče

Slovenia

stanislava.letonja@os-hoce.si



Učenci morajo pridobiti veščine za učenje problemskih ter

POVZETEK procesnih znanj tudi z uporabo digitalne tehnologije. Najprej jih

V zadnjih letih se je potrebno hitro in učinkovito prilagoditi je potrebno seznaniti z novimi oz. drugačnimi pristopi za uspešno

na nove izzive, ki nam jih narekuje življenje. Učitelje in učence reševanje problemov iz realnega sveta. Ob tem nas ne sme biti

je potrebno usposobiti, opolnomočiti z znanji, ki so v sodobnem strah uporabljati UI oz. ChatGPT. Pri reševanju problemov se

času še kako pomembna. V učni proces vključujemo učenci učijo povezovati znanje znotraj matematike in tudi širše

raziskovanje, opazovanje ter sklepanje. Za sodobnejše načine (interdisciplinarno). Naučijo se postavljati ključna raziskovalna

poučevanja, je potrebno pridobiti tudi ustrezna znanja, ki jih vprašanja, ki izhajajo iz življenjskih situacij in poskusiti z

lahko nato suvereno predajamo naprej. Na različnih usvojenim znanjem rešiti zastavljene naloge.

usposabljanjih, ki smo se jih udeležili, smo usvajali nova znanja V okviru medpredmetnega povezovanja naj učenci

za vpeljavo različnih oblik digitalnega izobraževanja. Dosedanje uporabljajo informacijsko-komunikacijsko tehnologijo (UN -

pridobljene ter usvojene digitalne kompetence smo še nadgradili. matematika, 2011).

Pri posameznih poglavjih v učnem procesu pri pouku Na strokovnih srečanjih ter usposabljanjih smo učitelji

matematike poskušamo prenašati digitalne kompetence tudi med pridobili znanja, le-ta pa lahko z novimi izzivi ter idejami

učence. Z novimi, drugačnimi načini poučevanja lahko v učnem koristno uporabimo v praksi.

okolju dosežemo večjo motiviranost in zanimanje za nekoliko



drugačne izzive. 2 PRIMER DOBRE PRAKSE



KLJUČNE BESEDE Aktivnosti, ki se navezujejo na učno vsebino o večkotnikih,

Digitalne kompetence, načini poučevanja smo preizkusili z uporabo programskega jezika Python in

Scratcha. Predstavljeno aktivnost smo vključili kot nadgradnjo



ABSTRACT osvojenih znanj pri matematiki. Digitalne kompetence, ki jih pri



In recent years, we have had to adapt quickly and effectively matematiko na nekoliko drugačen način. Spoznali so lahko, da tem predstavljamo ter razvijamo, učencem omogoča spoznavati

to the new challenges that life dictates to us. Teachers and se matematika ter programiranje medsebojno prepletata. Brez

students need to be trained and empowered with skills that are znanja matematike tudi ni uspeha pri programiranju ter obratno.

still very important in modern times. The learning process Na koncu poglavja večkotnikov smo se odločili, da poskusimo

involves research, observation and reasoning. For more modern osvojene cilje utrditi ter preveriti na drugačen način. Predhodno

methods of teaching, it is also necessary to acquire relevant je bilo potrebno ponoviti o kotih v večkotnikih, velikosti enega

knowledge, which can be confidently passed on. In the varios notranjega kota v pravilnem n-kotniku, o številu diagonal iz

trainings we attended, we acquired new skills for the introduction enega oglišča, o številu vseh diagonal,... Seznanili smo jih s



skills so far. In the individual chapters of the learning process in posameznih primerih so se spoznali Python ter ob tem razvijali the teaching of mathematics, we try to transfer digital digitalne kompetence. Za učence smo pripravili gradivo, ki so ga of various forms of digital education. We have developed digital programskim jezikom Python, saj ga niso poznali. Ob



teaching, we can achieve greater motivation and interest in zapisali napačno codo v Pythonu. Želeli smo, da učenci sami slightly differnet challenges in the learninig environment. ugotovijo napako in jo poskušajo odpraviti oz. popraviti. Vedno competences among the students. With new, different ways of reševali (slika 1). Namerno smo pri nekaterih nalogah v gradivu



KEY WORDS in poiskati je potrebo napako ter pot, da zadevo rešimo. Prav v življenju ne gre kot po maslu, vmes se kdaj pa kdaj tudi zatakne

Digital skills, teaching methods veseli smo bili njihovega razmišljanja, saj znajo usvojeno znanje

med seboj povezovati (matematiko in programiranje).



1 UVOD





Slika 1. Primer naloge na učnem listu- Python.

Slika 3. Primer naloge na učnem listu- Scratch

Najprej so pripravljeno gradivo reševali na list s pomočjo

svinčnika, nato pa še z uporabo IKT tehnologije. Tako so učenci

lahko spoznali, kako nam računalnik lahko pomaga pri 3 ZAKLJUČEK

vsakdanjih nekoliko drugačnih težjih nalogah. V programskem Učence bomo seznanili z možnostmi reševanja problemov z

jeziku Python smo zapisali codo, po kateri je potem računalnik uporabo informacijsko-komunikacijske tehnologije. Na ta način

hitro izpisal rešitev. Vstavljali smo podatke za različne n-kotnike bodo pridobili tudi digitalne kompetence. Poskusili jih bomo

in računalniška tehnologija nam je v nekaj sekundah podala oz. naučiti spopasti se z izzivi, problemi na primerih iz vsakdanjega

izračunala velikost enega notranjega kota, vsoto vseh kotov, življenja. Ob predstavljenih primerih so spoznali, da digitalna

število diagonal, … tehnologija lahko tudi pripomore k hitrejšemu učenju in

V drugem delu smo se lotili še načrtovanja pravilnih usvajanju zastavljenih ciljev na drugačen način. V naslednjih

večkotnikov. Učenci so na list s pomočjo svinčnika, ravnila in letih bomo poskusili pri delu v razredu ponuditi tudi tablice, tako

šestila narisali pravilne večkotnike na dva načina ( podana da bi izdelki lahko nastajali in bi se znanje postopoma

dolžina stranice in izračunan pripadajoči notranji kot, polmer nadgrajevalo. Ugotovili smo, da vsak učenec lahko osvaja cilje v

očrtane krožnice in izračunana velikost središčnega kota) (slika svojem tempu. Učenci bi svoje osvojene digitalne veščine lahko

2). dopolnjevali ter preiskovali tudi druge načine. Zavedati se

moramo, da so takšne možnosti pridobivanja znanj izvedljive



tudi pri mlajših učencih. Na usposabljanjih smo bili mnenja, da



bi morali učencem omogočiti, da bi si pridobili ter razvijali

digitalne kompetence. V osnovni šoli bi naj vsi učenci opravili

plavalni tečaj in kolesarski izpit. Naša prizadevanja bodo

obrodila rezultate, v kolikor bomo na vseh stopnjah VIZ dela in

v različnih predmetnih področjih poskušali vnašati drugačne

oblike ter metode dela za ustrezno pridobivanje ter izboljšanje

usvojenih digitalnih kompetenc.



Slika 2. Primer naloge na učnem listu. VIRI



Rešene naloge, ki so jih predhodno narisali, smo potem [1] Gradivo s seminarja: S programiranjem v matematiko, Ljubljana, 2024

[2] https://www.gov.si/podrocja/izobrazevanje-znanost-in-

preverili še z uporabo Scratcha (Slika 3). Nekateri učenci so ga sport/osnovnosolsko-izobrazevanje/, 19. 7. 2024



že poznali, nekateri pa ne. Ponovno smo še enkrat predstavili ter [3] https://www.zrss.si/digitalna-bralnica/smernice-za-uporabo-digitalne- tehnologije/ , 19. 7. 2024 pojasnili, kako zapisati ustrezen program, ki nam pravilen

večkotnik tudi izriše. Ob tem so spoznavali različne ukaze, zanke Slike: lasten arhiv

in povezali znanje matematike ter programiranja. Računalniška

tehnologija nam omogoča, da zapisan program v Scratchu tudi

takoj preverimo, če nam izriše to, kar bi si želeli. Morebitne

napake sproti dopolnjujemo, popravljamo in ponovno preverimo

delovanje. Podana nam je takojšnja povratna informacija.



Učna analitika med visokošolskimi učitelji:



preliminarni rezultati



Learning Analytics among Higher Education Teachers:



Preliminary Findings



Marjeta Marolt Danijela Jakšić

Anja Žnidaršič Vanja Slavuj

Univerza v Mariboru, Fakulteta za Sveučilište u Rijeci, Fakultet informatike i

organizacijske vede digitalnih tehnologija

Kranj, Slovenia Rijeka, Croatia

marjeta.marolt@um.si danijela.jaksic@inf.uniri.hr

anja.znidarsic@um.si vslavuj@uniri.hr



POVZETEK

Uporaba učne analitike ima velik potencial za doseganje 1 UVOD



boljših študijskih dosežkov, spodbujanje nadaljevanja študija ter Dostop in uporaba digitalne tehnologije pušča velike

izboljšanje prakse učenja in poučevanja, vendar je ta potencial še podatkovne sledi. Trženjska industrija izkorišča podatke za

vedno dokaj neizkoriščen. Medtem ko je objavljenih veliko napovedovanje zanimanja kupcev za prihodnje izdelke, Netflix

raziskav o razvoju vizualizacij, nadzornih plošč in napovednih predlaga filme glede na naše priljubljene žanre, Amazon pa

modelov, pa je bilo le malo raziskav o dejanski uporabi učne poskuša napovedati vrste knjig, ki bi si jih lahko želeli. Potencial

analitike med visokošolskimi učitelji, o dejavnikih, ki vplivajo podatkovne analitike je pritegnil akademsko sfero, da izkorišča

na njeno uporabo in prednostih, ki jih pri uporabi zaznavajo. Ta podate in se poslužuje podobnih pristopov. Študije, ki so

kvantitativna študija predstavlja preliminarne rezultate 40 proučevale pristope učne analitike v visokem šolstvu, kažejo

visokošolskih učiteljev na Univerzi v Mariboru in Univerzi v pozitivne rezultate pri ugotavljanju potreb študentov,

Rijeki. izboljševanju učnih rezultatov in podpore nadaljevanju študija

[1], [2], [3], [4]. Z uporabo učne analitike na ravni predmeta

KLJUČNE BESEDE lahko visokošolski učitelji in sodelavci pridobijo vpogled v učno

Učna analitika, univerza, visokošolski učitelji in sodelavci, vedenje študentov na podlagi vzorcev, ki jih najdejo v podatkih,

uporaba, dejavniki, prednosti in se tako lažje odločajo o potrebnih ukrepih. Zdi se, da je stališče,

da lahko učna analitika izboljša izobraževanje, splošno sprejeto.



ABSTRACT Izraz učna analitika uporabljamo v skladu s splošno sprejeto

opredelitvijo, ki jo je na 1. mednarodni konferenci o učni

The use of learning analytics has great potential to improve analitiki in znanju podalo Društvo za raziskovanje učne analitike

student achievement, encourage further study and improve (SoLAR) in na katero se sklicujeta Long in Siemens [5, p. 34]:

learning and teaching practice, but this potential is still largely "Učna analitika je merjenje, zbiranje, analiza in poročanje o

unexploited. While much research has been published on the podatkih o učencih in njihovih kontekstih za namene

development of visualisations, dashboards and predictive models, razumevanja in optimizacije učenja in okolja, v katerem poteka."



analytics among higher education teachers, the factors that deskriptivna in diagnostična učna analitika, ki se osredotoča na influence its use and the perceived benefits of using it. This analizo preteklih podatkov, druga pa napovedna in predpisovalna there has been little research on the actual use of learning V grobem ločimo dve podskupini učne analitike. Prva je



education teachers at the University of Maribor and the pomaga pri odločanju o potrebnih ukrepih [6]. University of Rijeka. Kljub temu, da nekatere univerze, npr. The Open University quantitative study presents preliminary results from 40 higher učna analitika, ki napoveduje vedenje in/ali rezultate učencev ter



KEYWORDS ne izkoriščajo popolnoma [7]. Kritiki menijo, da je učna analitika UK, veliko vlagajo v učno analitiko, njenega potenciala še vedno

Learning analytics, university, higher education teachers and le orodje, ki pomaga pri nenehni digitalizaciji visokega šolstva in

associates, use, factors, benefits da je pri vpeljavi učne analitike potrebno upoštevati človeški

faktor, saj se rezultati pojavijo zaradi interakcije orodij z

različnimi deležniki, kot so študenti, učitelji in managerji [7], [8].

Ta študija se osredotoča na visokošolske učitelje in sodelavce in

njihovo uporabo učne analitike. S pomočjo kvantitativne



raziskave želimo ugotoviti, kako visokošolski učitelji in usmerjeno analizo z uporabo z zaznavanjem vzorcev delovanja

sodelavci uporabljajo učno analitiko, s katerimi izzivi se pri tem in analizo zaporedja ter (3) analizo vsebine z uporabo rudarjenja

soočajo in katere prednosti uporabe zaznavajo. V prispevku so besedila ali drugih tehnik za analizo artefaktov, ki jih je ustvaril

predstavljeni preliminarni rezultati. študent. Področja prekrivanja med pristopi so še posebej

pomembna za nove sinergije uporabe. Ker take analize zahtevajo

naprednejša analitična znanja, se v praksi za uporabnike razvijejo

2 PREGLED PODROČJA nadzorne plošče z omenjenimi analizami, ki jim pomagajo

V preddigitalni dobi se podatki v izobraževanju niso veliko prepoznati specifične učne težave. Pri tem je potrebno

uporabljali, tehnologija učnih sistemov pa se je razvijala na izpostaviti, da taka praksa ni pogosta.

behavioristični osnovi. Čeprav se morda zdi, da behavioristična

filozofija ni povezana z učno analitiko, je Rodriguez [9] poudaril,

da obsežen model spletnega izobraževanja - MOOC, kot so

Udacity, EdX in Coursera, temelji na behaviorističnem pristopu,

saj prevzema tradicionalne učne metode, kot sta učenje z Vsebinsko usmerjena analiza Procesno usmerjena analiza

obvladovanjem in utrjevanje konceptov z interaktivnimi vajami. Na podlagi artefaktov, ki Na podlagi dnevnikov

Uporaba vidnih in merljivih učnih ciljev zagotavlja, da je jih ustvarijo študenti dejavnosti.

(npr. besedila). Analiza zaporedja

mogoče določene ravni učenja izmeriti in celo oceniti s strojem, Uporabljajo se tehnike (npr. procesno

kot je računalnik [10]. rudarjenja besedil. rudarjenje).

Pedagoške raziskave razlikujejo ocenjevanje in vrednotenje

predmetov. Formativno vrednotenje se uporablja za izboljšanje

učnega izdelka, ko je ta še v fazi razvoja, medtem ko se

sumativno vrednotenje uporablja za oceno učinkovitosti končne

različice izdelka. Najprej so se zbirali le kvantitativni podatki, Analiza socialnih omrežij

kasneje tudi kvalitativni in sicer se je preučevalo vedenje Na podlagi socialnih odnosov ali odnosov

med akterji in artefakti.

učencev iz opazovanj, uporabljaje so se rubrike ali kontrolni Meri se centralnost in kohezivnost

seznami. Ocenjevanje in vrednotenje predmetov imata danes še skupin.

vedno ključno vlogo v izobraževanju. Raziskovalci iščejo načine

uporabe vseh vrst povezanih podatkov za izboljšanje poučevanja

in učenja.

Čeprav se je digitalne tehnologija v izobraževanju začela

uporabljati že v začetku prejšnjega stoletja, se je uporaba Slika 1: Trojica metodoloških pristopov v učni analitiki po

spletnega izobraževanja razširila šele v devetdesetih letih [12]

prejšnjega stoletja [10]. Pojav sistemov za upravljanje učenja je

zagotovil platforme z veliko različnimi kanali in delovnimi

prostori, preko katerih lahko udeleženci komunicirajo in 3 METODOLOŠKI PRISTOP

izmenjujejo informacije v okviru predmeta. Učitelji lahko delijo Ker smo želeli raziskati dejansko uporabo učne analitike med

svoja predavanja, pripravljajo naloge in izpite ter vključujejo visokošolskimi učitelji, dejavnike, ki vplivajo na njeno uporabo

študente v spletne razprave, ki omogočajo sodelovalno učenje in prednosti, ki jih pri uporabi zaznavajo. Na podlagi pregleda

itd. Razmah spletnega učenja, zlasti uporaba sistemov za literature smo ugotovili, da nekatere obstoječe raziskave

upravljanje učenja, omogoča priročen način zbiranja in beleženja uporabljajo različne teoretične modele, kot so Model

velikih količin podatkov o učenju študentov. Podatki, ki jih sprejemanja tehnologije (TAM) (Davis, 1989) in Univerzalna

ustvarja sistem, vključujejo sledove interakcije učencev s spletno teorija sprejemanja in uporabe tehnologije (UTAUT) [13] ter

učno platformo. Te sledi podatkov povezujejo akterje (predvsem modeli sprejemanja, razviti za področje izobraževanja, kot je

študente in visokošolske učitelje in sodelavce) in vsebino Model sprejemanja, ki temelji na pomislekih (CBAM). Pri

predmeta (npr. videoposnetke, spletne strani, kvize, spletne razvoju vprašalnika smo izhajali UTAUT, ker se je izkazal, da

razprave) z interakcijskim vedenjem (npr. pregledovanje, najbolje pokriva dejavnike, ki so bili najpogosteje omenjeni v

odgovarjanje, klikanje). Beleženje postane samodejen proces, relevantni literaturi. Model smo prilagodili in razširili tako, da

kar je nujen pogoj za razvoj učne analitike. smo lahko bolj natančno merili uporabo učne analitike in

Čeprav univerze že nekaj let v veliki meri uporabljajo sisteme prednosti njene uporabe. Uporabili smo 5-stopenjsko lestvico

za upravljanje učenja, ustvarjanje in posredovanje vsebine Likertovega tipa. Vprašalnik smo pripravili v 1ki. Spremno

študentom še vedno v veliki meri temelji na intuiciji visokošolski pismo s povezavo smo konec junija poslali prodekanom za

učiteljev in sodelavcev. Ker sistemi za upravljanje učenja izobraževalno dejavnost vseh fakultet na Univerzi v Mariboru in

večinoma ponujajo je osnovne analize, je potrebna dodatna Univerzi v Rijeki, s prošnjo, da naše elektronsko sporočilo

analiza podatkov, da bi v celoti izkoristili razpoložljive podatke posredujejo naprej vsem visokošolskim učiteljem in sodelavcem

in jih uporabili za izboljšanje prakse poučevanja in učenja. Tu so na njihovi fakulteti. Zaradi slabšega odziva, podatke še vedno

v ospredju predvsem analiza družbenih omrežij, analiza čustev, zbiramo. Za namen tega prispevka smo podatke izvozili in jih

analiza vpliva in napovedovanje uspeha študentov [11]. Hoppe analizirali s pomočjo programa Microsoft Excel.

[12] predlaga trojico metodoloških pristopov v učni analitiki

(slika 1), ki vključuje: (1) analizo socialnega omrežja, vključno

z analizo odnosov med akterji ter akterji in artefakti, (2) procesno



4 REZULTATI

V raziskavi je sodelovalo 54 visokošolskih učiteljev in

sodelavcev, od tega jih je 40 vprašalnik v celoti izpolnilo. Samo Podatke zbiram prek sistemov

te smo uporabili za našo analizo. Z Univerze v Mariboru so 16% 23% 21% 35% 6% za upravljanje učenja (npr.

oddaja nalog, kviz, interakcija

vprašalnik izpolnili visokošolski učitelji in sodelavci osmih na forumu).

fakultet, z Univerze v Rijeki so vprašalnik izpolnili visokošolski

učitelji in sodelavci štirih fakultet. Največji delež respondentov Podatke zbiram z uporabo

ima akademski naziv redni profesor (28%) in docent (28%), spletnih orodij 2.0 (npr. MS 35% 28% 23% 10%5%

sledijo respondenti z nazivom izredni profesor (23%) in asistent Teams, družbeni mediji).

(15%) in višji predavatelj (8%) (slika 2).



Podatke zbiram s pomočjo študentske ankete. 26% 22% 28% 17% 9%

Redni profesor 28%



Izredni profesor 23% Podatke zbiram s senzorji

(npr. kamerami, mikrofoni)

Docent 28% ali/in z nosljivimi napravami 83% 13% 4%



Asistent (npr. očala s tehnologijo 15% sledenja oči).

Lektor 0%



Višji predavatelj Analiziram vsak zbrani 8% 33% 24% 27% 12% 6% podatek o študentu ločeno.

Predavatelj 0%



0% 10% 20% 30%

Združujem podatke o

30% 19% 27% 16% 10%

študentih v enaki obliki.

Slika 2: Delež respondentov po akademskih nazivih



Kar 70% sodelujočih visokošolskih učiteljev in sodelavcev Združujem podatke o

ima več kot 15 let delovne dobe, sledijo tisti, ki imajo 1-5 let študentih iz več virov 42% 16% 23% 16% 4%

delovne dobe (13%), nato 11-15 let delovne dobe (10%) in 6-10 podatkov z različnimi formati.

let delovne dobe (8%). Prevladuje starost med 45 in 54 let (44%),

sledi 55-64 let (23%), 35-44 let (15%), 25-34 let (10%) in 65 let Uporabljam poročila, ki so na



ali več (8%). Spola sta enakomerna prisotna. voljo v sistemih za 29% 24% 17% 27% 5% upravljanje učenj, spletnih Sodelujoči visokošolski učitelji in sodelavci uporabljajo učno orodjih/aplikacijah. analitiko, vendar v dokaj omejenem obsegu (slika 3).

Najpogosteje uporabljajo funkcionalnosti, ki jih ponujajo sistemi Uporabljam nadzorne plošče



za upravljanje učenja in sicer podatke zbirajo preko sistema za pripravljene z Excelom ali 63% 17%14% 4% 3% orodji za poslovno obveščanje upravljanje učenja, kot so npr. oddaje nalog, kvizi, …uporabljajo (npr. Power BI, Tableau). poročila, ki so na voljo v sistemih za upravljanje učenja, spletnih

orodjih/aplikacijah, ki jih uporabljajo. Nekoliko manj pogosto

zbirajo podatke s pomočjo študentskih anket in spletnih orodij, Za napovedovanje učnih

npr. MS Teams, ki jih uporabljajo pri svojem delu, zelo redko pa strojnega učenja. izidov uporabljam tehnike 76% 10%12% 3% zbirajo podatke s senzorji in nosljivimi napravami. Podatke tudi

združujejo, občasno združujejo podatke v enaki obliki, redko 0% 20% 40% 60% 80% 100%

združujejo podatke iz več virov z različnimi podatki. Pri analizah

občasno uporabljajo funkcionalnosti, ki so jim na voljo v Nikoli Redko Občasno Pogosto Vedno

sistemih in orodjih, redko uporabljajo nadzorne plošče,

pripravljene z Excelom ali orodji za poslovno obveščanje. Za

napovedovanju učnih izidov redko uporabljajo tehnike strojnega Slika 3: Uporaba učne analitike

učenja.

Preverili smo tudi dejavnike, ki so se izkazali v relevantni Pri trditvah, ki se navezujejo na družbeni vpliv so deleži med

literature za pomembne. Predvsem smo se osredotočili na strinjanji nekoliko bolj raznoliki in kažejo, da družbeni vpliv ni

pričakovan napor, družbeni učinek, predvidena tveganja in tako pomemben dejavnik (slika 4). Od vseh treh trditvah se

spodbujevalne pogoje. Pri pričakovanem naporu se slaba sodelujoči visokošolski učitelji in sodelavci najbolj strinjajo, da

polovica (49%) niti ne strinja niti, niti strinja s trditvijo, da se je je uporaba učne analitike trenutno zelo priljubljena, sledi

zdi uporaba učne analitike enostavna. S to trditvijo se ne strinja poznavanje ljudi, ki učno analitiko uporabljajo, medtem ko se

oz. močno ne strinja 27%, strinja oz. močno strinja pa 20%. najmanj strinjajo, da jim ljudje, ki vplivajo na njihove vedenje,

Podobni deleži so pri trditvah, da se je naučiti uporabljati učno priporočajo uporabo učne analitike.

analitiko enostavno in da zlahka pridobijo veščine za uporabo

učne analitike.



Kljub temu, da sodelujoči visokošolski učitelji in sodelavci



Menim, da je uporaba UA uporabljajo učno analitiko v dokaj omejenem obsegu pa se jih pri trenutno zelo priljubljena. 8% 10% 44% 37% 2% večini trditvah o prednostih uporabe učne analitike največ

Ljudje, ki imajo vpliv na moje opredeljuje, da se z njimi strinjajo (slika 6). Najbolj se strinjajo s

vedenje, mi priporočajo 20% 25% 35% 19% 2% trditvijo, da učna analitika pomaga pri sprejemanju informiranih

uporabo UA. izobraževalnih odločitev, sledita trditvi, da uporaba učne



Poznam ljudi, ki uporabljajo analitike povečuje učinkovitost poučevanja predmeta in 22% 22% 25% 29% 4% UA. predstavlja pomembno orodje za izboljšanje učnih praks.

0% 20% 40% 60% 80% 100% Najmanj pa se strinjajo s trditvijo, da uporaba učne analitike

prispeva k boljši izkušnji študentov.

Močno se ne strinjam

Ne strinjam se

Niti se ne strinjam niti se strinjam

Strinjam se UA predstavlja pomembno

Močno se strinjam refleksivno orodje za 4% 4% 40% 49% 4%

izboljšanje učnih praks.



Slika 4: Družbeni vpliv Uporaba UA prispeva k boljši 4% 10% 48% 36% 4% izkušnji študentov.

Pri predvidenih tveganjih uporabe učne analitike je ponovno



neopredeljenih oz. se s trditvami niti ne strinja, niti strinja. Je pa učinkovitost poučevanja 6% 4% 36% 49% 6% predmeta. večji delež tistih, ki se jim uporaba učne analitike zdi prej največji delež sodelujočih visokošolskih učiteljev in sodelavcev Uporaba UA povečuje

priložnost, kot tveganje. Pri trditvi »Zaupam informacijam, ki jih Uporaba UA pomaga pri

zagotavlja učna analitika« se 45% respondentov niti ne strinja, pravočasnem sprejemanju 4% 6% 44% 45% 2%

niti strinja, s trditvijo se strinja oz. popolnoma strinja 31%, 18% izobraževalnih odločitev.

se ne strinja oz. popolnoma ne strinja s trditvijo. S trditvijo »Učna Uporaba učne analitike (UA)



analitika zagotavlja prave informacije za sprejemanje ustreznih pomaga pri sprejemanju 4% 6% 30% 57% 4% informiranih izobraževalnih odločitev« se niti ne strinja, niti strinja 51%, 31% se s trditvijo odločitev. strinja oz. popolnoma strinja, le 7% se ne strinja oz. popolnoma 0% 20% 40% 60% 80% 100% ne strinja s trditvijo. Visokošolski učitelji in sodelavci pa se

najbolje strinjajo s trditvijo, da »Učna analitika zagotavlja

informacije v realnem času«.

Pri spodbujevalnih pogojih (slika 5) lahko vidimo, da se Močno se ne strinjam

respondenti najbolj strinjajo, da imajo na voljo ustrezna orodja

za analizo podatkov, nekoliko manj se strinjajo s trditvama, da je Ne strinjam se

na voljo dovolj informacij o uporabi učne analitike in da so na Niti se ne strinjam niti se strinjam

voljo usposabljana, najmanj pa se strinjajo s trditvijo, da so na Strinjam se voljo ustrezne smernice, ki regulirajo in urejajo dostop do

podatkov. Močno se strinjam



smernice, ki regulirajo in… Na voljo so ustrezne Slika 6: Prednosti uporabe učne analitike 10% 30% 42% 17% 2%



Na voljo so ustrezna orodja za 4% 13% 45% 31% 7% 5 SKLEP

analizo podatkov.

Na voljo so usposabljanja o Preliminarni rezultati raziskave kažejo, da je uporaba učne

uporabi UA. 12% 25% 41% 20% 4% analitike na univerzah še vedno v zgodnji fazi razvoja in s tem

potrjuje ugotovitve preteklih raziskav na tem področju. Kljub

Na voljo je dovolj informacij 16%15% 46% 20% 4% temu, da se vse več institucij in deležnikov zaveda potenciala, ki

o uporabi UA.

ga ponuja učna analitika za izboljšanje učnih izidov, analiza kaže,

0% 20% 40% 60% 80% 100% da visokošolski učitelji in sodelavci uporabljajo predvsem

funkcionalnosti, ki jih ponujajo sistemi za upravljanje učenja. Ti

Močno se ne strinjam podatki se zbirajo preko oddaj nalog, kvizov in drugih aktivnosti,

Ne strinjam se redko te podatke združujejo z drugimi viri ali uporabljajo

Niti se ne strinjam niti se strinjam napredne analitične tehnike, kot so strojno učenje.

Strinjam se Visokošolski učitelji in sodelavci soočajo z izzivi pri uporabi

učne analitike. Iz analize dejavnikov, kot so pričakovan napor,

Močno se strinjam družbeni učinek, predvidena tveganja in spodbujevalno okolje, je

razvidno, da so visokošolski učitelji in sodelavci razmeroma



Slika 5: Spodbujevalni pogoji skeptični do enostavnosti uporabe učne analitike. Prav tako



družbeni vpliv ne igra pomembne vloge pri njihovi uporabi učne LITERATURA IN VIRI

analitike. Pri predvidenih tveganjih so učitelji previdni in učni [1] T. Denley, ‘How Predictive Analytics and Choice Architecture Can Improve

analitiki popolnoma ne zaupajo. Najbolj pa se strinjajo s Student Success’, Research & Practice in Assessment, vol. 9, pp. 61–69,

trditvami glede spodbujevalnih pogojev za uporabo učne 2014.

[2] D. Gašević, S. Dawson, T. Rogers, and D. Gasevic, ‘Learning analytics

analitike. should not promote one size fits all: The effects of instructional conditions

Kljub omejenemu obsegu uporabe učne analitike, so in predicting academic success’, The Internet and Higher Education, vol. 28,

pp. 68–84, Jan. 2016, doi: 10.1016/j.iheduc.2015.10.002.

visokošolski učitelji in sodelavci prepoznali prednosti uporabe [3] R. Cobo-Rendon, K. Lobos Peña, J. Mella-Norambuena, N. Cisternas San

teh orodij. Največjo podporo imajo trditve, da učna analitika Martin, and F. Peña, ‘Longitudinal Analysis of Teacher Technology

pomaga pri sprejemanju informiranih izobraževalnih odločitev, Acceptance and Its Relationship to Resource Viewing and Academic

Performance of College Students during the COVID-19 Pandemic’,

povečuje učinkovitost poučevanja in predstavlja pomembno Sustainability, vol. 13, no. 21, Art. no. 21, Jan. 2021, doi:

orodje za izboljšanje učnih praks. 10.3390/su132112167.

[4] R. Kaliisa, A. I. Mørch, and A. Kluge, ‘“My Point of Departure for Analytics

Kljub pozitivnim vplivom učne analitike na visokošolsko is Extreme Skepticism”: Implications Derived from An Investigation of

izobraževanje je njena uporaba med učitelji še vedno omejena. University Teachers’ Learning Analytics Perspectives and Design Practices’,

Za popolno izkoriščanje potenciala učne analitike je potrebno Tech Know Learn, vol. 27, no. 2, pp. 505–527, Jun. 2022, doi:

10.1007/s10758-020-09488-w.

zagotoviti ustrezno podporno okolje, ki vključuje naprednejša [5] G. Siemens and P. Long, ‘Penetrating the Fog: Analytics in Learning and

analitična orodja in prilagojene nadzorne plošče, dostop do Education’, EDUCAUSE Review, vol. 46, no. 5, p. 30, 2011.

[6] M. Bamiah, S. N. Brohi, and B. Bashari, ‘Big data technology in education:

informacij in usposabljanj ter jasne smernice za uporabo Advantages, implementations, and challenges’, 2018. Accessed: Jun. 03,

podatkov. Prav tako je pomembno vključiti visokošolske učitelje 2024. [Online]. Available: https://www.semanticscholar.org/paper/BIG-

DATA-TECHNOLOGY-IN-EDUCATION%3A-ADVANTAGES%2C-

in sodelavce kot aktivne deležnike v procesu razvoja in AND-Bamiah-Brohi/f3f8c935382d3762aaa2c71d16b155ea6bb9d41b

implementacije učne analitike, saj lahko njihova neposredna [7] T. Olney, S. Walker, C. Wood, and A. Clarke, ‘Are We Living In LA (P)LA

vključenost pripomore k boljšemu razumevanju in uporabi Land? Reporting on the Practice of 30 STEM Tutors in their Use of a

Learning Analytics Implementation at the Open University’, Journal of

analitičnih orodij. Learning Analytics, vol. 8, no. 3, Art. no. 3, Oct. 2021, doi:

Nadaljnje raziskave bi morale vključevati bolj obsežne in 10.18608/jla.2021.7261.

[8] R. Ferguson, D. Clow, D. Griffiths, and A. Brasher, ‘Moving Forward with

longitudinalne študije ter upoštevati različne vidike in deležnike Learning Analytics: Expert Views’, Journal of Learning Analytics, vol. 6,

v izobraževalnem procesu. Prav tako je potrebno razviti boljše no. 3, Art. no. 3, Dec. 2019, doi: 10.18608/jla.2019.63.8.

metodološke pristope, ki bodo omogočali bolj generalizirane in [9] O. Rodriguez, ‘The concept of openness behind c and x-MOOCs (Massive

Open Online Courses)’, Open Praxis, vol. 5, no. 1, Jan. 2013, doi:

prenosljive ugotovitve. Učna analitika ima potencial, da postane 10.5944/openpraxis.5.1.42.

pomembno orodje za izboljšanje izobraževalnih praks, vendar so [10] D. Ye, ‘The History and Development of Learning Analytics in Learning,

Design, & Technology Field’, TechTrends, vol. 66, no. 4, pp. 607–615, Jul.

potrebne nadaljnje raziskave in razvoj podpornega okolja, ki bo 2022, doi: 10.1007/s11528-022-00720-1.

zagotovil uspešen prenos teorije v prakso. [11] G. Siemens and R. S. J. d. Baker, ‘Learning analytics and educational data

mining: towards communication and collaboration’, in Proceedings of the

2nd International Conference on Learning Analytics and Knowledge, in



ZAHVALA LAK ’12. New York, NY, USA: Association for Computing Machinery, Apr. 2012, pp. 252–254. doi: 10.1145/2330601.2330661.

Projekt sofinancirata Republika Slovenija, Ministrstvo za [12] H. U. Hoppe, ‘Computational methods for the analysis of learning and



visoko šolstvo, znanost in inovacije, in Evropska unija – knowledge building communities’, in Handbook of learning analytics, 2017, pp. 23–33. NextGenerationEU. Projekt se izvaja skladno z načrtom v okviru [13] V. Venkatesh, M. G. Morris, G. B. Davis, and F. D. Davis, ‘User Acceptance

razvojnega področja Pametna, trajnostna in vključujoča rast, of Information Technology: Toward a Unified View’, MIS Quarterly, vol.

komponente Krepitev kompetenc, zlasti digitalnih in tistih, ki jih 27, no. 3, pp. 425–478, 2003, doi: 10.2307/30036540.

zahtevajo novi poklici in zeleni prehod (C3 K5), za ukrep

investicija F. Izvajanje pilotnih projektov, katerih rezultati bodo

podlaga za pripravo izhodišč za reformo visokega šolstva za

zelen in odporen prehod v družbo 5.0: projekt Pilotni projekti za

prenovo visokega šolstva za zelen in odporen prehod.





Ustvarjanje in vrednotenje labirintov




Creating and Evaluating Mazes



Anja Mejač Irena Nančovska Šerbec

Pedagoška fakulteta, Univerza v Ljubljani Pedagoška fakulteta, Univerza v Ljubljani

Ljubljana, Slovenija Ljubljana, Slovenija

anja.mejac@pef.uni-lj.si irena.nancovska@pef.uni-lj.si



POVZETEK KEYWORDS

V naši raziskavi se posvečamo algoritmom za ustvarjanje Mazes, algorithms, evaluation of difficulty, interactive

popolnih labirintov, ocenjevanju njihove težavnosti ter learning environment

razvrščanju teh algoritmov glede na zahtevnost ustvarjenih

labirintov. Raziskali smo teoretično ozadje izbranih algoritmov,

kot so Primov, Kruskalov, Aldous-Broder, Iskanje v globino in 1 UVOD

Bakterijska rast. Ustvarili smo aplikacijo za učenje Primovega in Algoritmično generiranje labirintov je področje računalništva,

Kruskalovega algoritma za ustvarjanje labirintov. Težavnost ki združuje teorijo grafov, diskretno matematiko in proceduralno

ustvarjenih labirintov vrednotimo s pomočjo Markovske verige generiranje vsebin. Ta pristop omogoča ustvarjanje kompleksnih

glede na povprečno število korakov, ki jih naredi program pri struktur, ki se uporabljajo v različnih kontekstih, kot so

iskanju poti v labirintu po Monte Carlo metodi. Raziskali smo računalniške igre, učenje programiranja in razvijanje

primernost uporabe interaktivnega učnega okolja za samostojno računalniškega mišljenja. Labirinti ne predstavljajo le izziv za

učenje algoritmov za ustvarjanje labirintov med učenci zadnjega učence, temveč so tudi odlična priložnost za raziskovanje in

vzgojno-izobraževalnega obdobja osnovne šole in začetka razumevanje pomembnih matematičnih konceptov.

srednje šole (učenci stari med 12 in 16 let). Aplikacija uči Razumevanje matematičnega ozadja, vključno s teorijo grafov,

uporabnike novih algoritmov ter omogoča ustvarjanje in prispeva k razvoju učinkovitih algoritmov za ustvarjanje

vrednotenje labirintov, ki jih lahko uporabijo pri uvodnem labirintov. Grafi omogočajo jasno ponazoritev labirintov, kjer

programiranju. vozlišča predstavljajo celice labirinta, povezave pa prehode med

njimi. Ta matematična abstrakcija poenostavi analiziranje

KLJUČNE BESEDE labirintov, saj omogoča uporabo uveljavljenih algoritmov za

Labirinti, algoritmi, vrednotenje težavnosti, interaktivno učno iskanje poti.

okolje Ukvarjamo se tudi z metodami za objektivno vrednotenje

težavnosti labirintov, saj to omogoča primerjavo različnih



based on the complexity of the mazes they produce. We prispeva k izobraževalnemu procesu, še posebej za mlajše učence, ki se šele spoznavajo z osnovami računalništva in matematike, examined the theoretical foundations of selected algorithms, ter k razvoju njihovega računalniškega mišljenja. including Prim’s, Kruskal’s, Aldous-Broder, Depth-First Search, Poleg tega bomo predstavili samostojno razvito interaktivno and Bacterial Growth. To support this, we developed an učno okolje, ki bo učencem omogočilo, da se seznanijo z application that facilitates learning Prim’s and Kruskal’s maze algoritmi za ustvarjanje labirintov in tudi sami ustvarijo labirint generation algorithms. The difficulty of the generated mazes is mazes, evaluating their difficulty, and ranking these algorithms In our research, we focus on algorithms for generating perfect učencev. Razvoj učinkovitih in učencem prijaznih učnih okolij za učenje algoritmov, ki ustvarjajo labirinte, lahko pomembno ABSTRACT algoritmov in prilagoditev labirintov specifičnim potrebam

assessed using a Markov chain, which measures the average po korakih izbranega algoritma za ustvarjanje labirintov. To

number of steps the program takes to find a path through the okolje je namenjeno učencem od 7. do 9. razreda osnovne šole

maze, following the Monte Carlo method. Additionally, we ter dijakom 1. letnika srednje šole (učenci stari med 12 in 16 let).

explored the suitability of using an interactive learning



environment for independent study of maze generation 2 TEORETIČNA IZHODIŠČA algorithms among students in the later stages of primary

education and the early years of secondary education (ages 12 to V teoretičnem delu bomo najprej opisali pojem labirinta in

16). The application not only teaches users new algorithms but ga povezali z zgodovinskim vidikom. Nato bomo obravnavali

also enables them to create and evaluate mazes, which can be povezavo med labirinti in teorijo grafov. Na koncu bomo opisali

used in introductory programming. nekaj znanih algoritmov za ustvarjanje popolnih labirintov ter

način, kako smo ovrednotili ustvarjene labirinte.



2.1 Labirinti 𝐻 povezanega grafa 𝐺, kjer je 𝐻 drevo, se imenuje vpeto drevo

Labirint je pot oziroma skupek poti, ki vodijo od vhoda v grafa 𝐺.



labirint do izhoda iz njega. V splošnem delimo labirinte na Pri reševanju problemov iz vsakdanjega življenja za reprezentacijo problema pogosto uporabljamo utežene grafe. razvejane in nerazvejane. Razvejani labirinti (angl. “branching Takšni grafi se pojavljajo v številnih kontekstih, na primer pri tour puzzles”) so tisti, skozi katere mora reševalec najti ustrezno reševanju problema najkrajše poti – zelo znan je problem pot, nerazvejani ali enosmerni labirinti (angl. “non branching” ali trgovskega potnika. “unicursal”) pa nedvoumno vodijo skozi zapleteno postavitev do Definicija 2.8. Uteženi graf je graf, v katerem je vsaki cilja [1]. povezavi dodeljeno število (običajno je to število pozitivno), ki V grški mitologiji je labirint predstavljal kompleksno mu rečemo utež in označimo z 𝑤(𝑒) . Če je 𝐺 uteženi graf in 𝐻 strukturo, ki jo je zasnoval in zgradil legendarni umetnik Dedal njegov podgraf, potem je utež 𝑤(𝐻) podgrafa 𝐻 enaka vsoti za kralja Minosa iz Krete. Njegova funkcija je bila, da je uteži povezav podgrafa 𝐻 . zadrževal Minotavra, pošast, ki jo je nazadnje ubil junak Tezej s Pri labirintih sicer navadno nimamo takšnih težav, saj je pomočjo boginje Ariadne [2]. odstranjevanje sten iz določenega materiala enakovredno. Če pa Labirinte lahko delimo tudi glede na sistem prehoda. Osnovne bi imeli, na primer, stene iz različnih materialov in bi za tri vrste so: popolni labirint (angl. perfect maze), spleteni labirint odstranjevanje različnih materialov porabili različno veliko (angl. braid maze) in enosmerni labirint (angl. unicursal maze). energije, bi za tak problem potrebovali utežen graf. Recimo, da Obravnavali smo zgolj popolne labirinte, ki morajo ustrezati želimo porabiti čim manj energije za odstranjevanje sten, v tem naslednjim zahtevam: nimajo nobenih zank, nimajo izoliranih primeru želimo izbrati takšno množico povezav vpetega drevesa, celic in med vsakim parom celic obstaja natanko ena pot [3]. Taki da bo njihova vsota uteži najmanjša. Torej ne bo zadostovalo labirinti so ekvivalentni drevesom iz teorije grafov. zgolj katerokoli izbrano vpeto drevo. Temu rečemo, da želimo Ustvarjanje labirintov je načrtovanje razporeditve prehodov poiskati minimalno vpeto drevo tega grafa. in sten v labirintu. Obstaja veliko različnih pristopov k Definicija 2.9. Minimalno vpeto drevo uteženega grafa je ustvarjanju labirintov z različnimi algoritmi za ustvarjanje množica povezav z minimalno skupno utežjo, ki tvorijo vpeto labirintov, bodisi ročno bodisi samodejno z računalnikom. Za drevo grafa. Kadar graf ni obtežen, je vsako vpeto drevo ustvarjanje labirintov se uporabljata dva glavna mehanizma. Pri minimalno. izrezovanju prehodov se sprva ustvari mrežo v kateri nato Kot smo že omenili, teža na splošno ni pomembna za odstranjujemo stene, pri dodajanju sten pa se na odprtem predstavitev labirinta, kljub temu pa nekateri algoritmi za območju določi niz ovir [1]. Pri raziskavi smo se osredotočili na ustvarjanje labirintov temeljijo na algoritmih, ki iščejo mehanizem izrezovanja prehodov oziroma odstranjevanja sten. minimalna vpeta drevesa, zato je dobro poznati vsaj osnovno



2.2 definicijo uteženega grafa in minimalnega vpetega drevesa. Povezava s teorijo grafov



Definicije in izreki, ki so podani v tem podpoglavju, so 2.3 Algoritmi za ustvarjanje popolnih labirintov



povzeti iz [4]. Definicija 2.1. Graf Obstaja veliko različnih algoritmov za ustvarjanje labirintov. 𝐺 = (𝑉, 𝐸) je urejen par množic 𝑉 in 𝐸 , Opisali bomo le nekaj najbolj pogosto uporabljenih in najbolj



množice neurejenih parov različnih vozlišč iz znanih. Začeli bomo z algoritmoma, ki se uporabljata za kjer je 𝑉 neprazna množica vozlišč in 𝐸 neka podmnožica

𝑉 . llementom

reševanje problema iskanje minimalne poti v grafu – torej iščeta

množice 𝐸 rečemo povezave grafa G. Če je (𝑢, 𝑣) ∈ 𝐸, rečemo,

minimalno vpeto drevo (to sta Primov in Kruskalov algoritem).

da sta 𝑢 in 𝑣 sosednji vozlišči, kar označimo z 𝑢 ∼ 𝑣.

Nato pa si bomo pogledali še nekaj takšnih, ki ustvarijo le tako

Labirint lahko predstavimo z grafom. Vsaka celica labirinta

imenovani naključni sprehod, kar je za potrebe ustvarjanja



natanko takrat, kadar sta pripadajoči celici sosednji in med njima labirintov zadostno, saj potrebujemo le enega izmed možnih je predstavljena kot vozlišče grafa, dve vozlišči pa sta povezani

obstaja prehod oziroma pot. Popolni labirint ustreza enakim vpetih dreves grafa.

zahtevam, kot jih ima drevo, zato pravimo, da so popolni labirinti 2.3.1 Primov algoritem

ekvivalentni drevesom iz teorije grafov. Algoritem deluje tako, da na začetku izbere poljubno vozlišče

Izrek 2.6. Graf 𝐺 je drevo natanko tedaj, ko za vsaki dve 𝑣 grafa 𝐺 (celico v labirintu). Primovo drevo 𝑇 začne graditi z

vozlišči grafa G, obstaja natanko ena pot med njima. vozliščem 𝑣. Nato ustvari prazno množico 𝑚, ki bo predstavljala

Dokaz izreka bomo izpustili. vse mejne povezave. Mejne povezave so tiste, ki sovpadajo z

Začetno mrežo, v kateri bomo izrezovali prehode, mejnimi stenami vseh tistih celic, ki so že v drevesu 𝑇. Dokler

predstavimo z grafom mreže (kartezični produkt grafov poti). Da Primovo drevo 𝑇 ne doseže vseh vozlišč (celic labirinta)

bomo lahko iz začetne mreže, kjer ima vsaka celica vse štiri stene, ponavlja naslednje: posodobi množico mejnih povezav, izbere

dobili popoln labirint, moramo nekatere stene odstraniti. To povezavo 𝑒 iz množice 𝑚, ki ima najmanjšo utež, in v drevo 𝑇

bomo naredili tako, da bomo določili vpeto drevo, saj so popolni doda tisto vozlišče izbrane povezave 𝑒, ki še ni v drevesu 𝑇. Ko

labirinti ekvivalentni drevesom. Ko bomo izbirali vpeto drevo, doseže vsa vozlišča, dobi minimalno vpeto drevo danega grafa.

bodo izbrane povezave predstavljale odstranjene stene, kar bo Čeprav so utežene povezave v nekaterih primerih pomembne,

omogočilo prehodnost v labirintu. za ustvarjanje labirintov niso potrebne. Če bi želeli izvesti pravi

Definicija 2.7. Naj bosta 𝐺 1 = (𝑉 1, 𝐸1) in 𝐺2 = (𝑉 2, 𝐸2) Primov algoritem, bi morali vsem povezavam vnaprej dodeliti

grafa. Tedaj je 𝐺 1 podgraf grafa 𝐺2, če je 𝑉 1 ⊆ 𝑉2 in 𝐸1 ⊆ 𝐸2. naključne uteži, predno bi zagnali algoritem. Vendar to ni

Če je 𝑉 1 = 𝑉 2, je 𝐺1 vpet podgraf grafa 𝐺2. Vsak vpet podgraf potrebno, saj je pravzaprav zadosti, da na vsakem koraku kar



naključno izberemo eno izmed povezav iz množice 𝑚 (lahko si sledenje. Pomemben vidik je naključnost. Že na prvem koraku

predstavljamo, da imajo vse povezave enako utež) [5]. izbere naključno celico. Druga naključna izbira pa se nanaša na

izbiro soseda oziroma otroka vozlišča, do katerega se bo ustvaril

2.3.2 Kruskalov algoritem



Algoritem deluje tako, da ima na začetku množico možnih Lahko bi tudi rekli, da algoritem Iskanje v globino deluje prehod iz trenutne celice [3].

povezav, to so vse povezave v grafu, iz katere na vsakem koraku podobno kot algoritem Lovi in ubij. Oba namreč ustvarjata

izbere tisto povezavo, ki ima najmanjšo utež – če obstaja več naključni sprehod po mreži, razlika pa je v tem, kako se

povezav z enako utežjo, potem naključno izbere eno izmed njih. izkopljeta iz slepih ulic. Algoritem Lovi in ubij bo prešel v fazo

Če z dodajanje te povezave ne ustvari cikla, jo doda v vpeto “lovljenja”, algoritem Iskanje v globino pa se vrne nazaj in sledi

drevo in jo nato odstrani iz množice možnih povezav. To svojim korakom, dokler ne najde celice, ki ima še vsaj enega

ponavlja, dokler ne ustvari vpetega drevesa, torej dokler množica neobiskanega soseda. Za namen beleženja celic, ki smo jih že

povezav ni prazna. Na koncu dobimo minimalno vpeto drevo obiskali, uporabimo kar strukturo, ki se imenuje sklad. danega grafa.

Podobno kot pri Primovem algoritmu, tudi tokrat utežene 2.3.6 Algoritem Bakterijska rast

povezave niso potrebne. Za ustvarjanje popolnih labirintov je Algoritem Bakterijska rast ustvarja labirint na zelo podoben

dovolj, da na vsakem koraku kar naključno izberemo eno izmed način kot Primov algoritem. Oba gradita vpeto drevo iz enega

povezav [5]. vozlišča naprej na razširljiv način. Razlika je v “intenzivnosti”



2.3.3 Aldous-Broder algoritem znotraj zanke na vsakem koraku doda po eno povezavo, medtem dodajanja novih povezav in vozlišč v drevo. Primov algoritem

Algoritem deluje tako, da najprej naključno izbere začetno ko algoritem Bakterijska rast doda povezavo in ustrezno vozlišče

vozlišče 𝑣 . To vozlišče označi za trenutno in ga doda med vsaki celici z razpoložljivimi, tj. še neobiskanimi, sosedi. Število

obiskana vozlišča. Dokler ne označi vseh vozlišč kot obiskanih obiskanih celic se lahko torej po vsakem koraku največ podvoji.

ponavlja naslednje: izbere naključnega soseda 𝑢 trenutnega Najverjetneje od tod izvira ime algoritma, saj populacija bakterij

vozlišča 𝑣 in preveri, ali je to vozlišče 𝑢 že med obiskanimi. Če običajno narašča eksponentno [3]. to vozlišče 𝑢 še ni med obiskanimi, potem izbere povezavo med

vozliščema 𝑣 in 𝑢 (odstrani steno med pripadajočima celicama) 2.4 Vrednotenje labirintov in označi vozlišče 𝑢 kot obiskano. Nato označi vozlišče 𝑢 kot

Reševanje labirintov je proces, pri katerem je cilj najti pot od

trenutno vozlišče. Če pa je bilo vozlišče 𝑢 že obiskano, potem



samo označi vozlišče 𝑢 kot trenutno vozlišče [6]. razpotij in posledično slepih ulic (saj obravnavamo le popolne vhoda v labirint do izhoda iz njega. Če labirint ne vsebuje

2.3.4 Algoritem Lovi in ubij labirinte), bo takšen labirint manj zahteven za reševanje. V ta

Algoritem deluje tako, da najprej naključno izbere začetno namen se je veliko raziskovalcev spraševalo, kateri algoritmi

vozlišče 𝑣 . To vozlišče označi za trenutno in ga doda med ustvarjajo zahtevnejše labirinte in kateri manj zahtevne. Kako se

obiskana vozlišča. Dokler ne označi vseh vozlišč kot obiskanih lotiti vrednotenja labirintov in merjenja njegove težavnosti?

ponavlja naslednje: izbere naključnega še neobiskanega soseda u Mane idr. v [7] opisujejo, da je reševanje labirintov težje

(če le-ta obstaja) trenutnega vozlišča kadar imamo več razpotij in slepih ulic. Pri vrednotenju so poleg 𝑣 in izbere povezavo med

vozliščema števila razpotij in slepih ulic upoštevali tudi hitrost različnih 𝑣 in 𝑢 (odstrani steno med pripadajočima celicama).

Če tako sosednje vozlišče agentov, ki so reševali labirinte. Agenti so računalniški programi, 𝑢 ne obstaja, torej so vsi njegovi

sosedi že obiskani, potem vstopi v fazo “lovljenja” in naredi ki iščejo pot v labirintu na različne načine, tj. z različnimi

naslednje: pregleduje vozlišča enega za drugim, dokler ne naleti algoritmi za iskanje poti, kot so Iskanje v globino, Iskanje v

na neobiskano vozlišče širino in Dijkstrov algoritem. Po njihovih kriterijih so težji 𝑤 , ki je sosednje z vsaj enim obiskanim

vozliščem. Ko najde takšno vozlišče labirinti tisti, ki so ustvarjeni z algoritmom Aldous-Broder, 𝑤 , ga označi za trenutnega

in obiskanega ter izbere povezavo med vozliščem medtem ko so lažji tisti, ki so ustvarjeni z algoritmoma Lovi in 𝑤 in

katerimkoli izmed njegovih sosednjih že obiskanih vozlišč ubij ter Iskanje v globino. Do Podobnih ugotovitev je prišel tudi

(odstrani steno med pripadajočima celicama). Ko označi vsa Gabrovšek v [8].

vozlišča za obiskana, se algoritem ustavi. Bellot idr. v [9] vrednotijo labirinte na podoben način, vendar

Če se tak algoritem izvaja na labirintu, lahko fazo “lovljenja” se osredotočajo na to, kateri algoritmi ustvarijo najbolj “zabavne”

oziroma iskanja strukturiramo na naslednji način: algoritem labirinte za reševanje. Razvili so dva nova algoritma, imenovana

začne pri celici v zgornjem levem kotu in pregleduje celice po Twist&Merge ter Prim&Kill. Njihovo ocenjevanje “zabavnosti”

vrsti od leve proti desni, od zgoraj navzdol. Nadaljuje pregled, labirintov temelji na številu nepomembnih zidov, kot jih

dokler ne naleti na neobiskano celico, ki je sosednja vsaj eni že definirajo v svojem članku.

obiskani celici [6].



2.3.5 Algoritem Iskanje v globino 3 EMPIRIČNI DEL

Algoritem iskanja v globino raziskuje graf tako, da se usmerja V tem delu članka bomo predstavili način vrednotenja

v globino. To pomeni, da je naslednje obiskano vozlišče, če le-to labirintov, ustvarjenih z različnimi algoritmi. Nato bomo

obstaja, kot otrok prejšnjega vozlišča. Če ne obstaja, se algoritem predstavili aplikacijo, ki smo jo razvili z namenom učenja

vrne do prejšnjega vozlišča in poskusi z drugim sosedom. korakov algoritmov za ustvarjanje labirintov. Opisali bomo tudi,

Algoritem se konča, ko obišče vsa vozlišča grafa. Podobno deluje kako so izbrani učenci, stari med 12 in 15 let, sprejeli aplikacijo.

tudi naš algoritem Iskanje v globino, ki ustvari sprehod od

začetne točke po celotnem grafu in po potrebi uporabi povratno



3.1 Vrednotenje labirintov z Markovsko verigo Tabela 1: Težavnost labirinta glede na razpon povprečja

Predstavljamo vrednotenje labirintov, ki temelji na uporabi

diskretne Markovske verige s prehodnimi verjetnostmi. Težavnost labirinta Razpon povprečja

Definicija 3.1. Naj bo Lahek povprečje 2 ≤ 𝑛 𝑉 množica stanj, 𝑣 ∈ 𝑉 pa stanje.

Markovska veriga je usmerjen graf Srednji 2 2 𝐺 = (𝑉, 𝐸), katerega 𝑛 < povprečje < √𝑛 · 𝑛

povezave so označene s prehodnimi verjetnostmi Težek povprečje 2 𝑝 ∶ 𝐸 → (0,1] ≤ √𝑛 · 𝑛

z lastnostjo, da za vsako stanje 𝑣 ∈ 𝑉 velja, da je vsota

verjetnosti na povezavah, ki grejo iz stanja 𝑣 v druga stanja, Za testiranje smo izbrali Kruskalov algoritem, Primov

enaka 1, kar zapišemo kot ∀𝑣 ∈ 𝑉 ∶ 𝑝(𝑒) = 1. algoritem, algoritem Lovi in ubij ter algoritem Iskanje v globino.

Markovska veriga sicer lahko vsebuje zanke, ne pa tudi Za vsakega izmed teh štirih algoritmov smo ustvarili 100

večkratnih povezav. Vendar pa te lastnosti za namen vrednotenja labirintov velikosti 10x10. Za vsak labirint je računalnik

težavnosti popolnih labirintov ne potrebujemo, saj popolni izračunal njegovo vrednost povprečja in podal stopnjo težavnosti

labirinti ne vsebujejo zank in vemo, da je naš labirint drevo, kar v skladu s tabelo Tabela 1. Povprečje smo izračunali na 10000

pomeni, da med vsakima dvema vozliščema obstaja natanko ena ponovitvah. Na podlagi naših meritev smo prišli do naslednjih

pot [10]. ugotovitev. Kot je razvidno iz tabele Tabela 2 ima Kruskalov

Kadar rešujemo problem z Markovskimi verigami, nas algoritem najvišje povprečno število korakov, ki jih je agent

običajno zanima, kolikšna je verjetnost, da bomo v 𝑛-tem koraku naredil pri iskanju poti. Za njim sledi Primov algoritem, nato

prišli v stanje 𝑠 . V primeru reševanja labirintov pa nas bo algoritem Lovi in ubij in kot zadnji algoritem Iskanje v globino.

zanimalo, kolikšno je povprečno število korakov, tj. premikov iz

enega stanja v drugega oziroma iz ene celice v drugo sosednjo

celico, da bomo prišli iz začetne celice labirinta v končno celico Tabela 2: Povprečno število korakov glede na algoritem

labirinta.



lahko preprosto ustvarimo ustrezno Markovsko verigo. Labirint Iz danega labirinta in predpisanimi prehodnimi verjetnosti Povprečno število korakov za Algoritem labirinte velikosti 10x10

pretvorimo v graf, oziroma v našem primeru v drevo, vsaki Kruskalov 354,4

povezavi pa priredimo dano verjetnost [10]. Nato v labirint Primov 309,4

pošljemo našega agenta, ki šteje število korakov, ki jih potrebuje, Lovi in ubij 258,8

da najde pot iz labirinta. Naj bo celica (0,0) začetno stanje, tj. Iskanje v globino 224,9

celica, ki leži v levem zgornjem kotu, in celica (𝑛 − 1, 𝑛 − 1)

končno stanje, tj. celica, ki leži v desnem spodnjem kotu, kjer je Na opisani način smo ovrednotili kompleksnost manjših

𝑛 dolžina oziroma širina kvadratnega labirinta. Definirajmo labirintov dimenzij 9x9, 8x8 in 7x7. Pri vseh dimenzijah se je

sedaj premikanje našega agenta po labirintu. izkazalo, da agent pri labirintih, ki jih ustvari Kruskalov

Definicija 3.2. Naj bo 𝐴 agent, ki se giblje po labirintu. Začne algoritem, naredi največje povprečno število korakov pri iskanju

v celici (0,0) in se premika po labirintu, vse dokler ne pride do poti v labirintu. Nato je sledil Primov algoritem, za njim

končne celice algoritem Lovi in ubij ter kot zadnji algoritem Iskanje v globino. (𝑛 − 1, 𝑛 − 1) . Agent 𝐴 pri premikanju skozi

labirint upošteva predpisane prehodne verjetnosti. Na vsakem

koraku izbere nadaljno smer gibanja, ki ni enaka smeri iz katere 3.2 Predstavitev aplikacije

je prišel, razen v primeru, ko se znajde v slepi ulici. Takrat agent V tem podpoglavju bomo predstavili samostojno izdelano

𝐴 nadaljuje svoje gibanje nazaj v smeri iz katere je prišel, tj. interaktivno učno okolje (aplikacijo) za učenje algoritmov, ki

njegova edina možna smer premikanja. ustvarjajo labirinte. Aplikacija ima dve glavni funkcionalnosti:

Čeprav smo vnaprej določili verjetnosti, s pomočjo katerih se učenje algoritmov, ki je del sklopa Poglej, in samostojno

bo gibal agent 𝐴, to gibanje še vedno temelji na naključni izbiri. ustvarjanje labirintov z izbranim algoritmom, ki je del sklopa

Zaradi tega bomo gibanje agenta 𝐴 po labirintu ponovili večkrat, Ustvari. Aplikacija ponuja ponazoritev delovanja izbranih

na primer 1000-krat, in vzeli povprečno število korakov, ki jih je algoritmov, s katerimi lahko ustvarjamo labirinte. Po drugi strani

agent naredil čez vse ponovitve. Temu načinu ustvarjanja pa uporabnikom nudi možnost ustvarjanja svojih labirintov,

naključnih poti s ciljem vrednotenja strukture imenujemo tudi vendar le, če so njihovi posamezni koraki ustrezni glede na

Monte Carlo. Glede na to povprečno število obiskanih celic (ne izbrani algoritem. Za vsak izvedeni korak dobi uporabnik

nujno različnih), v nadaljevanju povprečje, določi težavnost takojšnjo povratno informacijo glede pravilnosti oziroma

labirinta. Odločili smo se, da bomo labirinte razdelili v 3 nepravilnosti koraka. Ob morebitni nepravilnosti koraka

kategorije: lahek, srednji in težek. uporabnik dobi pojasnilo, zakaj izvedeni korak ne ustreza koraku

Algoritem vrednotenja smo preizkusili na labirintih velikosti izbranega algoritma. Natančnejši opis aplikacije je podan v [11].

3x3, 4x4 in 7x7. Na podlagi teh primerov smo določili meje V sklopu Poglej ponazorimo delovanje izbranih algoritmov s

izračunanega povprečja za razporeditev labirintov v tri različne katerimi lahko ustvarimo labirinte. Odločili smo se, da v

kategorije težavnosti. Ta razporeditev velja za labirinte velikosti aplikacijo vključimo naslednje štiri algoritme: Kruskalov

𝑛x𝑛, kjer je 𝑛 > 5. algoritem, Primov algoritem, algoritem Lovi in ubij ter algoritem

Iskanje v globino. Princip delovanja aplikacije je pri vseh štirih

algoritmih enak. Uporabnik ima na voljo tri gumbe s katerimi

upravlja vizualizacijo delovanja algoritmov za ustvarjanje

popolnih labirintov, in sicer “Poženi”, “Korak” in “Ponastavi” (v



tem vrstnem redu). S klikom na gumb “Poženi”, se bo na mreži 3. Ali si že kdaj prej slišala za besedo algoritem?

začel izvajati izbrani algoritem, dokler ne bo labirint dokončno 4. Ali bi imela takšno aplikacijo tudi doma na osebnem

ustvarjen. Gumba “Poženi” in “Korak” sta do nadaljnjega oziroma družinskem računalniku?

onemogočena, tj. dokler uporabnik ne klikne na gumb 5. Ali bi se ti zdela uporaba aplikacije v šoli primerna, če

“Ponastavi”. S klikom na gumb “Korak”, se izvede naslednji bi pri pouku ali pri šolskem krožku obravnavali

korak izbranega algoritma. Ko uporabnik prvič klikne nanj, se delovanje Kruskalovega ali Primovega algoritma?

izvede prvi korak algoritma in gumb “Poženi” je do nadaljnjega 6. Kaj bi pri aplikaciji izboljšala?

onemogočen, tj. dokler uporabnik ne klikne na gumb “Ponastavi”. Pri prvem vprašanju so testiranke povedale, da jim je všeč, da

Po vsakem naslednjem kliku na gumb “Korak” se v labirintu lahko samostojno ustvarjajo labirinte. Všeč jim je bila tudi

odstrani ena izmed sten. To lahko ponavljamo, dokler ne dobimo sprotna razlaga korakov posameznega algoritma v sklopu poglej

popolnega labirinta. S klikom na gumb “Ponastavi” se ponovno in olajšava z barvnimi celicami pri začetniku. Pri drugem

naloži prikazana scena, tj. kot da bi kliknili gumb “Nazaj” in nato vprašanju so dejale, da ni nič takšnega, kar jim ne bi bilo všeč.

zopet izbrali enak algoritem. Na ta način lahko uporabnik prekine Na tretje vprašanje sta dve testiranki odgovorili pritrdilo. Besedo

usvajanje delovanja algoritma na labirintu ali preneha s algoritem sta že srečali pri pouku matematike, vendar pa se nista

sledenjem algoritmu po korakih. spomnili kaj beseda pomeni. Na četrto vprašanje so vse tri

V sklopu Ustvari lahko uporabnik sam izdela labirint po testiranke odgovorile pritrdilno. Povedale so, da bi aplikacijo z

korakih enega izmed algoritmov ali preveri, če razume delovanje veseljem uporabljale, ena izmed testirank je še dejala, da bi jo

enega izmed algoritmov. V ta del aplikacije sta vključena dva uporabila za potrebe šolskega časopisa, da bi lahko vstavila svoj

izmed prejšnjih štirih algoritmov iz sklopa Poglej. Ta dva labirint pod razvedrilo, kjer bi ga rešili še ostali učenci. Tudi na

algoritma sta: Kruskalov in Primov algoritem. Uporabnik po peto vprašanje so vse tri testiranke odgovorile pritrdilno.

izbiri algoritma nato izbere tudi težavnost: začetnik ali Povedale so, da vidijo uporabo aplikacije v okviru šolskega

poznavalec. Pri začetniku so celice labirinta obarvane tako, da krožka. Pri zadnjem vprašanju pa je ena izmed testirank kot

uporabniku pomagajo pri izbiri ustreznih dveh celic, med izboljšavo navedla dodatno stopnjo težavnosti.

katerima dvema bo ustvaril prehod. Na ta način je začetniku lažje Vse tri testiranke so bile prijetno presenečene in navdušene

ugotoviti med katerimi celicami že obstaja pot v labirintu in med nad aplikacijo in njeno uporabo ter uporabnostjo. Navigacija po

katerimi ne, med katerimi celicami lahko ustvari prehod in kdaj aplikaciji je potekala brez vmesnih prekinitev in pomoči.

je labirint končan. Pri poznavalcu pa so vse celice pobarvane z Ustvarjale so predvsem srednje in težke labirinte z malo

belo barvo. Ta način je za uporabnika zahtevnejši, saj mora sam napakami. Napake so se večinoma nanašale na ponesreči dvojno

dobro premisliti, kateri dve celici lahko poveže med seboj, tj. kliknjeno isto celico labirinta ali na ponesreči kliknjeno sosednjo

ustvari prehod med njima, in med katerimi celicami že obstaja celico od zamišljene zaradi slabše koordinacije računalniške

pot v labirintu. miške. Ob prvotnem vrednotenju uporabnosti aplikacije za

Po dokončno ustvarjenem labirintu se uporabniku izpiše tudi ustvarjanje labirintov se zavedamo individualnih razlik med

stopnja težavnosti njegovega ustvarjenega labirinta. Poleg tega testirankami. Z rezultati vrednotenja smo zadovoljni.

pa se tudi omogoči gumb “Izvedi več”. Ob kliku nanj, se

uporabniku prikaže matematično ozadje ustvarjanja labirintov. V

labirintu se izrišejo vozlišča grafa oziroma drevesa in povezave 4 ZAKLJUČEK

ter prikaže za učence prilagojen opis definicije grafa. Poleg tega V prvem delu prispevka smo predstavili osnovne značilnosti

pa se omogoči tudi gumb “Izvozi”. Ob kliku nanj se uporabniku labirintov in njihovo povezavo s teorijo grafov. Predstavili smo

prenese slika njegovega ustvarjenega labirinta v formatu .png. nekaj algoritmov s katerimi lahko ustvarimo popolne labirinte.

Uporabniku se odpre novo okno, kjer izbere mesto v svojem Naredili smo tudi pregled vrednotenja težavnosti algoritmov, ki

računalniku, kamor bo shranil to sliko. so jih izvedli drugi avtorji. Nato smo opisali empirični del, kjer

smo predstavili svojo metodo za vrednotenje težavnosti

3.3 Preizkus aplikacije ustvarjenega labirinta. Predstavili smo tudi ugotovitve glede tega,

Aplikacijo smo preizkusili s tremi priložnostno izbranimi kateri algoritmi ustvarijo težje in kateri lažje labirinte. Nato smo

učenkami različnih starosti, ki jih naprej imenujemo testiranke. predstavili izdelano aplikacijo s katero se uporabnik uči

Prva je bila učenka sedmega razreda osnovne šole, druga učenka delovanja algoritmov za ustvarjanje labirintov. Na koncu pa smo

osmega razreda osnovne šole, tretja pa je bila dijakinja prvega predstavili rezultate preizkusa aplikacije, ki smo ga izvedli na

letnika gimnazije. Vse tri prihajajo iz osrednjeslovenske regije. treh priložnostno izbranih učencih.

Vsaka od testirank je preizkusila aplikacijo. Preizkus je potekal Primer se osredotoča na področje algoritmičnega ustvarjanja

tako, da smo testiranki najprej predstavili aplikacijo in njene labirintov, vendar si želimo, da bi naša aplikacija in pridobljena

funkcionalnosti, nato pa je tudi sama preizkusila uporabo. spoznanja vplivala tudi na širši razvoj računalniškega mišljenja,

Predstavitev aplikacije z naše strani je potekala frontalno in je zlasti v dimenzijah algoritmičnega mišljenja in validacije.

trajala 10 minut. Po končani predstavitvi so testiranke same

preizkušale aplikacijo. Preizkušanje je v povprečju trajalo 30 LITERATURA

minut. Po končanem preizkusu smo z vsako od testirank izvedli [1] „Maze,” Wikipedia. Pridobljeno: 19. marec 2024. Dostopno na:

polstrukturiran intervju o njeni izkušnji z aplikacijo. Tekom http://en.wikipedia.org/wiki/Maze

[2] „Labyrinth,” Wikipedia. Pridobljeno: 19. marec 2024. Dostopno na:

pogovora smo zastavili naslednja vprašanja: https://en.wikipedia.org/wiki/Labyrinth

[3] M. Foltin, „Automated maze generation and human interaction,”

1. Kaj ti je bilo pri aplikaciji (najbolj) všeč? magistrsko delo, Faculty Of Informatics, Masaryk University, Brno, 2011.

2. Kaj ti pri aplikaciji ni bilo všeč?



[4] A. Benjamin, G. Chartrand, in P. Zhang, The fascinating world of graph

theory, 1. izd. Princeton University Press, 2015.

doi:10.1515/9781400852000.

[5] J. Buck, Mazes for programmers: code your own twisty little passages,

The Pragmatic Bookshelf, 2015.

[6] „Maze generation algorithm,” Wikipedia. Pridobljeno: 25. marec 2024.

Dostopno na: https://en.wikipedia.org/wiki/Maze_generation_algorithm

[7] D. Mane, R. Harne, T. Pol, R. Asthagi, S. Shine, in B. Zope, „An extensive

comparative analysis on different maze generation algorithms,”

International Journal of Intelligent Systems and Applications in

Engineering, let. 12, št. 2s, str. 37-47, 2023. Dostopno na:

https://ijisae.org/index.php/IJISAE/article/view/3557

[8] P. Gabrovšek, „Analysis of Maze Generating Algorithms,” IPSI

Transactions on Internet Research, let. 15, št. 1, str. 23-30, 2019.

[9] V. Bellot idr., „How to generate perfect mazes?“, Information sciences,

let. 572, str. 444-459, 2021, doi: 10.1016/j.ins.2021.03.022.

[10] A. Nusser, „The Simultaneous Maze Solving Problems,” magistrsko delo,

Institute of Formal Methods in Computer Science, University of Stuttgart,

2016. Dostopno na: https://elib.uni-

stuttgart.de/bitstream/11682/9656/1/ausarbeitung.pdf

[11] A. Mejač, „Ustvarjanje in vrednotenje labirintov,” magistrsko delo,

Univerza v Ljubljani, Pedagoška fakulteta, Ljubljana, 2024. Dostopno:

https://repozitorij.uni-lj.si/IzpisGradiva.php?id=159506.



Šolska knjižnica v koraku s časom



School Library Keeping up with the Time



Tina Mlakar

Gimnazija Moste

Ljubljana, Slovenija

tina.mlakar@gmoste.si



wait until all stakeholders agree on a long-term strategy

POVZETEK regarding e-books in education.

V prispevku smo opisali uvajanje novosti s področja

knjižničarstva v okolje šolske knjižnice. Avtomatizirana KEYWORDS

sistemska sporočila kot takšna sicer niso novost, vendar so na Automation, applications, book locator, Cobiss Ela, Cobiss

novo vpeljana v našo šolsko knjižnico in so prinesla takojšno Lib, e-books, inventory, mInventura

izboljšanje pri rednem vračanju gradiva. Tudi pri inventuri smo

v preteklem šolskem letu uvedli novost in sicer uporabo

aplikacije mInventura, ki omogoča zajem inventarne številke 1 UVOD

preko mobilnega telefona, kar je pripomoglo k hitrejšemu Novosti s širšega področja knjižničarstva posegajo tudi v

odčitavanju gradiva. Vklop Lokatorja gradiva sicer zahteva kar šolske knjižnice. V preteklem šolskem letu smo pri delu v šolski

veliko dela s strani knjižničarja pri vnašanju podatkov a ko je knjižnici na Gimnaziji Moste uvedli kar nekaj novosti. Nekatere

enkrat vse vneseno, nam ni potrebno več nič dodajati, razen so pripomogle k lažjemu in hitrejšemu izvajanju postopkov in

sprememb, če karkoli premikamo v postavitvi knjižnice, kar pa morda niti niso tako več nove, smo jih pa letos prvič s pridom

se zgodi zelo redko ali pa gre le za spremembe v manjšem obsegu. koristili.

Nova programska oprema Cobiss Lib je sicer še v prehodnem

obdobju, vendar smo se v večini šolskih knjižnic že seznanili z

njo in jo tudi uporabljamo vzporedno s Cobiss3. Predstavljena 2 AVTOMATSKA SISTEMSKA SPOROČILA

platforma Cobiss Ela pa zaenkrat poraja še ogromno vprašanj Cobiss omogoča vnos e-poštnega naslova člana knjižnice, s

glede vključevanja šolskih knjižnic in bomo morali počakati, katerim se lahko uporabniki prijavijo v Cobiss+ ali na mobilno

dokler se vsi deležniki ne dogovorijo o dolgoročni strategiji glede verzijo mCobiss, kjer sami podaljšujejo in rezerviraj o gradivo.

e-knjig v šolstvu. Vnos e-poštnih naslovov poteka ročno, tako da uporabnika

odpremo in vtipkamo podatke. To se po navadi uporablja bolj za

KLJUČNE BESEDE popravljanje ali vnos novih naslovov, sicer pa koristimo paketni

Avtomatizacija, aplikacije, Cobiss Ela, Cobiss Lib, e-knjige, vnos preko exelove tabele. V tem načinu torej lahko vnesemo

inventura, lokator gradiva, mInventura veliko količino naslovov in drugih podatkov v le nekaj sekundah.

Ob paketnem vnosu program sam pošlje obvestilo uporabniku z

ABSTRACT več informacijami o nadaljnji nastavitvi gesla in prijavi v Mojo

We described how proffesional librarianship novelties are knjižnico. Ravno to je razlog, da sem ugotovila da je bolj

introduced into the environment of the school library. Automated smiselno vnašati vsak oddelek posebej in sicer tik pred uro

system messages as such are not a novelty, but they have been knjižnično informacijskega znanja. Na uri jim povem, da so

newly introduced in our school library and have brought an dobili e-pošto in že preverim neveljavne ali nove naslove, če kdo

immediate improvement in regular returns. In the past school morda ni dobil pošte. Nato pri uri vodimo dijake, da si nastavijo

year, we also introduced an innovation in the inventory, the use geslo in zakaj je to uporabno. Pri nas se dijaki rezervacije redko

of the mInventura application, which enables the capture of the poslužujejo, tudi podaljševanje največkrat opravijo z dogovorom

inventory number via a mobile phone, which helped to scan the pri knjižničarki.

inventory faster. Incorporating the Locator does require a lot of Prav tako vneseni e-poštni naslovi omogočajo dodatne opcije,

work on the part of the librarian when entering data, but once kot je avtomatsko obveščanje. Lahko povemo, da je bil učinek

everything is entered, we do not need to add anything, except for obvestil viden takoj, saj se dijaki redno oglasijo v knjižnici z

changes, if we move anything in the layout of the library, which vračilom gradiva, ki bi ga sicer pozabili vrniti dokler jih nismo k

happens very rarely or it is only changes on a smaller scale. The temu pisno ali osebno pozvali. Podobno ugotovitev glede porasta

new Cobiss Lib software is still in the transition period, but we podaljševanja gradiva so ugotovili tudi v drugih šolskih

have already familiarized ourselves with it in most school knjižnicah [1].

libraries and are also using it in parallel with Cobiss3. The Avtomatsko obveščanje smo na Gimnaziji Moste vklopili za

presented Cobiss Ela platform still raises a lot of questions vse vpisane dijake. Za zaposlene ta opcija ni vklopljena oz. je

regarding the inclusion of school libraries, and we will have to vklopljena le pri nekaterih testno. Vklop avtomatskega

obveščanja je sicer zelo preprost. Izberemo lahko več parametrov

avtomatskih sporočil, mi smo vklopili možnost avtomatskih

opomnikov sporočil o skoraj pretečenem roku izposoje. To

možnost najprej vklopimo znotraj segmenta »Izposoja« na

»Domači knjižnici«, alineja »Elektronsko obveščanje«, kar

prikazujeta sliki 1 in 2. Seveda lahko označimo tudi druga

avtomatska obvestila, ki jih želimo aktivirati za pošiljanje (o

poteku rezervacije, o prispelem rezerviranem gradivu, o

skorajšnjem opominu, o poteku članstva, ipd.





Slika 3: Izbira podatkov o članih



3 MINVENTURA

Letos nas je ponovno obiskalo leto v katerem je previdena

inventura, ki se izvaja na vsakih 5 let ali manj. Podvig, ki poleg

fizičnega pregleda polic zahteva tudi poznavanje zakonske

podlage [5]. Pri fizičnem delu so na Izum-u pripravili aplikacijo,

ki zadevo olajša. Aplikacija mInventura sedaj omogoča, da

Slika 1: Elektronsko obveščanje

pametni telefon uporabljamo namesto skenerja, saj kamera

zajame kodo inventarne številke in jo zabeleži. Aplikacija je res

preprosta, enostavno si jo naložimo na telefon, prijavimo se z



uporabniškim imenom in geslom, ki ga imamo odprtega pod našo



knjižnico in nato lahko že pričnemo s skeniranjem (prikaza na

sliki 4).





Slika 2: Izbor avtomatskih obvestil





Ko smo sistemsko uredili avtomatska sporočila, moramo to

možnost omogočiti tudi članom. V tem primeru lahko zopet

urejamo in vnašamo e-poštni naslov posamezno, preko kartice

odprtega člana. Lahko pa uporabimo paketno metodo za Slika 4: Aplikacija mInventura in primer odčitavanja

spreminjanje podatkov o članih [3], pri kateri je sedaj dodana

možnost spreminjanja nastavitev e-obveščanja (kot prikazuje Aplikacija je zelo dobrodošla, saj je odčitavanje s telefonom

slika 3). Torej podatke o članih najprej izvozimo, pri tem preprosto za uporabo, ne potrebujemo dodatnih skenerjev niti

izberemo podatek »Obv. o poteku roka izposoje – E-Pošta«. Nato prenosnikov, hkrati lahko odčitava več zaposlenih, ki jim

v Exel datoteki, ki jo program ustvari, vpišete pri tem podatku predhodno seveda uredimo dostope preko Izum portala

pri vseh članih številko »1«, to pomeni potrjevanje prejema Izobraževanje.

sporočil. Tako urejeno datoteko nato uvozimo in s tem so podatki Pri odčitavanju s telefoni smo opazili le eno slabost.

spremenjeni in obvestila omogočena. Seveda lahko hkrati Aplikacija namreč ne javlja izposojenega gradiva na polici (to

izberemo in urejamo več podatkov. vidimo šele naknadno po pregledu izpisov in moramo ponovno

poiskati gradivo, ki smo ga že imeli v rokah) enako je pri gradivu

s statusom izgubljeno ali odpisano. Z našega vidika izguba je to

časa, saj zahteva, da po pregledu seznamov ponovno iščemo to

gradivo na polici in bi si želeli v prihodnosti tudi to izboljšavo.

Kljub temu pa je inventura na ta način olajšana in tekoče deluje.

Pri nas smo inventuro izvedli v kombinaciji, torej uporaba

aplikcije mInventura in direktno odčitavanje s skenerjem v

okolju Cobiss3.



4 LOKATOR GRADIVA

Vzpostavitve Lokatorja gradiva smo se lotili predvsem, ker

nas je zanimalo kako poteka vnos v aplikacijo in končna uporaba.

Prav nas je zanimalo ali bo končni izdelek izboljšal uporabniško

izkušnjo dijakov, ter nenazadnje tudi pripomogel pri razlagi in

predstaviti uporabe knjižnice pri urah knjižnično informacijskih

znanj.

Aplikacijo sestavljata dva dela. Prvi je namenjen knjižnici oz.

knjižničarju, ki pripravi tloris prostora. Drugi del je namenjen

uporabniku, ki ob iskanju lahko izbere vizualni prikaz kje točno

se nahaja izbrano gradivo. V Lokator gradiva lahko vnašamo več

nadstropij, predvidevam, da bi lahko namesto nadstropij vnašali

tudi različne lokacije, če je šolska knjižnica razporejena v

različne prostore ali stavbe.

Aktivacija Lokatorja zahteva kar nekaj predpriprave in časa.

Najprej si je potrebno premeriti in zrisati prostor, površino tal, Slika 6: Dodatni vnosi starih UDK vrstilcev

vse police in druge elemente, ki jih bomo dodali ter sledenje

UDK po policah. V veliko pomoč je bil tloris knjižnice, ki smo Seveda smo te dodatne vnose preverili, kako delujejo v praksi.

ga pridobili pri hišniku, tako da so bile potrebne le meritve Spodaj je prikazano kako se izkaže takšna rešitev pri iskanju

velikosti polic. Velikost miz, stolov in rastlin je v aplikaciji gradiva s strani uporabnika (sliki 7 in 8).

dodeljena, ostale elemente lahko prilagajate. Vse osnovne

informacije glede uporabe aplikacije vnosa podatkov pridobimo

v navodilih, ki jih je pripravil IZUM [2]. Po izrisu tlorisa, polic

in ostalih elementov (prikaz na sliki 5) je potrebno vnašanje

podatkov na police, to so tisti podatki, ki bodo iskanje v

COBISS+ pravilno usmerili v 3D prikazu.





Slika 7: Omogočanje prikaza s staro UDK oznako





Slika 5: Vnos tlorisa in pohištva





UDK in interne oznake so tisti ključni podatki, na podlagi Slika 8: Prikaz z novo oznako UDK



katerih se bo kasneje aplikacija povezala z bazo COBISS+ za

V obeh primerih gre za avtorja s črko B in spada pod

prikaz iskanega gradiva. Pri nas se je pojavila težava, namreč v

slovensko leposlovje. Prvi ima številčni UDK »821.163.6-3«,

uporabi so nove UDK postavitve, še vedno pa so na velikem

drugi pa novejšo oznako, ki vsebuje tudi besedilni opis »82 slo.

številu gradiva dodeljeni stari UDK vrstilci. Sčasoma smo s

proza« . Če starega UDK nebi vnesli, gradivo poleg nebi

preizkušanjem aplikacije našli prehodno rešitev. V aplikacijo

prikazovalo ikono Lokatorja. S takšno rešitvijo pa je obeh

smo kot dodatni UDK dodali stari številčni UDK vrstilec k vsaki

primerih lokator povezan s COBISS+ in prikazuje pravo lokacijo

polici, kjer se takšno gradivo lahko nahaja. Torej na primeru

knjige tudi v 3D izrisu (slika 9).

leposlovja pri slovenski prozi smo na vse police vključila dodatni

UDK (prikaz na sliki 6).



Novo okolje je vizualno prenovljeno, čeprav je logika dela

zelo podobna, le da tu vidimo vse možnosti že odprte. Kot pri

vsaki novosti, nas tudi tu čaka obdobje izboljšav in nadgradnje

na podlagi izkušenj in uporabe aplikacije v praksi.



6 COBISS ELA

Da bo slovenski prostor dočakal uresničitev platforme, ki bi

združila vse založnike in ponudbo e-knjig za izposojo smo prvič

slišali na Cobiss dnevih decembra 2023, ki naj bi začela delovati

takoj pomladi. Junija 2024 je IZUM na online konferenci

Slika 9: Vizualni prikaz iskanja gradiva Novosti v knjižničarstvu podrobno predstavil novo platformo

imenovano Cobiss Ela [7]. Novomeška knjižnica Mirana Jarca je



5 COBISS LIB in vsebinsko pomoč, pripravili so tudi navodila za bralce pri nacionalna koordinatorka platforme in knjižnicam nudi poslovno

Še ena novost, ki smo jo v knjižnicah letos sprejemali različno, izposoji in namestitvi aplikacije [6]. IZUM pa je prevzel tehnično

saj je zaenkrat še v prehodnem obdobju in to je novo okolje podporo.

Cobiss Lib. Vse kar smo sedaj izvajali preko Cobiss3 »Platforma COBISS Ela omogoča celovit tokokrog e-knjige

programske opreme, se sedaj deli na dva segmenta, in sicer od založnika do bralca« [4]. Platforma namreč založnikom

Cobiss Cat in Cobiss Lib. Prvi je namenjen katalogizatorjem za omogoča dodajanje e-knjig v formatu ePUB, pri čemer vsaki e-

obdelavo gradiva, drugi pa zajema ostale funkcije, ki jih šolske knjigi posebej določi pogoje izposoje (število točk za izposojo,

knjižnice potrebujemo pri vsakdanjem delu [8]. Glavna novost je kar predstavlja vrednost posamezne izposoje in veljavnost

ta, da Cobiss Lib ne potrebuje namestitve vmesnika, ker je licence. Izposoja e-knjig že deluje in sicer imamo možnost branja

dostopen preko kar preko brskalnika. Torej v spustnem seznamu na osebnem računalniku in na pametnih telefonih oz. mobilnih

poiščemo našo šolo, nato pa se vpišemo z uporabniškim imenom napravah, tudi bralniki naj bi bili omogočeni. Uporabiti je

za vstop v pravo okolje in že lahko pričnemo z delom (prikazuje potrebno aplikacijo. V praksi je trenutno branje preko telefona

slika 10). ali druge mobilne naprave onemogočeno. Ta ukrep so izvedli

zaradi varnostnih očitkov uporabe aplikacije Cantook, zato bo



potrebna nadgradnja ali omogočanje nove aplikacije za odpiranje

datotek e-knjig. Projekt ima še kar nekaj težav, ki se nam zdijo

tako kot pri vsakem novem projektu, sploh tako obsežnem,

normalni in verjamemo, da bodo ovire sčasoma presežene.

Posodobitve in nadgradnje so objavljene na spletni strani

namenjeni bralcem https://www.cobiss.si/bralci/ela/, kjer je

IZUM dodal tudi pomoč v obliki video navodil.

Četudi Cobiss Ela ne deluje še s polno paro, se nam že poraja

vprašanje kako bo z e-knjiga v šolskih knjižnicah. Vse kar smo

uspeli izvedeti je, da za področje šolskih knjižnic zaenkrat

potekajo dogovarjanja z ministrstvom. V juniju 2024 je šolski

Bbiblos še vedno deloval, vendar je logična poteza, da bi v

naslednjem letu v šolskih knjižnicah omogočili vsaj domača

branja preko Cobiss Ela.



7 ZAKLJUČEK

Slika 10: Vstop v Cobiss Lib preko brskalnika V šolskih knjižnicah je torej kar nekaj možnih novosti, ki jih

uvajamo glede na potrebe in uporabo. V praksi so se nekatere

Izum je za pomoč knjižničarjev pripravil tudi navodila za delo zadeve pokazale kot zelo praktične npr. avtomatska sporočila, ki

v novi aplikaciji, predstavitvene posnetke, online izobraževanja, pripomorejo k rednemu vračanju izposojenega gradiva. Nekatere

ter anketo o uporabi in predlaganih izboljšavah. zadeve se sprva ne zdijo zelo uporabne v šolski knjižnici, to je

Pri delu s Cobiss Lib smo knjižničarju ugotovili in sporočili lokator gradiva, a je vzpostavitev vseeno smiselna, saj se zadeva

kar nekaj predlaganih optimizacij in prilagoditev, še vedno pa je lahko nadgradi in se uporabnost pokaže šele v prihodnjih letih.

zadeva prehodnem obdobju do konca novega leta, ko se stari Je pa tu potrebna vključitev v pouk, da se dijake nauči na to

program Cobiss3 preneha uporabljati. Kar nekaj komentarjev je možnost iskanja in prikazovanja kot pomoč pri samostojnem

prišlo na počasno odzivanje, najverjetneje zaradi obremenjenost iskanju gradiva, kar vidimo kot največjo dodano vrednost. V

mreže, prav tako so nekatere funkcije sedaj drugačne zato ker gre knjižnicah smo kar navajeni novosti, zato menimo, da se tudi

za brskalnik. Kljub temu se hitro lahko navadimo delu v novem nove programske opreme kaj hitro privadijo in predlagamo

okolju, je pa res, da je potrebnih več klikov za ukaze in za vnos izboljšave iz prakse. V prihodnosti pričakujemo predvsem

podatkov. novosti in spremembe na področju e-knjig.





LITERATURA IN VIRI


[1] Česen Šink, Tadeja. 2024. Tudi v šolskih knjižnicah lahko člani preko

Cobiss+ podaljšujejo in rezervirajo gradivo. Dostopno na naslovu

https://knjiznicarske-novice.si/tudi-v-solskih-knjiznicah-lahko-clani-

preko-cobiss-podaljsujejo-in-rezervirajo-

gradivo/?utm_medium=email&utm_source=newsletter_1445&utm_cam

paign=nova-welcome-to-nuk&utm_id=1445 (14. 8. 2024)

[2] Institut informacijskih znanosti IZUM. 2022. Navodila za uporabo

aplikacije Lokator gradiva. Dostopno na naslovu:

https://home.izum.si/izum/e-prirocniki/COBISS-

plus/Navodila_za_uporabo_aplikacije_Lokator_gradiva.pdf (14. 8. 2024)

[3] Institut informacijskih znanosti IZUM, 2022. Videoposnetek Paketni vpis

E-naslova za e-obveščanje. Dostopno na naslovu

https://www.youtube.com/watch?v=1AsezT9T_GA&list=PLeBQsIfZEz

UMwkG857vqKdB_T_aGxf1X1&index=17 (14.8.2024)

[4] Institut informacijskih znanosti IZUM, 2024. Cobiss Ela. Dostopno na

naslovu: https://www.cobiss.si/bralci/ela/ (14. 8. 2024)

[5] Jesih Šterbenc, Katarina. 2023. Inventura v šolski knjižnici – primer dobre

prakse. Dostopno na naslovu: https://knjiznicarske-novice.si/inventura-v-

solski-knjiznici-primer-dobre-prakse/ (14. 8. 2024)

[6] Knjižnica Mirana Jarca, 2024. Nova prva javna platforma za izposojo e-

knjig. Dostopno na naslovu:

https://www.nm.sik.si/si/aktualno/detajl/?id=5651 (14. 8. 2024)

[7] Narodna in univerzitetna knjižnica, 2024. Novosti v knjižničarstvu.

Predstavitev platforme Cobiss Ela za izposojo e-knjig. Dostopno na

naslovu:

https://www.youtube.com/watch?v=iOEor8UxuJQ&list=PLEGos7sOX-

2GdE505kpRa5hbRZW4iNxCA&index=4 (14. 8. 2024)

[8] Stoviček, Jerneja. 2023. Nova programska oprema Cobiss Lib. Dostopno

na naslovu: https://knjiznicarske-novice.si/nova-programska-oprema-

cobiss-lib/ (14. 8. 2024)



Odkrivanje naravoslovnih konceptov skozi igro:



Kombinacija tradicionalnih in digitalnih metod v vrtcu



Discovering Science Concepts through Play: A Combination of



Traditional and Digital Methods in Kindergarten



Renata Mlinar

Vrtec Litija,

Litija, Slovenija

renata.mlinar@vrtec-litija.si



POVZETEK KEYWORDS

Digitalna tehnologija ima številne pozitivne učinke na Preschool children, digital technology, new approaches, early

vzgojno-izobraževalni proces v vrtcu, kjer postaja nepogrešljiv science, safe use of the Internet, creativity, children's motivation,

pripomoček pri raziskovalnem delu z otroki. Varna in digital literacy

premišljena raba interneta, računalnika, pametnega telefona,

fotoaparata in projektorja omogoča vzgojiteljem boljšo

motivacijo otrok ter spodbujanje njihovega zanimanja za 1 UVOD

odkrivanje novih vsebin. Digitalna tehnologija je postala del našega življenja. Na nek

V vrtcu prepoznavamo več prednosti uporabe digitalne način nas spremlja in v veliki meri pomaga pri naših opravilih v

tehnologije. Otroci so naravno radovedni in tehnologija lahko to prostem času in v službi. Tako je tudi v vrtcu nepogrešljiva pri

radovednost še dodatno spodbudi. Zelo pomembno je zagotoviti izvedbi mnogih nalog. Če primerjam delo pred nekaj leti in sedaj,

uravnotežen pristop k digitalnim aktivnostim. S skupinskim si ga sedaj zelo težko predstavljam brez digitalne tehnologije.

delom na projektih, kjer otroci uporabljajo različne digitalne Digitalni pripomočki, kot so fotoaparat, računalnik, mobilni

naprave, se učijo sodelovanja, komunikacije in reševanja telefon in projektor, so nepogrešljivi pripomočki, ki nam

problemov. Internet je bogat vir informacij, ki jih lahko vzgojiteljem pomagajo pri ustvarjanju spodbudnega učnega

vzgojitelji uporabimo za razširitev učnih vsebin. V prispevku okolja v vrtcu. V Kurikulu, ki je naš nacionalni dokument,

bom opisala varno rabo digitalne tehnologije v vrtcu, povezano s imamo zapisane cilje - le te uresničujemo na različne načine.

tradicionalnimi načini podajanja znanja. Iščemo zanimive in spodbudne teme, v načrtovanje vključimo

ideje otrok. Odprtost Kurikula zahteva avtonomnega

KLJUČNE BESEDE strokovnega delavca, ki zna ustvarjati spodbudno učno vzgojno

Predšolski otroci, digitalna tehnologija, novi pristopi, zgodnje okolje, v katero spada tudi smiselna uporaba digitalne

naravoslovje, varna raba interneta, kreativnost, motivacija otrok, tehnologije.[1]

digitalna pismenost



ABSTRACT 2 PREDŠOLSKI OTROCI V DIGITALNI DOBI Digital technology has many positive effects on the

educational process in kindergarten where it is becoming an Otroci v digitalni dobi imajo drugačne potrebe, kot so jih

indispensable tool in research work with children. The safe and imeli pred leti, ko se je tehnologija začela razvijati. Naša naloga

thoughtful use of the internet, computers, smartphone, cameras je, da otroke vzgajamo v duhu digitalnega napredka in jim

and projector enables educators to better motivate children and pomagamo izbrati dobro od slabega. Vsak otrok mora imeti stik

stimulate their interest in discovering new content. s tehnologijo, pomembno je le, v kolikšni meri mu bomo to

In kindergarten, we recognize several advantages of using dopustili. Postaviti moramo meje. Mediji imajo pomembno

digital technology. Children are naturally curious, and vlogo v življenju otroka, vendar ne morejo razviti zaupanja,

technology can further encourage that curiosity. It is very navezanosti, veselja in odkrivanja pravljičnih svetov, saj se vse

important to ensure a balanced approach to digital activities. to lahko pojavi le ob komunikaciji med dvema osebama. [2]

Group work on projects where children use different digital Otroke moramo usmerjati v uporabo digitalne tehnologije v učne

devices teaches them cooperation, communication and problem namene, ne le zgolj zabavne. Digitalna tehnologija nam nudi

solving. The internet is a rich source of information that širok spekter pripomočkov, ki jih moramo uporabiti in izkoristiti

educators can use to expand learning content. In this paper, I will v namene izobraževanja otrok. Nemalokrat se pri otrocih poraja

describe the safe use of digital technology in kindergarten, kakšno vprašanje, na katerega nimamo primernega odgovora. V

connected with traditional ways of imparting knowledge. takšnem primeru ni nič narobe, če otrokom odkrito povemo, da

odgovora ne poznamo. Lahko jih spodbudimo k razmišljanju, kaj 3.2 IZDELAVA PLAKATA



naj v takšnem primeru naredimo. V večini primerov otroci O vsem, kar so videli na posnetku, smo se pogovorili in predlagajo, da pobrskamo po internetu. Takšna vprašanja lahko izdelali plakat z značilnostmi vode, ki so jih že poznali (slika 2). privedejo do izvedbe raziskovalnih nalog, v katerih otroci z Otroci potrebujejo zelo konkretno razlago, zato sem se odločila, veseljem sodelujejo, ker so bili s svojim vprašanjem pobudniki da vsa agregatna stanja vode spoznajo s pomočjo preprostih raziskovalne teme. Delam v skupini otrok, starih od 2-6 let. naravoslovnih eksperimentov. Skupina te starosti je z vzgojnega vidika zelo zahtevna, tako pri

načrtovanju vzgojnega dela, kot pri izvedbi dejavnosti. Smo zelo

povezani in delujemo kot družina, starejši pomagajo mlajšim,

mlajši se učijo od starejših in hitreje pridobivajo na samostojnosti.

Kljub razlikam v starosti smo se lotili kar nekaj raziskovalnih

nalog, v katere smo vključili tudi uporabo digitalne tehnologije.

Pri načrtovanju sem sledila ciljem iz Kurikula in upoštevala

vzgojna načela.



3 AKTIVNOSTI OTROK

Opisala bom primer raziskovalne naloge, kjer so otroci preko

aktivnega sodelovanja spoznavali naravne zakonitosti in se

seznanili z uporabo digitalnih pripomočkov.



3.1 IZDELOVANJE OPTIČNIH IGRAČ-

SLIKOFRCEV

V vrtcu smo imeli predstavitev zavoda Vizo, ki se ukvarja s

filmsko vzgojo v vrtcih. Skupaj z otroki so izdelali optične Slika 2: Plakat-značilnosti vode

igrače-slikofrce in predstavili, kako posnamemo enostaven film.

Starejšim otrokom je bila vsebina zanimiva, zato so si želeli 3.3 POSKUS IZHLAPEVANJA VODE

izdelati več slikofrcev, s pomočjo katerih so predmeti na sličicah Vse, kar smo zapisali na plakat, so tudi preizkusili. Izvedli so

oživeli (slika 1). Naslednji dan je padal dež. Ena izmed deklic v poskus izhlapevanja vode, tako da so merili lužo in jo opazovali,

moji skupini je tiho opazovala dežne kapljice in me vprašala, kako se suši. Isti poskus so naredili tudi z vodo, ki so jo postavili

zakaj pada dež. To vprašanje je bilo naše izhodišče za začetek na teraso. V dva enaka kozarca so natočili vodo in označili nivo

raziskovalne naloge. Nekaj otrok, ki so slišali dekličino vode. Enega so zaprli s plastično folijo, drugega so pustili

vprašanje, je začelo podajati odgovore. Vprašala sem jih, če vedo, odprtega (slika 3). Poskus je trajal nekaj dni. Na plastični foliji

kako in zakaj pada dež na zemljo in kaj se s to vodo zgodi. so se nabrale kapljice, na odprtem kozarcu pa je bil nivo vode

Odgovori otrok so bili zelo zanimivi, vseeno pa smo si skupaj nižji. Vse to so otroci že predvideli, ker so nekaj dni prej

ogledali zelo nazoren posnetek o skrbi za čisto vodo in kroženje opazovali lužo in znali razložiti, da je s pomočjo sonca voda

vode v naravi, ki smo ga poiskali na internetu [2]. izhlapela (slika 4).





Slika 1: Igra s slikofrci



Slika 3: Izhlapevanje vode v kozarcu





Slika 4: Merjenje luž

Slika 6: Risanje z ledom, opazovanje taljenja ledu.



3.4 POSKUS ZMRZOVANJA VODE

3.5 POSKUS IZPAREVANJA VODE



odgovorili, da nastane led. Seveda so tudi to preizkusili. Vodo so izparevanja vode. V večjo posodo s steklenim pokrovom smo nalili vodo in jo zavreli. Otroci so lahko opazili paro, ki se je natočili v vrečke za izdelovanje ledenih kock in v dva lončka, nabirala na steni pokrovke (slika 7). Naslednji dan, ko se je vse enega so pustili odprtega, drugega so zaprli. Ugotovili so, da zgodi z vodo, če jo damo v zmrzovalnik. Vsi starejši otroci so Z njimi sem izvedla tudi tako imenovan »umetni dež« poskus Za četrto, trdno agregatno stanje, sem otroke vprašala, kaj se

ohladilo, so kapljice, ki so se nabrale na pokrovu, kapljale dol kot

voda zmrzne enako v zaprtem, kot odprtem lončku (slika 5). Z

dež.

ledenimi kockami so nato risali po toplih tleh in opazovali

taljenje ledu (slika 6).



Slika 7: Izparevanje vode-umetni dež.



S pomočjo eksperimentov in filma o kapljici so otroci

razumeli, da voda v naravi kroži. Odločili so se, da skupaj

napišemo zgodbo o kapljici vode. Zgodba je nastajala nekaj dni

v jutranjem bralnem kotičku. Najprej so izbrali ime glavne





Slika 5: Poskus zmrzovanja vode. zanimive ideje, v ospredju je bilo vseskozi kroženje vode. V junakinje kapljice-Plavka,ki je živela v močvirju. Nastajale so



pogovoru sem otroke vodila z vprašanji odprtega tipa. Vse



njihove odgovore sem zapisala. Kapljica Plavka je potovala po

vseh mogočih krajih in nastala je zgodba, na katero so bili zelo

ponosni. Naslov so ji dali Plavkino potovanje.

3.6 umirjeni. Starejši otroci so s tehniko risanja s črnim flomastrom Zgodba: PLAVKINO POTOVANJE

(AVTORJI: otroci skupine Metulji in in barvanja s flomastri narisali vse like in predmete iz zgodbe

vzgojiteljica Renata Mlinar) (slika 9). Jaz sem jih izrezala. Potrebovali smo več kosov enake

figure, zato so bili otroci zelo zaposleni in ustvarjanje je trajalo

Kapljica Plavka je živela v močvirju. Tam je živela tudi več dni.

žabica Regica. Rada je regala in skakala iz lokvanja na lokvanj

in tam srečala prijateljico Plavko. »Skoraj si me pohodila!«

»Oprosti. Nisem te videla.« Žabica je bila osamljena in se je zelo

rada igrala s Plavko. Nekega dne pa je močno sijalo sonce in

posušilo kapljico. Z meglico je odšla do oblakov, kjer je spoznala

veliko prijateljic kapljic. Spoznala je tudi dva glasna moža –

Bliska in Groma. Začelo jo je zelo hladiti, ropotalo in bobnelo je.

Začelo je deževati. Plavka je padla na rito v ledeno deželo, kjer

je srečala pingvine. Bila je zelo vesela. Spomnila se je na

pesmico »Si že videl kdaj, pingvine piti čaj« in jih to pesem tudi

naučila. Pingvini so bili zelo veseli in so Plavko spravili v iglu.

Tam se je spremenila v led. V deželo pingvinov so prišli tudi

ribiči, ki so potrebovali led za ribe. Prišli so do pingvinov in jih

vprašali: » Nam odstopite malo ledu?« »Seveda.« Dali so jim kos

ledu v katerem je bila Plavka. Odnesli so ga na ribiško ladjo. Led

so zdrobili na koščke in ga posuli po ribah. Ladja je priplula v

Afriko, kjer je zeloooo vroče. Ribe so odpeljali na tržnico. Sonce

je močno sijalo in Plavka se je zopet stopila. Odpotovala je nazaj

v oblake k prijateljicam kapljicam. Skupaj so se veselile, da bodo

kmalu zopet odšle na potovanje proti zemlji.



3.7 PRIPRAVA MATERIALA ZA FILM Slika 9: Risanje likov za film

Preko vseh poskusov, ki so jih izvedli in s pomočjo zgodbe o

kapljici so otroci pridobili veliko znanja. To znanje so povezali s Ko je bil ves material za film pripravljen, so otroci pričeli s

predstavljeno animacijo zavoda Vizo. Odločili smo se, da polaganjem sličic po vsebini zgodbe in fotografirali vsak prizor

zgodbo oživimo. Spomnila sem jih na gibljive slike-slikofrce in (slika 10). Pri tem sem jim pomagala, kadar je bilo potrebno. Pri

animacijo s sličicami. Razdelili smo si delo. Mlajši otroci so s rokovanju s fotoaparatom so bili zelo spretni.

pomočjo odraslega pripravili podlago -mlako.





Slika 8: Ustvarjanje-osnova za pripravo animacije Slika 10: Polaganje sličic, fotografiranje





Uporabili so akvarelne barve. S tehniko odtiskovanja s Možnost fotografiranja sem omogočila vsem starejšim



penasto gobico so ustvarili mlako, senčili so jo s temnejšimi otrokom. Pokazala sem jim, kako fotografijo oblikujemo, jo

odtenki modre barve (slika 8). Dodali so še nekaj mivke in s tem povečamo, zmanjšamo. Ob tem so pridobivali izkušnje z

ustvarili plastičnost. Pri tem so zelo uživali, bili so vztrajni in upravljanjem miške (slika 11).

svoje delo in trud, kar pomembno vpliva na njihov osebnostni

razvoj.

Vse te aktivnosti dokazujejo, da lahko digitalna tehnologija,

če je pravilno uporabljena, pomembno prispeva k izobraževanju

in razvoju predšolskih otrok in jih hkrati pripravlja na digitalno

pismenost, ki je vse bolj pomembna v sodobnem svetu.



LITERATURA

[1] Smernice za uporabo digitalne tehnologije v vrtcu, ZRSŠ 2021, dostopno

na naslovu: https://www.zrss.si/pdf/DTsmernice_vrtci.pdf

[2] Predšolski otrok v digitalni dobi (Jurka Lepičnik in Maja Hmelak) 2022

Lepši svet-kroženje vode(animacija): https://youtu.be/e8x4AXYpiI0





Slika 11: Urejanje fotografij z računalniško miško



Kar nekaj otrok je imelo težave pri manipulaciji z miško. Ob

večkratni uporabi se je stanje izboljšalo. Nastale so odlične

fotografije, ki smo jih vstavili v program Stop motion in zgodba

o kapljici Plavki je oživela.



4 ZAKLJUČEK

Dejavnosti z uporabo digitalne tehnologije so predšolskim

otrokom zelo vabljive, zato jih je pomembno usmerjati k smiselni

a omejeni uporabi. Popolna prepoved ni primerna, saj lahko

tehnologijo koristno vključimo v izobraževalne procese.

Digitalni pripomočki, kot so računalniki in fotoaparati,

pritegnejo otroke z visokim nivojem koncentracije in pozornosti,

saj vključujejo več čutil hkrati. Z uporabo tehnologije lahko

izobraževalne teme nadgradimo in naredimo bolj privlačne.

Skozi temo naravoslovja so otroci razvijali medsebojno

komunikacijo in komunikacijo z odraslimi ter s tem izboljšali

skupinsko povezanost. Naučili so se poslušati in spoštovati

vrstnike, kar je ključnega pomena za njihov socialni razvoj.

Pri ustvarjalnih dejavnostih so krepili svoj besedni zaklad in

domišljijo. Ena izmed uspešnih aktivnosti je bila izdelava poučne

zgodbe, kjer so otroci uporabili različne likovne tehnike za

upodobitev živali in predmetov, ki so jih nato vključili v nastanek

filma. Pri tem so sodelovali vsi otroci, od najmlajših do

najstarejših. Pridobili so občutek pripadnosti skupini in uspeh pri

doseganju skupnega cilja.

Zelo učinkovit pristop pri pridobivanju novega znanja in

veščin so bili naravoslovni eksperimenti, ki so jih izvajali.

Pridobili so sposobnost samostojne razlage enostavnega kroženja

vode v naravi. Naučili so se rokovanja z računalniško miško in

osnov urejanja fotografij. Ob teh aktivnostih so bili visoko

motivirani ter miselno in fizično aktivni. Razvijali so

koordinacijo oko-roka-ekran in akomodacijo oči. Spoznali so

osnove snemanja filma, dejavnost, ki jo bodo lahko še naprej

razvijali. Ponosni so bili na svoje izdelke, okrepili svojo

samozavest in občutek, da jim je uspelo. Naučili so se vrednotiti



Stolp v Pisi z vidika likovne umetnosti, matematike in



informatike



The Tower of Pisa from the Perspective of Visual Arts,



Mathematics, and Ccomputer Science



Alenka Močnik

Srednja šola Veno Pilon Ajdovščina

Cesta 5. maja 12,

Ajdovščina, Slovenija

alenka.mocnik@ss-venopilon.si



the degree of the tower's tilt or predict when it would collapse

POVZETEK without proper restoration. We utilized appropriate technology to

Že vrsto let na Srednji šoli Veno Pilon Ajdovščina v 1. facilitate the work. Later, in visual arts, the students created

letnikih gimnazijskega programa izvajamo ekskurzijo v Piso in models of the Leaning Tower of Pisa and made animations of the

Firence. Arhitekturno in zgodovinsko zelo pomembni mesti smo tower's construction.

poleg likovne umetnosti dijakom predstavili še z matematičnega

vidika ter pri tem vključili pridobljena znanja iz informatike. KEYWORDS

Stolp v Pisi so gradili med 12. in 14. stoletjem, vendar se je zaradi Digital literacy, information and communication technology,

mehkega terena skozi celotno fazo gradnje nagibal. Kasneje so informatics, fine art, mathematics, interdisciplinary course,

ga ustrezno utrdili in s tem preprečili, da bi se porušil. Dijaki so mathematical modeling

na terenu spoznali zgodovinsko in likovno ozadje stolpa ter se

preizkusili v fotografiji. Pri pouku matematike pa smo narisali

graf nagnjenosti stolpa v odvisnosti od leta ter zapisali ustrezno 1 UVOD

prilagoditveno funkcijo. Zapisali smo najustreznejši Medpredmetno in timsko poučevanje učiteljem predstavlja

matematični model in tako npr. odčitali za koliko stopinj je stolp velik izziv. Poleg iskanja idej za predstavitev snovi na drugačen,

nagnjen ali kdaj bi se brez ustrezne sanacije podrl. Pri izvedbi avtentičen način, je tovrsten pristop vsakič inovativen. Povezati

smo se posluževali ustrezne tehnologije, ki nam je delo olajšala. se s predmeti, ki se zdijo sprva nezdružljivi in kjer se zdi, da je

Kasneje so dijaki pri likovni umetnosti izdelovali modele nemogoče poiskati skupni imenovalec zato zahteva podrobnejšo

poševnega stolpa v Pisi ter ustvarjali animacije izgradnje stolpa. organizacijsko in dobro snovno pripravo.

V 1. letniku gimnazijskega programa dijaki običajno obiščejo

KLJUČNE BESEDE Firence in Piso z namenom spoznati arhitekturno in kulturno

Digitalna dediščino. Pri pouku likovne umetnosti se dijaki seznanijo z pismenost, informacijsko-komunikacijska

tehnologija, informatika, likovna umetnost, matematika, zgodovinskim ozadjem ter umetniškim pomenom toskanskih

medpredmetna povezava, matematično modeliranje mest. S profesorico likovne umetnosti Damjano Plešnar in

profesorico informatike Urško Kompara Žvokelj smo želele, da

ABSTRACT so še posebej pozorni na poševni stolp v Pisi. Dijaki so imeli

For many years, at Veno Pilon Hight School in Ajdovščina, nalogo, da fotografirajo stolp iz različnih zornih kotov in pri tem

we have been organizing an excursion to Pisa and Florence for upoštevajo osnovna načela in pravila fotografijo. Fotografije smo



to the students from a mathematical viewpoint, incorporating dinamično geometrijo in za risanje grafov, in sicer prosto dostopna programa GeoGebra in Graph. Kasneje so pri likovni their knowledge of computer science. The Tower of Pisa was umetnosti ter informatiki ustvarjali animacije gradnje stolpa v constructed between the 12th and 14th centuries, but due to soft Pisi s pomočjo aplikacije StopMotion Studio. ground, it leaned throughout the entire construction phase. Later, Dijaki so na ta način pridobili nova znanja, tudi v kontekstu it was appropriately reinforced to prevent it from collapsing. On- vseživljenjskega učenja. Tak način dela ima tudi močan site, the students learned about the historical and artistic introducing these architecturally and historically significant ustrezen model za prilagoditveno funkcijo, ki opisuje nagnjenost stolpa v odvisnosti od leta. Dijaki so uporabili programe za cities from the perspective of visual arts, we also presented them first-year students of the gymnasium program. In addition to potem uporabili za matematično modeliranje, saj smo iskali

background of the tower and practiced photography. In math motivacijski učinek, saj dijakom predstavimo neko snov

class, we drew a graph showing the tower's inclination over time multiperspektivno.

and derived the corresponding regression function. We identified Drugo in tretje poglavje sta povzeti po [7].

the most suitable mathematical model to, for instance, determine



2 Z medpredmetnim sodelovanjem omogočimo, da zadane cilje MEDPREDMETNO POUČEVANJE



2007) sta tudi povezovanje predmetov in disciplin ter holističnih Med načeli in cilji posodabljanja učnih načrtov (Smernice, dosežemo lažje, saj jih posamezen profesor v okviru svojega predmeta ne more doseči tako dobro in poglobljeno, kot kadar sodeluje s profesorjem drugega predmetnega področja. Tovrsten pristop učenja in poučevanja. Martin-Kneip, Fiege in Soodak način dela poveča motivacijo, izboljša komunikacijo ter (1955) opredeljujejo medpredmetno povezovanje kot primer omogoča rast na profesionalnem področju, saj s tem sodelujoči holističnega učenja in poučevanja, ki kaže realen interaktiven razširi in poglobi lastno znanje. svet, njegovo kompleksnost, odpravlja meje med posameznimi

disciplinami in podpira načelo, da je vse znanje povezano.



Medpredmetno povezovanje ne pomeni le razvijanja 3 MATEMATIČNO MODELIRANJE konceptualnega povezovanja (povezovanje sorodnih pojmov pri

različnih predmetih), ampak razvija pri učencih tudi generične Modeliranje pomeni konstruirati strukturo ali teorijo, ki

veščine, ki so neodvisne od vsebine in so uporabne v različnih vključuje lastnosti objekta, sistema ali procesa. Model

okoliščinah (npr. kritično mišljenje, obdelava podatkov, uporaba reprezentira fizično ali abstraktno situacijo (Abrams, 2001).

IKT…). Matematično modeliranje pomeni prevesti situacijo iz resničnega

Dejavnosti, povezane z medpredmetnim povezovanjem, sveta v matematični jezik (model).

vodijo k doseganju kompleksnih znanj in h kompleksnim V raziskavi PISA 2006 (Repež, A., Straus, M., 2008) je

pričakovanim modeliranje opredeljeno kot: » Prenos realnosti v (matematični) rezultatom. Medpredmetne povezave

uresničujemo in izvajamo na različnih ravneh in z različnimi svet, interpretacija modelov v smislu skladnosti z realnostjo, delo

cilji: z modelom, vrednotenje, analiziranje in kritika modela,

a) posredovanje podatkov o modelu ter spremljanje in Na ravni vsebin: obravnava oz. reševanje

interdisciplinarnih problemov. Pri teh dejavnostih uporabljamo nadzorovanje procesov modeliranja.«

specifična znanja posameznih disciplin in tudi generične veščine Matematični modeli poskušajo dati posnetek pomembnih

in spretnosti, ki predstavljajo aplikacijo specifičnega znanja na značilnosti dejanskega stanja situacije, objekta ali procesa v

avtentične probleme. matematičnem jeziku. Pri sestavljanju matematičnega modela

b) upoštevamo fizikalne ali katere druge zakone, npr. v Na ravni procesnih znanj: učenje in uporaba procesnih

znanj (npr. iskanje virov, oblikovanje poročila ali miselnega družboslovju.

vzorca, govorni nastop, delo v skupini,…). Matematično modeliranje je primer celostnega učenja pri

c) pouku matematike. Predstavlja razvoj problemskih znanj ter Na konceptualni ravni: obravnava pojmov iz različnih

predmetnih perspektiv z namenom poglabljanja in razumevanja reševanje problemskih situacij na eni stani, ter bogati vsebinska

(npr. naravna rast pri biologiji v povezavi z eksponentno funkcijo znanja na drugi strani. V tematskih sklopih kot je pojem funkcije

pri matematiki, eksponentno pojemanje v povezavi z upadanjem ali področje statistike ter obdelave podatkov je nepogrešljiv

vrednosti dobrin na trgu idr.). Primeri naj bodo kot pomembni gradnik, kjer teoretična znanja podkrepimo z uporabo različne

zgledi, ki so namenjeni razumevanje matematike in osmišljanju informacijsko-komunikacijske tehnologije. Hkrati lahko

matematičnih vsebin. vključujemo tudi kontekste drugih področij (npr. naravoslovnih

in družbenih).



Pri tovrstnih dejavnostih dijaki pridobivajo izkušnje in se Glavne faze modeliranja:

učijo matematike ter tudi generičnih znanj, ki naj bi se v končni 1. seznanjanja s problemom (uvid v problem),

fazi kazala kot kompleksni pričakovani rezultati, kot npr., da 2. oblikovanje predpostavk in matematična formulacija

dijaki: (prevajanje realističnih situacij v matematični kontekst/problem,

- analiziranje in razumevanje situacije, postavitev predpostavk, prepoznajo vlogo in pomen matematike in drugih

disciplin v realnih situacijah in se učijo matematiziranja; določanje povezav med spremenljivkami idr.),

- 3. postavitev modela (zapis funkcije, formule, postavitev uporabljajo matematiko v matematičnih kontekstih in

v realnih situacijah, geometrijskega modela idr.),

- 4. ugotavljanje veljavnosti modela (z vidika matematike modelirajo, primerjajo modele ter rezultate različnih

modelov in interpretirajo njihove rešitve z vidika matematike in in realistične situacije),

realnih situacij idr. 5. uporaba in interpretacija modela (interpretacija rešitev

z vidika matematike in realistične situacije, primerjava med



Didaktični vidiki medpredmetnega povezovanja iz modeli).

perspektive matematike:



- obravnavati matematične pojme iz različnih 4 IZVEDBA DEJAVNOST

predmetnih perspektiv;

- Naš projekt medpredmetnega povezovanja se je pričel, ko prepoznati matematični kontekst v realnih situacijah in

modelirati; smo z dijaki 1. letnika gimnazijskega programa obiskali

- svetovno znan poševni stolp v Pisi (slika 1). Stolp je bil zgrajen reševati interdisciplinarne probleme in matematizirati;

- med 12. in 14. stoletjem in je prvotno služil kot zvonik za razvijati uporabo IKT kot možnosti za razvoj

matematičnega znanja ter kot podporo pri učenju in poučevanju; sosednjo katedralo. Zaradi mehkega terena se je nagibal skozi

- celotno fazo gradnje, zaradi skrbi, da bi se porušil, so razvijati generične veščine in spretnosti.

strokovnjaki začeli meriti njegovo nagnjenost.



V nadaljevanju učne ure so dijaki samostojno reševali učni

list na temo modeliranja z linearno funkcijo. Podane podatke

nagnjenosti stolpa v danem letu, so vnašali v tabelo in kasneje

poiskali ustrezni model. Po končanem samostojnem

raziskovanju so bili podatki prikazani še na interaktivni tabli, kjer

smo skupaj z dijaki poiskali ustrezno prilagoditveno funkcijo še

s pomočjo programa Graph. Dijaki dobljeni predpis primerjajo s

tistim, ki so ga izpeljali sami s svojimi izračuni. Pri zadnjem delu

izvedbe dijaki uporabijo svoje pametne telefone ali tablične

računalnike ter program GeoGebra (slika 3). Ta del tako deluje

na dijake motivacijsko, saj svoje predpise linearne funkcije

primerjajo s tisto, ki jo izpiše program. Nalogo lahko rešujejo

dijaki samostojno ali pa v parih oziroma skupinsko in na ta način

krepimo socialne ter tudi digitalne kompetenca, in sicer

prepoznavanje potreb in problemov, reševanje konceptualnih

problemov in problemskih situacij v digitalnih okoljih. V tem



Slika 1: Obisk Pise. delu so bile uporabljene digitalne kompetence, in sicer reševanje

konceptualnih problemov in problemskih situacij v digitalnih

Med ogledi znamenitosti so se dijaki preizkusili v fotografiji okoljih ter vključeni naslednji cilji, kjer dijaki:

s pametnim telefonom. V pomoč so jim bili nasveti - modelirajo primere iz vsakdanjega življenja z linearno

profesoricUrške Kompara Žvokelj o kompoziciji, svetlobi in funkcijo,

obdelavi fotografij, ki so jih osvojili pri pouku informatike. - presojajo o veljavnosti modela,

Ustvarjanje in urejanje digitalnih vsebin ter umetniško izražanje - kritično interpretirajo rezultate,

z uporabo digitalnih sredstev so bile tiste digitalne kompetence, - razvijajo kritični odnos do informacij oziroma

ki so jih dijaki pridobili v tem delu dejavnosti. Cilji, ki so bili pri podatkov.

tem vključeni so bili, da dijaki:

- spoznajo osnovna načela in pravila fotografije (pravilo

tretjin, ostrenje in optimizacija osvetlitve …),

- samostojno pregledajo nastavitve svojega pametnega

telefona (znajo vključiti mrežo, način HDR …),

- samostojno prenesejo aplikacije za obdelavo fotografij

s spletne trgovine aplikacij,

- posnamejo fotografije na izbrano temo in jih

posredujejo učitelju prek spletnega okolja MS Teams.

Dejavnost smo nadaljevali pri pouku matematike, ko smo

obujali spomine na potepanje po znamenitih toskanskih mestih.

Zgodovinska dejstva in zanimiva arhitekturna spoznanja o

poševnem stolpu v Pisi smo želeli nadgraditi z matematičnimi Slika 3: Uporaba pametnih telefonov.

modeli. Kljub temu, da je stolp trenutno utrjen, nas je zanimalo

za koliko je nagnjen in kako se je nagibal skozi leta. Pričeli smo Dijaki so pri likovni umetnosti s profesorico Damjano Plešnar

s frontalnim prikazom nagnjenosti stolpa na interaktivni tabli s izdelovali modele poševnega stolpa v Pisi, kar zahteva velik

pomočjo odprtokodnega programa GeoGebra (slika 2). S nabor sposobnosti in spretnosti, med drugim tudi prostorsko

pomočjo orodij in fotografije stolpa smo določili za koliko vizualizacijo, ki je bistvenega pomena za matematično pismenost.

stopinj je stolp nagnjen. Dijaki so primer reševali vsak na svojem Pri tej dejavnosti smo uporabili predstavitev pojmov z modeli, ki

telefonu s pomočjo istoimenske aplikacije. je eden od dejavnosti izkustvenega učenja.

Če je fotografija umetnost, ki ujame trenutek in ga naredi

večnega, pa je animacija tehnika, ki ustvari iluzijo gibanja in



oživi predmete. Naše sodelovanje se je tako nadaljevalo z



izdelavo animacije poševnega stolpa v Pisi z različnimi

likovnimi pristopi na slikovnem polju ali v prostoru.

Konstrukcija stolpa je nastajala z rezanjem papirja ter čipke,

oblikovanjem gline in risanjem. Dijaki in dijakinje so s pomočjo

pametnih telefonov in aplikacije StopMotion Studio izdelali

krajše videoposnetke.

Dijaki pri tem usvajajo naslednje cilje:

- samostojno prenesejo aplikacijo s spletne trgovine

aplikacij,



Slika 2: Razlaga s pomočjo interaktivne table. Motion Studio, - spoznajo osnovne ukaze pri delu z aplikacijo Stop

- izdelajo animacijo v prilagojenih pogojih dela,

- z animacijo učinkovito prikažejo primer,

- uspešno izvozijo ustvarjeno animacijo na pametni

telefon.





Slika 4: Aplikacija StopMotion Studio.



5 ZAKLJUČEK

Izkazalo se je, da se lahko skozi tri leta gimnazijskega

izobraževanja večkrat vračamo k istemu izhodišču. Tokrat smo

se podali v sosednjo Italijo in preko zgodovine, umetnostne

zgodovine, likovne umetnosti, matematike in nenazadnje

uporabe informacijsko-komunikacijske tehnologije na prijeten

način povezali na videz zelo različna predmetna področja. Vse

prevečkrat pa se srečamo z nepremostljivo oviro v našem

prenatrpanem učnem načrtu, ki od posameznika zahteva nenehno

uresničevanje splošnih, procesnih in vsebinskih učnih ciljev.

Največja ovira se zdi pomanjkanje časa in energije za timsko

poučevanje in medpredmetna povezovanja. Pa vendar je končni

izdelek potem za posameznika neprecenljiv in je lahko učitelju v

zadovoljstvo in ponos. Pri tovrstnih sodelovanjih je odziv

dijakov pozitiven in je zato motivacija za delo večja, tudi zaradi

uporabe tehnologije, ki pri dijakih vzbudi zanimanje. Učitelji

smo tisti, ki s skrbno načrtovanimi dejavnostmi spodbujamo

razvoj digitalnih kompetenc, saj so ključne za razvoj sodobne

družbe.



LITERATURA IN VIRI

[1] Kocjan, M., Kotnik, B. Opačak, Ž. in Rau, P., 2015. Likovna umetnost, i-

učbenik za likovno umetnost v gimnazijskem programu. Ljubljana: Zavod

Republike Slovenije za šolstvo. Dostopno na naslovu:

https://eucbeniki.sio.si/lum/index.html (17. 8. 2024).

[2] Oreški, M.: Nasveti za fotografiranje s pametnim telefonom, 2017.

Dostopno na naslovu: https://omisli.si/nasvet-

strokovnjaka/fotograf/nasveti-za-fotografiranje-pametni-telefon/ (17. 8.

2024).

[3] Rugelj M., Šparovec J., Pavlič G., Kavka D., 2018. Linea Nova: Učbenik

za matematiko za 1. letnik gimnazije, Modrijan Izobraževanje, Ljubljana.

[4] Učni načrt. Informatika: Splošna, klasična in strokovna gimnazija. 2008.

Dostopno na naslovu:

http://portal.mss.edus.si/msswww/programi2011/programi/media/pdf/uc

ni_nacrti/UN_INFORMATIKA_gimn.pdf (17. 8. 2024).

[5] Učni načrt. Matematika. Splošna, klasična in strokovna gimnazija. 2008.

Dostopno na naslovu

http://eportal.mss.edus.si/msswww/programi2010/programi/media/pdf/u

n_gimnazija/un_matematika _gimn.pdf (17. 8. 2024).

[6] Vuorikari R., Kluzer S. in Punie Y., 2023. DigComp 2.2: Okvir digitalnih

kompetenc za državljane.

[7] Žakelj., A., Pustavrh. S. in drugi, 2010. Posodobitve pouka v gimnazijski

praksi. Matematika. Kmetič, S., Sirnik., M. (ur.). Ljubljana: Zavod RS za

šolstvo.





Ali bo umetna inteligenca spremenila raziskovanje v




osnovni šoli



Will Artificial Intelligence Change Research in Primary School?



Vesna Mrkela Rafaela Voglar

Osnovna šola Draga Kobala Maribor Osnovna šola Draga Kobala Maribor

Tolstojeva 3, 2000 Maribor, Slovenija Tolstojeva 3, 2000 Maribor, Slovenija

vesna.mrkela@osdk.si rafaela.voglar@osdk.si



da bi moralo poučevanje temeljiti na aktivnostih učencev in

POVZETEK načinih, kako lahko učenci z raziskovanjem pridobijo znanje.

V prispevku želimo prikazati, kako pomembno je Kot navaja Merriam Webster [3] je raziskovanje večplastnega

raziskovanje. Zelo pomembno je, da se učenci z raziskovanjem pomena. Med raziskovanje uporabljamo različne raziskovalne

srečajo že v osnovni šoli. Učenci spoznavajo osnovne korake metode kot so zbiranje informacij o nekem predmetu ali pojavu,

raziskovanje, med katerimi je tudi uporaba in iskanje ustrezne pazljivo in prizadevno iskanje, premišljeno poizvedovanje ali

literature. Med iskanjem in brskanje po spletu zagotovo naletijo proučevanje, preiskovanje ali eksperimentiranje za odkrivanje ali

tudi na umetno inteligenco. Pa se je sploh zavedajo? Pojavi se tolmačenje dejstev, preverjanje sprejetih teorij ali zakonov v luči

vprašanje, ali lahko in kako lahko umetna inteligenca vpliva na novih podatkov, praktična uporaba takih novih ali preverjenih

raziskovanje. teorij in zakonov. Že prva metoda je velikokrat vključena pri

Pri raziskovanju je pomen mentorja zelo velik, saj lahko pouku, pa se niti učitelji niti učenci tega sploh ne zavedamo. Če

veliko pripomore k pravilnemu usmerjanju in uporabi ustreznih želimo nekaj opisati, moramo znati zbati prave informacije o

virov. Ugotovili smo, da zaenkrat v osnovni šoli umetna predmetu ali pojavu, ki ga želimo opisati. Velikokrat pa

inteligenca še ne posega v raziskovanje, zavedamo pa se, da raziskovanje povezujemo z eksperimentiranjem [2], odkrivanjem

bomo morali zagristi tudi v to jabolko in učencem predstaviti in preverjanjem teorij, vendar se osnove skrivanji že v samem

prednosti in slabosti uporabe umetne inteligence. iskanju informacij. V osnovni šoli le redki pridejo na nivo

samostojnega eksperimentiranja za odkrivanje dejstev. Vendar



KLJUČNE BESEDE prav pri pisanju raziskovalnih nalog, lahko učence privedemo tudi do tega. Raziskovanje, umetna inteligenca, osnovna šola Pomembni procesi raziskovanja so pravzaprav glej, misli in



ABSTRACT pomeni analiziraj dobljene podatke, prepoznaj bistvene podatke, deluj [2]. Glej kot zbiraj informacije, jih zapiši in prebiraj. Misli

In this article, we want to show how important research is. It ki so za raziskavo pomembni in jih uporabi – deluj – za

is very important that students are exposed to research already in oblikovanje rešitev. Pa lahko tukaj pomaga tudi umetna

primary school. Students learn the basic steps of research, inteligenca? Menimo, da do neke mere ja, vendar je potrebno

including the use and search for relevant literature. While pazljivo in veliko preverjanja podatkov pridobljenih s pomočjo

searching and browsing the web, they are bound to come across različnih brskalnikov umetne inteligence. artificial intelligence. Are they even aware of it? The question

arises as to whether and how artificial intelligence can influence

research. 2 PREGLED STANJA

In research, the importance of a mentor is high, as she/he can V učnih načrtih za osnovno šolo je se učenci raziskovanja le



resources. We established that artificial intelligence has not postopke. Na šoli se nam zdi spodbujanje k pisanju raziskovalnih interfered with research in primary schools yet, but we are aware nalog in raziskovanju nasploh velikega pomena. Učenci začno greatly contribute to the correct direction and use of appropriate dotaknejo oziroma postavijo neke temelje in spoznajo osnovne

that we will have to bite into this apple and present students the razvijati različne kompetence, ki jih nato lahko uporabijo tudi v

advantages and disadvantages of using artificial intelligence. vsakdanjem življenju. Na šoli zato učenci začno z prvimi koraki



KEYWORDS nadgrajujejo raziskovanja v petem razredu z opisom mesta in raziskovanja že v četrtem razredu, kjer opisujejo najljubšo žival,

Research, artificial intelligence, primary school tako naprej do opisovanja znanih osebnosti pri pouku tujega

jezika. Pa vendar je raziskovanje več, kot le povzemanje iz spleta



1 UVOD V tretjem vzgojno-izobraževanem obdobju veliko naših in knjig.

Pred več kot sto leti se pojavijo prve ideje, kako bi moralo biti učencev stopa v svet raziskovanja, kjer spoznavajo čare besede

poučevanje povezano z raziskovanjem in ne le s prenašanjem zakaj. Sodelujemo na natečaju Mladi za napredek Maribora, ki je

znanja iz učitelja na učenca. Prav raziskovalec Dewey [1] je trdil, regijsko tekmovanje in odskočna deska za napredovanje na



državni nivo Mladi raziskovalci Slovenije, na katerega se redno umetne inteligence, jih spodbuditi in jim pokazati, kako nam

vsako leto uvrščamo. umetna inteligenca lahko delo olajša, ne pa tudi opravi namesto

Hitro se pojavi vprašanje, ali bo na radovednost učencev nas. Seveda moramo uporabo umetne inteligence najprej

vplival razvoj umetne inteligence in raziskovanje učencev več ne spoznati mentorji raziskovalcem, saj le tako bomo lahko

bo tak pritegnilo kot do sedaj. Menimo, da ne. Upamo, da bo učencem pomagali in jih usmerili na pravo pot.

umetna inteligenca pripomogla k večjemu raziskovalnemu



napredku že v mladosti. Učenci se ne bodo učili samo 4 ZAKLJUČEK



inteligence. Menimo, da mora biti umetna inteligenca samo Zavedamo se, da umetna inteligenca bo vplivala na raziskovanje tudi že v osnovni šoli. Na mentorjih oziroma raziskovalnih metod dela, ampak tudi ustrezne uporabe umetne

pomoč pri raziskovanju in odkrivanju novega, neznanega, ne pa



da jo uporabimo za pisanje celih nalog. uporabo umetne inteligence. Zaenkrat vsaj na naši šoli se učenci učiteljih pa je, kako bomo učencem prikazali smiselno in varno

tega orodja še ne poslužujejo. Verjamemo, da se ga bomo v



3 prihodnje, saj nas bo v to prisililo okolje. Zavedati se moramo RAZISKAVA

vseh pasti in prednosti, ki nam jih umetna inteligenca prinaša. Ne

V raziskavah, ki so jih izvedle fakultete [4] lahko zasledimo, zapirajmo vrat, ugriznimo v izzive, ki jih prinaša prihodnost, in

da veliko študentov uporablja ChatGPT. Po zadnjih podatkih več stopimo izven okvirjev ugodnega. Verjamemo, da nam bo uspelo

kot polovica študentov. Od vseh teh študentov jih več kot 90% in da bo umetna inteligenca pripomogla k še boljšim in

uporablja umetno inteligenco za iskanje informacij, ki bi jih kakovostnejšim raziskovalnim nalogam.

lahko dobili iz drugih strokovnih virov. Vendar gre z umetno

inteligenco lažje in hitreje. Študentje poročajo: Bolj kot študenti LITERATURA IN VIRI

uporabljajo ChatGPT za preiskovanje tega, kaj umetna [1] John Dewey’s learning theory: How we learn through experience,

inteligenca zmore, bolj verjetno poročajo o negativnih učinkih DOI: https://www.growthengineering.co.uk/john-dewey/.

rabe ChatGPT na njihov študijski uspeh. Prav zaradi tega nas je [2] Ernie Stringer: Akcijsko raziskovanje v izobraževanju, Kranj 2008,

zanimalo, koliko naših raziskovalcev v osnovni šoli uporablja Dravska tiskarna

[3] Spletni slovar Merriam Webster Online Dictionary.

kakršnokoli umetno inteligenco pri pisanju raziskovalnih nalog. DOI https://www.merriam-webster.com/.

V anketi, ki smo jo izvedli med učenci, ki so v zadnjih dveh letih [4] Raziskovalni dan na FDVju.

pisali raziskovalne naloge, smo dobili zanimive odgovore. DOI: https://www.fdv.uni-lj.si/raziskovanje/institut-za-druzbene-

Učenci so poznali umetno inteligenco, ki deluje preko družabnih vede/obvestila/raziskovalni-dan-fdv-2023-o-umetni-inteligenci-v-

druzboslovnem-raziskovanju-2

omrežij, kot so SnapChat, Instagram in podobno. Z umetno

inteligenco komunicirajo, a le te ne uporabljajo za šolsko delo.

Primer Copilot in ChatGPT so spoznali pri pouku, a jih po

njihovih besedah ni pritegnil tako, da bi jo redno začeli

uporabljati.

Na šoli smo v zadnjih dveh letih izdelali 26 raziskovalnih

nalog, ki smo jih prijavili na prej omenjen natečaj, in tako smo

lahko anketirali le 29 učencev. Od vseh anketiranih učencev niti

eden ni uporabljal ChatGPT, Copilot ali podobnih orodij umetne

inteligence pri pisanju raziskovalne naloge. Seveda je vzorec

zelo majhen. Anketiranje pred prihodom ChatGPT se nam ni

zdelo smiselno, saj se umetna inteligenca takrat še ni pojavljala

v javnosti. Tisti učenci umetne inteligence zagotovo niso

uporabili.

Na šoli se zavedamo, da je raziskovanje izjemnega pomena

za razvoj otroka. Pri pisanju raziskovalnih nalog učenci

spoznavajo različne načine učenja, s katerim lahko nadgradijo

klasičen pouk v učilnici. Pridobijo kompetence, ki so pomembne

za njihov nadaljnji razvoj in življenje.

Skozi anketo smo spoznali, da učenci še niso vešči uporabe

umetne inteligence in v zadnjih dveh letih s pomočjo le te niso

pisali raziskovalnih nalog. Verjamemo pa, da se bodo rezultati

spremenili.

Ker so nas rezultati anketiranih učencev presenetili, smo

anketirali tudi mentorje raziskovalnih nalog. Teh je manj kot

učencev, zato smo mentorjem postavili vprašanja v obliki

intervjuja. Rezultati intervjujev mentorjev raziskovalnih nalog

na šoli so enaki kot rezultati pri učencih. Mentorji si pri pregledu

nalog in pomoči učencem niso pomagali z umetno inteligenco,

čeprav jo poznajo, saj je bila predstavljena na različnih

konferencah in strokovnih srečanjih. Mentorji raziskovalcem se

zavedajo, da bo potrebno ugrizniti tudi v to jabolko in spoznati

ter raziskati, kako lahko umetna inteligenca pomaga pri

raziskovanju. Učence je potrebno naučiti smiselne uporabe



Poučevanje algoritmov in programiranja z Arduino



platformo: Praktični primer iz srednješolskega okolja



Teaching Algorithms and Programming with the Arduino



Platform: A Practical Example from a High School



Environment



Simon Muha Simon Konečnik

Elektro in računalniška šola / Šolski center Velenje Elektro in računalniška šola / Šolski center Velenje

Velenje, Slovenija Velenje, Slovenija

simon.muha@scv.si simon.konecnik@scv.si



POVZETEK

V prispevku predstavljamo projektno usmerjen pristop k 1 Uvod



poučevanju algoritmov in programiranja z uporabo platforme V današnjem svetu, kjer tehnologija igra ključno vlogo, je

Arduino, ki smo ga izvajali na Elektro in računalniški šoli, poučevanje algoritmov in programiranja že v zgodnjih fazah

Šolskega centra Velenje, v okviru projekta KATARINA. Skozi izobraževanja ključnega pomena. Srednješolski dijaki, zlasti tisti

praktično delo na pametnem merilniku temperature in vlažnosti v tehničnih programih, potrebujejo trdno osnovo na področju

so dijaki pridobili trdno osnovo v programiranju in razumevanju računalniškega razmišljanja in programiranja, da lahko

ključnih konceptov, kot so zanke, pogoji in funkcije, ter razvili učinkovito delujejo v sodobnem tehnološkem okolju. Arduino,

dragocene praktične veščine. Kljub izzivom, kot so razlike v priljubljena odprtokodna platforma za elektroniko in

predznanju dijakov in kompleksnost povezovanja elektronskih programiranje, predstavlja odlično orodje za povezovanje

komponent, so dijaki uspešno premagovali ovire in dosegali cilje. teoretičnega znanja z resničnim svetom.

Arduino se je izkazal kot učinkovito orodje za poučevanje in Elektro in računalniška šola, Šolskega centra Velenje je

spodbujanje inovativnosti, kar je bistveno pripomoglo k vključena v project KATARINA, ki ga vodi Fakulteta za

motivaciji dijakov in njihovi pripravljenosti za nadaljnje računalništvo in informatiko Univerze v Ljubljani. Na šoli smo v

izobraževanje. šolskem letu 2023/24 vključili dijake prvih letnikov programa

elektrotehnik in programa tehnik mehatronike. Za poučevanje

KLJUČNE BESEDE področja Algoritmi in programiranje smo izbrali platformo

Arduino, Arduino. V nadaljevanju bomo predstavili naš pedagoški pristop, algoritmi, programiranje, srednješolsko

izobraževanje, praktično učenje cilje izobraževanja, opisali praktični projekt, ter delili izkušnje in

izzive, s katerimi smo se srečali.



ABSTRACT 1.1 Projekt KATARINA The paper presents a project-oriented approach to teaching Projekt KATARINA, ki ga vodi Fakulteta za računalništvo in algorithms and programming using the Arduino platform, informatiko Univerze v Ljubljani, v sodelovanju z desetimi implemented at the Electro and Computer School, School Center partnerskimi srednjimi šolami in gimnazijami, sistematično Velenje, as part of the KATARINA project. Through practical ustvarja podporna okolja za vključevanje temeljnih vsebin work on a smart temperature and humidity sensor, students računalništva in informatike v srednješolsko izobraževanje. gained a solid foundation in programming and understanding key Projekt poudarja izmenjavo izkušenj in dobrih praks, kar bo concepts such as loops, conditions, and functions, while prispevalo k nadaljnjemu razvoju in vključevanju RIN vsebin v developing valuable practical skills. Despite challenges such as slovenske srednje šole. varying levels of student prior knowledge and the complexity of

connecting electronic components, students successfully



overcame obstacles and achieved their goals. Arduino proved to 2 Pedagoški pristop in cilji be an effective tool for teaching and fostering innovation,

significantly contributing to student motivation and their Pri poučevanju algoritmov in programiranja smo uporabili

readiness for further education. projektno usmerjeno metodo, kjer so dijaki skozi ves proces

pridobivanja znanja delali na enem glavnem projektu –



KEYWORDS Pametnem merilniku temperature in vlažnosti. Ta pristop je omogočil, da so dijaki postopoma, korak za korakom, vključevali Arduino, algorithms, programming, secondary education, novo znanje v svoj projekt, hkrati pa so se naučili tudi reševanja hands-on learning konkretnih problemov iz resničnega sveta.



2.1 problemov. Projektno zasnovano poučevanje je dijakom Učni cilji



Glavni učni cilji so bili: omogočilo celovit vpogled v proces razvoja programske opreme, od začetne ideje do končnega izdelka, ter hkrati poudarilo

 Razumevanje osnovnih konceptov algoritmov, kot pomembnost povezovanja teoretičnega znanja z njegovo

so zaporedje ukazov, pogojni stavki in zanke. praktično uporabo.

 Pridobitev veščin za delo s spremenljivkami,



 3 Pametni merilnik temperature in vlažnosti Uporaba Arduina za implementacijo teh konceptov vhodnimi in izhodnimi podatki ter uporabo funkcij.

v praktičnih projektih. Pametni merilnik temperature in vlažnosti predstavlja

osrednjo projektno nalogo, v kateri so dijaki uporabili različne

2.2 Metodologija poučevanja koncepte programiranja, pridobljene tekom izobraževanja.

Dijake smo skozi 24 učnih ur vodili od osnov razumevanja Projekt temelji na uporabi Arduina, ki omogoča zajemanje

algoritmov do razvoja kompleksnega programa, ki uporablja okoljskih podatkov prek senzorjev za temperaturo in vlažnost.

različne konceptualne elemente. Poučevanje je bilo strukturirano Zbrani podatki so bili nato obdelani in prikazani na LCD zaslonu,

tako, da je vsaka ura temeljila na že pridobljenem znanju, kar je kar je dijakom omogočilo praktično uporabo znanja s področja

omogočalo progresivno gradnjo veščin in samozavesti pri algoritmov, dela s podatki ter implementacije vhodno-izhodnih

reševanju programskih problemov. operacij. Projekt je bil zasnovan tako, da povezuje vse ključne

Učne enote smo razdelili tako, da so pokriti vsi cilji. elemente programskega razvoja, kar dijakom omogoča celovit

vpogled v proces ustvarjanja funkcionalnega sistema.



Tabela 1: Pokritost vsebin z učnimi cilji 3.1 Projektna naloga

Osrednji projekt je bil zasnovan kot razvoj pametnega



Koncept / merilnika temperature in vlažnosti, ki temelji na platformi Učni cilji vsebina Arduino. S pomočjo senzorjev so dijaki zbirali podatke o

- Dijak razloži pojem algoritma. temperaturi in vlažnosti, jih obdelali ter prikazali na LCD

01 - Dijak pokaže primer problema, kjer obstaja zaslonu. Projekt je vključeval vse ključne koncepte

Pojem več algoritmov za reševanje. programiranja, ki smo jih predhodno obravnavali v okviru

algoritma - Dijak prikaže različne oblike zapisov teoretičnih vsebin.

algoritma. Dijaki so začeli z enostavnim merjenjem temperature in

Dijak uporabi spremenljivko za predstavitev vlažnosti ter takojšnjim prikazom vrednosti na LCD zaslonu.

spreminjajočega se podatka. Nato so uporabili funkcije, tako da so prejšnji program razdelili

- Določi ustrezni tip spremenljivke oz. tip na funkciji preberiPodatke() in prikaziNaLCD(). V nadaljevanju

vsebine (npr. celo število, decimalno število, so v program vključili pogojni stavek, s katerim so dodali

niz). opozorilo, ki se prikaže na LCD zaslonu, če temperatura preseže

- Rokuje s spremenljivko (priredi in spremeni določeno mejo. Števne zanke so uporabili za merjenje

03 vrednost, prebere vrednost, uporabi jo v temperature in vlažnosti vsakih 10 sekund, pri čemer so izvedli

Zaporedje ustreznih operacijah). 10 meritev, rezultate pa prikazali in shranili v tabelo. Zanke s

ukazov - Spreminja podatke enega tipa v drugega (npr. pogojem so uporabili za neprekinjeno spremljanje in prikaz

niz -> število, število -> niz, decimalno število - podatkov, dokler ni bil pritisnjen gumb za zaustavitev. > celo število).

- Dijak s primerom pokaže pomembnost

vrstnega reda ukazov. Tabela 2: Primer za števne zanke

- Dijak sledi izvajanju algoritma in napove

njegov izid. Tema Števne zanke

05 Naloga Dijaki bodo uporabili števne zanke za

- Dijak sestavlja preproste programe, s katerimi

Vhod kot periodično merjenje temperature in

prebere podatke, jih uporabi za novo

zaporedje vlažnosti.

informacijo in informacijo izpiše.

podatkov Aktivnosti Razlaga uporabe števnih zank, pisanje

kode, ki periodično meri in shranjuje

Za celoten proces poučevanja smo izbrali projektno nalogo, podatke.

ki je predstavljala osrednji element izobraževalnega procesa. Primer naloge Uporabite for zanko za merjenje

Dijaki so projekt začeli razvijati že v uvodnih učnih urah, pri temperature in vlažnosti vsakih 10

čemer so postopoma uvajali in uporabili vsako novo pridobljeno sekund za 10 meritev. Izračunajte

znanje ter veščine. Vsaka učna enota je bila zasnovana tako, da povprečno vrednost. Rezultate prikažite

je gradila na predhodno obravnavanih vsebinah, kar je in shranite rezultate v tabelo

omogočalo kontinuiran napredek in poglabljanje razumevanja. S Pričakovani Dijaki obvladajo uporabo števnih zank

tem pristopom so dijaki lahko postopoma dograjevali svoj rezultati za izvajanje ponavljajočih se nalog.

projekt, kar jim je omogočilo praktično uporabo teoretičnih

konceptov, razvijanje kompleksnih programskih rešitev ter

krepitev njihove samozavesti pri reševanju programskih



3.2 komponent, zagotavljanju ustreznih napetosti in pravilnemu Faze projekta

delovanju senzorjev, preden so dobili želene rezultate v svojih

Načrtovanje programih.

V začetnih fazah projekta so dijaki spoznali osnovne pojme Pogosto se je zgodilo, da tudi pravilno napisana programska

algoritmov in jih uporabili pri načrtovanju osnovnega algoritma koda ni delovala, dokler ni bila vezava popolnoma pravilna, kar

za merjenje temperature in vlažnosti. To je vključevalo določitev je zahtevalo dodatno potrpežljivost in vztrajnost pri odpravljanju

zaporedja ukazov in razumevanje, kako lahko vhodni podatki napak. Učitelji smo morali prilagoditi tempo učenja glede na

(podatki iz senzorja) vplivajo na izhodne rezultate. sposobnosti posameznih dijakov, kar je omogočilo boljše

Razvoj kode individualno napredovanje. Raznolikost v predznanju dijakov je

V naslednjih urah so dijaki začeli razvijati svojo kodo. predstavljala še en izziv; v razredu smo imeli dijake, ki so se

Naučili so se, kako uporabiti spremenljivke za shranjevanje doma že ukvarjali s področjem elektronike ali mehatronike in so

podatkov, kako delujejo različne vrste zank in pogojnih stavkov imeli določene izkušnje, ter tiste, ki so jih to področje zelo

ter kako razdeliti kodo na funkcije, ki olajšajo ponovno uporabo zanimalo, a še niso imeli praktičnih izkušenj. To je zahtevalo

in razumevanje programa. Vsaka faza razvoja kode je bila prilagajanje učnega procesa in dodatno individualno podporo, da

zasnovana tako, da je bila kompleksnost postopoma povečana. so vsi dijaki lahko dosegli pričakovane cilje.

Razlike v znanju in izkušnjah so se še posebej pokazale pri

Integracija in testiranje izdelavi projektne naloge, kjer so nekateri dijaki začeli z

Ko so dijaki obvladali posamezne dele kode, so jih združili v osnovnimi nalogami, kot je preprosto merjenje temperature in

celovit program. Nato so program testirali, da bi zagotovili prikaz rezultatov na zaslonu, medtem ko so drugi napredovali do

njegovo pravilno delovanje in da so izpolnjeni vsi zahtevani cilji. bolj kompleksnih rešitev. Nekateri dijaki so tako iz osnovnih

Testiranje je bilo ključnega pomena za razvoj veščin iskanja nalog prešli na razvoj naprednejših funkcij, kot je avtomatizacija

napak in izboljševanja programske opreme. meritev in vodenje evidenc. Ta raznolikost v dosežkih je bila

izziv, a tudi priložnost, da so dijaki s pomočjo

Rezultati projekta



Dijaki so skozi delo na projektu razvili trdno razumevanje nivoju. individualiziranega pristopa razvijali svoje veščine na lastnem

osnovnih konceptov programiranja. Poleg tega so razvili



praktične veščine, kot so delo s senzorji, prikaz podatkov na 4.3 Prednosti uporabe Arduina zaslonu in uporaba različnih programskih struktur. Projekt je bil

uspešen v tem, da je dijakom omogočil neposredno uporabo Arduino se je izkazal za izjemno učinkovito orodje za

naučenih teoretičnih znanj v praktičnem kontekstu. poučevanje programiranja. Dijakom je omogočil, da so

abstraktne koncepte, kot so zanke, pogoji in funkcije, takoj videli

v delovanju, kar je prispevalo k globljemu razumevanju.

4 Izkušnje in izzivi Pomembnost praktičnega dela pri učenju elektronike in

mehatronike je velika, saj dijaki skozi praktične projekte

4.1 Izkušnje dijakov pridobivajo ne samo teoretično znanje, ampak tudi konkretne

Dijaki so pokazali veliko zanimanje za projekt in izrazili veščine, ki jih lahko neposredno uporabijo v realnih situacijah.

zadovoljstvo z načinom učenja. Kombinacija teoretičnega znanja Praktično delo z Arduinom omogoča dijakom, da povežejo

in praktičnega dela je pripomogla k boljšemu razumevanju snovi teoretične koncepte s fizičnimi rezultati, kar izboljšuje njihovo

in večji motivaciji za učenje. Praktično delo je omogočilo razumevanje in utrjevanje znanja. Poleg tega pa ta praktični

dijakom, da so si pridobili dragocene veščine pri sestavljanju pristop ponuja pomembno izkušnjo v današnjem izobraževalnem

elektronskih vezij, rokovanju s senzorji in povezovanju okolju, kjer se dijaki pogosto srečujejo predvsem z virtualnimi

komponent, kar je bistveno nadgradilo njihovo teoretično znanje. vsebinami. Praktične izkušnje, ki jih ponuja delo z Arduinom, so

Poleg tega so dijaki razvili veščine reševanja problemov, saj ključne za razvoj veščin reševanja problemov, kritičnega

so morali samostojno odpravljati napake v vezavah in mišljenja in inovativnosti.

programskih kodah. Delo na projektu je spodbujalo kritično Za dijake, ki se usmerjajo v področja elektronike in

razmišljanje in ustvarjalnost, hkrati pa jim je omogočilo, da so se mehatronike, je tovrstno praktično delo nepogrešljivo, saj jim

naučili sistematičnega pristopa k razvoju kompleksnih sistemov. omogoča, da pridobijo izkušnje s povezovanjem komponent,

Praktično vezanje in integracija različnih modulov sta dijakom razumevanjem delovanja električnih tokokrogov in ustvarjanjem

omogočila tudi izboljšanje spretnosti načrtovanja in organizacije delujočih sistemov. Vendar pa so te praktične izkušnje dragocene

dela, kar je ključnega pomena za njihovo nadaljnje izobraževanje tudi za dijake na drugih področjih, saj jim omogočajo stik z

in poklicno pot. realnostjo in razumevanje, kako lahko njihova teoretična znanja

vplivajo na svet okoli njih.

4.2 Poleg tega je Arduino zelo prilagodljiv in dostopen, kar Izzivi



razumevanje zaporedja ukazov in uporaba zank. Vendar pa so te uporabnikov ponujata bogato podporo in vire, kar dodatno olajša težave skozi praktično delo in dodatno razlago uspešno učenje in omogoča dijakom, da raziskujejo in razvijajo svoje premagali. Eden izmed ključnih izzivov je bil tudi proces vezave projekte nad okviri učnih načrtov. Tako je Arduino ne le orodje Med poučevanjem smo se srečali z nekaterimi izzivi, kot so delujejo. Njegova odprtokodna narava in široka skupnost omogoča dijakom vseh stopenj znanja, da z njim uspešno



komponent in preverjanje delovanja sistemov, saj je uspešen za učenje, temveč tudi platforma za inovacije in ustvarjalnost. rezultat odvisen od mnogih dejavnikov. Dijaki so morali

posvetiti veliko pozornosti pravilnemu povezovanju elektronskih



5 Zaključek

Izkušnje s poučevanjem algoritmov in programiranja z

uporabo Arduina so pokazale, da je ta pristop izjemno učinkovit

pri povezovanju teoretičnega znanja z njegovo praktično

uporabo. Dijaki so skozi projektno usmerjeno delo na pametnem

merilniku temperature in vlažnosti pridobili ne le osnovne

koncepte programiranja, ampak tudi dragocene praktične veščine,

ki so ključne za njihovo nadaljnje izobraževanje in poklicno pot.

Praktično delo z Arduinom je dijakom omogočilo, da so

teoretične koncepte, kot so zanke, pogoji in funkcije, videli v

delovanju in jih uporabili za reševanje konkretnih problemov.

Pomembnost praktičnega dela v izobraževanju, še posebej na

področju elektronike in mehatronike, je neprecenljiva. Dijaki so

skozi proces dela na projektu razvili veščine, kot so reševanje

problemov, kritično razmišljanje, načrtovanje in organizacija

dela. Poleg tega je praktično delo pomembno tudi v širšem

kontekstu, saj dijakom omogoča, da premostijo vrzel med

virtualnimi in resničnimi izkušnjami, kar je v današnjem

izobraževalnem sistemu pogosto zanemarjeno.

Arduino je s svojo prilagodljivostjo, dostopnostjo in široko

skupnostjo uporabnikov idealno orodje za spodbujanje

inovativnosti in ustvarjalnosti pri dijakih. Ne glede na njihovo

predznanje, so dijaki lahko uspešno delovali z Arduinom, kar je

dodatno pripomoglo k njihovemu samozaupanju in motivaciji za

nadaljnje raziskovanje in učenje.



ZAHVALA

Zahvaljujemo se Fakulteti za računalništvo in informatiko

Univerze v Ljubljani za vodstvo projekta KATARINA in

podporo pri implementaciji vsebin v učni proces. Zahvaljujemo

se tudi dijakom Elektro in računalniške šole, Šolskega centra

Velenje, za njihovo sodelovanje in motivacijo, ter učiteljem in

mentorjem, ki so s svojo strokovnostjo omogočajo uspešno

izvedbo projekta.



REFERENCES

[1] Pogačnik, Borut. 2017. Uporaba razvojnega sistema Arduino UNO v

programu elektrotehnik v srednjem strokovnem izobraževanju (SSI). In

Zbornik šestindvajsete mednarodne Elektrotehniške in računalniške

konference ERK 2017. Ljubljana: Univerza v Ljubljani, 535-538.

[Elektronski vir]. Dostopno na: http://erk.fe.uni-

lj.si/2017/papers/pogacnik(uporaba_razvojnega).pdf

[2] Pesek, Igor. 2022. Didaktika računalništva 1: zbrano gradivo. Maribor:

Fakulteta za naravoslovje in matematiko. [Elektronski vir]. Dostopno na:

https://estudij.um.si/course/view.php?id=21120

[3] Arduino 2: Začetni koraki in praktični Arduino projekti za vsakogar!.

2021. Ljubljana: AX elektronika. Priročnik. ISBN 978-961-6680-20-2.



Uporaba umetne inteligence v osnovni šoli



The Use of Artificial Intelligence in Elementary School



Tina Pajnik Sonja Strgar

Osnovna šola Vide Pregarc Osnovna šola Antona Martina Slomška Vrhnika

Ljubljana, Slovenija Vrhnika, Slovenija

tpajnik@gmail.com sonja.strgar@guest.arnes.si



POVZETEK

V prispevku so predstavljeni primeri uporabe umetne 1 DIGITALNA STRATEGIJA IN UMETNA

inteligence v okviru pouka na dveh osnovnih šolah. Po pregledu INTELIGENCA

strokovne literature in priporočil na področju razvijanja V okviru prenovljenih evropskih smernic, ki poudarjajo

digitalnih strategij, s katerimi se načrtno razvija digitalne pomen vseživljenjskega učenja, se je oblikovalo osem ključnih

kompetence učiteljev in učencev, ugotavljamo, da strategije sicer kompetenc [1]:

ponujajo modele in razvojne prioritete, a ne naslavljajo posebej - večjezičnost,

uporabo umetne inteligence. Na eni strani Evropski parlament - pismenosti,

naslavlja umetno inteligenco kot prednostno nalogo Evropske - kulturna ozaveščenost in izražanje,

unije, na drugi strani šole načrtujejo strategije razvoja digitalnih - podjetnostne kompetence,

kompetenc znotraj lastnih omejitev. Glede na to, da umetna - državljanske kompetence,

inteligenca postaja del vsakdana sodobnega človeka, bi se učenci - osebnostna, družbena in učna kompetenca,

morali seznaniti z umetno inteligenco že v osnovni šoli. Prav s - digitalne kompetence,

tem namenom predstavljamo nabor različnih tematskih sklopov, - naravoslovje, tehnologija, inženirstvo, matematika.

kjer so se v osnovni šoli izvajale dejavnosti z uporabo umetne Če kompetence pomenijo skupek znanja, spretnosti in stališč,

inteligence. potem sama uporaba IKT opreme še zdaleč ni dovolj za

razvijanje digitalne kompetence. DigComp ali okvir digitalnih

KLJUČNE BESEDE kompetenc za državljane podaja enotno izrazoslovje in ponuja

Umetna inteligenca, osnovna šola, ChatGPT, digitalne referenčni model, ki se ga lahko vpelje tudi na področje vzgoje

kompetence in izobraževanja. Konceptualni model okvira DigComp [1]

razdeli razvoj digitalne pismenosti v pet kategorij:



ABSTRACT - informacijska in podatkovna pismenost,

- komuniciranje in sodelovanje,



intelligence in the context of lessons at two elementary - ustvarjanje digitalnih vsebin, The article presents examples of the use of artificial schools.

After reviewing the professional literature and recommendations - varnost,

in the field of developing digital strategies, which are used to - reševanje problemov.

systematically develop the digital competence of teachers and Šole pristopajo k razvoju digitalne kompetence na različne

students, we note that the strategies offer models and načine, saj so pogoji za izvedbo raznoliki. Po priporočilih stroke

development priorities, but do not specifically address the use of šole oblikujejo svojo digitalno strategijo, ki je odvisna od

artificial intelligence. On the one hand, the European Parliament razvojnih prioritet [2]: od digitalne didaktike, opremljenosti šole

addresses artificial intelligence as a priority task of the European z digitalno tehnologijo, stopnje digitalne kompetentnosti



development of digital competences within their own limitations. so pokazali, da je umetna inteligenca ena od tehnologij, ki bi Given that artificial intelligence is becoming a part of modern morala postati del strategije, saj nas nagovarja v vsakdanjem Union, on the other hand, schools plan strategies for the učiteljev in učencev. Vendar to ni dovolj. Izzivi sodobnega časa

man's everyday life, students should be introduced to artificial

intelligence already in primary school. It is for this purpose that življenju. Evropski parlament [3] umetno inteligenco definira kot

we present a set of different thematic sections where activities zmožnost stroja, da izkazuje človeške lastnosti, kot so mišljenje,

using artificial intelligence were carried out in elementary učenje, načrtovanje in kreativnost, pri čemer omogoča tehničnim

school. sistemom, da zaznavajo okolje, obdelajo, kar zaznajo, in rešijo

problem. Še več, Evropski parlament [3] umetno inteligenco

KEYWORDS postavlja kot prednostno nalogo Evropske unije. Nekatere

Artificial Intelligence, Elementary School, ChatGPT, digital tehnologije umetne inteligence so prisotne že več desetletij, npr.

competences spletno oglaševanje in nakupovanje, iskanje na spletu,

prevajalniki, novejše tehnologije umetne inteligence pa že

spreminjajo naša vsakdanja življenja.

Brezplačno nevronsko omrežje OpenAI na svoji spletni strani



[4] pojasnjuje, da je ChatGPT klepetalni robot z umetno

inteligenco, ki lahko:

- odgovori na preprosta in zapletena vprašanja,

- ustvari različna besedila in reši naloge,

- sestavi natančne poizvedbe,

- preoblikuje in predeluje različne medije (knjige, filme),

- opravlja rutinske naloge oz. piše poročila,

- pomaga pri programiranju.

Poleg ChatGPT se na spletu najdejo številne aplikacije, ki

vsebujejo tehnologijo umetne inteligence: Text, Otter, Cleo,

Lensa, Alexa, Gemini, Replika, Calm, FaceApp, Siri, Google

Assistant, Duolingo, Socratic, Cleo, Google Map, Leonardo, Slika 1: Rešena naloga iz učnega lista Dall-E, Craiyon. Nekatere aplikacije so bolj splošne narave in

imajo širok spekter uporabe, medtem ko so se druge osredotočile



na prepoznavanje obrazov, učenje tujih jezikov, branje

zemljevidov, ustvarjanje grafičnih podob, ohranjanje zdravja,

izobraževanje, iskanje informacij, vodenje financ in podobno.

Umetna inteligenca postaja sopotnik sodobnega človeka. Ne gre

zgolj za razvijanje digitalnih kompetenc, ampak tudi za

prepoznavanje in razumevanje prednosti in pomanjkljivosti

umetne inteligence. Kljub temu, da imajo osnovne šole že

izdelane načrte za razvijanje digitalnih kompetenc, ostaja umetna

inteligenca izziv za učitelje. Poleg pomanjkanja ustrezne

digitalne opreme, izpada kadra, pomanjkljivega znanja na

digitalnem področju, številčnih oddelkov in drugih

organizacijskih omejitev strokovni delavci iščejo načine, kako

umetno inteligenco integrirati v redni pouk.

V nadaljevanju so predstavljeni primeri dejavnosti z uporabo Slika 2: Uporaba pametnega telefona

umetne inteligence na dveh osnovnih šolah pri pouku

matematike, izbirnega predmeta šah, angleščine in knjižnično- Učenci so povedali, da so pri matematiki do sedaj uporabljali

informacijskih znanj. PhotoMath, ne pa ChatGPT. Pri učenju matematike si pomagajo

tudi s spletno platformo https://astra.si/ai/, ki nudi napredna

orodja umetne inteligence za podporo učenju matematike.

2 UMETNA INTELIGENCA V OSNOVNI Uporabnikom omogoča, da rešujejo matematične naloge,

ŠOLI analizirajo in vizualizirajo probleme ter prejmejo pomoč pri

razumevanju matematičnih konceptov. S funkcionalnostmi, kot

2.1 UPORABA UMETNE INTELIGENCE ZA so avtomatsko preverjanje odgovorov in razlaga postopkov,

UTRJEVANJE SNOVI PRI MATEMATIKI platforma olajša učenje in izboljša učne izkušnje. ChatGPT

V 9. RAZREDU uporabljajo predvsem pri predmetih angleščina, biologija, fizika,

Za utrjevanje celoletne snovi pri matematiki smo se odločili geografija, kemija, slovenščina in zgodovina.

uporabiti umetno inteligenco. Učenci 9. razreda so k pouku Za prednosti uporabe ChatGPT so navedli hitrejše iskanje

prinesli pametne telefone. Predhodno smo preverili, ali imajo podatkov, pomoč z razlago, enostavne in kratke odgovore,

možnost uporabe mobilnih podatkov in dostop do aplikacije prihranek časa ter brezplačno uporabo.

ChatGPT ter se dogovorili, da bodo telefoni napolnjeni. Kot slabosti so izpostavili možnost pridobivanja napačnih

Učiteljica je pripravila učni list, na katerem so bile tri podatkov, nejasne postopke reševanja posameznih matematičnih

matematične naloge in evalvacijski vprašalnik. Navodilo je bilo, nalog, ki jih učenci še niso obravnavali in jih ne razumejo, dolge

da v parih rešijo vse tri naloge, nato pa posamezno nalogo odgovore in nepotrebne podatke, občasno nerazumevanje

preverijo s pomočjo ChatGPT-ja. Naloge so učenci reševali na vprašanja ter počasno delovanje ali zatikanje.

učni list, nato so zamenjali barvo pisala in preverili rešitev. Če je Učencem so bili zelo všeč gumbi v ChatGPT, kot so preberi,

bila rešitev pravilna, so preverili tudi postopek reševanja. Če je kopiraj, ustvari znova, slab odgovor in sprememba modela. Z

bila napačna, so s pomočjo ChatGPT-ja poskusili rešiti nalogo gumbom "preberi" so lahko dobili glasovno predstavitev

(Slika 1). Če jim to ni uspelo, je pomagala učiteljica. Na začetku besedila, kar jim je olajšalo razumevanje nalog. Gumb "kopiraj"

so učenci nalogo prepisali v ChatGPT, nato pa je en par našel jim je omogočil hitro in enostavno prenašanje besedila v druge

funkcijo snemanja, tako da so nato naloge le prebrali ali pa aplikacije ali dokumente. Z možnostjo "ustvari znova" so lahko

poslikali. Ob koncu ure smo izvedli evalvacijo (Slika2). dobili novo različico odgovora, kar je bilo koristno, kadar prva

razlaga ni bila dovolj jasna. Gumb "slab odgovor" so uporabljali

za označevanje neustreznih odgovorov, kar je pomagalo

izboljšati delovanje sistema. Funkcija "sprememba modela" jim

je omogočala preklapljanje med različnimi modeli umetne

inteligence, da bi našli tistega, ki je najbolj ustrezal njihovim učenci pri pouku reševali na svojih pametnih telefonih. Vsak

potrebam (Slika 3). učenec je reševal učne liste svojih sošolcev, kar je omogočilo

medsebojno izmenjavo znanja in idej. Po reševanju so drug

drugemu podajali povratne informacije, kar je vključevalo



ocenjevanje kakovosti vprašanj, ustreznosti odgovorov ter

splošne uporabniške izkušnje. Ta proces je spodbudil kritično

razmišljanje in izboljšal sodelovanje med učenci, saj so se naučili

prepoznati in popraviti morebitne napake ter pohvaliti dobro

zasnovane naloge.





Slika 4: Del interaktivnega učnega lista



Med izdelovanjem interaktivnih učnih listov smo veliko

poudarka namenili pogovorom o prednostih in slabostih uporabe

umetne inteligence, ki so jih učenci odkrivali med uporabo. Na

koncu smo izvedli tudi evalvacijsko uro, kjer so učenci podali

splošne komentarje na nalogo (Slika 5). Večina učencev je

izpostavila, da naloga ni bila enostavna, saj so se morali naučiti

uporabljati ChatGPT in spletno stran LiveWorksheets. Sprva je

bila naloga za mnoge zahtevna, vendar so si pomagali z

videovodiči za izdelavo interaktivnih učnih listov, ki jih je

posredovala učiteljica. Učenci so povedali, da jim je naloga vzela

Slika 3: Uporaba umetne inteligence pri matematiki veliko časa, a so se pri tem naučili mnogo novega in uporabnega.



Posebej so poudarili, da je ChatGPT pripomoček, ki pomaga



pri delu, ne sme pa delati nalog namesto njih. Učna ura jim je bila

zelo všeč, saj smo se veliko pogovarjali o prednostih in slabostih

uporabe umetne inteligence.



2.2 UPORABA UMETNE INTELIGENCE PRI

IZDELAVI INTERAKTIVNIH UČNIH

LISTOV PRI IZBIRNEM PREDMETU

ŠAH

Slika 5: Splošen komentar na nalogo

Pri izbirnem predmetu šah so učenci prejeli nalogo, da

izdelajo interaktivni učni list s pomočjo spletne strani

Ugotovili so, da ChatGPT ni vseveden in da lahko naredi

LiveWorksheets (https://www.liveworksheets.com/).

napake. Naloga je bila težka, vendar tudi zabavna. Ko so se

LiveWorksheets je spletna platforma, ki omogoča učiteljem in

naučili, kako izdelati interaktivni učni list, so bili ponosni na

učencem ustvarjanje interaktivnih učnih listov. S tem orodjem

svoje izdelke. Učenci so imeli težave z vpisom v ChatGPT in

lahko učitelji pretvorijo tradicionalne papirnate naloge v

prijavo na spletno stran LiveWorksheets, vendar so te težave

digitalno obliko, kjer učenci lahko neposredno rešujejo naloge na

uspešno rešili s pomočjo sošolcev in učiteljice. Predlagali so, da

spletu. Platforma podpira različne tipe nalog, vključno z izbiro

bi še kdaj izvajali podobne naloge, kjer bi delali z različnimi

med več možnimi odgovori, dopolnjevanjem besedila in povleci

programi umetne inteligence in primerjali rezultate.

in spusti nalogami, kar pripomore k bolj dinamičnemu in

Poseben poudarek so namenili učenju pravilnega

angažiranemu učenju. Vsak učenec je prejel eno temo, ki se je

formuliranja ukazov za ChatGPT, saj so pogosto prejeli napačna

nanašala na pravilnike Šahovske zveze Slovenije. Najprej so

vprašanja in odgovore zaradi nepravilno oblikovanih ukazov.

morali temeljito prebrati svojo temo. Nato so s pomočjo umetne

Učenci so ugotovili, da je natančnost in jasnost ukazov ključnega

inteligence, najpogosteje ChatGPT, pripravili osem vprašanj

pomena za pridobivanje ustreznih informacij od umetne

različnih tipov, pri čemer je moral vsak sklop vsebovati vsaj štiri

inteligence. Napačno oblikovani ukazi so pogosto vodili do

možne odgovore, od katerih je bil vsaj en pravilen. Pri tem so

nesmiselnih ali neustreznih odgovorov, kar je povzročilo

morali preveriti smiselnost vprašanj in odgovorov v pravilnikih.

frustracije in podaljšalo čas, potreben za dokončanje naloge.

Sprva so učni list pripravili v urejevalniku besedil, se prijavili na

Skozi proces so se naučili, kako strukturirati svoje poizvedbe na

spletno stran LiveWorksheets in izdelali interaktivni učni list

način, ki je bil bolj razumljiv za ChatGPT. To vključuje uporabo

(Slika 4). Ko so bili interaktivni učni listi pripravljeni, so jih

natančnih ključnih besed, jasnih vprašanj in, kjer je potrebno, Po opravljenem izboru nastalih izdelkov so učenci natisnjene

dodatnih pojasnil za kontekst. slike izrezali in ustvarili svoj plakat z naslovom Our Pets, ki je

Poleg tega so spoznali, da je koristno preveriti odgovore, ki postal izhodišče za obravnavo in utrjevanje angleškega besedišča

jih pridobijo, in jih primerjati z informacijami iz zanesljivih pri naslednjih urah. Učenci so se strinjali, da je spletna stran

virov, kot so pravilniki Šahovske zveze Slovenije. Ta proces je Craiyon uporabna tudi za izdelavo didaktičnega materiala in

učence naučil kritičnega mišljenja in preverjanja informacij, kar slikovnih prikazov. Ugotovili so, da Craiyon izvrši ukaz in

je pomembna veščina v današnjem digitalnem svetu, kjer so ustvari sliko tudi pri nelogičnih ali neresničnih ukazih, kot na

informacije lahko netočne ali zavajajoče. Učenci so skozi to primer »mouse cooking« (Slika 6), kar pomeni, da je uporabnik

izkušnjo pridobili dragocene veščine, ki jim bodo koristile tudi tisti, ki mora vedeti, kako in zakaj uporablja neko spletno orodje.

pri drugih predmetih in nalogah, kjer je uporaba umetne

inteligence lahko v pomoč.

Na koncu so se strinjali, da je naloga, čeprav zahtevna, 2.4 UPORABA UMETNE INTELIGENCE PRI

prispevala k njihovemu boljšemu razumevanju uporabe umetne POUKU KNJIŽNIČNO-

inteligence in jim omogočila pridobitev dragocenih veščin za INFORMACIJSKIH ZNANJ prihodnost.

Učenci četrtega razreda se pri pouku slovenščine učijo tvoriti



2.3 smiselno in zaokroženo besedilo. Tvorjenje besedil je ena težjih UPORABA UMETNE INTELIGENCE PRI

POUKU ANGLEŠČINE nalog, pri čemer so poleg slovničnih znanj potrebna tudi druga

znanja in veščine. Pred samim tvorjenjem besedila pri pouku

Učenci tretjega razreda so pri pouku angleščine obravnavali slovenščine smo izvedli uro knjižnično-informacijskih znanj, pri

sklop Pets ali domače živali. Najprej so ponovili osnovno kateri smo preverjali poznavanje slovenskega besedišča. Najprej

besedišče o domačih živalih, nato so pripravili kratke govorne smo uporabili klasične slovarje v tiskani obliki, nato smo si na

nastope o domačih živalih. Ker vsi niso imeli hišnih ljubljenčkov, spletu ogledali slovarja Fran in Franček. Zanimalo nas je, kako

smo se dogovorili, da bodo izdelali razredni plakat s hišnimi dobro učenci poznajo pomen preprostih besed in kako bo besede

ljubljenčki, pri čemer pa ne bodo sami risali, ampak jim bo razložil ChatGPT v odprti slovenski različici

pomagala umetna inteligenca. Ker pa spletna stran uporablja (https://talkai.info/sl/chat/).

ukaze le v angleškem jeziku, je bilo potrebno pripraviti pravilne Učenci so dobili primer krajšega umetnostnega besedila in ga

besedne zveze, pri čemer se je učenje angleščine izkazalo za prebrali. Izbrali so pet besed, ki so jih zapisali na listič in jih

uporabno. razložili po slovarskem sistemu (Slika 7).

Na levo stran table so učenci zapisali poimenovanja za živali

v tujem jeziku, nato pa so ponovili tudi angleške glagole, ki so

jih zapisovali na desno stran table. Samostojno so tvorili različne

kombinacije živali in dejanj ter jih zapisali na listič, npr. guinea

pig eating, bunny jumping, dog running, cat sleeping. Sledilo je

spoznavanje spletne strani Craiyon ( https://www.craiyon.com/),

ki omogoča brezplačno ustvarjanje slik s pomočjo umetne

inteligence. V nižjih razredih se vedno znova izkaže, da je veliko

lažje, če učenci dostopajo do aplikacij brez uporabniških računov.

Spletna stran Craiyon je preprosto oblikovana, zahteva vpis Slika 7: Primer reševanja



podobe: umetniška (Art), fotografija (Photo), risba (Drawing). Svoje lističe so prinesli k računalnikom v kotičkih in najprej ukaza in omogoča izbor različnih oblik prikazov grafične



Učenci so svoje angleške ukaze prenesli s papirja na spletno stran preverili, na kakšen način ChatGPT razloži pomen besed. Ker pa se tudi ChatGPT zmoti ali pa je njegova razlaga nepopolna, so in oživeli svoje domače živali (Slika 6). morali pomen besede preveriti še v spletnem slovarju. V

nekaterih primerih odgovori ChatGPT niso bili več v slovenščini,



saj je besedo prepoznal kot hrvaško ali srbsko. Učenci so svojo



razlago primerjali z razlago ChatGPT in razlago slovarja. Po

pregledu naključno izbranih besed so ugotovili, da ima ChatGPT

omejeno razlago slovenskih besed, ki ni vedno pravilna, kar pa

ne velja za spletne slovarje. Učenci so kot prednost ChatGPT

navedli preprosto uporabo in ponujen odgovor, ki ga je potrebno

samo prebrati. ChatGPT je pravilno razložil vsakdanje

enopomenske slovenske besede, pri večpomenskih ali

nevsakdanjih besedah pa je bila njegova razlaga nepopolna ali

celo napačna, npr. nevsakdanji, nagel, ugnati (Slika 8).



Slika 6: Primer izdelka z uporabo Craiyon

tehnologije ter njene uporabnosti. Opisani primeri kažejo, da je



uporaba tehnologij z umetno inteligenco preprosta in se jo da

uporabiti na različnih področjih. Učenci z uporabo tehnologije

dobijo vpogled v delovanje umetne inteligence na vsakodnevni

ravni in spoznajo njene omejitve. Vzporedno s samo uporabo pa

je potrebno ozavestiti vzgojni moment odgovornosti uporabnika



Slika 8: Primer razlage besede s ChatGPT uporabljal določena spletna orodja. Navkljub omejitvam, s umetne inteligence, ki se mora zavedati, v kakšen namen bo

katerimi se šole soočajo, digitalne strategije in načrti za

Ob zaključku ure smo na tablo zapisali deset besed po izbiri



razložil besedo. Ko smo prešteli označbe, se je izkazalo, da so postala del življenja sodobnega otroka, ki bo odraščal v hitro spletni slovarji bolje razložili posamezne besede. Po pregledu spreminjajoči se tehnološki družbi. Prav zato bi šole morale najti nabora izbranih besed so učenci prišli do sklepa, da se ChatGPT način, kako integrirati nova znanja v že obstoječi učni program, učencev in označili, kateri spletni pripomoček je pravilno opremo ali strokovna usposabljanja. Umetna inteligenca je razvijanje digitalnih kompetenc, niso vedno vezani na samo



bolje obnese pri razlagi besed iz neumetnostnih besedil, kar so pokazati uporabnost in omejitve novih tehnologij, ki jih sodobni tudi preizkusili. Neumetnostna besedila vsebujejo več preprostih človek srečuje na vsakem koraku. oz. enopomenskih besed, medtem ko umetnostna besedila



vsebujejo ekspresivne in zastarele besede ter besede, ki imajo več LITERATURA IN VIRI pomenov. Najbolj zanimiva pa je bila ugotovitev, da nobena od

tehnologij ne more nadomestiti branja, kjer bralci spoznavajo [1] Vourikari, R., Kluzer, S., Punie, Y. 2023. Dig Comp 2.2: Okvir digitalnih

kompetenc za državljane. Z novimi primeri rabe znanja, spretnosti in

pomen besedišča v različnih kontekstih. stališč. Ljubljana: Zavod Republike Slovenije za šolstvo. DOI:

https://www.zrss.si/pdf/digcomp_2_2_okvir_digitalnih_kompetenc.pdf

[2] Sambolić Beganović, S. et al. 2023. Zakaj digitalna strategija? Priporočila



3 snovalcem digitalnih strategij. Ljubljana: Zavod Republike Slovenije za ZAKLJUČEK

šolstvo. DOI: https://www.zrss.si/wp-content/uploads/2023/08/Zakaj-

digitalna-strategija.pdf

Umetna inteligenca v šoli je uporabna na različnih [3] Kaj je umetna inteligenca in kako se uporablja v praksi? DOI:

predmetnih področjih in na različnih razvojnih stopnjah. Ne https://www.europarl.europa.eu/topics/sl/article/20200827STO85804/kaj

samo, da bi umetna inteligenca morala postati del digitalnih -je-umetna-inteligenca-in-kako-se-uporablja-v-praksi

[4] ChatGPT: Klepetalni robot TalkAI v slovenščini. DOI:

strategij šole, učenci bi morali poznati prednosti in slabosti nove https://talkai.info/sl/





Using Digital Analysis Skills for Description of Simple




Pendulums – Interdisciplinary Example of Combining



Physics, Mathematics and Computer Science



Uporaba digitalne analize pri opisu enostavnih nihal –



Interdisciplinarni primer povezovanja fizike, matematike in



računalništva



Luna Pestotnik Stres

Gimnazija Bežigrad

Peričeva 4,

1000 Ljubljana

luna.pestotnik@gimb.org



ABSTRACT 1 INTRODUCTION



observation of physics experiments. In the case of pendulum computer science – into educational system is still in experiments where the oscillations take place on short time scale development. It is important to make a connection between of tenth of a second, the use of ICT help us to track the pendulum informational technology and science education, as the youth is movements in time. A video recording of a pendulum from a the leading force of the world. In this context it is essential to distant point was made and then analysed using different ICT present informational technology as a way of connecting tools. We studied the impact of different tension in the pendulum education and science. for analysis of a physics experiments. ICT skills are essential in crucial for the development of the society. However, the establishing the potential for automated or digital analysis in the integration of different fields of science – physics, chemistry, In this paper, we describe creation of a digital environment In today’s world, digital skills are becoming increasingly

supporting string on the motion of pendulums and compared it In this paper, we describe an experiment done as part of the

with the theoretical prediction. With the help of independently International Baccalaureate at the Gimnazija Bežigrad. There



library, using analytical tools such as ROOT, we showed the how changing the tension in the supporting string of a coupled interdependence of the independent variable on the coupled pendulum affects the behavior of the two pendulums; secondly, created automated Python program and the OpenCV graphical were two aims to the experiment: Firstly, we wanted to determine

pendulum movement. to learn about digital tools needed to conduct the experiment as



KEYWORDS The goal of this research was to determine the relationship a regular study.

Digital technologies, digital skills, data analysis, coupled between the tension in the supporting string and the period of

pendulum, video recording, smart phone energy exchange. We also explored the energy transfers between

the pendulums and the effect of damping on the system.

POVZETEK The conducted experiment is based on the theory of coupled

V tem prispevku opisujemo vzpostavitev digitalnega okolja pendulums. In the high-school physics curriculum the equations

za analitiko v fizikalnih eksperimentih. IKT spretnosti so of motion for a single mathematical pendulums are derived from

bistvenega pomena pri ugotavljanju možnosti avtomatizirane ali a single differential equation. An interesting extension is an

digitalne analize pri opazovanju fizikalnih eksperimentov. extension to two pendulums, which are connected by a common

Dokazali smo, da ta trditev velja v primeru sklopljenega nihala. support string which produces a driving force depending on the

Analiza videoposnetka sklopljenega nihala je bila narejena. position of one pendulum on another and enables the energy

Rezultati so bili med seboj primerjani glede na silo v vrvici na transfer between pendulums. The motion is in general not

katero sta pritrjeni obe nihali. S pomočjo neodvisno ustvarjenega periodic, for certain conditions the energy is transferred between

Python programa in grafične knjižnice OpenCV ter z uporabo the pendula in a repeated manner. An interesting dependence on

analitičnega orodja ROOT, smo dokazali soodvisnost med the tension in the string can be studied.

spremenljivkami v gibanju sklopljenega nihala. The movement of the pendulums was recorded using a smart

phone camera. The acquisitions had to be taken in an controlled



sklopljeno nihalo, video posnetek, pametni telefon Digitalna technologija, digitalne veščine, analiza podatkov, without any visible features were preliminary to eliminate any potential errors in the detection of the pendulums. KLJUČNE BESEDE environment without shadows and homogeneous background

During the analysis phase the positions of two masses were In a coupled pendulum, the situation is similar. However,

extracted from the video frames and analytical expression with there is an additional force of the supporting spring made by one

free parameters fitted to the data. The fit parameters were then pendulum on the other. On this case the differential equations of

used to study the above dependence. motion of the two oscillators are:



2 METHODS

The analysis of the movement of the coupled pendulum was

made in two phases. The first part of the experiment was

acquiring data in the form of video recording by a smart phone.

The second part of our research was analytical.

Constants ν1, ν2 and μ are introduced as follows:



2.1 Pendulum movement detection



2.1.1 A coupled pendulum setup



Figure1) consisted of two spheres of the same masses on strings For a pair of coupled oscillators, where the conditions for The setup of a coupled pendulum used in the experiment (see

of equal length, attached to a supporting string. The system is simple harmonic motion apply, there are two ways that the

symmetrical so that lengths a and b are equal. One side of the system can oscillate; equal in magnitude and direction or equal

string was passed over a pulley, under which a weight was hung. in magnitude but opposite in direction – called normal modes. In

The force in the string was varied using weights of different case of ν1 = ν2 we get:

masses.





The solutions for equations are therefore:



From the equation, we can see that each oscillator has two

Figure 1. A coupled pendulum setup oscillations; one with frequency 𝜈 , and another with frequency 1

𝜇2 , which we annotate as frequency f. The reciprocal of

2.1.2 Procedure 2𝜈1

frequency f is the period of energy exchange (P); the time one of

Firstly, it was ensured that the pendulums were of equal

the oscillators (when the motion starts at x1 = x0, and x2 = 0)

lengths and weights. One of the mass spheres was released from

takes to completely stop at x=0. It can be calculated as:

a small controlled angle, at different tensions in the string, and

left to swing. The motion of both mass spheres was recorded with

a camera so that exact position and energy dissipation could be

analyzed via video analysis using a custom Python code. A video



for tensions 2.5N, 5.0N, 7.5N, 10N, 12.5N, 15N, 17.5N, 20.0N, 2.1.4 Video recording of the pendulums 22.5N and 25.0N in the string were made. The detection of the motion was done by a digital camera. The

2.1.3 Theoretical background camera was placed vertically above the setup and positioned

Considering a simple harmonic motion of a simple mass- close to the pendulums, because the measurements on the video

spring pendulum. The differential equation driving the motion would otherwise not be detected accurately and would get

can be written as distorted. It was ensured that the material under the pendulums

was uniform and that the environment was controlled in a way

that all shadows were eliminated. The camera used in this



experiment was smart phone iPhone 13 camera.



2.2 Analysis



, where the general solution of it gives us the equation of The experiment was divided into three parts. Firstly, the

motion motion of both spheres was recorded with a smart phone camera

so that exact position and energy dissipation could be analyzed

via video analysis using a custom Python code.





2.2.1 Python video to image processing



The photos taken by the camera were then transferred to a



computer, where further analysis was carried out. The captures





were analyzed using OpenCV library [1] in Python, an open 3 RESULTS source software library for computer vision and machine learning.

The video was divided into images, whereas the capture of 3.1 Determining the displacement

photos was 33 frames per second. The pictures where then Our initial goal was to describe the movement of coupled

converted to greyscale, and a threshold was applied, so that the pendulums. We recorded 5 minutes long videos. After the

masses of pendulums could be recognized. The threshold was acquisition, the analysis was carried out. The images were

denoted in a binary system, where the colour of the picture was analyzed using the described digital tools. After the process of

either black (1) or white (0). The image was divided into two analyzing and converting the videos, a graph of the displacement

regions corresponding to images of two masses. The base code of each pendulum, through time was drawn by fitting the function

was found on the internet and was then changed so that it suited of a coupled pendulum equation on the graph. The graph drawn

our experiment’s needs. by the program is shown below:





Figure 2. A part of OpenCV Python code



2.2.2 Python [2] – calculating the center of the mass



The converted frame to a black-and-white image was then Figure 5. Displacement through time for Pendulum 1 and 2

further analyzed, using the black pixels to calculate the center of in x and y direction both masses:



The movement was analyzed for pendulum1 and pendulum 2

in x and y directions. The movement in x direction is as expected

a sinusoidal oscillation, while the movement in the y direction is

a also sinusoidal with much smaller oscillation due to imperfect

i initial conditions, when the mass was released. On the graph it runs over all the black pixels, and xi and yi are the pixel

coordinates of the i-th pixel. can be seen, that there are two frequencies with which the

pendulums oscillate, the zoom of the motion in x is shown in



Figure 6:



Figure 3. Pendulum analysis

Figure 6. Displacement through time for Pendulum 1

2.2.3 ROOT analysis

ROOT [3] was in this experiment used to determine: The first frequency v presents a normal oscillation of a



- pendulum, while frequency f is a waving behavior of period of The average time between two frames,

- energy exchange. the difference between the two times,

- the difference in displacement

3.2 Determining the velocity

Using the obtained data the velocities were calculated as dx/dt.

Further on, the kinetic and potential energies were calculated as From the graphs x(t) in the x and y direction, velocity of each

mv2 pendulum was determined as the first derivative of displacement /2 and mgh.

using ROOT.





Figure 4.A part of ROOT code





Figure 3 and Figure 4 both show the energy of the Pendulums

1 and 2. Both pendulums are slowly decreasing in the energy, as

some of the energy gets transferred into other forms of energy.

The dissipation of energy is therefore clearly seen in the

following graph.



Figure 7. Velocity through time for Pendulum 1 and 2 in x

and y direction



As expected, the amplitude of velocity of both pendulums in

the x- direction is slowly decreasing, which is corresponding to

the decreasing velocity. In the y direction, the error from the

initial displacement becomes less evident, as the movement and

therefore velocity settles down.

Figure 20. Total energy of the system

3.3 Energy of a coupled pendulum

Both kinetic and potential energy were calculated for each 3.4 The effect of tension on the period of energy

frame, so that a total energy could be plotted. exchange



The system was tested under different initial conditions,

where the tension in the supporting string has been changed. The

effect of different tensions was tested in terms of period of energy

exchange, the time the first pendulum needed to completely stop.

The force in the string has therefore varied from 2.5N to 25N.

When analyzing the video recording and fitting the right function

on the graph, a period of energy exchange was also calculated for

each tension. The results were obtained as follows and plotted in

Excel:

140.00

s) 120.00

(

ge

an 100.00 h

xc

e 80.00

Figure 8. Energy through time for Pendulum 1 and 2 in x gyer 60.00 en f

and y direction o 40.00 d io

er 20.00 P

-

- 5.000 10.000 15.000 20.000 25.000 30.000

Tension in the string (N)





Figure 11. Period of energy exchange as a function of



tension in the string



The interdependence found from the graph is almost linear. It

can be seen that there are some errors to the result, which are a

consequence of different distortions of the results.



4 DISCUSSION

After completing the experiment we have been able to make

a few observations and address several improvements that could

be made to increase the accuracy of the experiment.

Firstly, the smart phone camera could be positioned in a way

that the distortion of the picture would not affect the results. This

Figure 9. Energy through time for Pendulum 1 and 2 in x could be done by making a special set up for the camera to

direction stabilize it, and insure the position of the camera was constant.

Further on, the length of the pendulums could be measured

more accurately, to eliminate the errors due to uneven pendulums.



Lastly, the background on the video could be more even, to

ease the conversion to black and white pixels.

In conclusion, the experiment could be performed more

professionally and accurately.. However, in a high school setting,

when students are limited by time and resources, we believe the

experiments have been performed optimally.



5 CONCLUSION

The purpose of the work was to describe the movement of the

coupled pendulum and to look at the correspondence between the

tension in the string and the period of energy exchange.

For the purpose of the experiment a simple custom program

has been used. The use of this program is a great base for further

explanations and demonstrations of kinematics of oscillating

systems. Basic knowledge is needed for the understanding of the

program, which makes it perfect for teaching high school

students, trying to connect their knowledge of physics, computer

science and mathematics.



REFERENCES

[1] OpenCV: Open source computer vision [internet]. [cited on August 22,

2024]. Available from:

[2] Python. [internet]. [cited on August 22, 2024]. Available from:

https://www.python.org/.

https://docs.opencv.org/4.x/d6/d00/tutorial_py_root.html.

[3] ROOT analysis programme. [internet]. [cited on August 19, 2024].

Available from: http://root.cern.ch.



Umetna inteligenca v izobraževalnem procesu – uporaba



virtualnih laboratorijev pri kemiji



Artificial Intelligence in the Educational Process – The Use of



Virtual Laboratories in Chemistry



Miha Povšič

Srednja ekonomska storitvena in

gradbena šola

Šolski center Kranj

Slovenija

miha.povsic@sckr.si



application offers significant opportunities to enrich the

POVZETEK educational process and enhance the personalized learning

Vpliv umetne inteligenca (UI) v izobraževalnem procesu je approach.

vedno večji, saj se je vključila že skoraj v vse pore šolstva.

Elemente umetne inteligence lahko s pridom izkoristijo tako KEYWORDS

učitelji za pripravo učnih gradiv, kot učenci za samostojno delo, Artificial intelligence, virtual laboratories, personalized

seveda pod budnim očesom učitelja. Članek poudarja prednosti learning

personaliziranega učenja, interaktivnosti in takojšnih povratnih

informacij, ki jih umogoča uporaba UI z uporabo virtualnih 1 Vloga UI v izobraževanju

laboratorijev pri poučevanju kemije in sicer ChemCollective, za Tehnologije UI že vplivajo na številne vidike izobraževanja,

poučevanje kemijskih reakcij. S pomočjo raziskave v katero je od administrativnih nalog do personaliziranih učnih izkušenj. Na

bilo vključenih 57 dijakov, razdeljenih v testno in kontrolno nekaterih ključnih področjih je UI prinesla pomembne izboljšave.

skupino, je bilo ugotovljeno, da uporaba aktivnega laboratorija Personalizirano učenje je ena od teh, kjer sistemi UI analizirajo

pri poučevanju kemijskih reakcij spodbuja boljšo angažiranost podatke o učencih in tako omogočajo prilagojene učne izkušnje,

učencev, boljše razumevanje kompleksnih konceptov in boljše ki naslovijo posamezne prednosti in slabosti [1]. Povečana

učne rezultate. Vključevanje UI v šolah zahteva temeljit angažiranost učencev je še ena prednost, saj interaktivna orodja,

premislek, vendar s smiselno in premišljeno uporabo prinaša kot so virtualni laboratoriji, spodbujajo motivacijo in večjo

veliko možnosti za obogatitev učnega procesa in nadgradnjo vključenost učencev [2]. Poleg tega UI omogoča dostop do

personaliziranega pristopa učenja. širokega nabora izobraževalnih virov, kar olajša tako poučevanje

kot učenje [3].

KLJUČNE BESEDE

Umetna inteligenca, virtualni laboratoriji, personalizirano 1.1 Uporaba virtualnih laboratorijev pri pouku

učenje kemije

Za dodatno obogatitev pouka z možnostjo večkratne

ABSTRACT ponovitve, so virtualni laboratoriji postali neprecenljiva orodja

The impact of artificial intelligence (AI) in the educational za učence. Eden izmed takšnih virtualnih laboratorijev, ki je

process is growing, as it has become integrated into nearly every brezplačen je ChemCollective, ki ponuja realistične simulacije

aspect of schooling. AI elements can be effectively utilized by kemijskih eksperimentov v angleškem jeziku. Številne šole se

both teachers for preparing instructional materials and students dandanes soočajo s prostorsko stisko in nimajo kemijskih

for independent work, all under the careful supervision of the laboratorijev, prav tako pa zaradi kadrovske podhranjenosti

teacher. This article highlights the advantages of personalized marsikatera šola nima asistentov za pripravo in pomoč pri

learning, interactivity, and immediate feedback enabled by AI izvajanju vaj. Določene vaje organske kemije, so za dijake

through the use of virtual laboratories in teaching chemistry, nevarne, saj se pri reakcijah sproščajo strupeni plini. Aplikacija

specifically ChemCollective, for instructing on chemical omogoča učencem izvajanje poskusov v varnem in

reactions. A study involving 57 students, divided into test and nadzorovanem okolju. To je še posebej koristno pri poučevanju

control groups, found that the use of active laboratories in kemijskih reakcij, kjer je praktična izkušnja ključnega pomena

teaching chemical reactions promotes better student engagement, za razumevanje.

improved understanding of complex concepts, and enhanced Z uporabo virtualnih laboratorijev zagotovimo varno okolje

learning outcomes. While the integration of AI in schools za učenje, saj lahko učenci izvajajo eksperimente brez tveganja

requires careful consideration, its thoughtful and deliberate za poškodbe ali škodo. Prav tako pa lahko eksperiment v primeru

neuspeha ponavljajo toliko časa, dokler ne osvojijo določenega

znanja [4]. Realistične simulacije pomagajo učencem bolje

razumeti zapletene koncepte kemijskih reakcij [5]. Ena izmed

prednosti je tudi dejstvo, da lahko do virtualnih laboratorijev

učenci pristopajo kjerkoli in kadarkoli – vse kar potrebujejo je

pametni telefon, tablico ali računalnik. S tem je omogočeno, da

učitelji prilagodijo učno uro glede na učilnico v kateri se nahajajo

istočasno pa lahko učenci dostopajo do virtualnih laboratorijev

tudi od doma, ko želijo določeno učno snov dodatno utrditi ali

zgolj preveriti kakšno zakonitost [1]. Vritualni laboratorij, ki

uporablja UI omogoča tudi takojšnje povratne informacije o

uspešnosti eksperimentov, kar učencem pomaga pri učenju in

odpravljanju napak v realnem času [6]. Slika 2: Prikaz uporabe ChemCollective

Uporaba virtualnih laboratorijev je v šolah smiselna, vendar

se mora tudi smiselno uporabljati. Namen virtualnih 1.2 Implementacija virtualnih laboratorijev pri

laboratorijev ni popolna nadomestitev fizičnih laboratorijev, saj pouku kemijskih reakcij

ne omogočajo urjenja ročnih spretnosti učencev, ki jih lahko Da bi ugotovil uspešnost in smiselnost uporabe virtualnih



dodobra naučijo določenega orodja/aplikacije, ki ga imajo namen razdeljenih v dve skupini: testno skupino, ki je za izvajanje uporabljati pri pouku. Na spletu lahko najdemo številna kemijskih eksperimentov uporabljala virtualni laboratorij usposabljanja/vodiče za virtualne laboratorije, ki prikažejo ChemCollective, in kontrolno skupino, ki je delala v fizičnem učinkovito uporabo virtualnih laboratorijev in njihovo vključitev laboratoriju. Testna skupina je štela 28 učencev, kontrolna pa 29. v svoj učni načrt [2]. Pri pripravi učnega načrta, naj bo glavna Študija je potekala v času ene učne konference. Učinke uporabe smernica ta, da v učni proces vključuje virtualne laboratorije na virtualnega laboratorija sem meril z izvedbo predtestov in pripomočkov za izvajanje kemijskih eksperimentov. Pred izvedel raziskavo med dvema skupinama srednješolskih učencev uporabo virtualnih laboratorijev je smiselno, da se učitelji pri pouku kemije. V raziskavi je sodelovalo 57 učencev, pridobijo le z rokovanjem dejanske steklovine in uporabe ostalih laboratorijev na znanje in posledično učne rezultate učencev, sem

način, ki dopolnjuje tradicionalne metode poučevanja [7]. zapoznelih testov, pri čemer sem uporabil orodje Kahoot za

ocenjevanje znanja v obeh skupinah.





Slika 1: Prikaz Virtualnega laboratorija ChemCollective





Kljub očitnim prednostim virtualnih laboratorijev je potrebno



pomisliti tudi na izzive. Kljub dejstvu, da imajo dandanes skoraj

vsi učenci pametne telefone, ki brez težav omogočajo uporabo Slika 3: Prikaz uporabe ChemCollective v razredu

virutalnih laboratorijev, ne smemo pozabiti na tiste, ki zaradi

takšnih ali drugačnih razlogov pametnih telefonov nimajo. V ta Rezultati, pridobljeni s pomočjo zapoznelega testa,

namen, je smiselno, da ima šola na voljo prenosne računalnike opravljenega preko aplikacije Kahoot, so pokazali statistično

ali tablične računalnike, ki učencem omogočajo sledenje učni uri. pomembne razlike v znanju testne skupine v primerjavi s

V kolikor šola takšnih sredstev ne more zagotoviti, lahko učence kontrolno skupino in sicer za kar 7%. Vprašanja, ki so se

razporedimo v dvojice ali manjše skupine, kjer na manjšem nanašala na odnos in ne na znanje učencev o kemijskih reakcijah

številu telefonov izvajajo vaje. [4]. Ker uporaba UI vključuje pa so prikazale še večje odstopanje, saj je bilo dijakom takšno

zbiranje in analizo podatkov o učencih, je pomembno, da ima delo zanimivejše in so pokazali pozitivnejši odnos do učne snovi

šola zapisan informacijsko varnostni dokument, v katerem so – učenci v testni skupini so bili bolj angažirani in motivirani.

določeni ukrepi za varstvo podatkov in za zaščito informacij o 92 % učencev testne skupine je poročalo, da jim je uporaba

učencih [6]. Zaradi algoritmov UI je potrebna tudi previdnost in virtualnih laboratorijev naredila učenje bolj zanimivo in

ozaveščanje učencev o uporbi “piškotkov”, saj se UI uči od dostopno. Rezultati so pokazali, da so imeli učenci testne skupine

uporabnikov in shranjuje njihove pretekla iskanja. Previdno je globlje konceptualno razumevanje, saj so zaradi večjega števila

treba obravnavati tudi potencialno prekomerno odvisnost od ponovitev v virtualnih laboratorijih imeli bolje reševali probleme

tehnologije [1]. in imeli večjo sposobnost uporabe kemijskih konceptov v

resničnih scenarijih. Povratne informacije iz testne skupine so



poudarile učinkovitost virtualnih laboratorijev pri zagotavljanju

takojšnjega odziva.

Znaten napredek pri testni skupini in višja angažiranost med

učenci testne skupine kažejo na (začetno) učinkovitost virtualnih

laboratorijev pri omogočanju boljših učnih rezultatov. Eden

izmed ključnih vidikov, ki prispevajo k temu uspehu je

nedvomno interaktivno učenje, ki omogoča prilagajanje učenja

potrebam učencev in prostoru in takojšnje povratne informacije.

Dejstvo, da lahko virtualni laboratoriji prilagodi učno izkušnjo

posameznim potrebam, omogoča učencem napredovanje v

lastnem tempu in po potrebi tudi ponavljanje zahtevnih

konceptov [1]. Realistične simulacij v virtualnih laboratorijih

omogočijo interaktivno učno okolje, ki abstraktne kemijske

koncepte naredilo bolj oprijemljive in razumljive [4].



1.3 Zaključek

Vključitev umetne inteligence v izobraževalni proces je zgolj

vprašanje časa, saj ga že dandanes uporabljajo številni učenci za

lažje pridobivanje informacij pri pouku ali učenju. Uporaba

virtualnih laboratorijev pri poučevanju kemijskih reakcij pri

pouku kemije, ponuja pomembne prednosti v smislu

personaliziranega učenja, angažiranosti in konceptualnega

razumevanja. Rezultati raziskave, ki jo opisujem v tem članku

prikazujejo pozitiven vpliv virtualnih laboratorijev na učne

rezultate učencev, kar spodbudno kaže na njihov potencial za

prilagoditev tradicionalnih metod poučevanja. Trditi, da lahko

virtualni laboratoriji nadomestijo fizične je seveda neprimerno in

nesmiselno, saj ne nadomestijo pridobivanja ročnih spretnosti

učencev.

S tem, da učitelji sprejmejo UI in so jo pri pouku pripravljeni

tudi uporabljati kot dodatno spodbudo oz. način za boljše

poučevanje, naredijo prvi korak k pripravi učencev na prihodnost,

kjer je tehnološka pismenost ključnega pomena.



LITERATURA

[1] Holmes, W., Bialik, M., & Fadel, C. (2019). Artificial Intelligence in

Education: Promises and Implications for Teaching and Learning. Boston:

Center for Curriculum Redesign.

[2] Luckin, R., Holmes, W., Griffiths, M., & Forcier, L. B. (2016).

Intelligence Unleashed: An argument for AI in Education. London:

Pearson.

[3] UNESCO. (2021). AI in Education: Challenges and Opportunities for

Sustainable Development. Pariz: United Nations Educational, Scientific

and Cultural Organization.

[4] Zawacki-Richter, O., Marín, V. I., Bond, M., & Gouverneur, F. (2019).

Sistematični pregled raziskav o aplikacijah umetne inteligence v

visokošolskem izobraževanju: Predhodni rezultati. International Journal

of Educational Technology in Higher Education, 16(1), 39.

doi:10.1186/s41239-019-0179-2

[5] Roll, I., & Wylie, R. (2016). Evolucija in revolucija umetne inteligence v

izobraževanju. International Journal of Artificial Intelligence in

Education, 26(2), 582-599. doi:10.1007/s40593-016-0110-3

[6] Russell, S., & Norvig, P. (2020). Artificial Intelligence: A Modern

Approach (4. izd.). Pearson.

[7] Popenici, S. A. D., & Kerr, S. (2017). Raziskovanje vpliva umetne

inteligence na poučevanje in učenje v visokošolskem izobraževanju.

Research and Practice in Technology Enhanced Learning, 12(1), 22.

doi:10.1186/s41039-017-0062-8



Prednosti IKT tehnologije pri pouku športne vzgoje



Advantages of ICT Technology in Physical Education Lessons



Polona Rajher

III. Gimnazija

Maribor, Slovenija

polona.rajherl@gmail.com



particular product. As a result, unhealthy foods are often

POVZETEK mistaken for healthy foods and eaten in good faith. In order to

V prispevku je predstavljeno, kako se lahko s pomočjo IKT approach adolescents in a different way the importance of a

tehnologije lotimo izziva prekomerne telesne teže dijakov in healthy diet for their development and metabolism, a study was

dijakinj ter spregovorimo o pomenu zdrave prehrane in presnove, conducted in which students first tested their knowledge of

ki je v času odraščanja za mladostnike ključnega pomena. healthy eating, completed a survey of their eating habits, and then

Mladostniki se v našem učnem sistemu pogosto srečajo s participated in workshops on the importance of healthy eating

pomenom zdrave prehrane, prav tako pa je zdrava prehrana and metabolism. Through the Energy for life application,

pogosto tema v medijih. Kljub pogostemu omenjanju pomena students learned the importance of a healthy diet and how

zdrave prehrane, je raziskava pokazala, da mladostniki nutrients are metabolised in our bodies, and then tested their

velikokrat posežejo po nezdravi hrani. Več kot polovica knowledge with a questionnaire. At the end of the workshop,

anketiranih mladostnikov je mnenja, da se prehranjujejo zdravo, students used the application Veš kaj ješ? and made examples of

vendar so rezultati pokazali, da je v resnici takšnih zgolj 28%. healthy menus.

Potrošniki pogosto v trgovini kupujemo hrano glede na izgled

embalaže, mladostniki pa so glede na vizualni izgled in KEYWORDS

oglaševanje določenega izdelka še bolj dovzetni. Posledično ICT technology, healthy nutrition, overweight, workshop,

nezdravo hrano pogosto zamenjajo za zdravo in jo v prepričanju, metabolism, aplication

da je zanje ustrezna, jedo. Da bi mladostnikom na drugačen način

približali pomen zdrave prehrane, je bila opravljena raziskava, v

kateri so dijaki najprej preverili svoje znanje o zdravi prehrani, 1 UVOD

izpolnili anketo o njihovih prehranjevalnih navadah in bili nato Zdrava prehrana je nekaj o čemer se zadnje čase zelo pogosto

vključeni v delavnice o pomenu zdrave prehrane in presnove. Na govori. Pomen zdrave prehrane je ključen za zdrav razvoj, saj

delavnicah so dijaki s pomočjo aplikacije Energy for life na telo za izgradnjo notranjih ogranov, mišic, kostnine in živčevja

interaktivni način spoznali pomen zdrave prehrane in kako se nujno potrebuje hranila, ki jih vsebuje hrana. V zgodnjem

hranila presnovijo v našem telesu. Svoje znanje so nato preverili otroštvu se običajno prehranjujemo bolj zdravo in uživamo

z vprašalnikom. Na koncu delavnice so dijaki s pomočjo raznoliko prehrano, ki nam omogoča zdrav razvoj. Otrokov

aplikacije Veš kaj ješ? naredili primere zdravih jedilnikov. jedilnik določajo starši oz. skrbniki in posledično ne moremo kot

posamezniki vplivati na izbor hrane in uživati zgolj hrane, ki nam

KLJUČNE BESEDE je všeč. V času pubertete, pa mladostniki počasi začnejo sami

IKT tehnologija, zdrava prehrana, prekomerna telesna teža, določati svoj jedilnik in pri tem večkrat ne upoštevajo smernic

delavnice, presnova, aplikacija zdravega prehranjevanja. Nepravilen izbor živil in njihovo

uživanje pa lahko pusti posledice, ki se pokažejo kot nepravilna,

ABSTRACT neenakomerna rast za vse življenje, prav tako pa lahko vpliva na

The article presents how we can tackle the challenge of oblikovanje prehranjevalnih navad v kasnejšem življenju. [6]

excessive body weight in male and female students and address Zaradi upada upoštevanja smernic zdrave prehrane skozi



eating in our learning system, and healthy eating is often a topic so sprva izpolnili vprašalnik, ki je preveril njihove prehranske navade, nato pa so preko didaktične delavnice, pri katerih so in the media. Despite the frequent mention of the importance of uporabljali interaktivni aplikaciji, ki mladostnike spodbudita k a healthy diet, research has shown that adolescents often resort boljšem razumevanju pomena zdrave prehrane in njene presnove, to unhealthy foods. More than half of the adolescents surveyed spoznavali načela zdravega prehranjevanja. Z vidika varovanja said they were eating healthy, but the results showed that only zdravja je namreč zdrava prehrana za mladostnike izredno 28% actually did. Consumers often buy food at the store based technology, which is crucial for adolescents during their growth. namen je ugotoviti prehranske navade mladostnikov, v kateri so sodelovali dijaki tretjih letnikov III. gimnazije Maribor. Dijaki Adolescents are often confronted with the importance of healthy the importance of healthy diet and metabolism using ICT obdobje pubertete je bila posledično opravljena raziskava katere

on the appearance of the packaging, and adolescents are even pomembna, saj imajo ravno v tem obdobju veliko večje potrebe

more receptive to the visual appearance and advertising of a po beljakovinah, vitaminih in mineralih ter energiji. [1]



2 Ker so bolezni, ki jih povzroča nezdrava prehrana velik POMEN ZDRAVE PREHRANE



gospodinjstva srečajo s pomenom zdrave prehrane, kjer se učijo Mladostniki se že v času osnovne šole pri pouku javnozdravstveni problem, se veliko pozornosti namenja k omilitvami posledic le teh.

o prehranski piramidi in različni pripravi hrane. Nato se o



prehrani in presnovi učijo pri pouku biologije, kjer največ 3 INTERAKTIVNO POUČEVANJE poudarka namenijo presnovi ter vgradnji hranil v človeško telo

in posledicah nepravilne prehrane in prehranskih navad. Njihove Ustaljena praksa poučevanja mladostnikov o pomenu zdrave

prehranske navade so v času osnovne šole običajno dobre, vsaj v prehrane, očitno ne prinaša dobrih rezultatov, saj se delež

prvi in drugi triadi. V tretji trijadi pa mladostniki med glavnimi mladostnikov, ki imajo slabe prehranske navade in posledično

obroki pogosto uživajo prigrizke za katere je značilno, da zdravstvene težave, iz leta v leto povečuje. Razlog za slabe

vsebujejo veliko energije, zaradi prisotnosti sladkorjev in navade mladostnikov, bi lahko bili tudi zastareli načini

maščob, ne vsebujejo pa hranilnih snovi, ki jih telo potrebuje za poučevanja v šoli, saj so učenci zaradi uporabe interneta,

krepitev odpornosti. Ena izmed praks mladostnikov v tretji triadi mobitelov in televizije navajeni na interaktivno pridobivanje

je opuščanje zajtrka in neustrezen ritem prehranjevanja ter informacij. Frontalen način poučevanja, katerega uporablja

premajhne količine zaužite zelenjave. Vendar se prehranske večina učiteljev, v večini primerov ne doprinese k trajnostnem

navade osnovnošolskih otrok v tretji triadi v veliki meri znanju, saj so učenci do takšnega pridobivanja informacij vedno

razlikujejo od navad srednješolskih otrok, saj le ti velikokrat manj motivirani pri učnem procesu [4]. Učitelj mora dan danes

nimajo primerne šolske malice. [5] učence spodbujati, da sami pridejo do določenih spoznanj, ne da

Zdravo prehrano bi lahko opredelili z naslednjimi jim zgolj posreduje znanje. Z uporabo in upoštevanjem

smernicami: konstruktivističnega pristopa pri izkustvenem učenju lahko

- redna prehrana (vsaj 5 obrokov dnevno), učitelj učencem pomaga povezati ugotovitve s praktičnimi

- uživanje hrane glede na prehransko piramido, zakonitostmi. Uporaba IKT (informacijsko komunikacijske

- raznolikost hrane, tehnologije) lahko omogoči posredovanje učne snovi na bolj

- počasno uživanje hrane, ki jo dobro prežvečimo, atraktiven način. Pri pouku lahko uporabimo tako računalnike,

- izogibanje sladkim in mastnim jedem, kot tudi mobilne telefone, sploh v primeru, ko imamo v vzgojno

- izogibanje prigrizkom [2]. izobraževalnem zavodu težave zagotoviti primerno število

Iskanje krivca slabih prehranskih navad mladostnikov ni računalnikov za večjo skupino učencev. Uporaba mobilnih

enostavno, saj na njihove navade vpliva mnogo različnih telefonov pri pouku ima tudi svoje omejitve in predstavlja

dejavnikov. Živimo v okolju, ki spodbuja debelost, saj smo kot določene izzive, sploh od učiteljev. Prednost mobilnih telefonov

družba izredno potrošniko naravnani in nam mediji dnevno je v tem, da jih ima večina učencev in da so običajno zmogljivejši

posredujejo vedno nove in nove proizvode. Poleg tega, pa nas od šolskih računalnikov [3].

okolje spodbuja k temu, da smo vedno manj aktivni (električna



kolesa, električni skiroji, uporaba osebnega vozila...). Podatki za 4 APLIKACIJE ZA POUČEVANJE O

Slovenijo kažejo naraščajoče gibanje prekomeno hranjenih in ZDRAVI PREHRANI IN PRESNOVI debelih mladostnikov in odraslih pri nas, kar kaže, da se navade

pogosto prenašajo s starejše na mlajšo generacijo [7]. Debelost Na spletu je množica različnih aplikacij in spletnih strani, ki

pri mladostnikih je namreč eden izmed največjih problemov opisujejo zdravo prehranjevanje vendar je zgolj peščica

današnjega časa, saj prekomerna telesna teža v mladostniškem didaktično primernih za uporabo v učnem procesu. Večina

obdobju zelo pogosto nakaže na prekomerno telesno težo tudi v aplikacij je namenjena ali izobraževanju ali zabavi, vendar le

odraslem obdobju. Pri mladostniku neuravnotežena prehrana malo je takšnih, ki bi združevale oba spektra. Ena izmed aplikacij,

povzroča manjšo delovno storilnost, utrujenost, manjšo ki je tako poučna, da uporabnik pridobi in utrdi svoje znanje, kot

sposobnost pomnenja in kronične bolezni. Strokovnjaki so tudi zabavna, da se ne naveliča in je pripravljen dlje časa ostati

ugotovili, da so številne bolezni neposredno povezane s hrano, motiviran, se imenuje Energy for life. Aplikacija sicer ne ponuja

kar pomeni da je od prehrane odvisno naše zdravje. Nepravilna uporabe v slovenskem jeziku, temveč je zgolj v angleškem in

prehrana pa je večkrat povezana z različnimi slabimi navadami španskem jeziku. Posledično se lahko aplikacijo uporabi kot

kot je telesna neaktivnost in kajenje. Bolezni katere lahko primer medpredmetnega povezovanja biologije in angleščine.

neposredno povežemo s prehrano so: Prednost aplikacije je tudi v tem, da je popolnoma brezplačna in

- beri-beri (posledica, dolgoročnega pomankanja vitamina omogoča več različnih učnih modulov. Aplikacija s pomočjo

B1), igrifikacije – uporaba dinozavra, ki pritegne pozornost

- skorbut (nastane zaradi pomanjkanja vitamina C v daljšem uporabnika, uči novih konceptov in ga motivira za interakcijo,

časovnem obdobju), kar izboljša učni proces. Zaradi enostavnega umesnika je ta

- hipertenzija (povišan krvni tlak zaradi hrane, ki vsebuje aplikacija dostopna vsem učencem različnih starosti, tako v

visoke vrednosti nitratov), osnovni, kot tudi v srednji šoli. Vizualen prikaz in strokovna

- zobna gniloba (posledica prekomernega uživanja sladke natančnost izrazov, se določi s pismenimi sposobnostmi

hrane in slabe ustne nege). uporabnika. Velik poudarek je na znanstveni pismenosti,

Obstaja še množica drugi bolezni, vendar njihovega nastanka povezan z vizualnimi dražljaji, ki uporabnika spodbujajo k

ne morejo povezati izključno z slabimi prehranskimi navadami. aktivni uporabi aplikacije. Aplikacija je zasnovana tako, da se jo

[8] lahko uporablja ob spremstvu vodje učnega procesa ali pa se jo

uporablja samostojno. Slika 1 prikazuje dinozavra, ki

mladostnika nagovarja, kaj vse lahko s pomočjo aplikacije izve. premajhnih količinah. Pri sliki 4 lahko vidimo, kako je razčlenjen

Uporabnik lahko izbira med različnimi področji, ki jih lahko s prehranski semafor, ki uporabniku sporoča še dopustno vrednost

pomočjo aplikacije osvoji in utrdi svoje znanje. Slika 2 prikazuje določenih hranil v izdelku. Aplikacija Veš kaj ješ? ni namenjana

modul prebavil, katerega uporabnik lahko izbere v aplikaciji in s zabavi in je posledično manj atraktivna kot aplikacija Energy for

pomočjo njega pregleda kaj vse njegovo telo potrebuje za life, zato je vsaj uvodna uporaba aplikacije priporočljiva pod

delovanje. Nato uporabnik preveri kaj se zgodi, če tega hranila vodstvom učitelja. Učitelj mora pred uporabo mobilnih telefonov

ne dobi (nastanek različnih obolenj). Prav tako, ga aplikacija kot učnega pripomočka med poukom oceniti smiselnost, pri

spodbuja k nadaljnjem razmišljanju in preizkušanju naučenega kateri učni snovi ga bo uporabil in katere učne cilje bo s pomočjo

znanja. njih usvojil. Prav tako ni namen vsakodnevne uporabe mobilnih

telefonov pri učnem procesu, saj bi s tem njegova atraktivnost

izzvenela, prav tako pa bi lahko postal moteči faktor v razredu

zaradi morebitne nedovoljene uporabe iz strani učencev.





Slika 1: Prikaz različnih modulov aplikacije



Slika 3: Prikaz aplikacije Veš kaj ješ?





5 RAZISKAVA



Da bi ugotovili odnos do zdrave prehrane mladostnikov in

preverili znanje o pomenu zdravega prehranjevanja in presnove,

je bila opravljena raziskava, v kateri so sodelovali dijaki tretjih

letnikov III. gimnazije Maribor. Za namen raziskave je bilo

izbranih 42 dijakov, ki so sprva izpolnili vprašalnik, ki je bil

sestavljen iz 14 vprašanj, katerih namen je bil ovrednotiti znanje

dijakov o zdravi prehrani in presnovi, prav tako pa je bil namen

vprašalnika preveriti njihove prehranske navade. Nato so bile

izvedene 3 učne ure, pri katerih so dijaki s pomočjo aplikacije

Energy for life ugotovili pomen zdrave prehrane in presnove,

nato pa so s pomočjo aplikacije tudi preverili svoje znanje. Pri

učni uri je bila vloga učitelja zgolj kot usmerjevalec učnega

procesa in ne posredovalca znanja. Po uporabi aplikacije so bili

Slika 2: Prikaz modula hranila dijaki pozvani kakšne so njihove prehanske navade in opisu

njihovega tedenskega jedilnika. Ker so imeli pri jedilniku večje

Aplikacija Veš kaj ješ? omogoča uporabniku preveriti težave, so nato s pomočjo aplikacije Veš kaj ješ? preverili kakšne

energijsko vrednost hrane. Prav tako uporabniku omogoča, da s jedi so primerne za določen del dneva in katerih jedi se morajo

pomočjo prehranskega semaforja ugotovi, katerih živil naj se izogibati pri svoji prehrani. Po zaključku dela z aplikacijo so

zaradi prekoračenih vrednosti določenih hranil izogiba. Namen dijaki s pomočjo razgovora posredovali kaj so se naučili in

aplikacije je pomoč pri sestavi jedilnika, katerega lahko vsak izročke oblikovane skupaj z učiteljem zapisali v zvezke. Nato so

posameznik sestavi glede na svoje prehranske navade. Slika 3 ponovno izpolnjevali vprašalnik, s katerim je bilo preverjeno, kaj

prikazuje primer uporabe aplikacije na pametnem telefonu med so se naučili. Določeni rezultati vprašalnika so bili pričakovani,

opravljanjem nakupov. Aplikacija nam omogoča uporabo saj se je tako znanje kot tudi odnos do zdrave prehrane izboljšal.

kamere pri pametnem telefonu in s tem skeniranje črtne kode Presenetljivo je bilo to, da je bilo 54% odstotkov dijakov mnenja,

izdelka, katerega prehransko vrednost želi uporabnik preveriti. da se prehranjuje zdravo, vendar je bilo nato v nadaljnjem delu

Aplikacija nam nato s pomočjo barvnega semaforja prikaže rezultatov vprašalnika prikazano, da so zmotno prepričani v svoj

katerih hranil je v izdelku v zadostnih, prekomernih ali zdrav način prehranjevanja. Kar 48% dijakov navaja premalo



časa kot glaven krivec za njihove slabe prehranske navade, 35%

dijakov pa navaja, da stres močno pripomore k njihovim

prehranskim navadam. Kot glaven razlog zakaj kupijo določen

proizvod so dijaki podali izgled in njegovo oglaševanje. Večina

dijakov ne pregleda hranilnih vrednosti pred nakupom produkta.

Pri vprašanju katera pijača ima najvišjo vrednost sladkorja, je

sprva od 42 dijakov pravilno odgovorilo zgolj 5 dijakov, kar

nakazuje na njihovo zmotno znanje o hranilih v sladkanih pijačah.

Pri zapoznelen vprašalniku pa je pravilno odgovorilo 18 dijakov.

Večina dijakov, kar 34 od 42 je odgovorila pri zapoznelen

vprašalnikov, da bodo več pozornosti namenili svojim

prehanskim navadam.



6 ZAKLJUČEK

Ključno je, da se o zdravi prehrani začnemo pogovarjati čim

prej v začetku izobraževalnega procesa. Vendar zgolj frontalno

poučevanje ne prinaša željenih rezultatov, saj se delež

prekomerno debelih otrok in mladostnikov v zadnjem času

drastično povečuje. V raziskavi je predstavljen eden od načinov,

kako lahko izobrazimo mlade o pomenu zdrave prehrane.

Aplikacije, ki jih dan danes učenci uporabljajo pri pouku, so

dostopne preko mobilnih telefonov, kar pomeni, da lahko tudi

vzgojno izobraževalni zavodi z omejitvami pri omogočanju

dovolj velikega števila računalnikov za vse učence, omogočijo

svojim učencem izvedbo pouka na takšen način. Ali bo uporaba

takšnih aplikacij drastično vplivala na prehranske navade

učencev je neodgovorno trditi, vendar njihova uporaba lahko

pripomore k boljšem ozaveščanju zdravega načina

prehranjevanja in razumevanju presnove hranil ter zdravemu

načinu življenja.



VIRI

[1] Gavin, L. M., Dowshen, S. A., Izenberg, N.(2007). Otrok v formi.

Praktični vodnik za vzgojo zdravih otrok – od rojstva do najstniških let.

Ljubljana: Mladinska knjiga.

[2] Kodele, M., Suwa – Stanojević, M., Gliha, M. Prehrana. Ljubljana: DZS,

2000.

[3] Lobe, B. in Muha, S. Internet v vsakdanjem življenju slovenskih otrok in

mladostnikov (online). 2012. (citirano 15. 5. 2023). Dostopno na naslovu:

http://www.safe.si/uploadi/editor/1297947005MLADINAN

ETU_porocilo.pdf.

[4] Ormrod, J.E. (2014). How Motivation Affects Learning and Behavior.

Education Psychology, Developing Learners’ (online). 2014. (citirano 21.

5. 2023). Dostopno na naslovu:

https://www.education.com/reference/article/motivation- affects-

learning-behavior/.

[5] Polajžar, A, Razgor, E, Regner, A. (2009). Zakon spreminja prehranske

navade srednješolcev »bolje pozno kot nikoli«. Celje.

[6] Sizer, F., Whitney, E. (2006). Nutrition. 10th ed. Belmont: Thomson

Wadsworth.

[7] Zupančič, A., Hoyer, S. Prehranjevalne navade študentov.

[8] Žitnik, N. Trtžnokomunikacijski program zdrave prehrane med dijaki in

študenti. Ljubljana: 2009.





Informacijska tehnologija pri inkluzivnem pouku tujega




jezika



Information technology in inclusive foreign language lesson



Tina Rajhman

Šolski center Kranj

Slovenija

tina.rajhman@sckr.si



razrede, kar je v preteklosti vodilo k uvedbi koncepta integracije

POVZETEK v naše šole. Koncept integracije pravzaprav izvira z Danske, kjer

Vključevanje digitalnih orodij kot so Animoto, Book Creator je bila posledica sprememb v vzgoji in izobraževanju oseb z

ter animacija lutk ali predmetov v ure angleščine omogoča motnjami v razvoju. Takratni danski minister je integracijo

ustvarjanje vključujočega učnega okolja za učence z različnimi opredelil kot vključevanje oseb z razvojnimi izzivi v normalno

potrebami. Animoto omogoča enostavno ustvarjanje in okolje v največji možni meri. To je pomenilo selektivno

prilagajanje videoposnetkov, medtem ko animacija ponuja vključevanje otrok in mladostnikov s posebnimi potrebami v

praktičen način za vizualno predstavitev idej v lastnem tempu. redne vrtce in šole Error! Reference source not found..

Book Creator omogoča ustvarjanje in deljenje digitalnih knjig ter Opozoriti velja, da inkluzija ne vključuje le otrok, ki so

spodbuja sodelovanje med učenci. Ta orodja spodbujajo prepoznani kot otroci s posebnimi potrebami, temveč tudi druge.

kreativnost, učencem pa omogočajo prilagajanje učnih izkušenj To so otroci, ki izvirajo iz socialno depriviligiranih in

lastnim individualnim potrebam, kar povečuje njihovo subkulturnih okolij, otroci priseljencev ali beguncev, otroci, ki

angažiranost in omogoča učinkovitejše učenje. Tako lahko imajo težave s socialno integracijo Error! Reference source not

učitelji ustvarijo okolje, kjer se vsi učenci lahko izražajo na njim found., otroci iz etničnih manjšin ter otroci iz družin z nižjim

najprimernejši način. socialno-ekonomskim statusom Error! Reference source not

found.. Ti učenci se pogosto soočajo z izzivi v šolah zaradi



KLJUČNE BESEDE jezikovnih ovir, pomanjkanja spodbud doma, drugačnih vrednot

in zato potrebujejo znatno podporo tako s strani šole kot učiteljev

Inkluzija, tuji jeziki, informacijska tehnologija, Animoto,

Error! Reference source not found.. Lesar Error! Reference

animacija, Book Creator.

source not found. navaja, da inkluzivno izobraževanje zajema



ABSTRACT tudi otroke, katerih starši začasno delajo v tujini, ter otroke z

različnimi spolnimi usmeritvami. Zakon o spremembah in

Integrating digital tools such as Animoto, Book Creator and dopolnitvah Zakona o osnovni šoli Error! Reference source not

stop-motion animation into English lessons creates an inclusive found. omenja tudi nadarjene učence, za katere mora šola

learning environment for students with diverse needs. Animoto zagotoviti ustrezno izobraževanje s prilagoditvijo učnih metod,

allows for easy creation and customization of videos, while stop- oblik dela, vsebin in omogočiti njihovo sodelovanje v dodatnih

motion animation offers a hands-on approach to visually urah ter drugih oblikah šolskih dejavnosti. presenting ideas at one’s own pace. Book Creator enables the

creation and sharing of digital books, encouraging collaboration 2 Inkluzija pri pouku tujega jezika among students. These tools foster creativity and allow students

Čeprav je poučevanje tujega jezika z upoštevanjem inkluzije

to tailor their learning experiences to their individual needs,

lahko zelo zahtevno, primer v nadaljevanju pokaže, kako

thereby increasing their engagement and enabling more effective

inkluzijo lahko omogočimo s pomočjo informacijske tehnologije.

learning. This way, teachers can create an environment where all

Primer ure angleškega jezika v srednji šoli se osredotoča na

students can express themselves in the way that suits them best.

roman "Skrivnostni primer ali Kdo je umoril psa" britanskega



KEYWORDS pisatelja Marka Haddona. Roman se osredotoča na protagonista

in pripovedovalca, Christopherja Boonea, 15-letnika z avtizmom.

Inclusion, foreign languages, information technology, Cilj pri načrtovanju ure je obravnavanje kompleksnosti avtizma,

Animoto, stop-motion animation, Book Creator. da bi se dijaki lažje spoprijeli z izzivi, ki jih postavlja ta zanimiva

tema.

1 Integracija in inkluzija Ura je zasnovana z namenom, da bi dijaki prebrali kratek

Integracija in inkluzija sta precej podobna koncepta, ki se odlomek iz romana, izbrani izsek besedila pa izrecno poudarja,

razlikujeta glede na različne poglede na obravnavano temo, da ima glavni junak posebne lastnosti. Predstavljeni odlomek je

nedvomno pa gre v obeh primerih za vključevanje otrok s tako za dijake kot za bralce tega članka v izvornem jeziku, torej

posebnimi potrebami v šolo, pri čemer se jim nudi pomoč, v angleščini.

prilagoditve in podporo na vseh ravneh.

Zakonska podlaga Error! Reference source not found.

določa vključevanja učencev s posebnimi potrebami v redne



tako pa se poglablja v širše vprašanje, ali se lahko vživijo in

potencialno vzpostavijo povezave z nekom, ki ima avtizem.



2.1 Inkluzija z informacijsko tehnologijo

Kot učitelji smo dobro seznanjeni z edinstvenimi močmi in

preferencami naših učencev ali dijakov. Glede na to da poučujem

angleščino v programih, usmerjenih v računalništvo, pouk lahko

popestrim z navodilom dijakom, naj pripravijo predstavitev

Christopherja ali pa dela romana z uporabo enega od

razpoložljivih informacijskih orodij. V nadaljevanju

predstavljena orodja so zanimiva alternativa PowerPointu, saj so

uporabniku prijazna in ponujajo zanimive rešitve.



2.1.1 Animoto

Animoto je vsestransko in brezplačno orodje za ustvarjanje

videoposnetkov. Njegovi uporabniki lahko ustvarijo in delijo

neomejeno število videoposnetkov na svojih računih, brez da bi

bilo treba naložiti dodatno programsko opremo. Uporabniki

projekt začnejo tako, da dodajo vse potrebne slike in posnetke,



nato pa jih preprosto povlečejo in spustijo v nastajajoči video.

Platforma ponuja vrsto oblikovalskih orodij, ki uporabnikom

omogočajo, da prilagodijo barve, pisave, glasbo in še več. Error!

Reference source not found.



2.1.2 Stop Motion Animation

Tako imenovana stop-motion animacija je tehnika, ki se

uporablja v animaciji za oživitev statičnih predmetov na zaslonu.

To se doseže z gibanjem predmeta v majhnih korakih, medtem

ko se v vsakem koraku posname en kader. Ko se vsi kadri

predvajajo v zaporedju, prikazujejo gibanje. V opisani animaciji

se pogosto uporablja glinene figure, lutke in miniaturne figure,

saj jih je mogoče enostavno upravljati in premikati. Error!

Reference source not found.



2.1.3 Book Creator

Book Creator je digitalno orodje, ki uporabnikom omogoča

ustvarjanje in branje multimodalnih digitalnih knjig, torej tudi

spodbuja angažiranost učencev oziroma dijakov, ki lahko

ustvarjajo zgodbe in jih multimedijsko oblikujejo. Pri ustvarjanju

svojih knjig lahko uporabniki ustvarijo lastno vsebino (s

pisanjem, risanjem ali nalaganjem lastnih videoposnetkov,

zvočnih datotek in slik) ali pa izberejo spletne multimedijske vire

z interneta. Book Creator se lahko uporablja tudi pri projektno

usmerjenih učnih dejavnostih, saj omogoča sodelovanje več

uporabnikov v realnem času pri ustvarjanju ene knjige. Error!

Slika 1: odlomek iz romana "Skrivnostni primer ali kdo je Reference source not found.

umoril psa"



Za nalogo, ki vključuje predstavitev Christopherja, lahko



ključne besede, ki opisujejo glavnega junaka, in ki poudarijo Ko dijaki besedilo preberejo, dobijo nalogo, da identificirajo dijaki torej prilagodijo videz svojega izdelka, da se ujema s tematiko in njihovo individualno ustvarjalno vizijo.

njegovo edinstvenost. Dijaki v parih primerjajo svoje sezname in



svoje izbire predstavijo drugemu paru. V zadnjem delu so 3 Zaključek dijakom predstavljene tri možnosti, ki temeljijo na inkluziji, in

sicer morajo ustvariti ilustracijo, ki prikazuje glavnega junaka Vključevanje digitalnih orodij kot so Animoto, Stop-motion

Christopherja in njegov svet; odgovoriti na niz vprašanj; ali pa animation in Book Creator v ure angleščine na sodoben in

napisati kratek odstavek o srečanju s Christopherjem. zanimiv način omogoči inkluzijo dijakov z različnimi

Te tri različne naloge so prilagojene različnim sposobnostim sposobnostmi. Ta orodja ne omogočajo le kreativnega izražanja,

in preferencam dijakov. Izbrani pristop upošteva, ali so dijaki temveč tudi opolnomočijo dijake, da prilagodijo svoje učne

bolj spretni pri izražanju skozi vizualne ali besedne vsebine, prav izkušnje svojim individualnim potrebam.



Uporabniku prijazen vmesnik Animoto omogoča enostavno

ustvarjanje in prilagajanje multimedijskih predstavitev, s čimer

zadovoljuje različne učne stile. Stop-motion animacija ponuja

praktičen pristop, ki učencem omogoča, da vizualno predstavijo

ideje v svojem tempu, kar spodbuja avtonomijo v učnem procesu.

Book Creator s svojo vsestranskostjo ne podpira le ustvarjanja

personaliziranih digitalnih knjig, temveč tudi spodbuja

sodelovanje in omogoča različne načine izražanja. Z uporabo teh

digitalnih orodij lahko učitelji omogočijo vključujoče učno

okolje, v katerem lahko učenci in dijaki ne glede na svoje

sposobnosti sprostijo svojo ustvarjalnost in se vključijo v učni

proces na način, ki jim najbolj ustreza.



VIRI

[1] Animoto. (2024). https://animoto.com/education/classroom (Pridobljeno

18. 8. 2024)

[2] Bela knjiga o vzgoji in izobraževanju v Republiki Sloveniji. (2011).

Ljubljana: Zavod Republike Slovenije za šolstvo.

[3] Book Creator (2024). https://edtechbooks.org/onlinetools/book-creator

Pridobljeno 18. 8. 2024)

[4] Lesar, I. (2013). Ideja inkluzije – med različnimi koncepti pravičnosti in

etičnimi teorijami. Sodobna pedagogika, 64(2), 76–95.

[5] Medveš, Z. (2003). Integracija/inkluzija v vrtcu, osnovni in srednji

šoli. Sodobna pedagogika, 54(posebna izdaja), 278–286.

[6] Opara, B. (2015). Dodatna strokovna pomoč in prilagoditve: pri vzgoji in

izobraževanju otrok s posebnimi potrebami. Ljubljana: Centerkontura.

[7] Stop Motion Animation (2024).

https://www.techopedia.com/definition/109/stop-motion-animation

(Pridobljeno 18. 8. 2024)

[8] Zakon o spremembah in dopolnitvah Zakona o osnovni šoli (ZOsn-

H) (2011). Uradni list Republike Slovenije, št. 87/2011. Pridobljeno 18.

8. 2024, http://www.uradni-

list.si/1/objava.jsp?urlid=201187&stevilka=3727



Pomanjkanje kritičnega mišljenja je nevarnejše



od umetne intelligence



Lack of Critical Thinking is More Dangerous than



Artificial Intelligence



Vladislav Rajkovič

Univerza v Mariboru, Fakulteta za

organizacijske vede

Kranj, Slovenija

vladislav.rajkovic@gmail.com



POVZETEK ABSTRACT

Kritično mišljenje igra ključno vlogo pri interpretaciji, Critical thinking plays a key role in interpreting, evaluating

ocenjevanju in uporabi informacij. Medtem ko umetna and applying information. While artificial intelligence (AI)

inteligenca (UI) predstavlja tehnološke priložnosti in potencialne presents technological opportunities and potential dangers, such

nevarnosti, kot so zloraba podatkov, pristranskost algoritmov in as data misuse, algorithmic bias, and job automation, the real risk

avtomatizacija delovnih mest, se pravo tveganje skriva v lies in the inability of humans to properly judge and use this

nezmožnosti ljudi, da te informacije ustrezno presodijo in information.

uporabijo. Critical thinking enables people to separate facts from

Kritično mišljenje omogoča ljudem, da ločimo dejstva od misinformation, recognize ulterior motives, and analyse different

dezinformacij, prepoznavamo skrite motive in analiziramo perspectives. Without these skills, we are more susceptible to

različne perspektive. Brez teh veščin smo bolj dovzetni za manipulation, conspiracy theories and wrong decisions, which

manipulacije, teorije zarot in napačne odločitve, kar lahko vodi can lead to wider social and political consequences.

do širših družbenih in političnih posledic. AI is a tool that we can design and direct, but its use and

UI je orodje, ki ga lahko oblikujemo in usmerjamo, vendar impact depend heavily on the critical thinking of users. If users

njegova uporaba in vpliv sta močno odvisna od kritičnega lack these skills, the technology can exacerbate already existing

mišljenja uporabnikov. Če uporabniki nimajo teh veščin, lahko problems such as polarization, the spread of fake news, and the

tehnologija poveča že obstoječe težave, kot so polarizacija, strengthening of prejudice.

širjenje lažnih novic in krepitev predsodkov. Education and the promotion of critical thinking at all levels

Ključnega pomena je izobraževanje in spodbujanje kritičnega are of key importance, as it enables the responsible and

mišljenja na vseh nivoji, saj omogoča odgovorno in premišljeno thoughtful use of UI and reduces the risks associated with the

uporabo UI ter zmanjšuje tveganja, povezana z napačno uporabo incorrect use of technology. A lack of these skills can lead to

tehnologije. Pomanjkanje teh veščin lahko vodi v večje in bolj bigger and more far-reaching consequences than the technology

daljnosežne posledice kot sama tehnologija. itself.

V prispevku bodo izpostavljeni pomembni pedagoški vidiki The contribution will highlight important pedagogical aspects

kritičnega mišljenja, kot so: vloga učitelja, kurikuluma in of critical thinking, such as: the role of the teacher, the

interdisciplinarnega pristopa. Posebej bo poudarjen pomen curriculum and the interdisciplinary approach. Special emphasis

veščin za 21. stoletje, kot so: digitalna pismenost, sistemsko will be placed on the importance of 21st century skills such as:

mišljenje in razumevanje UI. Govor bo tudi o ustvarjanju digital literacy, systems thinking and UI understanding. The talk

podpornega okolja za kritično mišljenje. will also be about creating a supportive environment for critical

thinking.



Uporaba izobraževalnih aplikacij za prepoznavo travniških



rastlin



The use of educational applications to identify meadow plants



Suzana Rebec

I. osnovna šola Celje

Celje, Slovenija

suzana.rebec@guest.arnes.si



POVZETEK 1 UVOD

V prispevku je predstavljena učinkovita uporaba Digitalna tehnologija je že vrsto let močno vpeta v delo

izobraževalnih aplikacij za spoznavanje rastlin in primerjava le- učitelja, uporaba informacijsko-komunikacijske tehnologije

teh glede na njihovo uporabnost in učinkovitost. Aplikacije so se (IKT) je postala tudi nepogrešljiv del pouka. Pri mlajših učencih

izkazale kot odličen način za prepoznavo travniških rastlin in v 1. triletju osnovne šole je vnašanje digitalnih vsebin prisotno v

dober pripomoček pri obravnavi učne snovi. Kot najboljša manjši meri kot pri starejših učencih, vendar je kljub temu, da

aplikacija izmed preizkušenih se je izkazala aplikacija Pl@ntNet. šestletniki še niso vešči uporabe digitalne tehnologije in vsebin

Enostavna za uporabo, jasno in pregledno zastavljena, shranjuje le-te, mogoče pri pouku uporabljati IKT. Pri tem je potrebno

vnesene rastline in je zelo bogato opremljena s fotografijami premišljeno izbirati snov in vsebine, ki jih lahko kvalitetno

primerov, ki jih nalagajo uporabniki. Obenem ponuja tudi imena povežemo z uporabo tovrstne tehnologije.

rastlin v slovenskem jeziku z možnostjo lastnega vnašanja Pomembno je tudi premišljeno zastaviti celotno obravnavo

predlogov in popravkov. Ob vpisu se je mogoče povezati z določene učne snovi in predvideti, v katerem delu obravnavane

uporabniki s celega sveta. Zelo uporabna je tudi spletna stran snovi bodo določene digitalne vsebine najbolj pripomogle k

aplikacije, ki ponuja dodatno izobraževanje s spoznavanjem boljšemu razumevanju in najboljši možni zapomnitvi snovi.

rastlin po celem svetu. S pomočjo aplikacij je mogoče kvalitetno Prepletanje digitalnih vsebin s preostalimi načini

razvijati tehnike izkustvenega učenja. poučevanja je ob skrbnem načrtovanju lahko odličen način

podajanja snovi in doseganja razumevanja naučenega pri

KLJUČNE BESEDE učencih. Razvijanje računalniške pismenosti je z razpoložljivo

Izobraževalne aplikacije, travniške rastline, izkustveno tehnologijo možno v vseh fazah učnega procesa.

učenje Izobraževalne aplikacije so pri delu z učenci eden od

učinkovitih in zanimivih načinov uporabe digitalnih vsebin, s

ABSTRACT pomočjo katerih je lahko kvalitetno posredovano znanje ter

The article presents the effective use of educational dosežen željen nivo razumevanja.

applications for learning about plants and their comparison in Pri podjetju Lingvist, estonski jezikovni aplikaciji, trdijo, da

terms of their usefulness and effectiveness. The applications have je v določenih primerih lahko učenje preko matematično

proven to be an excellent way to identify meadow plants and a optimiziranih aplikacij bolj učinkovito kot klasično učenje v

good aid in dealing with the learning content. The Pl@ntNet učilnicah. [7]

application proved to be the best application among those tested. Aktivno učenje je način, s katerim dosežemo dobre rezultate

Easy to use, clearly and transparently laid out, stores entered pri znanju učencev. Učenci se ne naučijo veliko samo s tem, da

plants and is richly equipped with photos of examples uploaded sedijo v razredih in poslušajo učitelje, si zapomnijo vnaprej

by users. At the same time, it also offers the names of plants in pripravljene naloge in podajajo odgovore. O tem, kar se učijo, se

the Slovenian language with the possibility of entering your own morajo pogovarjati, o tem pisati, to povezovati s preteklimi

suggestions and corrections. Upon registration, it is possible to izkušnjami in uporabljati v vsakdanjem življenju. To, česar se

connect with users from all over the world. The website of the naučijo, morajo ponotranjiti. Aktivno učenje se lahko zgodi tudi

application is also very useful, offering additional education by izven učilnice. [1]

learning about plants around the world. With the help of Izkustveno učenje temelji na izkušnjah. To je oblika učenja,

applications, experiential learning techniques can be developed ki skuša povezati neposredno izkušnjo, opazovanje, spoznavanje

qualitatively. (kognicijo) in ravnanje v neločljivo celoto. Tovrstno učenje se je

razmahnilo v zadnjih letih, kot odgovor na pretežno na knjigah

KEY WORDS in učiteljevi besedi zasnovano učenje, ki razvija predvsem

Educational applications, meadow plants, experiential abstraktno simbolično znanje in kot izraz želje po tesnejši

learning povezavi teorije in prakse. S pomočjo tehnologije je tehnike

izkustvenega učenja smiselno izvajati tudi v e-izobraževanju.



Permission to make digital or hard copies of part or all of this work for personal

or classroom use is granted without fee provided that copies are not made or

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this

work must be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s). Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia S. Rebec



2 Vizija kreatorjev aplikacije iNaturalist je svet, v katerem APLIKACIJE ZA PREPOZNAVO RASTLIN

Prvošolci so si na učnem sprehodu na travniku v času lahko vsakdo razume in vzdržuje biotsko raznovrstnost s prakso

podaljšanega bivanja ogledali različne travniške rastline. Nekaj opazovanja divjih organizmov in izmenjavo informacij o njih.

različnih vrst rastlin so tudi nabrali. S pomočjo različnih aplikacij Njihovo poslanstvo je povezati ljudi z naravo ter pospešiti

smo prepoznavali njihova imena. Med vrsto ponujenih aplikacij znanost in ohranjanje biotske raznovrstnosti. na spletu so bile uporabljene tri aplikacije: Plantify, iNaturalist iNaturalist.org se je leta 2008 začel kot zadnji magistrski projekt,

in Pl@ntNet. šest let kasneje je postal pobuda Kalifornijske akademije

znanosti in leta 2017 skupna pobuda z National Geographic

Society. Lani je iNaturalist postal neodvisna neprofitna

organizacija. Na mednarodni ravni sodeluje z več različnimi

2.1 Plantify organizacijami prek mreže iNaturalist, da bi zagotovil

Osnovna ideja podjetja, ki je ustvarilo aplikacijo Plantify, je lokalizirano izkušnjo skupaj z večjim dosegom in vplivom. [3]

iskanje načinov za vzpostavljanje stika z naravo. Aplikacija se je izkazala kot uporabna pri prepoznavi

V svetu, kjer se vse več časa preživi v zaprtih prostorih, je travniških rastlin. Za bolj natančno poimenovanje rastlin je ravno

zlahka pozabiti na stik z naravo. Pozablja se, kako pomembna je tako potrebna registracija, ob prijavi pa lahko za rastline, ki jih

skrb za rastline, ki nas obdajajo. Kaj pa, če bi se lahko z njimi ne prepoznamo, le-te poimenujejo drugi uporabniki, ki določeno

pogovarjali? Kaj, če bi imeli način komunikacije z rastlinami in rastlino prepoznajo in s tem širijo mrežo znanja in izpopolnjujejo

vzpostavitev osebne povezave z njimi? To je točno tisto, kar virtualno knjižnico rastlin s celega sveta.

postaja resničnost z novo in inovativno aplikacijo "Plantify". [5]

Aplikacija je edinstven način povezovanja in komuniciranja

z rastlinami. Preko aplikacije se je mogoče pogovarjati s sobnimi,

vrtnimi, pa tudi divjimi rastlinami. Na vprašanje, kako se

počutijo, kaj potrebujejo, se odgovor prikaže v aplikaciji v samo

nekaj sekundah. Za izvedbo aplikacija uporablja edinstveno

tehnologijo za prevajanje jezika rastlin in jih naredi razumljive

ljudem. [5]

Delovanje aplikacije je bilo preizkušeno tudi v praksi. Pri

prepoznavi rastlin ni bilo vedno možno natančno določiti vrste

rastline. Ponuja samo latinska in angleška imena rastlin. Ne

ponuja možnosti popravka določanja vrste rastline v brezplačni

različici uporabe, kot je to mogoče pri drugih aplikacijah.

Uporabniki, ki želijo aplikacijo izkoristiti za informacije o vzgoji

določene sobne ali vrtne rastline, pa jo lahko koristno uporabijo.





Slika 2: Aplikacija iNaturalist. Lastni vir.





2.3 Pl@ntNet

Pl@ntNet je platforma za znanost, ki uporablja umetno

inteligenco (AI) za lažjo identifikacijo in popis rastlinskih vrst.

Je eden največjih svetovnih observatorijev biotske raznovrstnosti

z več milijoni sodelavcev v več kot 200 državah. [6]

Omogoča prepoznavanje več deset tisoč rastlinskih vrst.

Najbolj usposobljeni s prispevki in pregledovanjem opazovanj

prispevajo največ strokovnega znanja, povprečni poznavalci pa

to znanje pridobijo z uporabo aplikacije Pl@ntNet. Pl@ntNet je

zaradi aktivnih uporabnikov zbral že več kot milijardo slik

rastlin.

Aplikacija je učinkovito orodje za prepoznavanje rastlin s

pomočjo fotografij. Organizirana je po različnih tematskih in

Slika 1: Aplikacija Plantify. Lastni vir. geografskih florah. Na seznamu je mogoče izbrati tisto, ki ustreza

uporabnikovi regiji ali območju zanimanja. Če uporabnik ne ve,



2.2 kaj izbrati, izbere "Svetovna Flora", ki ima najširšo pokritost, iNaturalist

iNaturalist pomaga prepoznati rastline in živali, hkrati pa vendar bo dobil manj natančne rezultate kot z bolj natančno

ustvarja podatke za znanost in ohranjanje narave. S snemanjem določeno lokacijo. [6]



in deljenjem opažanj uporabniki ustvarjajo podatke raziskovalne Uporaba aplikacije je bila zelo pozitivna, saj je enostavna za

kakovosti za znanstvenike, ki si prizadevajo za boljše prepoznavanje rastlin in ponuja širok spekter rastlinskih vrst in

razumevanje in zaščito narave. [3] dodatno nadgradnjo znanja z vsebinami, povezanimi z

rastlinami. Uporabljena je bila na začetku obravnave učne snovi,

The use of educational applications to identify meadow plants Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



kot tudi v nadaljevanju. Učencem se je zdela zanimiva, bili so

dodatno motivirali za iskanje rastlin, njihovo spoznavanje in

uporabnost.





Slika 5: Izdelava herbarija. Lastni vir.



Igrali so se tudi igro »spomin«, na kateri so bile fotografije

rastlin in njihova imena. Pri nastajanju igre so sodelovali tudi

Slika 3: Aplikacija Pl@ntNet. Lastni vir. učenci. Z vsemi aktivnostmi in igrami, vključenimi v učni

proces, so učenci zelo dobro sledili vsebini in si dobro zapomnili



rastline, zanimala jih je tudi njihova uporabnost, zato so v



knjigah in na spletu ob pomoči in dodatnih napotkih poiskali tudi

te podatke.





Slika 6: Značilnosti rastlin. Lastni vir.

Slika 4: Aplikacija Pl@ntNet. Lastni vir.

S pomočjo naučenega je sledilo ustvarjanje – slikanje



3 travnika in izdelava slik iz posušenih travniških rastlin. AKTIVNOSTI

V nadaljevanju je sledilo spoznavanje rastlin na spletni strani Ustvarjene slike travnika so krasile učilnico do konca šolskega

Pl@ntNet. Na zemljevidu sveta so si učenci lahko ogledali, kje leta, uokvirjene slike s posvetilom pa so učenci podarili učiteljem

vse določena rastlinska vrsta uspeva. Izbrali so si tudi določeno ob koncu šolskega leta.

državo in z navdušenjem spoznavali, kakšne rastline tam rastejo. Poleg izvedenih dejavnosti je možnosti za aktivno učenje in

Zanimivo je bilo ugotavljati, ali kakšna travniška rastlina uspeva poustvarjanje zelo veliko. Izbira didaktičnih iger in načina

tudi zelo daleč stran od Slovenije. Ugotovili so na primer, da vključevanja aktivnega sodelovanja učencev je odvisna tudi od

navadni rman raste celo na nekaterih območjih Sibirije. starosti učencev, saj je mogoče s skupinami starejših učencev

Na spletu so si ogledali Digitalni herbarij, ki ga je ustvarila učenje še bolj prenesti na učečega. Ravno tako je mogoče že

slovenska ljubiteljica narave in fotografije in primerjali nabrane izbrane dejavnosti in igre nadgraditi in prilagoditi starostni

rastline ter si prebrali, kakšne lastnosti imajo. stopnji.

Nabrane rastline so učenci zložili med časopisni papir,

kasneje pa z nekaj pomoči ustvarili »Herbarij travniških rastlin«.

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia S. Rebec



[7] Razcvet računalniških aplikacij. Dostopno na naslovu



https://racunalniske-novice.com/razcvet-izobrazevalnih-aplikacij-po-

svetu/ [3. 7. 2024]



Slika 7: Poustvarjanje in izdelava slik. Lastni vir.



4 ZAKLJUČEK

Kljub starosti učencev, omejitvam zaradi začetnega

opismenjevanja ter nepoznavanju digitalne tehnologije so učenci

lahko sledili uporabi le-te, vsi deležniki uporabljene tehnologije

so pripomogli k boljšemu razumevanju naučenega ter boljšemu

pomnjenju snovi. Uporabljene izobraževalne aplikacije in druga

digitalna tehnologija se je izkazala kot učinkovito učno orodje,

ki odpira kritično razmišljanje, hkrati pa motivira in omogoča

boljšo zapomnitev.

Predstavljen način spoznavanja novih vsebin se je izkazal

kot zelo spodbuden in učinkovit, kar se je kazalo pri njihovi

motivaciji, napredku in navdušenju za delo.

Izkustveno učenje povezuje teoretični in praktični del ter

omogoča učinkovitejše dojemanje in spoznavanje snovi, učenci

pa so v tem procesu v aktivni vlogi. Obenem učenci razvijajo tudi

veščine komunikacije, sodelovanja, poslušanja. Ta način

poučevanja zajema izkušnje, opazovanje, spoznavanje pojmov in

uporabo znanja v novih situacijah. K temu zelo pripomore tudi

telesno gibanje, uporaba pantomime, »brainstorming«,

opazovanje, delo na terenu, različne didaktične naloge in igre.

Digitalno opismenjevanje je skozi igro mogoče v vseh fazah

učenja. Uporaba IKT je lahko dobra motivacija skozi celoten

učni proces, saj učenci radi posegajo po elektronskih napravah in

lahko to s pridom izkoristimo za pritegnitev pozornosti za učenje

novih vsebin pri pouku.



LITERATURA IN VIRI

[1] Chickering, A. W., Gamson, Z. F. (1987). Seven Principles for Good

Practise in Undergraduate Education. AAHE Bulletin, 40 (7). Dostopno

na: https://www.lonestar.edu/multimedia/SevenPrinciples.pdf

[3. 7. 2024]

[2] Digitalni herbarij 2. Dostopno na naslovu

https://maligaj.wordpress.com/2018/06/10/digitalni-herbarij-2-poletne-

cvetlice/ [3. 7. 2024]

[3] iNaturalist. Dostopno na naslovu https://www.inaturalist.org/pages/about

[3. 7. 2024]

[4] Lapuh B., J. (2009). Učinkovitost učenja iz spletnih učnih virov.

Doktorska dizertacija. Pedagoška fakulteta. Univerza v Ljubljani.

[5] Plantify. Dostopno na naslovu https://plantify-app.com/ [3. 7. 2024]

[6] Pl@ntNet. Dostopno na naslovu https://identify.plantnet.org/sl [2. 7.

2024]





Uporaba informacijsko komunikacijske tehnologije pri




poučevanju slepih in slabovidnih otrok, mladostnikov ter



izzivi, ki nastajajo pri tem



The Use of Information and Communication Technology in



the Teaching of Blind and Partially Sighted Children and



Adolescents and the Challenges that Arise in This



Marko Ribič

Center IRIS

Langusova ulica 8

1000 Ljubljana

marko.ribic@center-iris.si



POVZETEK 1 IKT V ŠOLAH

IKT oprema ima pri poučevanju slepih in slabovidnih otrok Uporaba IKT opreme pri poučevanju je v slovenskem šolstvu,

zelo velik pomen, saj se v današnjem času brez nje poučevanja predvsem kar se tiče učiteljev, kar dobro uveljavljena na vseh

teh otrok ne da izpeljati. Uporaba računalnikov, prenosnih ravneh izobraževanja. Veliko manj jo uporabljajo učenci v

računalnikov, brajevih zaslonov oziroma vrstic, raznih samem učnem procesu. Pri učencih, ki so slabovidni ali celo

elektronskih lup in drugih elektronskih lup je nujna za učinkovito slepi pa je IKT oprema ključnega pomena pri njihovem

pedagoško delo in za napredek otrok s primanjkljaji vida. Ker izobraževanju tako v vrtcu, osnovni šoli, srednji šoli in naprej na

imajo sodobni računalniki vse manj USB priključkov in ker je fakulteti. V današnjem času si namreč ne moremo predstavljati,

težava pri shranjevanju podatkov in zapiskov na te medije, smo da otroci s primanjkljaji vida ne bi uporabljali prilagojenega

se lotili reševanja tega izziva. V prispevku predstavljamo opremo, tabličnega računalnika, prenosnega računalnika s posebno

ki se uporablja pri poučevanju slepih in slabovidnih otrok in programsko opremo, raznih elektronskih lup za povečavo in

rešitev za shranjevanje podatkov in zapiskov, ki jo bodo izbiro kontrasta in pa seveda elektronskih brajevih vrstic.

uporabljali učenci pri pouku.

2 TABLIČNI RAČUNALNIKI S TAKTILNO

KLJUČNE BESEDE MREŽO

IKT oprema, slepi in slabovidni ortoci, brajev zaslon, Slepi otroci, pa tudi otroci, ki so slabovidni in imajo diagnozo

elektronska lupa, intune, oblačne storitve poslabšanja vida že v vrtcu, uporabljajo tablične računalnike, ki

so nadgrajeni s posebno taktilno mrežo, ki uporabniku omogoči,

ABSTRACT da začuti kar je na zaslonu. To je predvsem uporabno v vrtcu,

ICT equipment plays a very important role in teaching blind kjer imajo slepega otroka, saj lahko na ta način sam igra različne

and partially sighted children, because nowadays it is impossible igre (npr. spomin), ki jih drugače ne bi mogel igrati oziroma bi

to teach these children without it. The use of computers, laptops, jih lahko igral samo v primeru, da bi imeli za to prilagojene

braille displays, various electronic magnifiers and other igrače. Na tabličnih računalnikih s taktilno mrežo je tudi veliko

electronic magnifiers is necessary for effective pedagogical work aplikacij, ki so prilagojene za slepe otroke. Otroci s tem vadijo

and for the progress of children with visual impairments. Since svoje zaznavne sposobnosti, ki jih bodo v življenju še kako rabili.

modern computers have less and less USB ports and because it Tablica s taktilno mrežo je prav tako v veliko pomoč pri slepih

is a problem to save data and notes on these media, we set out to otrocih, ki obiskujejo prvi razred osnovne šole, saj se na njih

solve this challenge. In this article, we present the equipment lahko učijo brajeve pisave.

used in teaching blind and partially sighted children and the

solution for storing data and notes, which will be used by 3 PROGRAMI ZA PREVAJANJE BESEDILA

students during lessons. V BRAILLOVO PISAVO

Pri slepih in slabovidnih otrocih z diagnozo poslabšanja vida

KEYWORDS že v prvi triadi osnovne šole začnejo z uporabo računalnika

ICT equipment, blind and visually impaired individuals, (stacionarnega ali prenosnega), ki ima prilagojeno programsko

Braille display, electronic magnifier, Intune, cloud services opremo za vsakega uporabnika posebej. Pri slepih je v prvi vrsti

računalnik z operacijskim sistemom Windows v slovenskem

jeziku, programska oprema JAWS (plačljiva programska

oprema), ki prevaja besedilo na zaslonu in tudi opise slik v 4 TEŽAVE S SHRANJEVANJEM

brajevo pisavo. Potrebno je prilagoditi različne nastavitve, da PODATKOV lahko učenci poslušajo kaj je na zaslonu. V ta namen se lahko

Pri uporabi računalnika s prilagojeno opremo v šoli pa se

uporablja eBralec, ki je prilagojen za slovenski jezik, ali v

srečujemo z velikimi težavami pri shranjevanju podatkov

zadnjem času tudi kar Windows govorec.

oziroma vsebin, zapiskov, ki jih slepi in slabovidni učenci

vnašajo v računalnik.



Na Centru IRIS, kjer izvajamo prilagojene izobraževalne

programe za slepe in slabovidne od vrtca do srednje šole, smo do

pred kratkim uporabljali USB ključke, na katere so si učenci

shranjevali datoteke, mape itd., ki so jih potrebovali ali ustvarili

pri pouku. Ker so se ključki izgubljali, kvarili, poleg tega pa je

prihajalo tudi do poškodovanja USB vtičev na računalniku, smo

pred dobrim letom in pol začeli razmišljati kako bi zadevo rešili.

V ta namen smo za učence pridobili in jim dodelili Microsoft

Office 365 licence A3, njihove računalnike povezali z njihovim

Office365 računom ali enostavno dodali napravo v Intune.

Otroke smo naučili kako se prijaviti v računalnik z Intunom, kako

se prijaviti v spletno različico Office365 in kako le tega

uporabljati. Na računalniku smo nastavili avtomatsko

Slika1: Tablični računalnik s taktilno mrežo Feelife pro shranjevanje. Tako je odpadlo dodatno shranjevanje na zunanje

medije, saj se vse, kar je narejeno s pomočjo programov, ki se



nahajajo v okviru Microsoft Office avtomatsko shranjuje v oblak.

Pri slepih in tudi slabovidnih učencih pride v poštev predvsem

Microsoft Word, saj si med poukom pišejo na računalnik zapiske

iz katerih se nato s pomočjo brajeve vrstice (slepi učenci) tudi

učijo. Seveda je predpogoj uporabe tega, da se učenci naučijo

uporabljati računalnik, da se znajdejo v okolju Windows, da

poznajo ukaze za premikanje med mapami oziroma poznajo delo

na računalniku brez miške in nenazadnje da se naučijo slepega

tipkanja. Pri učencih, ki so slepi in nimajo pridruženih motenj,

slepo tipkanje ne predstavlja neke velike težave, saj imajo razvite

druge čute. Več težav je s slabovidnimi, ki se težje navadijo na

tipkanje, predvsem v takih primerih, kjer so še kot otroci bolje

videli, sedaj pa se jim vid slabša.

Slika 2: Program JAWS



V slovenskih šolah se za predvajanje besedila na elektronsko

brajevo vrstico uporablja program JAWS, ki je zelo drag, vendar

omogoča prikaz vseh znakov in funkcij, ki se uporabljajo pri

poučevanju. Obstaja tudi odprtokodna različica te rešitve,

imenovana NVDA, ki je prevedena tudi v slovenščino. Ima pa ta

program določene pomanjkljivosti. Prav zaradi tega so se na

Zvezi slepih in slabovidnih Slovenije odločili, da se v šolstvu raje

uporablja plačljiv program.

Slika 4: zaslonska slika Intune





Center IRIS je kot center za slepe in slabovidne otroke



pristojen tudi za posojanje opreme za slepe in slabovidne otroke,

ki se šolajo v šolah v domačem kraju. Oprema se posodi šoli in

ne otroku, tako da imajo to opremo učenci v šolah in jo praviloma

ne nosijo domov. V teh primerih mi kot center ne moremo

dodeliti licenc Office 365 otrokom, saj je za to pristojna matična

šola. V večini primerov se v teh šolah še vedno za shranjevanje

uporabljajo USB ključki, ki se prenašajo iz šole domov in

obratno. Ko se pripravi oprema za šolo, se namesti na prenosni

računalnik ali program JAWS za slepe ali ZoomText za



Slika3: Brajev zaslon oz vrstica slabovidne učence. Uporabnikom se pusti administrativne pravice, da lahko njihovi računalnikarji naložijo določene

programe, če bi se jih potrebovalo za delo v šoli.

Praksa je pokazala, da računalnikarji na matičnih šolah ne

naredijo na teh računalnikih kaj več kot naložijo kak program,

zato jih poskušamo preko naših mobilnih učiteljev vzpodbujati,

da se lotijo tudi oblačnega shranjevanja podatkov. Dogaja se

namreč, da se veliko opreme polomi (predvsem USB priključki),

ki so vezani na matično ploščo, njihova zamenjava pa je draga

oziroma v določenih primerih, ko je polomljeno več USB

priključkov, tudi nesmiselna. USB priključki se namreč pri delu

uporabljajo predvsem za povezavo prenosnega računalnika z

brajevo vrstico in po navadi še za brezžično miško, ki je v pomoč

spremljevalcu slepega učenca.



5 OPREMA ZA SLABOVIDNE

Pri slabovidnih otrocih pa se poleg računalnika z ustrezno

programsko opremo uporablja tudi elektronska lupa, ki učencu

omogoča, da prebere vsebino iz zvezka sošolcev, knjige ali table.

Ker je slabovidnosti več vrst je pomembno, da taki učenci

čimprej dobijo elektronsko lupo. Na elektronski lupi se nato

nastavi kontrast, barva in druge specifike, ki učencem omogočajo,

da lahko sledijo pouku.





Slika 5: Elektronska lupa



6 ZAKLJUČEK

IKT oprema na vseh ravneh pomaga ljudem s

pomanjkljivostjo vida. Izkazalo se je, da so pri uporabi IKT

opreme zelo dobrodošle oblačne storitve, saj slepim in

slabovidnim učencem prihranijo marsikatero nevšečnost. Še

vedno pa je zelo pomembno sočasno z učenjem o različnih

predmetih učiti tudi, da je potrebno pomagati sočloveku, še

posebej ranljivim skupinam, kar slepi in slabovidni so. Vedno se

pojavijo kakšne težave, ki jih slepi učenci ne morejo rešiti sami.

Takrat je res zelo pomembno, da so na voljo ali spremljevalci ali

sošolci, ki pomagajo učencem pri reševanju težav.



Literatura

[1] https://www.feelif.com/si/feelif-pro/

[2] https://www.nvaccess.org/files/nvda/documentation/userGuide.html

[3] https://www.freedomscientific.com/products/software/jaws/

[4] https://www.helptech.eu/braille-displays-active-braille

[5] https://www.diopta.si/

[6] https://intuneeducation.portal.azure.com/



Izdelava videoposnetkov za lažje utrjevanje slovničnih



struktur v tujem jeziku



The Use of Self-Recorded Videos to Facilitate Grammar



Learning in an EAL classroom



Eva Rolih

Srednja vzgojiteljska šola, gimnazija

in umetniška gimnazija Ljubljana

Ljubljana, Slovenija

eva.rolih@svsgugl.si



POVZETEK

Cilj sodobnega izobraževanja ni le prenos informacij, temveč 1 UVOD



tudi vzbujanje zanimanja in angažiranosti učencev in učenk, kar Učitelji in učiteljice tujih jezikov se nemalokrat srečujemo z

je pogosto težje dosegljivo pri učenju in utrjevanju slovničnih dilemo, kako oplemenititi pogosto monotono učenje in utrjevanje

struktur v tujem jeziku. Uporaba inovativnih metod kot so slovničnih struktur in ga nadgraditi na način, da je zanimivo, a

snemanje videoposnetkov lahko ta proces naredi bolj dinamičen vseeno učinkovito. Rešitev je iskanje metod, ki spodbujajo

brez da bi ogrozil njegovo kvaliteto. Članek predstavlja, kako aktivno sodelovanje in praktično uporabo jezika. Ena izmed njih

lahko projekt snemanja videoposnetkov med poukom pripomore je zagotovo snemanje videoposnetkov, ki jih ustvarjajo učenci in

k celostnem razvoju dijakov in dijakinj. Obravnava prednosti učenke samostojno. Predstavljeni projekt snemanja potopisov

snemanja videoposnetkov kot metodo učenja pri pouku tujega osmišlja učenje slovnične strukture trpnika in združuje več

jezika in na praktičnem primeru iz razreda prikazuje, kako lahko področij povezanih s pisnim in ustnim izražanjem v tujem jeziku

pripomore k nadgradnji jezikovnih spretnosti, večji motivaciji in ter rabo tehnologije za snemanje in montiranje posnetkov.

angažiranosti ter razvoju digitalnih spretnosti. Prav tako na



kratko predstavi in evalvira različna orodja za samostojno 2 RABA VIDEOPOSNETKOV KOT snemanje in montažo.

METODA UČENJA PRI POUKU TUJEGA



KLJUČNE BESEDE JEZIKA

Angleščina kot tuj jezik, učenje slovnice, videoposnetki, Raba videoposnetkov pri pouku ni nič novega. Učitelji in

projektno delo učiteljice pogosto v pouk vključijo informativni video z

namenom poglobitve znanja in popestritve pouka. Tovrstna raba



ABSTRACT videoposnetkov postavlja dijake in dijakinje v pasiven položaj,

saj se od njih večinoma pričakuje pasivno sprejemanje informacij

The aim of modern education is not only to transmit ali urjenje slušnih spretnosti. Zaradi tega učenje na tak način ni

information, but also to arouse the interest and improve the vedno tako efektivno kot bi si želeli in posledično spravlja rabo

engagement of students. This can be difficult to achieve when videoposnetkov med poukom na slab glas.

acquiring or consolidating grammatical structures in a foreign Drugače je z videoposnetki, ki jih ustvarijo dijaki in dijakinje.

language. Using innovative methods such as recording videos Pri pouku tujega jezika si prizadevamo, da so aktivno udeleženi

can make this process more dynamic without compromising its v učni proces, kar od njih zahteva aktivno produkcijo jezika. S

quality. This article represents how self-recorded videos can projekti, kot je snemanje lastnih videoposnetkov, jim

contribute to the overall students’ development. Furthermore, it omogočimo prav to in raziskave kažejo, da je učenje bolj

discusses the benefits of video recording as a learning method in učinkovito, če so neposredno vključeni v proces ustvarjanja

an EAL classroom. It uses a practical example to portray how it videoposnetka. [1]

can help to improve language skills, increase motivation, and Lepota takšnih projektov je njihova vsestranskost in

participation. Moreover, it briefly introduces and evaluates zmožnost, da jih lahko navežemo na praktično katerokoli

different tools for self-recording and editing. tematsko enoto in slovnično strukturo. Dijakom in dijakinjam

omogoča, da urijo svoje jezikovne spretnosti na kreativen in



KEYWORDS kontekstualiziran način, zaradi česar so dokazano bolj motivirani

za učenje. [2] Pri pisanju scenarija urijo svoje pisne spretnosti in

EAL, grammar learning, self-recorded videos, project-based rabo slovničnih struktur, medtem ko pri snemanju vadijo

work govorno izražanje, s poudarkom na intonaciji in naglaševanju. S



tem ko večkrat poslušajo lasten govor ali govor sošolca oz.

sošolke, lahko kritično ovrednotijo svoje govorne spretnosti [3],

kar prispeva k boljši ozaveščenosti tujega jezika in njegovi rabi.

Na ta način nadgrajujejo tako pisno kot tudi ustno izražanje v

ciljnem jeziku.

Kljub temu, da je primarni cilj te dejavnosti nadgradnja

jezikovnega izražanja, ponuja tak projekt še veliko več, saj

obenem urijo svoje veščine uporabe tehnologije za snemanje,

obdelavo in montažo videoposnetkov in se spoznavajo s principi

dobrega snemanja (kadri, svetloba, zvok ipd.) ter reševanje

morebitnih tehničnih težav v vseh fazah projektnega dela. Prav

tako na različne načine razvijajo lastno ustvarjalnost na več

področjih. [4] Pri pisanju scenarija krepijo zmožnosti

kreativnega pisanja, pri čemer morajo upoštevati, da mora biti Slika 1: Del scenarija o Senegalu (VIR: lasten, zajem

sporočilo jasno, informativno, zanimivo in predstavljeno v zaslonske slike)

omejenem času. Enako morajo ubrati kreativen pristop pri

vizualni zasnovi kadrov, tempa in prehodov tako med



snemanjem kot tudi montažo.

Obenem razvijajo spretnosti sodelovanja in dela v skupini.

[4] Predvsem sodelujejo pri pisanju scenarija in montaži, kar od

njih zahteva učinkovito komunikacijo, delitev dela in

usklajevanje različnih mnenj. Poleg tega morajo reševati

potencialne konflikte in sprejemati skupne odločitve glede

končnega izdelka.

Vse to dokazuje, da gre za večplasten in kompleksen projekt,

ki združuje različne veščine tako na jezikovnem kot digitalnem

področju.



3 IZVEDBA PROJEKTA SNEMANJE

POTOPISA Slika 2: Del scenarija o izmišljeni deželi Centopia (VIR:

lasten, zajem zaslonske slike)



3.1 Določitev teme in navodila

Projekt snemanja potopisov smo izvedli z dijaki in 3.3 Pregled in popravljanje scenarija



dijakinjami 2. letnika programa umetniška gimnazija ob Že med samim pisanjem so dijaki in dijakinje med sabo

zaključku 8. učne enote, ki se je osredotočala na temo in sodelovali ter popravili napake in na podlagi medsebojne

besedišče potovanj ter slovnično strukturo trpnika. Projekt je povratne informacije nadgradili vsebino. Prav tako je vsaka

kulminacija njihovega znanja, kreativnosti in jezikovnih skupina dobila še mojo pisno povratno informacijo na podlagi

spretnosti, saj so morali uporabiti novo besedišče in naučeno katere so svoje scenarije izpopolnili. Sama sem se predvsem

slovnico v praktičnem in ustvarjalnem kontekstu. osredotočala na primerno rabo slovničnih struktur in

Skupine so izbirale med obstoječimi destinacijami (npr. edinstvenost napisanega. V primeru plagiatorstva so bili sestavki

Senegal), dovoljeno pa jim je bilo ustvariti tudi potopis o zavrnjeni.

izmišljeni destinaciji. Tako so nekatere skupine izbrale

fantazijske destinacije, kot so npr. Atlantida, dežela Smrkcev ali 3.4 Snemanje

pa ustvarile povsem svojo. Skupine so zavoljo večjega miru, svobode in časa potopise

Projekt je potekal v več fazah: pisanje scenarija, pregled in snemale po šoli v prostem času. Rok za oddajo končnega izdelka

popravljanje scenarija, snemanje, montaža in predstavitev. je bil dva tedna, kar jim je omogočilo dovolj časa za snemanje in

montažo videoposnetkov.

3.2 Pisanje scenarija Glede na to, da so v projekt bili udeleženi dijaki in dijakinje

V skladu z navodili so morali pripraviti scenarij, ki je za umetniške gimnazije, smer gledališče in film, večjih težav s

vsakega posameznika oz. posameznico vključeval vsaj minuto snemanjem ni bilo, saj so dovolj znanja pridobili pri pouku

govora. Prav tako so morali napisati uvod in zaključek. Besedilo praktičnih predmetov. Snemanja so se lotili na različne načine,

je moralo vsebovati trpnik v različnih časih in čim več različnih in sicer: snemanje s fotoaparati, ki so si jih lahko izposodili v

ostalih slovničnih struktur. Vsaka oseba je morala predstaviti eno šoli, s pametnimi telefoni, s kamerami na prenosnih računalnikih

od zanimivosti izbrane destinacije. Ker je šlo za potopis, so pisali preko ZOOMa.



izpopolnili tudi doma. Pisanju scenarija smo namenili eno šolsko uro, lahko pa so ga Pametni telefon je vsekakor najenostavnejši za snemanje videoposnetkov, saj omogoča enostavno snemanje na različnih v prvi osebi. 3.4.1 Snemanje s pametnimi telefoni

lokacijah. Dijaki in dijakinje so zaradi osebne rabe izkušeni v

njegovi uporabi, kar omogoča hitro in učinkovito snemanje brez 3.5.2 Montaža s TikTok-om

dodatnega usposabljanja. Prav tako trenutni pametni telefoni V nasprotju s privzetim urejevalnikom TikTok ponuja velik

zagotavljajo zadovoljivo kakovost slike. Vsekakor pa obstajajo spekter raznih vizualnih učinkov, ki lahko posnetek naredijo

različne pomanjkljivosti, kot so pomanjkanje stabilnosti pri privlačnejši. Njegova enostavna in intuitivna uporaba s

snemanju, slabša kakovost zvoka in pomanjkanje naprednih številnimi predlogami omogoča hitro ustvarjanje dinamičnih

snemalnih nastavitev (osvetlitev, ostrina ipd.) videov brez posebnega znanja montaže. Uporabniki in

uporabnice lahko dodajo tudi glasbeno podlago ali zvočne



preko ZOOMa učinke. TikTok je predvsem primeren za kratke in nezahtevne 3.4.2 Snemanje s kamerami na prenosnih računalnikih

Zaradi koronavirusne pandemije je ZOOM dijakom in videe z omejeno kakovostjo slike.

dijakinjam dobro poznan, zato je dobra izbira za snemanje

videoposnetkov. Njegova velika prednost je možnost

sodelovanja več oseb na daljavo in razne vgrajene funkcije. Pri

snemanju potopisov se je še za posebej uporabno funkcijo

izkazala možnost dodajanja virtualnih ozadij, kar je dijakom in

dijakinjam omogočalo, da z izbiro primernega ozadja ustvarijo

iluzijo, kot da so dejansko prisotni na izbrani destinaciji. Ker pa

je ZOOM vezan na internetno povezavo, se lahko zgodi, da je

kakovost slike in zvoka slabša, saj je odvisna od pasovne širine

in kompresije pri prenosu podatkov. Tu veliko vlogo igra tudi



kvaliteta kamere na prenosnem računalniku. Slika 4: Uporabniški vmesnik TikToka za obdelavo videov

3.4.3 Snemanje s fotoaparatom ali kamero (VIR: https://p16-va-tiktok.ibyteimg.com/obj/musically-

Snemanje s fotoaparatom ali kamero vsekakor zagotavlja maliva-obj/226f6a7aa05db59c0030aa434d2623ec.png)

visoko kakovost slike in zvoka ter ponuja večji spekter naprednih

funkcij, a zaradi svoje cene niso vsakomur dostopni. Prav tako 3.5.3 Montaža z naprednimi orodji

raba fotoaparata terja več znanja in izkušenj ter morebitno Profesionalna orodja kot so Adobe Premiere Pro ali DaVinci

dodatno opremo (stojala, mikrofoni, luči). Resolve ponujajo široko paleto funkcij za obdelavo videa na

visoki ravni in dopuščajo urejanje videoposnetkov visoke

3.5 Montaža in produkcija videa kakovosti. Prav tako podpirajo različne formate in dolžine

Po snemanju so skupine prešle na fazo montaže. Proces je videoposnetkov. Zaradi svoje kompleksnosti zahteva več znanja,

vključeval izbiro najboljših posnetkov, primernega zaporedja, še posebej za uporabo naprednih funkcij. V nasprotju s prej

urejanje zvoka in morebitno dodajanje ostalih grafik. Dijaki in omenjenima možnostma, napredni programi običajno niso v

dijakinje so uporabile različne programe za montažo, od najbolj celoti brezplačni in zahtevajo zmogljivo računalnik in veliko

enostavnih (privzeti urejevalnik na telefonu ali prenosnem prostora za shranjevanje podatkov. Delo s takšnimi programi je

računalniku, TikTok ipd.) do bolj zahtevnih (Adobe). primerno za tiste z več znanja in interesa za tovrstno področje.



3.5.1 Montaža s privzetim urejevalnikom na telefonu

Tudi v tej fazi se je pametni telefon dobro izkazal, saj že

njihovi privzeti urejevalniki dopuščajo brezplačno, hitro in

preprosto montažo brez potrebe internetne povezave ali

kakršnekoli druge opreme. A pogosto ponujajo zgolj osnovne

funkcije z omejeno kakovostjo, kar lahko vpliva na kvaliteto

končnega izdelka.





Slika 5: DaVinci Resolve (VIR:



https://www.newsshooter.com/wp-



content/uploads/2023/04/DaVinci-Resolve-18-Edit-2-



740x410.jpg)





3.6 Predstavitev potopisov

Ko so bili potopisi pripravljeni, smo eno šolsko uro namenili

predstavitvi in ogledu njihovih zaključnih izdelkov. Potopisi so

bili odlično sprejeti, dijaki in dijakinje so bili navdušeni in

Slika 3: Privzeti urejevalnik na pametnem telefonu

zadovoljni, da so se lahko učili na tak način.

Samsung (VIR: lasten)





VIRI IN LITERATURA


[1] Rakhmanina, L., & Kusumaningrum, D. (2017). The Effectiveness of Video

Blogging in Teaching Speaking Viewed From Students’ Learning

Motivation. Proceedings of ISELT FBS Universitas Negeri Padang, 5, 27–

34

[2] Nurul Afifiah, B. A., Sakinah, B. Z., & Yunus, M. M. (2019). Integrative

Task-Based Learning: Developing Speaking Skill and Increase Motivation

via Instagram. International Journal of Academic Research in Business and

Social Sciences, 9(1), 620–636. https://doi.org/10.6007/ijarbss/v9-i1/5463.

[3] Yousef, F.A.M. et al. (2014). The State of Video-Based Learning; A Review

and Future Perspectives. International Journal on Advances in Life

Sciences.6 (3&4), 122-135.



Slika 6: Primer potopisa o Finski (VIR: lasten, zajem https://www.iariajournals.org/life_sciences/lifsci_v6_n34_2014_paged.pd [4] Lam, J. Z., & Md Yunus, M. (2023). Student-produced video for learning: A zaslona) systematic review. Journal of Language Teaching and Research, 14(2), 386-

395. https://doi.org/10.17507/jltr.1402.14





Slika 7: Primer potopisa o Senegalu (VIR: lasten, zajem

zaslona)





Slika 8: Primer potopisa o Senegalu (VIR: lasten, zajem

zaslona)



4 ZAKLJUČEK

Vključevanje videoposnetkov, ki jih ustvarijo dijaki in

dijakinje, v učni proces pouka tujega jezika ima številne

prednosti. Ne samo, da so dijaki in dijakinje aktivno vključeni v

nadgrajevanje svojega znanja in posledično izboljšajo svoje

pisne, slušne in jezikovne spretnosti v angleščini, temveč tudi

razvijajo svoje digitalne spretnosti in krepijo lastno ustvarjalnost

ter zmožnosti timskega dela. Tak pristop k učenju jim dopušča

veliko svobode, zaradi česar so bolj motivirani in angažirani.

Obenem jim tudi omogoča, da urijo ciljne slovnične strukture v

angleščini na zanimiv in kontekstno bogat način. Vsekakor je

velika prednost, če dijaki in dijakinje že znajo uporabljati opremo

za snemanje in montiranje videoposnetkov, vendar lahko tudi

tisti brez predhodnih veščin s pomočjo natančnih navodil

usvojijo nove veščine. Uporabljajo lahko različna orodja in

programe, ki so prilagojeni njihovi stopnji znanja in izkušenj na

tem področju. Kljub temu, da gre za večplastni projekt, ki

vključuje različne veščine, je potrebno zagotoviti, da je primarni

cilj (tj. urjenje določenih jezikovnih spretnosti) dosežen in se ne

izgubi v številnih drugih aspektih projekta.



Izpopolnjevanje jezikovnih in digitalnih veščin s



projektnim delom: Ustvarjanje razrednega tabloida s



platformo Canva



Honing Language and Digital Skills through Project-based



Learning: Creating a Class Tabloid with Canva



Eva Rolih

Srednja vzgojiteljska šola, gimnazija

in umetniška gimnazija Ljubljana

Ljubljana, Slovenija

eva.rolih@svsgugl.si



osnovne šole dalje. [1] Zato je še toliko bolj pomembno, da jim

POVZETEK omogočimo ustrezne aktivnosti, ki spodbujajo razvoj njihovih

V čedalje bolj razvijajočem svetu je pomembno, da v pouk jezikovnih spretnosti v ustreznem avtentičnem kontekstu. Eno

premišljeno vključimo inovativne pristope in uporabna orodja, s izmed orodij je vsekakor projektno delo, ki zaobjema širok

katerimi dijaki in dijakinje razvijajo ključne veščine za spekter področij in kompetenc ter dopušča, da dijaki in dijakinje

prihodnost. Sem spada tudi spretna uporaba sodobnih tehnologij, uporabijo svoje praktično znanje v realnih situacijah. S

ki posameznikom in posameznicam pomagajo pri vsakodnevnih projektom izdelave razrednega tabloida preko platforme Canva,

izzivih. V prispevku bom predstavila ključne prednosti učenja s sem poskusila povezati dve področji, ki sta izredno pomembni v

projektnim delom, spletno platformo Canva in njene odlike ter današnjem svetu: pisno izražanje v tujem jeziku in digitalne

praktičen primer, kako lahko združimo učenje tujega jezika in spretnosti.

digitalnih spretnosti z implementacijo prej omenjene spletne



platforme za grafično oblikovanje. 2 UČENJE S PROJEKTNIM DELOM



KLJUČNE BESEDE Učenje s projektnim delom (ang. Project-based learning) je

Projektno delo, angleščina kot tuj jezik, pisne spretnosti, multidimenzonalen pristop k usvajanju znanja tujega jezika, ki

Canva, grafično oblikovanje vključuje aktivno raziskovanje in reševanje kompleksnih izzivov,

obenem pa udeležencem in udeleženkam omogoča, da poglobijo



ABSTRACT svoje jezikovne veščine in jih uporabijo v praktičnem kontekstu.

Poleg tega je idealna priložnost za povezovanje različnih

In an increasingly evolving world, it is imperative to

spretnosti in področij, v našem primeru področja angleščine in

thoughtfully integrate innovative approaches and useful tools

grafičnega oblikovanja.

into the classroom to assist students develop key skills for the

Projektno delo je pogosto premalo uporabljeno, saj je

future. This encompasses the skilful use of modern technologies

pogosto po krivici razumljeno kot zgolj priložnost za razvedrilo

to help individuals meet everyday challenges. This paper

ali premor od vsakodnevne razredne rutine. Posledično se učenci

presents the key benefits of project-based learning, the Canva

in učenke predvsem ukvarjajo z vizualno privlačnostjo svojih

online platform and its strengths, along with a practical example

projektov in pri tem zanemarijo razvoj svojih jezikovnih

of how to combine learning a foreign language with acquiring

spretnosti. A, če je projekt načrtovan premišljeno, ima neverjeten

digital skills through the use of the aforementioned online

potencial za razvoj jezikovnih veščin, vsebinskega znanja,

graphic design platform.

metakognitivne zavesti, kritičnega mišljenja in socialnih



KEYWORDS učenci in učenke prevzamejo odgovornost za svoje učenje in spretnosti. [2] Poleg tega narava tovrstnega pristopa zahteva, da

Project-based learning, EAL, writing skills, Canva, graphic sodelujejo z drugimi, s čimer izboljšajo svoje spretnosti

design raziskovanja in reševanja problemov. [3]

Ker projektno delo pogosto zaobjema veliko različnih



1 kompetenc, se ga je smiselno poslužiti ob zaključku obravnave UVOD

posamezne učne enote ali več njih. Pri pouku tujega jezika to

Pisno izražanje v tujem jeziku predstavlja svojevrsten izziv za dijakom in dijakinjam omogoča, da v kontekstu uporabijo čim

učence in učenke, ne glede na to, da se učijo angleščine že od več elementov tujega jezika (npr. tematsko besedišče,

obravnavane slovnične strukture ipd.). Na ta način lahko in se odločile za slednjo. Z obema tipoma so se že seznanile v

preverimo njihovo celostno obvladovanje tujega jezika. predhodnih urah, zato so znale razlikovati med vsebino,

naslovnikom in načinom pisanja, ki pritiče tabloidnim

publikacijam. Vsaka skupina je podala predlog naslova časopisa,

3 SPLETNA PLATFORMA CANVA ki je nato bil izbran z glasovanjem. Prav tako je vsaka

Canva je spletna platforma za oblikovanje, ki je začela posameznica sama izbrala vsebino njihovega lastnega prispevka.

delovati leta 2013 v Avstraliji in cilj katere je opolnomočiti Kljub temu, da je vsaka morala pripraviti lasten prispevek, so

vsakogar za samostojno oblikovanje vizualnih vsebin, med delale v skupinah, saj jim je to v skladu z načeli projektnega dela

drugim tiskovin, objav za socialna omrežja, predstavitve, plakate, omogočilo skupinsko refleksijo napisanega, izmenjavo mnenj,

videov ipd. [4] Zaradi svojega uporabniku prijaznega vmesnika medsebojno pregledovanje (ang. peer review) ter posledično

in številnih predlog je dobra izbira za vse, ki se z grafičnim izboljšanje končnega izdelka.

oblikovanjem srečujejo prvič. Platforma ponuja številna osnovna Sledil je kratek praktičen uvod v spletno platformo Canva.

in napredna orodja za urejanje, poleg tega pa nudi številne Ker se je večina dijakinj s platformo srečala prvič, sem jim

videoposnetke in informativne članke, ki uporabnikom in ponudila možnost uporabe že obstoječih predlog, do katere so

uporabnicam pomagajo izboljšati svoje spretnosti. Velika dostopale preko iskalnika, lahko pa so ustvarile lastno. Tako so

prednost platforme je možnost sodelovanja v realnem času, kar se seznanile, kako lahko prilagodijo že obstoječe predloge s



zaradi česar je dobra izbira za timsko delo. Na voljo sta so se spoznale tudi z drugimi funkcionalnostmi platforme kot so brezplačna in plačljiva različica. Slednja omogoča še več iskanje in nalaganje fotografij, oblikovanje besedila in raba pomeni, da lahko več ljudi hkrati ureja posamezen dokument, spremembo pisav, barvne sheme in ostalih elementov. Poleg tega

naprednih funkcij in večjo knjižnico grafičnih elementov, vendar

dodatnih elementov (oblike, okvirji, grafike, nalepke ipd.).

je brezplačna različica v večji meri povsem zadostna. Ob prijavi

lahko uporabnik izbira med več jeziki, vendar so dijakinje zaradi

narave predmeta za jezik vmesnika izbrale angleščino.





Slika 2: Primer že pripravljenih predlog (VIR: lasten,



zajem zaslonske slike)



Prednost Canve je zagotovo njena intuitivnost in ‘povleci in

izpusti’ način dela, zato so dijakinje usvojile osnovne funkcije

ter jih uspešno implementirale v svoje projekte. Bolj spretne

Slika 1: Uvodni zaslon v Canvi (VIR: lasten, zajem dijakinje so se preizkusile tudi v enostavni obdelavi fotografij, ki

zaslonske slike) so jih priredile, da so ustrezale vsebini njihovih člankov.

Posamezne strani so bile nato združene v celoto z ujemajočo



4 naslovnico. IZVEDBA PROJEKTA USTVARJANJA

RAZREDNEGA TABLOIDA Na ta način niso utrjevale zgolj lastnih pisnih veščin, temveč

so tudi izboljšale svoje digitalne spretnosti, saj so se naučile

Projekt izdelave razrednega tabloida smo izvedle z uporabljati različna orodja za oblikovanje in obdelavo digitalnih

dijakinjami 2. letnika programa predšolske vzgoje ob zaključku vsebin. Hkrati so pridobile znanje na področju iskanja virov,

10. učne enote, cilja katere sta med drugim bila usvojiti besedišče organizacije in načrtovanja projektov v Canvi ter na področju

povezano s področjem medijev in raba trpnih glagolskih oblik. grafičnega oblikovanja, saj so pridobile osnovno razumevanje

Projektu smo namenile štiri šolske ure, od tega dve uri za zakonitosti grafičnega oblikovanja, kot so kompozicija, kontrast,

načrtovanje dela in pisanje prispevkov ter dve uri za spoznavanje barvna shema, poravnava besedila ipd.

s programom Canva in grafično oblikovanje prispevkov.

Do tedaj so se dijakinje spoznale s sedanjimi, preteklimi in

prihodnjimi slovničnimi časi, glagolniki, nedoločniki in trpniki v

present simple, past simple in present perfect simple. Eden izmed

ciljev projekta je bil tudi ustrezna uporaba čim več prej

omenjenih slovničnih struktur v lastnih besedilih.

Ker je avtonomija učenca ena izmed pomembnih elementov

za kakovostno izvedeno projektno delo [5], so dijakinje prejele

možnost izbire ustvarjanja netabloidne ali tabloidne publikacije



uporabo Canve nadgradile svoje digitalne kompetence,

predvsem na področju urejanja besedil in grafičnega oblikovanja.

Zaradi svoje preglednosti in enostavnosti za uporabo je Canva

idealna za izdelavo tovrstnih projektov. V prihodnje bi tovrsten

projekt lahko nagradili še z medpredmetnim povezovanjem z

informatiko in likovno vzgojo, saj bi tako pridobile dodatno

znanje na področju oblikovanja. Tovrstno znanje lahko nato

unovčijo v vsakdanjem življenju, na primer za oblikovanje

predstavitev, tiskovin, življenjepisov in drugih dokumentov za

osebno in profesionalno rabo.



VIRI IN LITERATURA



Slika 3: Primer naslovnice in posameznih strani (VIR: [1] Argawati, N. O., & Suryani, L. (2020). Project-based learning in teaching

writing: The implementation and students’ opinion. English Review:

lasten, zajem zaslonske slike) Journal of English Education, 8(2), 219-226. doi:

10.25134/erjee.v8i2.2120.

[2] Bülent, A., Stoller L. F. (2005). Maximizing the Benefits of Project Work



in Foreign Language Classrooms. English Teaching Forum. Volume 43.

Number 4, 10-21.

[3] Bilgin, I., Karakuyu, Y., & Ay, Y. (2015). The effects of project based

learning on undergraduate students’ achievement and self-efficacy beliefs

towards science teaching. Eurasia Journal of Mathematics, Science and

Technology Education, 11(3), 469-477. doi: 10.12973/eurasia.2014.1015a

[4] Canva. https://www.canva.com/ (30. 7. 2024).

[5] Bülent, A., Stoller L. F. (2005). Maximizing the Benefits of Project Work

in Foreign Language Classrooms. English Teaching Forum. Volume 43.

Number 4, 10-21.



Slika 4: Ustvarjanje prispevkov v Canvi (VIR: lasten)





Slika 5: Izdelava fotomontaže (VIR: lasten)



5 ZAKLJUČEK



Projektno delo dopušča hkratno izpopolnjevanje pisnih in

digitalnih kompetenc, saj s svojim celovitim in integrativnim

pristopom k učenju povezuje različna področja. To je idealno za

pouk tujega jezika, ki že sam po sebi združuje različna področja,

teme in spretnosti. Obenem moramo biti pozorni, da projekt

primerno osmislimo, saj so na ta način dijaki in dijakinje bolj

motivirani za delo, končni izdelek pa je bolj kakovosten in v

skladu z zastavljenimi cilji. Tako so bile spretnosti pisnega

izražanja v tujem jeziku postavljene v ustrezen kontekst, zaradi

česar so bili učni rezultati uspešnejši. Prav tako so dijakinje z





Uporaba aplikacije Stanford Cable TV News Analyzer v




visokem šolstvu



Using the Stanford Cable TV News Analyzer in Higher



Education



Doc. dr. Dejan Romih Feliks Možina Asist. Nejc Fir

Univerza v Mariboru Univerza v Ljubljani Univerza v Mariboru

Ekonomsko-poslovna fakulteta Ekonomska fakulteta Ekonomsko-poslovna fakulteta

Maribor, Slovenija Ljubljana, Slovenija Maribor, Slovenija

dejan.romih@um.si fm9230@student.uni-lj.si nejc.fir@um.si

Red. prof. dr. Silvo Dajčman Izr. prof. dr. Arne Baruca

Univerza v Mariboru Texas A&M University-San Antonio

Ekonomsko-poslovna fakulteta College of Business San Antonio

Maribor, Slovenija Teksas, ZDA

silvo.dajcman@um.si abaruca@tamusa.edu



V literaturi srečujemo različne primere uporabe klepetalnega

POVZETEK robota ChatGPT v (visokošolskem) izobraževanju [4, 5, 6], ne

V tem prispevku podajamo primere uporabe aplikacije srečujemo pa različnih primerov uporabe aplikacije Stanford Ca-

Stanford Cable TV News Analyzer v visokem šolstvu, pri čemer ble TV News Analyzer (analizatorja kabelskih novic), ki tudi te-

se osre-dotočamo na predavanja iz ekonomije. Ugotovili smo, da melji na umetni inteligenci. Tako bomo v tem prispevku več po-

je apli-kacija koristen pripomoček za izobraževalce, pa tudi zornosti posvetili aplikaciji Univerze Stanford, ki so jo razvili

izobraže-vance. Hong idr. [7] in je prosto dostopna na https://tvnews.stanford.edu.

Trdimo lahko, da je ta aplikacija koristna za oblikovalce politike



KLJUČNE BESEDE in raziskovalce, pa tudi izobraževalce, ki jo lahko uporabljajo v izobraževalnem procesu. Dejstvo je, da z novimi tehnologijami Analiza, analizator, ekonomija, kabelska novica, predavanje nastajajo tudi nove priložnosti, ki jih lahko izkoriščamo ali pa



ABSTRACT tudi ne [8].

Ta prispevek je strukturiran tako, da ima še pet poglavij. V

In this paper, we provide examples of the use of the Stanford poglavju 2 predstavljamo namen, v poglavju 3 metode, v pogla-

Ca-ble TV News Analyzer in higher education, focusing on vju 4 rezultate, v poglavju 5 predloge, v poglavju 6 pa sklep.

econom-ics classes. We have found that the application is a

useful tool for educators, as well as students.

2 NAMEN

KEYWORDS Namen prispevka je srednje- in visokošolskim učiteljem

Analysis, analyzer, economics, cable news, lecture ekono-mije predstaviti tri primere uporabe analizatorja kabelskih

novic in zapolniti vrzel v literaturi o (visokošolskem)



1 UVOD njegovo uvedbo. To apli-kacijo smo izbrali zato, ker omogoča izobraževanju (v dobi umetne inteligence), ki je nastala z

Umetna inteligenca vpliva tudi na (visokošolske) analizo kabelskih novic o kakem aktualnem dogodku, zanimiva

izobraževalne ustanove (izobraževalni proces). Uvedba pa ni samo za srednje- in visokošolske učitelje ekonomije, ampak

klepetalnega robota ChatGPT je med njimi povzročila potrebo npr. tudi za srednje- in visokošolske učitelje statistike, ki lahko

po prilagajanju novim razmeram, kar smo pričakovali glede na podatke aplikacije upo-rabijo pri računanju srednje vrednosti ali

njegov uspeh. Izkušnje učijo, da ni (visokošolskega) drugih statistik.

izobraževalca, ki ne bi slišal za kle-petalni robot ChatGPT, ki ga Prednost aplikacije, ki jo uporabljamo v tem prispevku, je, da

je razvilo podjetje OpenAI. Hitro po njegovi uvedbi novembra omogoča interaktivno delo z dijaki in/ali študenti, tako domačimi

2023 se je pokazalo, da je klepetalni robot ChatGPT za kot tujimi, kar je razvidno tudi iz primerov, ki jih predstavljamo

(visokošolske) izobraževalce izziv, ki mu ni para, kar nas ne v nadaljevanju.

preseneča glede na njegovo zmogljivost [1, 2]. Nasprotujoča si

stališča do uporabe klepetalnega robota ChatGPT v

(visokošolskem) izobraževanju so (visokošolske) iz-obraževalce 3 METODE

razdelila v dva tabora: na zagovornike in nasprot-nike njegove V tem prispevku uporabljamo metodi kompilacije in študije

uporabe v (visokošolskem) izobraževanju [3]. pri-mera (angl. case study), pri čemer se omejujemo na

predstavitev aplikacije Stanford Cable TV News Analyzer, ki Analiza slike 1: Iz grafikona na sliki 1 lahko domnevamo

omogoča ana-lizo kabelskih novic od januarja 2010, in primerov (sklepamo), da so v opazovanem obdobju ameriški mediji, na-

njene uporabe na predavanjih iz ekonomije. Uporabljamo tudi tančneje kabelske televizije CNN, FOX in MSNBC, o brezposel-

metodi analize in sinteze. nosti največ poročale do (vključno) leta 2014 in od leta 2020 do

Posebnost aplikacije je, da temelji na uporabi umetne inteli- (vključno) leta 2022, tj. med zdravstveno krizo (epidemijo

gence, ki je tudi v visokem šolstvu aktualna tema. covida-19).

Vprašanja za študente: 1. Kaj je bil razlog za to, da so med

epidemijo covida-19 ameriški mediji veliko poročali o brezpo-

4 REZULTATI selnosti? 2. Ali in kako lahko podatki o času zaslona, v katerem

V tem poglavju predstavljamo tri primere uporabe je bila v podnaslovu navedena angleška beseda unemployment,

analizatorja kabelskih novic, za katere menimo, da lahko koristijo oblikovalcem politike? 3. Ali in kako lahko poročanje

koristijo srednje- in visokošolskim učiteljem ekonomije, saj ameriških medijev o brezposelnosti vpliva na razpoloženje Ame-

lahko ti z njegovo upo-rabo popestrijo pouk ekonomije in ričanov, pri čemer upoštevajte, da ločimo dobre in slabe novice?

nenazadnje pritegnejo pozor-nost dijakov oz. študentov. V ta

namen smo izbrali tri pojme: brezposelnost (angl.

unemployment), inflacijo (angl. inflation) in ekonomsko politiko

(angl. economic policy), ki so tudi navedeni v Predmetnem

izpitnem katalogu za splošno maturo 2024 – eko-nomija [9].

Vsak primer vsebuje eno ali več slik, njuno analizo in eno ali

več vprašanj za študente. Čeprav so primeri namenjeni visoko-

šolskim učiteljem, jih lahko uporabljajo tudi srednješolski učite-

lji, pri čemer morajo analizo prilagoditi srednješolskemu izobra-

ževanju.



4.1 Primer 1: Brezposelnost

Analizator kabelskih novic kaže čas zaslona (angl. screen

time) (v sekundah, minutah itd.), v katerem je bila v podnaslovu

nave-dena kaka angleška beseda (npr. economy) ali kaka besedna

zveza (npr. economic policy) oz. v katerem je bila na zaslonu pri-

kazana kaka znana oseba (npr. Donald Trump, Joe Biden,

Kamala Harris), kar pomeni, da ne kaže čas poročanja o kakem Slika 2: Posnetek zaslonske slike v primeru besede

(aktualnem) dogodku ali osebi. Lahko pa domnevamo, da daljši unemployment na dan 16. 8. 2024 (čas zaslona v sekundah

ko je čas zaslona, daljši je čas poročanja. po TV-postajah) [7, 10].

V tem podpoglavju grafično prikazujemo rezultate analize ka-

belskih novic o brezposelnosti, ki smo jo naredili s pomočjo apli- Analiza slike 2: Iz grafikona na sliki 2 lahko domnevamo

kacije Univerze Stanford. Grafikon na sliki 1 smo dobili tako, da (sklepamo), da sta v opazovanem obdobju o brezposelnosti

smo v iskalno polje vtipkali text=unemployment, grafikon na (naj)več poročali kabelski televiziji FOX in MSNBC.

sliki 2 pa tako, da smo v prvo iskalno polje vtipkali Vprašanje za študente: Ali in kako lahko razlike v poroča-nju

text="unemployment" AND channel="CNN", v drugo ameriških medijev o brezposelnosti vplivajo na razpoloženje

text="unemployment" AND channel="FOX", v tretje pa članov (podpornikov) demokratske in republikanske stranke?

text="unemployment" AND channel="MSNBC". Sliki 1 in 2 lahko analiziramo na predavanju in/ali seminar-

skih vajah, na katerih obravnavamo brezposelnost. Pri tem je po-

membno, da uporabimo interaktivno spletno stran, ki nam omo-



goča, da ne analiziramo samo gibanja časa zaslona, ampak tudi



kabelske novice. (S klikom na točko na krivulji se namreč prika-

žejo vse kabelske novice o kaki stvari, ki se obravnava – gl. sliko

3.) Poleg tega lahko uporabimo tudi druge, strožje (iskalne) kri-

terije, s čimer lahko vsebino predavanja približamo potrebam. To

je zlasti dobrodošlo pri delu v majhnih skupinah, ko lahko upo-

rabljamo tudi na študenta osredotočeni pristop.



Slika 1: Posnetek zaslonske slike v primeru besede

unemployment na dan 16. 8. 2024 (čas zaslona v minutah)

[7, 10].





Slika 4: Posnetek zaslonske slike v primeru besede inflation



na dan 16. 8. 2024 (čas zaslona v minutah) [7, 10].





Slika 5: Posnetek zaslonske slike v primeru besede inflation

na dan 16. 8. 2024 (čas zaslona v sekundah po TV-postajah)

Slika 3: Posnetek zaslonske slike v primeru besede [7, 10].

unemployment, potem ko smo kliknili na točko na krivulji

(april 2020) (čas zaslona v minutah) [7, 10]. Analiza slike 5: Iz grafikona na sliki 5 lahko domnevamo

(sklepamo), da je v opazovanem obdobju o inflaciji (naj)več po-

4.2 ročala kabelska televizija FOX. Primer 2: Inflacija



kabel-skih novic o inflaciji. Grafikon na sliki 4 smo dobili tako, (podpornikov) demokratske in republikanske stranke? da smo v iskalno polje vtipkali text=inflation, grafikon na sliki 5 Sliki 4 in 5 lahko analiziramo na predavanju in/ali seminar-pa tako, da smo v prvo iskalno polje vtipkali text="inflation" skih vajah, na katerih obravnavamo inflacijo, pri čemer lahko po-V tem podpoglavju grafično prikazujemo rezultate analize ameriških medijev o inflaciji vplivajo na razpoloženje članov Vprašanje za študente: Ali in kako lahko razlike v poroča-nju

AND chan-nel="CNN", v drugo text="inflation" AND

datke o gibanju časa zaslona, v katerem je bila v podnaslovu na-

channel="FOX", v tretje pa text="inflation" AND

vedena angleška beseda inflation, primerjamo s podatki o

channel="MSNBC".

gibanju stopnje inflacije v ZDA. To lahko naredimo tudi pri

Analiza slike 4: Iz grafikona na sliki 4 lahko domnevamo

brezpo-selnosti, kjer lahko podatke o gibanju časa zaslona, v

(sklepamo), da so ameriški mediji o inflaciji največ poročali po



letu 2020, v katerem je prišlo do izbruha epidemije covida-19. unemployment, pri-merjamo s podatki o gibanju stopnje katerem je bila v podnaslovu navedena angleška beseda



izbruhu epidemije covida-19 ameriški mediji veliko poročali o Vprašanja za študente: 1. Kaj je bil razlog za to, da so po brezposelnosti v ZDA.



inflaciji? 2. Ali in kako lahko podatki o času zaslona, v katerem 4.3 Primer 3: Ekonomska politika je bila v podnaslovu navedena angleška beseda inflation, kori-

stijo oblikovalcem politike? 3. Ali in kako lahko poročanje ame- V tem podpoglavju grafično prikazujemo rezultate analize

riških medijev o inflaciji vpliva na razpoloženje Američanov? kabel-skih novic o ekonomski politiki. Grafikon na sliki 6 smo

dobili tako, da smo v iskalno polje vtipkali text="economic

policy", grafikon na sliki 7 pa tako, da smo v prvo iskalno polje

vtipkali text="economic policy" AND channel="CNN", v drugo

text="economic policy" AND channel="FOX", v tretje pa tudi da dijake oz. študente usmerijo k njegovi uporabi pri pisanju

text="economic policy" AND channel="MSNBC". seminarskih nalog. Predlagamo jim tudi, da primerjalno

analizirajo poročanje kabelskih postaj CNN, FOX in MSNBC ter



da uporabijo tudi druge, strožje (iskalne) kriterije.



6 SKLEP

Nasploh lahko trdimo, da je analizator kabelskih novic

koristen pripomoček. V podkrepitev te trditve smo navedli tri

(praktične) primere njegove uporabe na predavanjih iz

ekonomije, ki lahko drugim visokošolskim učiteljem ekonomije

služijo kot ilustra-cija. To velja tudi za srednješolske učitelje

ekonomije. Upošte-vati moramo, da so pripadniki generacije Z

drugačni od pripadni-kov prejšnjih generacij, kar povzroča

potrebo po prilagajanju iz-obraževalnega procesa njihovim



economic policy na dan 16. 8. 2024 (čas zaslona v minutah) (dejanskim) potrebam in že-ljam. Izkušnje (iz Slovenije) učijo, Slika 6: Posnetek zaslonske slike v primeru besedne zveze

[7, 10]. da moramo visokošolski uči-telji na predavanjih več pozornosti

posvečati (praktični) uporab-nosti znanja, saj lahko s tem

Analiza slike 6: Iz grafikona na sliki 6 lahko domnevamo izboljšamo izobraževalni proces.

(sklepamo), da so ameriški mediji o ekonomski politiki največ

poročali leta 2016 in 2024. ZAHVALA

Vprašanje za študente: Kaj je bil razlog za to? Raziskavo je financirala Javna agencija za

znanstvenorazisko-valno in inovacijsko dejavnost Republike



Slovenije (BI-US/24–26–006).



VIRI

[1] Ayeni, O. O., Al Hamad, N. M., Chisom, O. N., Osawaru, B., & Adewusi,

O. E. (2024). AI in education: A review of personalized learning and edu-

cational technology. GSC Advanced Research and Reviews, 18(2), 261–

271. https://doi.org/10.30574/gscarr.2024.18.2.0062

[2] Elbanna, S., & Armstrong, L. (2024). Exploring the integration of

ChatGPT in education: adapting for the future. Management & Sustaina-

bility: An Arab Review, 3(1), 16–29. https://doi.org/10.1108/MSAR-03-

2023-0016

[3] Mogavi, R. H., Deng, C., Kim, J. J., Zhou, P., Kwon, Y. D., Metwally, A.

H. S., ... & Hui, P. (2024). ChatGPT in education: A blessing or a curse?

A qualitative study exploring early adopters’ utilization and perceptions.

Computers in Human Behavior: Artificial Humans, 2(1), 100027.

https://doi.org/10.1016/j.chbah.2023.100027

[4] Jo, H. 2023. Understanding AI tool engagement: A study of ChatGPT

Slika 7: Posnetek zaslonske slike v primeru besedne zveze usage and word-of-mouth among university students and office workers.

economic policy na dan 16. 8. 2024 (čas zaslona v sekundah Telematics and Informatics, 85, November 2023, 102067.

https://doi.org/10.1016/j.tele.2023.102067

po TV-postajah) [7, 10]. [5] Pittner, D., & Romih, D., & Dejan. (2023). Uporaba klepetalnega robota

ChatGPT med študenti ekonomije in medicine. Ekonomija plus : revija

Analiza slike 6: Iz grafikona na sliki 7 lahko domnevamo Društva ekonomistov Maribor. 2(1), 24–30. https://demb.si/publika-

cije/ekonomija-plus-2-1.pdf

(sklepamo), da sta v opazovanem obdobju o ekonomski politiki [6] Mahmud, A., Sarower A.H., Sohel A., Assaduzzaman, M.,& Bhuiyan, T.

(naj)več poročali kabelski televiziji FOX in MSNBC. (2024). Adoption of ChatGPT by university students for academic purpo-

ses: Partial least square, artificial neural network, deep neural network and

Vprašanje za študente: Ali in kako lahko razlike v poroča-nju classification algorithms approach. Array, 21, 100339.

ameriških medijev o ekonomski politiki vplivajo na razpolo- https://doi.org/10.1016/j.array.2024.100339

[7] Hong, J., Crichton, W., Zhang, H., Fu, D. Y., Ritchie, J., Barenholtz, J., ...

ženje članov (podpornikov) demokratske in republikanske & Fatahalian, K. (2021). Analysis of faces in a decade of us cable tv news.

stranke? V KDD'21: Proceedings of the 27th ACM SIGKDD Conference on

Sliki 6 in 7 lahko analiziramo na predavanju in/ali seminar- Knowledge Discovery & Data Mining (str. 3011–3021).

https://doi.org/10.1145/3447548.3467134

skih vajah, na katerih obravnavamo ekonomsko politiko. Po- [8] Patrício, M. R., & Gonçalves, B. F. (2024, January). ChatGPT: Systematic

datke o gibanju časa zaslona, v katerem je bila v podnaslovu na- review of potentials and limitations in education. V International Confe-

rence on Information Technology & Systems (str. 339–348). Springer Na-

vedena angleška besedna zveza economic policy, lahko primer- ture Switzerland. http://doi.org/10.1007/978-3-031-54256-5_32



jamo s podatki o gibanju ekonomskopolitične negotovosti v ZDA. [9] Državni izpitni center. (2022). Predmetni izpitni katalog za splošno ma- turo 2024 – ekonomija. Ljubljana: Državni izpitni center.

https://www.ric.si/mma/PIK_EKN-2024-

novi.pdf/2023011710020901/?m=1673946129



5 PREDLOGI [10] Stanford Cable TV News Analyzer. (2024). The Data. https://tvnews.stan-ford.edu/

Srednje- in visokošolskim učiteljem ekonomije predlagamo,

da razmislijo o uporabi analizatorja kabelskih novic pri pouku

eko-nomije in da skupaj z dijaki oz. študenti analizirajo

poročanje ka-belskih postaj o kakem (aktualnem) dogodku, pa



Razvijanje socialno-emocionalnih veščin z uporabo



animiranega filma v predšolskem obdobju



Developing Social-Emotional Skills through Animated Film in



Early Childhood



Luka Sterle

Predšolska vzgoja

Vrtec Litija

Litija, Slovenija

luka.sterle@vrtec-litija.si



interactions with peers, adults, and the environment.

POVZETEK Additionally, various media, such as animated films, can play a

Socialno-emocionalne veščine, kot so prepoznavanje in significant role in enhancing these skills.

izražanje At the beginning of the kindergarten year, educators observed čustev, empatija, reševanje konfliktov in

samoregulacija, so bistvene za uspešen razvoj otrok. V that children were not tidying up their toys, mishandling toys in

predšolskem obdobju otroci razvijajo te veščine skozi interakcijo different play areas, and lacking a sense of responsibility for the

z vrstniki, odraslimi in okoljem. Poleg tega lahko različni mediji, kindergarten materials. To address these challenges, the idea was

kot so animirani filmi, igrajo pomembno vlogo pri krepitvi teh conceived to mirror the children's actions, leading to the creation

veščin. of Trmica. Trmica is a crocheted round ball with legs, arms, and

V začetku vrtčevskega leta je bilo opaziti, da otroci ne a face, which does not speak but whispers both good and bad

pospravljajo igrač, neprimerno rokujejo z igračami v različnih things to the children. Through its mischievous behavior, Trmica

kotičkih in nimajo razvitega čuta za odgovornost do materiala v encourages children to recognize emotions and resolve conflicts.

vrtcu in ne navezujejo stikov z vrstniki. Da bi spremenil te vzorce, Trmica represents a stubborn child who believes they can do

sem se odločili zrcaliti otroška dejanja, kar je vodilo do ideje o anything without facing consequences.

Trmici. To je kvačkana okrogla žogica z nogami, rokami in The aim of this article is to investigate and monitor the

obrazom, ki ne govori na glas, ampak šepeta otrokom na uho, behavior of children after the introduction of Trmica into the

tako dobre kot slabe stvari. S svojim nagajivim vedenjem group, and to assess how the use of animated films featuring

spodbuja otroke k prepoznavanju čustev in reševanju konfliktov. Trmica can help in developing social-emotional skills in

Nered ustvarja le v času odsotnosti otrok iz vrtca. Trmica preschool children. The research focuses on observing children

predstavlja trmastega otroka, ki misli, da lahko počne karkoli while watching the animated films, analyzing their behavior, and

brez posledic za svoja dejanja. collecting feedback from parents and educators to evaluate the

Cilj tega članka je raziskati in spremljati vedenje otrok po effects of these films on the development of social-emotional

vključitvi Trmice v skupino ter oceniti, kako lahko uporaba skills.

animiranih filmov s Trmico pomaga pri razvijanju socialno-

emocionalnih veščin pri predšolskih otrocih. Raziskava se KEYWORDS

osredotoča na opazovanje otrok med gledanjem animiranih Social-emotional skills, preschool children, animated films,

filmov, analizo njihovega vedenja ter zbiranje povratnih Trmica, emotion recognition, self-regulation, empathy, conflict

informacij od staršev in vzgojiteljev, da bi ocenili učinke teh resolution, responsibility

filmov na razvoj socialno-emocionalnih veščin.



KLJUČNE BESEDE 1 UVOD

Socialno-emocionalne veščine, predšolski otroci, animiran Razvoj socialno-emocionalnih veščin v predšolskem obdobju

film, Trmica, prepoznavanje čustev, samoregulacija, empatija, je ključen za celostni razvoj otrok. Te veščine vključujejo

reševanje konfliktov, odgovornost sposobnost prepoznavanja in izražanja čustev, empatijo,

reševanje konfliktov ter samoregulacijo, ki so bistvene za

ABSTRACT uspešno interakcijo z vrstniki in odraslimi. V predšolskem



emotions, empathy, conflict resolution, and self-regulation, are okoljem in drugimi osebami [1]. Vendar pa sem v začetku leta opazil težave, kot so neprimerno rokovanje z igračami, essential for the successful development of children. In the Social-emotional skills, such as recognizing and expressing obdobju otroci te veščine razvijajo predvsem skozi interakcije z



preschool period, children develop these skills through neprijaznost do vrstnikov in vzgojiteljev ter pomanjkanje odgovornosti. Da bi učinkovito reševal te izzive, sem se odločil

preizkusiti uporabo animiranih filmov (stop motion pictures), kot Poleg tega ima dodano vrednost posnetek, ki je narejen v

orodja za izboljšanje socialno-emocionalnih veščin, pred igralnici. V kolikor lahko vključiš rekvizite s katerimi otroci

uporaba filma sem z otroki izvedel različne dejavnosti, ki dnevno rokujejo in z njimi preživijo večino dneva je to dodana

povezujejo skupino, krepijo samopodobo posameznika in vrednost. Takšne vsebine omogočajo vključitev tematik, ki so

celotno skupino ter širši krog ljudi s katerim se otroci dnevno trenutno najbolj zaželene, pereče ali bodo otrokom in

družijo. vzgojiteljem pomagale pri uspešni realizaciji zastavljenih ciljev.

V tem članku bom predstavil primere dejavnosti za Pomembno pa je poudariti, da posnetki, filmi in podobne

prepoznavanj neprimernega vedenja, prepoznavanje in izražanje vsebine ne smejo biti edini vir spodbude za otroke. Gre le za eno

čustev ter uporabe lika Trmica (slika 1), kvačkane žogice z izmed mnogih orodij, ki jih je treba skrbno vključiti v celoten

rokami, nogami in obrazom, ki sem ga uvedel v igralnico, da bi pedagoški pristop, kjer so na prvem mestu neposredne izkušnje

otroke spodbudil k refleksiji o njihovem vedenju in razvoju in interakcije z okoljem in sovrstniki.

potrebnih veščin. Proučil sem, kako so otroci reagirali na

spremembe v okolju, kako je Trmica vplivala na njihovo vedenje

ter kakšne so bile posledice za njihov socialno-emocionalni 3 UVAJANJE AKTIVNOSTI ZA RAZVOJ

razvoj. SOCIALNO-EMOCIONALNIH

SPRETNOSTI IN SAMOREGULACIJE



Pred začetkom uporabe lika Trmica v programu sem se

osredotočil na dejavnosti, ki so prispevale k razvoju socialno-

emocionalnih veščin otrok. Vpeljal sem različne aktivnosti, ki so

omogočile raziskovanje in izražanje čustev ter izboljšanje

sposobnosti samoregulacije in empatije.





Slika 1: Trmica se igra z igračami



2 RAZUMEVANJE ANIMIRANIH FILMOV

V PREDŠOLSKEM OBDOBJU

Animirani filmi lahko v predšolskem obdobju pripomorejo

pri razvoju otrokove socialno-emocionalne pismenosti,

kognitivnih veščin in ustvarjalnosti. Takšne vsebine omogočajo

otrokom, da se identificirajo z liki in situacijami v filmih, kar jim Slika 2: Moj prijatelj - plakat

pomaga razumeti in izražati svoja čustva ter razviti empatijo do



drugih. Vključevanje animiranih filmov v izobraževalne

dejavnosti lahko pozitivno vpliva na medosebne odnose otrok ter

sposobnosti reševanja konfliktov, saj jim ponuja vpogled v

različne čustvene izkušnje in socialne dinamike. Kakovost

izobraževalnih vsebin in pristopov, vključno z uporabo

animiranih filmov, ima pomembno vlogo pri zagotavljanju

visoke kakovosti vzgoje in izobraževanja v vrtcih in šolah, kar

potrjujejo smernice in raziskave na tem področju [2].

Poleg tega animirani filmi spodbujajo otroško ustvarjalnost in

kognitivni razvoj. Domišljijski elementi v animiranih filmih

omogočajo raziskovanje novih idej ter razvoj kritičnih miselnih

sposobnosti. Zgodbene vsebine, ki vključujejo reševanje

problemov in prepoznavanje vzorcev, pripomorejo k razvoju

otrokove sposobnosti razmišljanja in reševanja težav. Takšne

vsebine spodbujajo kognitivni razvoj in kreativnost otrok ter

prispevajo k njihovemu celostnemu razvoju. Vključevanje

kakovostnih animiranih filmov v izobraževalne programe je v

skladu z najboljšimi praksami za zagotavljanje kakovosti

izobraževalnega procesa, kot je navedeno v smernicah za

kakovost v vrtcih in šolah [2]. Slika 3: Srce – moja družina

Med njimi je bila izdelava plakatov o najboljših prijateljih, predlagali, da pripravimo past in ulovimo nepridiprava. Tako

kjer so otroci zapisali imena svojih prijateljev in opisali najljubše smo iz kartona spletli mrežo in jo postavili nad čokolado, ki je

trenutke z njimi, kar je spodbujalo refleksijo in razvoj empatije. bila na mizi. Naslednji dan je bila past sprožena, a o čokoladi ni

Po izdelanem plakatu (slika 2) so se odšli otroci igrati s svojim bilo ne sluha ne duha. Na mizi je bila le rdeča lisa.

najljubšim prijateljem. Seveda to ni bilo povsem izvedljivo, saj Nekaj dni, pred ogledom videa o Trmici, smo prebrali zgodbo

so se med seboj imena ponavljala. Težavo so rešili s skupno igro. "Zmajček in Trmica," kjer se je otrokom že začela sestavljati

Poleg tega so otroci ustvarili plakate v obliki srca (slika 3), kjer celotna slika. Ugotovili so, da bi bil nepridiprav lahko prav ta

so risali najljubše osebe v družini in med prijatelji, kar je krepilo Trmica iz zgodbe. Po več tednih ugotavljanja je napočil čas, ko

občutek ljubezni in pripadnosti. Med risanjem plakata so v kameri na listku ni bilo slike, ampak je pisalo "video." Otroci

opisovali dogodke, se spominjali lepih trenutkov in z nasmehom so ugotovili, da moramo kamero priklopiti na računalnik. To smo

opravili nalogo. Za prepoznavanje čustev sem uporabil slikovne naredili in začel se je predvajati posnetek (stop motion picture) s

kartice, ki so prikazovale različna čustva, ter aktivnosti, kjer so Trmico v glavni vlogi. Kamera jo je posnela, kako je razmetavala

otroci označevali svoje počutje na pripravljeni šabloni. To je stole v igralnici, raztresala različne igrače, plezala po omarah in

prispevalo k razvoju čustvene pismenosti in samorefleksije. brskala po predalih. V sobi je zavladala tišina. Otroci so nemo

Dejavnosti smo izvajali v dopoldanskem času med prihodom v gledali posnetek in komentirali dogodke. Ta dan so končno

vrtec in v popoldanskem času med odhajanjem domov. V ugotovili, kdo jim povzroča nered v igralnici.

primeru neželenega dogodka v skupini, sem med pogovorom

vprašal otroka o njegovem počutju, tako da so lahko te občutke

označili na listih. Sčasoma so jih usvojili z besedami, tako da niso 5 SPOZNAVANJE LIKA TRMICE

več potrebovali slikovne opore. V sklopu aktivnosti za razvoj Ko smo Trmico locirali in našli, je vsa prestrašena skočila

samoregulacije sem se osredotočil na učenje in vadbo besednih meni v naročje (slika 4). Z otroki smo si v krogu ogledali Trmico

ukazov ter gest za obvladovanje čustev in vedenja. Otroci so se in jo vprašali, od kod je prišla in ali je ona tista, ki je delala nered

učili uporabljati fraze, kot so "stop," "to mi ni všeč," "nehaj v naši igralnici. Nič ni odgovarjala, le gledala nas je. Kmalu smo

takoj" v konfliktnih situacijah in vadili uporabo geste z dlanjo za ugotovili, da govori le na uho, zato sem jaz prvi poslušal njene

prekinjanje neželenega vedenja. ideje in moram reči, da so bile zelo nagajive (Trmica govori na

Dnevno smo izvajali pretvarjanja situacij v skupini, ki so uho tako, da si sam izmišljaš kaj ti govori). Seveda nam je tudi

najpogosteje vodile v konflikt. Pretvarjanje je potekalo tako, da odgovorila na vsa naša vprašanja. Trmico smo si podajali naprej,

sva s sodelavko zaigrala njihov primer, se o njem pogovorili in tako da so jo vsi pozdravili in se z njo pogovorili. Otroci so imeli

ga nato še enkrat zaigrala s predlogi otrok, da rešiva nastalo zelo zanimive ideje o tem, zakaj je razmetavala sobo in

situacijo. Kasneje so tudi otroci igrali igre pretvarjanja. Uporabil prestavljala igrače. Po končanem pogovoru smo ji poiskali mesto

sem vizualne pripomočke, kot so plakati in kartice z navodili, da v igralnici in se dogovorili, da jo lahko vsak vzame v naročje, če

sem podprl učenje teh veščin, otroci pa so jih redno uporabljali v bo le lepo rokoval z njo, vendar se mora vedno vrniti na

vsakodnevnih situacijah znotraj učnega okolja. Te dejavnosti so dogovorjeno mesto. Tako je Trmica postala del naše skupine.

bile ključne za pripravo otrok na delo z likom Trmica, saj so

prispevale k njihovemu boljšemu razumevanju in izražanju



čustev ter spodbujanju pozitivnega vedenja v skupini.



4 ZRCALJENJE OTROŠKIH DEJANJ IN

ISKANJE KRIVCA: UPORABA LIKA

TRMICE

Da bi otroke pritegnil in jih spodbudil k razmišljanju o svojih

vedenjskih vzorcih, sem v skupino začel vnašati rahle

spremembe, kot so razmetane barvice, pomešane igrače,

neurejeni delovni zvezki, izginule igrače, prestavljeni predmeti

(stenkska ura, stoli, kotički, znaki otrok itd.). Vsako jutro, ko so

prišli v igralnico in naleteli na nered ali izginotje predmetov, sem

jih vprašal: "Kdo od vas ni včeraj pospravil igrač?" Vsak dan je

bil odgovor enak - vsi so trdili, da so vse pospravili. Ko so Slika 4: Našli smo Trmico

ugotovili, da se te stvari dogajajo v času njihove odsotnosti iz

vrtca, so predlagali, da ponoči postavimo kamero v sobo, da bi

videli, kdo povzroča ta nered. Kamero smo naredili iz kartona in 6 SODELOVANJE LUTKE V SKUPINI

jo postavili na najvišjo omaro, tako da je zajela celotno igralnico. Trmica je dnevno sodelovala v jutranjih krogih in v vseh

Nato smo čakali. Napetost je rasla in vsak dan smo preverjali, ali drugih delih dneva. Kljub temu, da smo jo vključili v našo rutino,

je v kameri kakšna slika. nam je še vedno nagajala; ponoči je razmetavala igralnico, zjutraj

Nekega dne smo dejansko našli sliko. Na njej je bil le del pa je tiho opazovala naše odzive. Kadar se kakšna igračka ali

Trmice in razdejanje, ki ga je povzročila. Otroci niso mogli igralni kotiček nista primerno uredila, smo prosili Trmico, da

ugotoviti, kdo je to, zato se je napetost še povečala. Ko so se slike nam pomaga urediti kotičke. Otroci so z njo bolj zagreto

vrstile več tednov in nihče ni mogel ugotoviti, kdo je to, so pospravljali in brez dodatnih spodbud izvajali vodene dejavnosti

v času bivanja v vrtcu. Trmica je postala kot nagrada za druženje postali bolj vestni in odgovorni pri svojih vsakodnevnih opravilih,

po dobro opravljenem delu, kar je še dodatno motiviralo otroke kar je prineslo pozitivne spremembe tako v vrtcu kot tudi doma.

pri njihovih nalogah. Negativni vedenjski vzorci otrok so se

skoraj popolnoma prekinili, saj so otroci začeli uporabljati

Trmico kot izgovor za neprimerno vedenje, češ da jim je Trmica 8 ZAKLJUČEK

rekla naj naredijo kakšno neumnost. Kljub temu so bili otroci Vpeljava lika Trmice v skupino predšolskih otrok se je

zelo ljubeznivi do lutke, kar je pripomoglo k temu, da je večina izkazala kot učinkovit pristop za spodbujanje razvoja socialno-

negativnih dejanj izvenela. V prvem tednu prisotnosti Trmice je emocionalnih veščin. Otroci so skozi interakcijo z lutko Trmico

bilo opaziti več skupne igre, saj so otroci vključevali lutko v in spremljanje njenih domnevnih dejanj pridobili globlje

svoje najljubše kotičke, kar je dodatno spodbujalo sodelovanje in razumevanje čustev, empatije, odgovornosti in pravilnega

pozitivno interakcijo med njimi (slika 5). vedenja.Ta inovativni pedagoški pristop je pripomogel k temu,

da so otroci začeli bolj zavzeto pospravljati svoje stvari, se bolj



angažirano vključevali v skupinske dejavnosti in razvijali

sposobnost reševanja konfliktov na miren način.

Starši so opazili, da so otroci doma bolj odprti za razprave o

svojih čustvih in vedenju, kar kaže na širši vpliv Trmice izven

vrtčevskih okvirov. Kljub temu, da so otroci včasih Trmico

uporabili kot izgovor za lastne pomanjkljivosti, so se pri tem učili

pomembnih življenjskih lekcij o odgovornosti in posledicah

svojih dejanj. S tem so postali bolj vestni in samostojni, kar se

kaže tudi v njihovem vedenju doma.

Uporaba animiranih filmov in zgodb, skupaj z uporabo lutke

kot interaktivnega orodja, se je izkazala kot učinkovit način za

podporo celostnemu razvoju otrok. Trmica je tako postala

pomemben del skupinskega življenja, ki ne le zabava, temveč

tudi uči in vodi otroke k bolj zrelim in odgovornim osebam.

Prepoznanje in naslavljanje otrokovih čustev ter vedenj preko

Slika 5: Trmica posluša zgodbo z otroki likov, kot je Trmica, ponuja vrtcem in šolam priložnost za

nadgradnjo tradicionalnih vzgojnih metod z inovativnimi

pristopi, ki so prilagojeni potrebam sodobnih otrok.

7 ODZIV STARŠEV PPRED IN PO

SPREJEMU TRMICE V SKUPINO. VIRI

Starši so pred uvedbo Trmice v skupino opažali, da se otroci [1] Brezovnik, M. (2021). Metode in tehnike za razvoj socialno-emocionalnih

veščin v predšolskem izobraževanju. Maribor: Fakulteta za izobraževalne

doma pogosto niso želeli soočiti z odgovornostmi, kot so vede.

pospravljanje svojih igrač, oblačil in opravljanjem osnovnih [2] ZRSŠ (2023). Smernice za kakovost v vrtcih in šolah. Znanstveno-

domačih nalog, kot je urejanje postelje ali obutve. Po sprejemu raziskovalni center Slovenske akademije znanosti in umetnosti.

Trmice v skupino so začeli opažati zanimivo spremembo v

vedenju svojih otrok. Kadar otroci doma naredijo kakšno

neumnost ali ne opravijo svojih nalog, krivijo Trmico, saj naj bi

jim ona rekla, naj ne opravijo svojih obveznosti. To je postal

pogost izgovor pri otrocih. Starši hkrati opažajo, da otroci zdaj

veliko več govorijo o neprimernih vedenjih Trmice in o tem, kaj

je prav in kaj ne.

Poleg tega so starši opazili, da otroci pogosto omenjajo, kako

se Trmica v vrtcu obnaša in kaj vse se je naučila. To je

spodbudilo zanimive pogovore doma, kjer otroci sami sebe

opozarjajo, da se "ne smejo obnašati kot Trmica." Starši

sporočajo, da otroci, ki so bili prej zadržani pri izražanju svojih

čustev, zdaj bolje ubesedijo svoje občutke in razpravljajo o

nastalih situacijah, kar nakazuje, da so razvili večjo zavest o

pravilnem in nepravilnem vedenju.

Zanimivo je tudi to, da so starši začeli opažati, da otroci zdaj

pokažejo več empatije do drugih članov družine. Včasih, ko pride

do nesoglasij, jih poskušajo reševati na mirnejši način, kot so to

počeli prej. Opažajo, da otroci zdaj več in bolj temeljito

pospravljajo svoje stvari kot prej, ter se trudijo slediti dogovorom

doma. Čeprav se morda še vedno poskušajo izogniti

odgovornosti, pa je očitno, da so sprejeli sporočilo Trmice in



Psihologija barv na tečaju grafičnega oblikovanja



The Psychology of Colors in Graphic Design Course



Iva Strniša

GFP Kranj

Kranj, Slovenija

iva.strnisa@gfp.si



graphic designers and psychologists, which could improve the

POVZETEK effectiveness of visual communication solutions.

V digitalni dobi ima grafično oblikovanje ključno vlogo pri

ustvarjanju vizualno prepričljivih sporočil, ki presegajo KEYWORDS

jezikovne meje in dosegajo globalno občinstvo. Projekt RaSTKo, Graphic design, colors, psychology, course, RaSTKo

sofinanciran s strani Ministrstva za vzgojo in izobraževanje RS

ter EU, si prizadeva povečati vključenost odraslih v

vseživljenjsko učenje in izboljšati njihove kompetence. V sklopu 1 UVOD

projekta se izvajajo brezplačni tečaji, vključno z izobraževanjem V digitalni dobi, ki jo živimo, je moč vizualne komunikacije

o grafičnem oblikovanju. Le ta je namenjen različnim poklicnim postala neprecenljiva. Grafično oblikovanje je postalo ključno

profilom, ter pokriva temeljne in napredne vidike oblikovanja in orodje za ustvarjanje močnih, prepričljivih sporočil, ki segajo

uporabo grafičnih programov. Poseben poudarek je na vplivu preko jezikovnih meja in nagovarjajo občinstvo na globalni ravni.

barv na čustva in vedenje uporabnikov, kjer se udeleženci Beseda »grafično« izvira iz grške besede »graphein« in

naučijo strateške uporabe barv za učinkovito vizualno pomeni pisanje oz. označevanje. Beseda oblikovanje oz. v

komunikacijo. angleščini »design« izvira iz francoskega izraza »dessiner«, kar

Empirični del, ki vključuje anketo med udeleženci tečaja, pomeni risanje, načrtovanje, skiciranje in oblikovanje. Iz tega

potrjuje, da barve močno vplivajo na čustvene reakcije in vedenje lahko sklepamo, da je v svojem primarnem pomenu grafično

uporabnikov, pri čemer sta bila rdeča in modra ocenjeni kot oblikovanje smiselno »pisanje« informacij in pretvorba le-teh [1].

najbolj učinkoviti barvi za jumbo plakate. Članek zaključuje s Od podjetij do posameznikov se vedno več ljudi zaveda

priporočilom za sodelovanje med grafičnimi oblikovalci in pomena estetike in funkcionalnosti, ki ju prinaša dobro

psihologi, kar bi lahko izboljšalo učinkovitost vizualnih oblikovanje. V tem kontekstu je znanje grafičnega oblikovanja

komunikacijskih rešitev. postalo izjemno cenjeno in iskano, saj omogoča ustvarjanje

vizualnih izdelkov, ki izstopajo in imajo trajen vpliv.

KLJUČNE BESEDE Kakovostni izobraževalni viri se včasih izkažejo za

Grafično oblikovnaje, barve, psihologija, tečaj, RaSTKo nedostopne zaradi finančnih omejitev. Tukaj pa marsikomu

lahko pomagajo brezplačni tečaji, ki so preko projektov



creating visually compelling messages that transcend language In the digital age, graphic design plays a crucial role in evropskih sredstev. Te ne le odpirajo vrata kakovostnemu izobraževanju, ampak tudi omogočajo inovativnost in kreativni ABSTRACT financirani ali sofinancirani s strani državnih inštitucij ali

barriers and reach a global audience. The RaSTKo project, co- razvoj med udeleženci, ki se izobraževanj sicer ne bi udeležili.

financed by the Ministry of Education of the Republic of



Slovenia and the EU, aims to increase adult participation in 2 PREDSTAVITEV PROJEKTA lifelong learning and enhance their competencies. As part of the

project, free courses are offered, including training in graphic RaSTKo - Razvoj spretnosti in temeljnih kompetenc je

design. These courses are designed for various professional projekt, ki ga sofinancirata Ministrstvo za vzgojo in

profiles, covering fundamental and advanced aspects of design izobraževanje Republike Slovenije in Evropska unija iz

and the use of graphic software. Special emphasis is placed on Evropskega socialnega sklada plus, njegov namen pa je povečati

the impact of colors on emotions and user behavior, where vključenost odraslih v vseživljenjsko učenje (VŽU) ter izboljšati

participants learn the strategic use of colors for effective visual kompetence, ki jih odrasli potrebujejo zaradi potreb na trgu dela,

communication. večje zaposljivosti in mobilnosti ter osebnega razvoja za

The empirical part, which includes a survey among course delovanje in odzivanje na tehnološke, demografske in podnebne

participants, confirms that colors significantly influence spremembe v sodobni družbi.

emotional reactions and user behavior, with red and blue being Ljudska univerza Kranj je na javnem razpisu Temeljne

rated as the most effective colors for billboards. The article kompetence 2023-2029 Ministrstva za vzgojo in izobraževanje

concludes with a recommendation for collaboration between pridobila finančna sredstva za izvajanje tega projekta, ki se bo s



pomočjo konzorcijskih partnerjev (Šolski center Kranj in  Teoretično znanje: Razumevanje zgodovine in teorije

Biotehniški center Naklo) izvajal do junija 2029. Med grafičnega oblikovanja ter obdelave fotografij, vključno

načrtovanimi kazalniki je izvedba 248 izobraževalnih programov, s preučevanjem pomembnih gibanj, slogov in

od tega 28 javnoveljavnih, vključenih pa naj bi bilo 1942 oblikovalskih trendov.

udeležencev.  Praktične veščine: Razvijanje tehničnih in praktičnih

Programi, ki se s strani vseh sodelujočih izvajajo v sklopu veščin, ki so potrebne za izvedbo različnih projektov,

projekta, so za udeležence brezplačni. Pri izvajanju programov vključno z oblikovanjem publikacij, embalaže, oglasnih

bodo sodelovali različni izvajalci in strokovni delavci, ki bodo sporočil, spletnih strani in drugih vizualnih medijev.

skušali pokriti različna področja za dvig spretnosti in kompetenc  Kritično razmišljanje: Spodbujanje analitičnega in

odraslih, še posebej ranljivih skupin. Ciljna skupina so odrasli, kritičnega razmišljanja pri evalvaciji lastnih del in del

med katerimi bodo zajeti tudi tisti iz ranljivih skupin: nižje drugih oblikovalcev, kar pomaga pri izboljšanju

izobraženi, aktivni, ki so stari 55 let ali več, ter tujci. kakovosti in učinkovitosti oblikovalskih rešitev.

 Etična in profesionalna vprašanja: Učenje o etičnih



3 standardih v industriji grafičnega oblikovanja in TEČAJ GRAFIČNEGA OBLIKOVANJA

obdelave fotografij ter razumevanje profesionalnih

Cilji projekta so pridobitev in izboljšanje temeljnih praks in norm.

kompetenc ter splošne izobraženosti odraslih. Za doseganje teh Končni cilj tečaja je bil, da znajo tečajniki izdelati svoj oglas

ciljev, je bilo pripravljenih več izobraževanj iz različnih področij. z uporabo teoretičnih in praktičnih veščin, ki so jih pridobili na

Eno od področij je bilo tudi pridobivanje digitalnih kompetenc, tečaju. Krmelj pravi, da nastanek enega oglasa zahteva ljudi z

v katero sodi tudi grafično oblikovanje. znanjem s področja ekonomije, oblikovanja, fotografije, filma,

Program je namenjen vsem odraslim osebam, ki že delajo v psihologije, jezikoslovja, umetnosti, kulture, računalništva... [2],

industriji oblikovanja ali umetnosti in bi lahko izobraževanje zato smo pri načrtovanju vsebin tečaja moči združili strokovnjaki

uporabili za izboljšanje svojih tehničnih veščin in obogatitev iz različnih področij.

svojega portfelja:



 podjetnikom, ki pogosto potrebujejo znanje grafičnega 4 VPLIV BARV NA ČUSTVA IN VEDENJE

oblikovanja za ustvarjanje marketinških materialov, UPORABNIKOV logotipov, spletnih strani in drugih vizualnih vsebin za

svoje podjetje, V tečaju grafičnega oblikovanja, ki so ga izvajali na Šolskem

 centru Kranj, sem vsebine obogatila z vlogo barv, ki so marketinškim strokovnjakom, ki bi znanje grafičnega

oblikovanja lahko uporabili za izdelavo privlačnih pomemben zunanji dejavnik pozornosti, in pomagala doseči cilj

marketinških materialov, kot so letaki, oglasi, digitalne razumevanja vizualne komunikacije, konkretneje, kako vizualne

objave in druge promocijske vsebine, elemente uporabljati za učinkovito komunikacijo sporočil in idej

 preko barv in njihovega odnosa do realnosti. fotografom, ki bi izobraževanje lahko izkoristili za



izboljšanje svojih veščin obdelave fotografij, kar jim 4.1 Barva kot fizični fenomen in njena vloga v



 kulturi učiteljem in izobraževalcem, ki poučujejo umetnost, pomaga pri izboljšanju kakovosti svojih del in trženju,

grafično oblikovanje ali sorodne predmete, ki bi lahko Pri zaznavanju realnosti so izredno pomembni predvsem:

izobraževanje uporabili za pridobitev novih orodij in oblika zaznanih predmetov, njihov relief, njihova voluminoznost

znanja za poučevanje, in njihovi prostorski odnosi z drugimi predmeti. Toda barva ima

 samostojnim ustvarjalcem, ki želijo razvijati svoje kljub temu mnogo globlji pomen, saj je njen vpliv sila

spretnosti iz hobija ali kot dodaten vir dohodka in bi razprostranjen v različnih sferah subjektovega psihičnega

lahko tečaj izkoristili za izboljšanje svojih veščin. (predvsem čustvenega) dogajanja [3].

Cilji izobraževalnega programa pa so bili sledeči: Barva, za razliko od oblike, dodaja predmetom še določeno



 Razumevanje osnov grafičnega oblikovanja: Udeleženci oblikah človekove kulturne dejavnosti, saj spodbuja subjektovo čustveno vsebino. Simbolika barv ima tako pomen v vseh

se naučijo osnovnih načel oblikovanja, kot so intuitivnost, ga sili k spoznavanju in omogoča doživljajsko

kompozicija, tipografija, barvna teorija in vizualna polnost, ki temelji na aktivaciji mnogih psihičnih procesov na

hierarhija. več ravneh [3].

 Obvladovanje grafičnih programov: Znanje uporabe

različnih grafičnih programov, kot so Adobe Photoshop, 4.2 Psihološko-simbolični učinki nekaterih barv

Illustrator, InDesign, Lightroom in druge specializirane ter vpliv na doživljanje in vedenje

programske opreme za grafično oblikovanje in obdelavo uporabnikov

fotografij.

 Rdeča je bila prva barva, ki jo je človek poimenoval [3]. Razvoj kreativnih sposobnosti: Spodbujanje kreativnega

razmišljanja in inovativnosti pri ustvarjanju vizualnih Psihološko-simbolični učinek doseže zaradi fizične povezanosti



 komunikaciji precej previdni, saj je zelo vpadljiva barva, ki lahko Razumevanje vizualne komunikacije: Poudarek na v določenih okoliščinah deluje precej nasilno. Opazovalec jo bo razumevanju, kako vizualne elemente uporabljati za rešitev za različne medije in platforme. z ognjem in krvjo. A z redečo moramo biti pri vizualni

učinkovito komunikacijo sporočil in idej. pogosto implicitno interpretiral kot nevarnost ali prepoved, saj



vedno izstopi iz svojega naravnega okolja. Presoditi moramo

torej, ali jo lahko uporabimo kot orodje za aktivnost, privlačnost,

dinamičnost ali pa bo v kontekstu delovala neresno in površinsko

in bo posledično opazovalec spregledal njeno sporočilo.

Zelena se v naravi pojavlja v izredno velikih količinah, a je

običajno stopnja njene zasičenosti nizka, tako da je njen vpliv na

človeka običajno ugoden [3]. Pogosto jo povezujemo s svobodo.

Nenazadnje tudi zelena luč pomeni prosto pot. Pri vizualni

komunikaciji moramo upoštevati pomirjevalni vpliv na

organizem, a z zavedanjem, da je vpliv na posameznika odvisen

predvsem od drugih barv, ki se pojavljajo skupaj z njo. Če jo

kombiniramo z modro in belo, sporoča pozitivne, s črno, rumeno

in vijoličasto pa negativne lastnosti.

Kljub temu, da rumeno barvo največkrat asociiramo s soncem,

ki simbolizira svetlobo, vedrino in optimizem, je potrebno biti

pazljiv pri uporabi te barve, saj njena simbolika ni vedno tako



optimistična. Pogosto so jo uporabljali kot nevarno opozorilo ali Slika 1: Kreativa jumbo plakata iz leta 2022 neprijetna čustva (npr. kugo, karanteno, izdajalstvo; izmed

čustev pa zavist, ljubosumnost in skopost). Z rumeno v

kombinaciji s črno bo učinek kar »zbodel v oči«, sploh, če bo



sporočilo v daljavi (npr. plakat), bo pa asociacija pri ljudeh

verjetno »nevarno, strupeno« [3].

Modro povezujemo z daljavo in neskončnostjo, saj z

oddaljevanjem skoraj vsi predmeti dobijo modrikast odtenek, saj

jih postopoma prekrijejo zračne plasti [3]. Poleg oddaljenosti

simbolizira prosojna modrina še velikost, saj optično povečuje

predmete. Dodaja pa modra implicitno atribute božanskosti, saj

jo najdemo v simboliki kar nekaj religij. Vseeno ne smemo mimo

občutka hladnosti, ki jo vzbudi moder prostor, ljudje, ki jih

osvetlimo z modro barvo pa izgledajo bledi in bolehni [3].



5 NEPRAVILNA IN PRAVILNA UPORABA

BARV NA PRIMERU

Na tečaju grafičnega oblikovanja, so bila tečajnikom jasno

predstavljena prej zapisana teoretična izhodišča v povezavi z Slika 2: Zelo ne opazen jumbo plakat zaradi pomanjkanja

vplivom barv na čustva in vedenje uporabnikov. Predstavljen je nevpadljivih barv

bil tudi primer na osnovi izdelanega jumbo plakata.



Sliki 1 in 2 prikazujeta primer, ko pri oblikovanju jumbo

plakata niso bili dovolj upoštevani psihološki in simbolni učinki

barv, da bi dosegli učinkovito vizualno komunikacijo. To je

privedlo do zelo ne opaznega plakata. Zeleni grafični elementi so

se preveč zlili z okolico (travo in nebom), kar je zmanjšalo

njegovo vidnost in učinkovitost pri pritegnitvi pozornosti.

Nasprotno pa Sliki 3 in 4 prikazujeta primer uspešnega

plakata, kjer so bile barve izbrane in uporabljene strateško. Še

vedno sicer prevladuje zelena barva, ki je povezana z celostno

grafično podobo šole. Pristovnik pravi, da je celostna grafična

podoba sestavljena iz imena, simbola in/ali logotipa, tipografije,

barve, slogana in pogostokrat tudi drugih grafičnih elementov [4].

Dodana pa je tudi rdeča barva, ki pritegne pozornost in plakat

loči od pokrajine.

Končni jumbo plakat je bil vizualno privlačen in čustveno

učinkovit. Uporabljene barve so pomagale pritegniti pozornost,

komunicirati sporočilo in vzbuditi pozitivne občutke pri Slika 3: Kreativa jumbo plakata iz leta 2023

gledalcih.

Udeleženci tečaja so skozi ta proces pridobili praktično

znanje o pomenu in uporabi barv v grafičnem oblikovanju, kar

bo koristilo njihovim prihodnjim projektom.



udeležencev o strateški uporabi barv za učinkovito vizualno

komunikacijo.



Tabela 1: Ocene barv glede na čustveni vpliv in zaznavanje



Čustveni Jasnost

Barva Privlačnost

vpliv sporočila



Rdeča 4.3 4.1 4.7



Zelena 3.7 3.5 3.4



Modra 4.0 4.0 4.2



Rumena 3.4 3.8 3.9



Črna 3.2 3.3 3.7



Bela 3.1 3.2 3.3



Slika 4: Zelo opazen jumbo plakat zaradi pravilne izbire

barv Tabela 2: Ocene kombinacij barv glede na učinkovitost

jumbo plakatov



6 EMPIRIČNI DEL Kombinacija Privlačnost Jasnost sporočila

Za boljše razumevanje vpliva barv na zaznavanje in

Rdeča in Bela 4.4 4.8

učinkovitost jumbo plakatov, smo izvedli kratko anketo med

udeleženci tečaja. Cilj ankete je bil raziskati, kako so različne Modra in Rumena 4.1 4.3

barve povezane s čustveno reakcijo in zaznavno interpretacijo Zelena in Črna 3.3 3.4



uporabo na jumbo plakatih glede na mnenja udeležencev. Rumena in Črna 3.5 4.1 uporabnikov ter ugotoviti, katere barve so najbolj primerne za

Anketa je bila sestavljena iz vprašanj, povezanih z Modra in Bela 4.0 4.2

zaznavanjem barv in njihovo povezavo s čustvi in subjektivno

interpretacijo. Udeleženci so ocenjevali občutke in vrednotili

različne barve in kombinacije barv na lestvici od 1 (zelo 7 ZAKLJUČEK

negativno) do 5 (zelo pozitivno). Skupno je sodelovalo 15 Tečaj grafičnega oblikovanja, izveden v sklopu projekta, je

udeležencev, ki so izpolnili anketo ob zaključku tečaja. poudaril pomembnost razumevanja barv in njihovega vpliva na

Rezultati ankete glede na percepcijo barv so navedeni v tabeli vizualno komunikacijo. Barve igrajo ključno vlogo pri

1, rezultati glede na kombinacijo barv, pa v tabeli 2. oblikovanju in prenašanju sporočil, saj imajo močne psihološke

Glede na vrednosti v tabelah, bi lahko zapisali naslednjo in simbolične učinke, ki vplivajo na čustva in vedenje

interpretacijo rezultatov: uporabnikov. Razumevanje teh učinkov je nujno za ustvarjanje

 učinkovitih vizualnih sporočil, vendar je pogosto zapleteno in Rdeča: Udeleženci so rdečo barvo ocenili kot zelo

čustveno močno in privlačno. Verjetno izstopa zaradi zahteva globoko strokovno znanje.

svoje visoke vidljivosti in jasnosti sporočila. Glede na kompleksnost vplivov barv na čustva in vedenje

 uporabnikov je priporočljivo, da se grafični oblikovalci pri svojih Modra: Modra je bila ocenjena visoko zaradi svoje

umirjenosti in profesionalnega videza. Kombinacije projektih posvetujejo s psihologi. Psihologi lahko ponudijo

modre in bele so bile po mnenju udeležencev zelo dragocen vpogled v miselne in čustvene odzive, ki jih različne

učinkovite. barve lahko povzročijo pri občinstvu, ter pomagajo oblikovalcem

 izbrati barvne sheme, ki bodo najbolje komunicirale želena Zelena: Medtem ko zelena barva prinaša občutek

svobode in narave, je bila ocenjena nižje zaradi nizke sporočila in vzbudile pričakovane reakcije.



 vodi do bolj premišljenih in učinkovitih vizualnih rešitev. Rumena in Črna: Kombinacija teh dveh barv je bila Psihološko utemeljeno oblikovanje ne le izboljša estetsko ocenjena kot zelo učinkovita za jasnost sporočila, vendar zasičenosti in očitno manjše privlačnosti. Sodelovanje med grafičnimi oblikovalci in psihologi lahko



kombinacije očitno povezane s specifično čustveno reakcijo in Rezultati ankete kažejo, da so različne barve in barvne nekoliko nižje za privlačnost, mogoče tudi zaradi vrednost, ampak tudi poveča učinkovitost komunikacije in dosega večji vpliv na občinstvo. Takšen interdisciplinarni pristop asociacij na nevarnost. omogoča oblikovalcem, da ustvarjajo vizualna sporočila, ki ne le izstopajo, ampak tudi trajno vplivajo na gledalce.

zaznavno interpretacijo uporabnikov. Barve, ki so po mnenju

udeležencev raziskave najboljša izbira za jumbo plakate, so bile

rdeča in modra, zlasti v kombinaciji z belo. Ugotovitve ankete

bodo uporabljene za nadaljnje izboljšanje vsebinskega dela

tečaja grafičnega oblikovanja, z namenom izobraževanja





LITERATURA IN VIRI


[1] U. Žalig. 2021. Vpliv grafičnega oblikovanja plakatov na opazovalce,

Diplomsko delo, Univerza v Mariboru, Fakulteta za elektrotehniko,

računalništvo in informatiko, Maribor.

[2] M. Krmelj. 2015. Prikrito oglaševanje, magistrsko delo, Univerza v

Mariboru, Ekonomsko-poslovna fakulteta, Maribor.

[3] A. N. Kovačev. 1997. Govorica barv, Prešernova družba, Ljubljana.

[4] G. Pristovnik. 2020. Lastna celostna grafična podoba, Diplomsko delo,

Univerza v Mariboru, Fakulteta za elektrotehniko, računalništvo in

informatiko, Maribor.



S kvizom spoznajmo živali na travniku



Let's Get to Know the Animals on the Meadow with a Quiz



Tina Šebenik

Župnijski vrtec Vrhnika

Voljčeva 21

1360 Vrhnika

malatiny@gmail.com



computer in the past months. Thus, in our group, we also devoted

POVZETEK an important place to the field of ICT technology and spread

V prispevku je predstavljeno, kako smo v skupini 3 letnikov knowledge and its meaningful applicability. We focused on

z uporabo IKT- tehnologije, popestrili vzgojno delo in otroke Google Chrome, where we looked at animals that live in the

usmerili v nov učni pripomoček, ki jih je motiviral in umiril. meadow and that children know. We played some songs of these

Otroci so že v preteklih mesecih pokazali veliko zanimanja za animals (butterflies, moles, ants, snails and ladybugs) on the

uporabo računalnika. Tako smo tudi v naši skupini na področju YouTube channel and played them several times, because

IKT- tehnologije namenili pomembno mesto in širili znanje, ter children love to listen to them. I prepared a short quiz for the

njeno smiselno uporabnost. Osredotočili smo se na Google children with questions about meadow animals, riddles and a

Crome, kjer smo si pogledali živali, ki živijo na travniku in jih memory game. First, they answered 6 questions about these

otroci poznajo. Na Youtube kanalu smo si zavrteli nekaj pesmi animals, then they answered 3 riddles, showed where individual

teh živali ( metulja, krta, mravlje, polža in pikapolonice) in jih animals live and ended with a memory. The questions were

večkrat predvajali, saj jih otroci zelo radi poslušajo. Otrokom prerecorded. We opened PowerPoint and indulged in a quiz that

sem pripravila kratek kviz z vprašanji o travniških živalih, took us to the end with a few correct clicks. The mentioned

uganke in igro spomin. Najprej so odgovorili na 6 vprašanj o teh didactic aid motivated the children, they were more collected and

živalih, potem so odgovorili na 3 uganke, pokazali kje focused on the tasks that awaited them. At first we solved them

posamezne živali živijo in zaključili s spominom. Vprašanja so together, but then the work was done individually with each

bila vnaprej posneta. Odprli smo PowerPoint in se prepustili individual. The teachers and I were present the whole time, as the

kvizu, ki nas je z nekaj pravilnih klikov pripeljal do konca. children were 3 years old and needed help handing ICT tools.

Omenjen didaktični pripomoček je otroke motiviral, bili so When solving the quiz, they listened to the teacher’s questions,

bolj zbrani in osredotočeni na naloge, ki so jih čakale. Sprva smo followed the sliders, gave the correct answer or pointed their

jih reševali skupaj, potem pa je delo potekalo individualno z finger at the computer screen where the answer to the question

vsakim posameznikom. Vzgojiteljici sva bili ves čas prisotni, saj was. They didn’t get tired of the rescue. We carried out activities

so bili otroci stari 3 leta in so potrebovali pomoč pri rokovanju z throughout the week. We also added their popular hero Marshall,

ikt orodjem. Pri reševanju kviza so poslušali vprašanja who appears in the Paw Patrol cartoon, to the quiz. He guided

vzgojiteljice, spremljali drsnike, podali pravilen odgovor ali s them through all the questions and further encouraged them to

prstom pokazali na računalniški ekran, kjer je bil odgovor na make the solution easier.

vprašanje. Reševanja se niso naveličali. Dejavnosti smo izvajali We were aware that ICT is part of the everyday environment

cel teden. V kviz smo dodali tudi njihovega priljubljenega junaka of preschool children, which makes sense to use in the

Marshalla, ki nastopa v risanki Tačke na patrulji. On jih je vodil educational process to teach them about its didactic value and

skozi vsa vprašanja in jih dodatno spodbudil, da je bilo reševanje take advantage of the popularity of electronic devices.

lažje.

Zavedali smo se, da je IKT del vsakdanjega okolja tudi KEYWORDS

predšolskih otrok, ki ga je smiselno uporabiti v učno-vzgojnem Learning through play, preschool period, ICT, collaboration

procesu in jih podučiti o njeni didaktični vrednosti in izkoristiti

priljubljenost elektronskih naprav.

1 UVOD

KLJUČNE BESEDE Sodobna tehnologija na vsakem koraku kroji naš vsakdan,

Učenje skozi igro, predšolsko obdobje, IKT, sodelovanje tako v naših službenih okoljih kot zasebnem življenju.

Posledično je tudi vsaka generacija naših najmlajših, ki jih vleče



ICT technology to diversify the educational work and direct the The article presents how, in a group of 3-year-olds, we used pri rokovanju s t. i. pametnimi napravami. Za uporabo moderne tehnologije so zelo motivirani, saj jih privlačijo številni ABSTRACT zgled staršev, starejših sorojencev, vzgojiteljev, vse bolj spretna

children to a new learning aid, which motivated and calmed them. multisenzorični dražljaji.

The children have already shown a lot of interest in using the

Digitalni mediji so kulturno orodje, ki močno vplivajo na naš

način razmišljanja. Delujejo kot okoljska spodbuda, ki oblikuje

našo možgansko strukturo. Preko računalnika otroci prejemajo

različne nasprotujoče iztočnice, ki vplivajo na njihovo vedenje in

posledično na razvoj njihovih možganov. Če digitalno

tehnologijo uporabljamo zmerno, le-ta otrokom zelo koristi.

Pozitivno vpliva na fizični, kognitivni in čustveni razvoj otroka

[1].

V kurikulumu za vrtce je navedeno, da se digitalna

tehnologija vedno bolj uporablja na različnih področjih in okoljih

(doma, v vrtcu…) in za različne namene (zabavo, učenje…).

Slika 2: Vprašanje o živalih

Predšolski otrok se dnevno srečuje z IK sredstvi (opazuje,

posluša odrasle, rokuje…), zato sem se kot vzgojiteljica odločila,

da otrokom ponudim različne možnosti in izzive (otroke učimo

uporabljati IK sredstva, jim približamo internet) in tako

ozaveščamo svoj odnos do nje [2].



2 POTEK DELA V SKUPINI



2.1 Uvodna motivacija

Z otroki smo se zbrali pred računalniškim zaslonom in si

preko Google Croma pogledali živali, ki živijo na travniku. Sami

so predlagali, katero žival naj poiščemo. Poiskali smo polža, Slika 3: Uganka

mravljo, pikapolonico, krta in metulja. Otroci so navdušeno

opazovali ekran in miško in ugotovili, da lahko na računalniku Po rešenih ugankah so otroci povedali, kje domuje čebela,

poiščemo vse, kar nas zanima. Tako smo odprli še YouTube mravlja in krt. Na koncu pa so se preizkusili še s kvizom. Vsem

kanal in predvajali pesmi o izbranih živalih. Izbire je bilo veliko. otrokom so bila vprašanja jasna in odgovori so bili pravilni.

Nekatere pesmi smo predvajali večkrat. Junak Marshall jim je bil posebno všeč (Slika 4). Motiviral je

tudi ostale, da so lažje sedeli na stolu in sledili celotnemu kvizu.

2.2 Glavni del dejavnosti- reševanje kviza Z vsakim dnem jim je šlo lažje. Postali so bolj samozavestni.

Tudi otroci, ki v prvih dneh niso bili dovolj aktivni, so postali

Osredotočili smo se na glavno dejavnost in začeli z

bolj sproščeni in so se lahko vključili med ostale otroke in

reševanjem kviza (slika 1). Sledila so vprašanja (slika 2): Klikni

sodelovali. Za njih je bilo to nekaj novega, neznanega. Učenje

na žival in poslušaj kako se imenuje, klikni na žival, ki ni

preko igre jih je navdušilo.

travniška, koliko travniških živali vidiš, katera žival prinaša

srečo, katera žival prideluje med, katera žival nosi vedno hiško s

seboj?

Za tem so sledile uganke (slika 3):

Ko je majhen, je buba, ko odraste, leti, na rožah poseda, poleti

živi?

Iz smrečja igel dom imajo, nič iger,- delo le poznajo?

V črni zemlji dan za dnem, rije varno vsem očem, če pokuka,

kje na plan, s kupčkom zemlje je izdan?





Slika 4: Otroci si ogledujejo Marshalla, ki nastopa v kvizu





3 ZAKLJUČEK



Projektno delo s kvizom spoznajmo živali na travniku, se je



izkazalo kot uspešen primer dobre prakse. Vsekakor je omenjeni

didaktični pristop popestril naše vzgojno delo. Otroci so se v

Slika 1: Reševanje kviza vrtcu prvič srečali z IKT-tehnologijo in vsekakor ne zadnjič.

Otroci so si s tem razširili splošno razgledanost, logično

mišljenje, ter krepili spomin. Vse to, jim lahko v prihodnje

pomaga in zmanjša ovire pri socialni interakciji. Tudi

najmlajšim tako zagotovimo vključitev v porajajočo se digitalno

pismenost in jim zagotovimo enake možnosti in zmanjšamo

razlike med njimi. Računalnik je močno motivacijsko sredstvo,

ki ne potrebuje posebnih stimulacij za delo. Otroci so videli, da

nam naprava ponuja veliko možnosti. Ob njej se lahko zabavamo,

sprostimo in kaj novega naučimo.





Slika 5: Igra spomin na koncu kviza



LITERATURA IN VIRI

[1] Gold J. 2015. Vzgoja v digitalni dobi. Ljubljana: Založba Didakta.

[2] Usar K., Jerše L. 2016. Smernice za vključevanje IKT v vrtcu. Ljubljana:

ZRSŠ.



E-branje in e-knjige: posebnosti in izzivi v šolski praksi



E-reading and E-books: Specifics and Challenges in School



Practice



Dr. Maja Šebjanič Oražem

Osnovna šola Puconci

Puconci 178

9201 Puconci

maja.sebjanic@gmail.com



pretvarjanja v glas, tj. izgovorjave. Gre za človeku specifično

POVZETEK dejavnost, ki nam ni dana, ampak se je (na)učimo. Da postane

V prispevku predstavljamo e-knjige in e-branje v šolski avtomatiziran proces, je potrebno (dovolj) vaje. Kot tako

praksi. Osredotočamo se na posebnosti in izzive, ki se porajajo predstavlja najučinkovitejše sredstvo usvajanja znanja. Učenje

pri njihovem vključevanju v pouk slovenščine. Povzemamo branja poteka stopenjsko (več o tem v [18]).

nekatere raziskave e-branja, njegove specifike in možnosti, ki jih Če se osredotočimo na namene prispevka, bomo v



nudi za bralno motivacijo pri današnjih šolajočih se generacijah, nadaljevanju s pojmom “branje” zajeli opredelitev, uveljavljeno 1 v slovenistični stroke. Ta branje razume kot temeljno dejavnost, ki so v digitalni svet vpete praktično na vseh korakih.

ki vodi k bralni pismenosti [5].



KLJUČNE BESEDE

1.2 Kratek pregled raziskav e-branja

E-knjiga, e-branje, pouk slovenščine, motivacija za branje,

digitalna generacija V slovenskem prostoru je relativno malo raziskav, ki bi se

osredotočale na raziskovanje razlik med “klasičnim” branjem, tj.



ABSTRACT branjem s papirja, in med “digitalnim” oz. e-branjem. Med



In this paper we present e-books and e-reading in school 2 je generacija Y raziskavami izstopa S. Pečjak [20], ki s sodelavko ugotavlja, da pri e-branju nekoliko hitrejša od branja s papirja,

practice. We focus on the specifics and challenges that arise,

kar pripisuje verjetnemu pogostemu pregledovanju družabnih

when we include them in Slovenian language lessons. We

omrežij, brskanju po spletu in branju informativnih vsebin, pri

summarize some research of e-reading, its specifics and the

čemer zaradi same narave tovrstnih besedil redkeje uporabljajo

possibilities, that e-books offer for reading motivation among

metakognitivne strategije in torej manj preverjajo svoje

today's school-age generations, who are embedded in the digital

razumevanje med samim branjem. Posledično lahko z

world at practically every step.

računalnika berejo hitreje kot s papirja, vendar manj natančno in



KEYWORDS več časa za naloge bralnega razumevanja in odgovarjanjem na imajo zato več težav s ponovnim iskanjem informacij, porabijo

E-book, e-reading, Slovene language lesson, reading vprašanja. Glede razumevanja prebranega pa med tistimi, ki so

motivation, digital generation brali s papirja, in tistimi, ki so brali na zaslonu, ni pomembnih

razlik [20].



1 Za primerjavo z rezultati slovenske raziskave naj navedemo BRANJE IN E-BRANJE: PODOBNOSTI IN



RAZLIKE Mangen [4]. Ugotovili so, da so učenci, ki so brali besedila s še primer raziskave na Norveškem, ki jo je s sodelavci izvedla E.



1.1 papirja, dosegli boljše rezultate pri nalogah bralnega Opredelitev (e-)branja

razumevanja od tistih, ki so brali besedila na zaslonih. V tem

Pri strokovnih (in znanstvenih) razpravah o (e-)branju je segmentu se raziskavi med seboj razlikujeta.

vedno potrebno vzeti v obzir čas in s tem družbene okoliščine, A. Kepic Mohar [1] pa ugotavlja, da je proces pridobivanja

znotraj katerih se pogovarjamo o branju. Prav tako je potrebno znanja in s tem celotnega razvoja posameznika, za katerega je v

jasno definirati, o katerem segmentu branja oz. stopnji razvoja preteklosti veljalo, da potekalo (s knjigami in) z branjem,

bralne zmožnosti razpravljamo. dandanes le še pot do pragmatičnega cilja, torej orodje za lažji

Če se dotaknemo opredelitve branja, gre za proces vstop na trg delovne sile. Novi digitalni mediji, ki praktično

dekodiranja/prepoznavanja zapisanih črk in njihovega vsakodnevno vstopajo v vse sfere našega življenja, vplivajo in



1 2 V prispevku se bomo osredotočili na branje digitalnega bralnega gradiva in Tudi t. i. digitalna generacija, ki je odraščala v drugačnih okoliščinah in je

ga mestoma primerjali s (klasičnim) tiskanim. Stroka pa poleg teh dveh med vrste bolj vajena branja z računalniškega zaslona, zato ima tudi najbrž temu primerno

bralnih gradiv uvršča še zvočne in interaktivne (za več glej razvojni projekt Bralna razvite bralne strategije [20]. pismenost in razvoj slovenščine – OBJEM (Ozaveščanje, Branje, Jezik, Evalvacije,

Modeli), ki je potekal pod okriljem Zavoda Republike Slovenije za šolstvo).



spreminjajo oblike sporazumevanja/komunikacije, jezik, ki ga knjige, ki v zadnjih letih (zlasti po epidemiji covida) vedno bolj

pri tem uporabljamo, značilnosti pisanja in podobno, nenazadnje prodirajo v vsakdanje življenje in šolsko prakso, zato je prav, da

pa tudi naše zmožnosti sprejemanja besedil različnih vrst, kamor nekaj mesta namenimo tudi razmejitvi teh dveh pojmovanj.

spada branje.

2.1 Opredelitev e-knjige

1.3 Posebnosti branja elektronskih besedil Slovar slovenskega knjižnega jezika (dostopen na fran.si v

Digitalizacija je spremenila naš vsakdan in močno posegla nadaljevanju SSKJ) opredeljuje e-knjigo oz. elektronsko knjigo

tudi v šolstvo. Ne le da je omogočila hiter dostop do velike kot knjigo v elektronski obliki, namenjeno zlasti za branje na e-

količine informacij, spremenila je tudi naš način razmišljanja ter bralniku ali drugih elektronskih napravah. Kot sopomenko e-

branja in posledično učenja, če ostajamo na predpostavki, da je knjigi najdemo tudi pojem digitalna knjiga (prim. [24]). Tukaj

branje (še vedno) najučinkovitejši način usvajanja znanja, čeprav najdemo še nekoliko obširnejšo opredelitev, saj gre za

se učenci več ne učijo samo s papirja, ampak tudi z zaslona.3 enciklopedičen članek, in sicer da je to knjižna publikacija,

Da bi lahko razumeli posebnosti branja elektronskih besedil izdana v digitalni obliki. Lahko je izdana samostojno ali kot

in jih upoštevali v svoji praksi, je potrebno poznati značilnosti različica tiskane knjige. Namesto v fizični obliki do nje

digitalnih virov. Ti imajo dve temeljni značilnosti: dostopamo s pomočjo računalniške naprave in se izpisuje na

zaslonu (za primerjavo glej [6]).

(1) nematerialnost oz. neoprijemljivost in

E-knjiga je torej prvenstveno namenjena branju (in

(2) večzaznavno naravo interakcij.

posledično v šolski praksi bralnemu razumevanju).

Učenec torej pri uporabi digitalnih virov nima konkretnih oz.



oprijemljivih predmetov za manipulacijo (prijemanje, 2.2 Opredelitev zvočne knjige premikanje …), poleg tega pa je v digitalnem okolju prekinjena

prostorsko-časovna povezanost med aktivnostjo učenca in V SSKJ ne najdemo zadetka pod iztočnico “zvočna”, ko

njegovim želimo najti definicijo zvočne knjige, prav tako ne pod iztočnico zaznavanjem. Digitalna/elektronska besedila

zmanjšujejo eno ključnih navigacijskih sposobnosti bralca, tj. “knjiga”. Ponujena nam je razlaga pod iztočnico “avdioknjiga”

nadzor nad branjem, ki je ena od osrednjih regulacijskih in je v SSKJ opredeljena kot “obsežnejše literarno besedilo, ki je

sposobnosti [19]. posneto na nosilec zvoka, navadno na CD-ploščo”. Nadalje jo

A. Kepic Mohar [1] (po Kovač in van der Well 2018) najdemo še pod iztočnico “zvočnica”, ki jo SSKJ opredeljuje kot:

povzema tri načine branja, ki so posledica digitalizacije in ki so (1) “zvočni posnetek zlasti literarnega besedila, ki ga

povezani z dolžino besedil ter medijem, na katerem beremo: navadno berejo profesionalni bralci, igralci” oz. kot

(1) (2) “knjig[o] za otroke, ki omogoča zaznavanje branje s preletom (angl. skimming );

(2) predstavljene vsebine (tudi) s sluhom”. vživeto branje (angl. immersive/absorbed reading ), ki

poteka pri branju žanrskega leposlovja; Iz slovarskega sestavka izvemo še, da je pojem “zvočnica”



(3) nastal iz “zvočne knjige” po zgledu angleškega izraza “audio poglobljeno/globoko branje (angl. in-depth/deep reading book” in da se zanj uporabljata še sinonimna pojma ), ki pa je kognitivno zahtevnejše oz. študijsko 5 branje in vključuje predznanje.

4 “avdioknjiga” in “zvočna knjiga”.

Da je e-branje drugačno od branja s papirja, je pokazala tudi Zvočna knjiga/zvočnica je torej knjiga, namenjena poslušanju

mednarodna raziskava (posledično v šolski praksi imenovanemu slušnemu Spretnosti odraslih (PIAAC), kakor

povzema njene rezultate P. Javrh [15]. Že sama narava razumevanju).

elektronskih besedil narekuje “drugačno” branje, je potrdila

raziskava. Poleg tega pa je za učinkovito obdelavo informacij v 2.3 Kaj pa (e-)bralnik?

digitalnem okolju predpogoj dobro usvojena bralna E-bralnik (pogosto tudi elektronski papir ali e-papir) je

tehnika/veščina. Ta je namreč potrebna, da lahko pri e-branju in naprava, zelo podobna tabličnemu računalniku, namenjena

obdelavi e-gradiv stečejo zahtevnejši kognitivni procesi. branju e-knjig. Zanj je značilen zaslon e-ink oz. zaslon s črnilom.

“Zaslon s črnilom je vrsta zaslonske tehnologije, ki natančno

posnema videz knjige. E-ink zaslon je tudi zelo varčen in manj

2 E-KNJIGA IN ZVOČNA KNJIGA: ISTA obremenjuje oči.” [10] (več o tem tudi [11])

STVAR, RAZLIČNO POIMENOVANJE?



V šolski praksi se pogosto dogaja, da se uporabljata dva 3 APLIKACIJE, NAMENJENE BRANJU E-

različna izraza za isto stvar, kar vnaša marsikatero zmedo in

nejasnost. Tudi kar zadeva e-branje (ali elektronsko branje), ki se KNJIG

odvija v e-knjigah (ali elektronskih izdajah knjig), se dogajajo V nadaljevanju navajamo pregled aplikacij, namenjenih

(napačne) interference z drugim pojmom. Gre za t. i. zvočne branju e-knjig. Pri tem se osredotočamo na e-knjige v



3 Za lažje razumevanje posebnosti e-branja (tudi digitalnega branja, kakor se rezultate raziskav, ki so pokazale, da „digitalne” generacije nimajo odpora do branja

pogosto imenuje v slovenistični stroki, primer [2]), na kratko po S. Pečjak [19] z zaslona, a lažje razumejo slišano/prebrano in prikličejo bistvo prebranega, če

povzemamo nekaj značilnosti generacij, ki so v času nastajanja prispevka vključene poslušajo oz. preberejo na papirju kot v e-obliki. v sistem formalnega izobraževanja. Gre za dve generaciji, in sicer za (1) generacijo 4 Na podlagi nove informacije si širimo besedni zaklad in s kognitivnimi Z (rojeni v letih 1995–2010), imenovano tudi digitalni pionirji, in (2) generacijo procesi gradimo znanje [1] (po Kovač in van der Well 2018).

alfa (rojeni po letu 2010), ki jo imenujejo tudi digitalni domorodci. »Za obe 5 Za več glej [12]. generaciji je značilno, da sta v svojem življenju doživeli največ družbenih

sprememb od vseh generacij pred njima, najbolj pa ju je zaznamovala digitalizacija

življenj, tj. povsod navzoča prisotnost tehnologije« (62) Nadalje avtorica povzema

slovenskem jeziku, saj želimo osvetliti priložnosti in ovire pri zmožnost literarnega branja kot specifično podvrsto

njihovi uporabi pri pouku slovenščine 6 (kot prvega oz. sporazumevalne zmožnosti ter se usposabljajo za branje in

maternega jezika): interpretacijo literarnih besedil [22] (7).

Eden od gradnikov bralne pismenosti je motiviranost za

(1) KOBI je aplikacija, namenjena začetnemu branju, tj.

branje, ki zajema interes za branje, pozitiven odnos do branja

učenju in treningu branja (bralne tehnike). Uporaba

različnih vrst besedil (za različne namene, in sicer branje za

aplikacije je brezplačna. Ker je namenjena učenju branja

znanje in branje leposlovja oz. razvojanje bralne kulture) ter

in je torej primerna za prvo triletje osnovne šole, je

izbiranje tiskanih in drugih informacijskih virov glede na namen

podrobneje ne bomo obravnavali.

branja v povezavi z uporabo različnih lokacij bralnih virov. Drug

(2) Biblos, spletna platforma, je prva slovenska e-knjižnica

(za naš prispevek relevanten) gradnik je še tekoče branje, ki med

(in e-knjigarna), nastala leta 2013 pod okriljem

drugim zajema tehniko branja (in njeno prilagajanje zahtevam

Beletrine, zavoda za za založniško dejavnost, in s

besedila).

sodelovanjem mreže slovenskih knjižnic ter inštitutom

IZUM, skrbnikom sistema COBISS [13]. Prvenstveno je 3.1.2 Zbirka e-knjig v slovenščini

bila namenjena izposoji e-knjig, kasneje tudi njihovemu Razpoložljiva baza e-knjig v slovenskem jeziku zajema zlasti

nakupu. Za uporabnike, člane splošnih knjižnic, je bila 8 literarna/umetnostna besedila, kot je razvidno s Slike 1 in ki jih

izposoja brezplačna. S 1. 6. 2024 jo je za člane splošnih obravnavamo pri pouku književnosti. Ker gre za branje literarnih

knjižnic nadomestila platforma COBISS Ela (več o njej besedil, to vrsto branja v slovenistični stroke imenujemo

v nadaljevanju). Beletrina ima od takrat naprej novo literarno branje9 (prim. [5, 3]).

platformo, Beletrina Digital, ki ob plačilu naročnine



nudi izposojo in nakup e-knjig, zvočnih knjig, filmov,

podkastov in prispevkov.

(3) sBiblos je Beletrinina spletna platforma, namenjena

izposoji knjig za učence zadnjega triletja osnovne šole in

srednješolce. Uporaba je zanje brezplačna, prijava je

mogoča z AAI računom.

(4) Mladinska knjiga PLUS je aplikacija založbe Mladinska

knjiga, ki ob plačilu izbranega paketa nudi različne

digitalne vsebine (e-knjige, zvočnice, animirane

pravljice in podkaste ter drugo, glede na izbran paket).

(5) COBISS Ela je platforma, ki nudi članom splošnih

knjižnic ob namestitvi ustrezne programske opreme

izposojo e-knjig.



3.1 Za šolsko prakso uporabne aplikacije



Pregled aplikacij za e-branje, ki so uporabne v šolski praksi, Slika 1: Tipologija e-knjig v COBISS Eli (zajem zaslona)

bomo naredili na podlagi temeljnih ciljev pouka književnosti, ki

jih opredeljuje učni načrt za osnovno šolo v zadnjem triletju (7.– Tudi baza e-knjig, ki so brezplačno na voljo za učence in

9. razred) in učni načrt za gimnazijo.7 Kot podlago smo vzeli še dijake v platformi sBiblos zajema literarna besedila, kot je

gradnike bralne pismenosti projekta OBJEM [8], ki se nanašajo razvidno s Slike 2.

na branje in bralno motivacijo.





3.1.1 Cilji pouka slovenščine (kot prvega oz. maternega

jezika)

Naj na kratko zaradi lažjega razumevanja in za argumentacijo

izbora povzamemo temeljne cilje pouka slovenščine, ki se

nanašajo (tudi) na (motivacijo za) branje.

Učenci in dijaki pri pouku slovenščine razvijajo

sporazumevalno zmožnost v slovenskem (knjižnem) jeziku, torej

zmožnost kritičnega sprejemanja (in tvorjenja) besedil različnih

vrst. Branje (zlasti v okviru pouka književnosti) prepoznavajo

kot užitek, prijetno doživetje in intelektualni izziv. “Pri

razvijanju sporazumevalne zmožnosti ob dejavnem stiku z

neumetnostnimi in umetnostnimi besedili z uporabo digitalne

tehnologije varno, ustvarjalno in kritično pridobivajo ter Slika 2: Tipologija e-knjig v sBiblosu (zajem zaslona)

uporabljajo podatke/informacije.” [21] (7) V gimnaziji razvijajo



6 8 Osredotočili se bomo na pouk slovenščine od 6. do 9. razreda osnovne šole in Podatke sicer navajamo za platformo COBISS Ela, vendar podobno velja tudi

pouk slovenščine v srednješolskem izobraževanju. za Biblos v začetku leta 2017 [13]: razmerje gre v prid leposlovja, in sicer je

7 Na ostale programe srednješolskega izobraževanja se nismo posebej predstavljajo 71 % vseh razpoložljivih e-knjig, strokovna literatura pa 29 %.

osredotočali, saj gre v večini za nadaljevanje in manjšo nadgradnjo osnovnošolskih 9 V stroki pojmujemo literarno branje kot najbolj kompleksno branje, ki nudi ciljev. največ možnosti oblikovanja vsestranskega bralca [5, 3].



E-knjige v slovenskem jeziku najdemo tudi na slovenskem 3.2.3 Čemu v šolsko prakso vključiti e-knjige?

Wikiviru [23] (in drugih spletnih mestih), kjer so digitalizirana Nanizali bomo zgolj nekaj argumentov, zakaj v šolsko prakso

dela slovenskih avtorjev različnih obdobij slovenske književnosti (občasno) vključiti e-knjige.

(prim. [14, 25]). Ker gre za postavitev literranih besedil na splet, Prvič se nam zdi takšen način zelo pragmatičen, saj

do katerih ne dostopamo preko aplikacije, jih v našem prispevku potrebujemo malo opreme, zadostuje že kakšen (pametni) telefon

ne bomo podrobneje obravnavali. ali tablični računalnik in lahko manjša skupina (predlagamo do

4) učencev/dijakov bere skupaj eno e-knjigo. Na ta način

3.2 Možnost uporabe izbranih aplikacij pri pokrijemo tako sodelovalno učenje, razvijanje digitalnih

pouku kompetenc učencev in še hkrati jih motiviramo za branje ter

Pri predstavitvi možnosti uporabe aplikacij za branje e-knjig pogovarjanje o prebranem (izmenjavo mnenj, različnih

smo se odločili za vključitev tistih, ki nudijo brezplačno uporabo interpretacij prebranega, razjasnjevanje manj jasnih mest ipd.).

za učence in dijake, saj na ta način omogočimo vsem učencem Na ta način lahko pokrijemo tudi skupno daljše branje izbranih

enake možnosti dostopa. To sta dve, in sicer COBISS Ela in literarnih del po delih z učencem zanimivo/privlačno tematiko.

sBiblos. Podoben način dela lahko uporabimo tudi za (obvezno)

domače branje, ki nemalokrat učencev ne pritegne in se ga

3.2.1 COBISS Ela izogibajo ali pa površno preberejo (pogosto zgolj obnove na

Platforma COBISS Ela na spletni strani [7] nudi pregledna spletu). Digitalno tehnologijo lahko v tem primeru uporabimo

navodila, katere naprave in aplikacije za branje podpirajo tudi za refleksijo o prebranem in analizo prebranega, in sicer v

izposojo in branje e-knjig. Kot prednost platforme gre izpostaviti, obliki e-dnevnika, lastne e-knjige, e-diagrama ali e-table. Ob tem

da lahko e-knjige beremo na mobilnih napravah, ki temeljijo na je potrebno učencem le še dati nekaj usmeritev/navodil, kako se

Androidu in tudi na sistemu iOS. Zraven je navedeno tudi ime lotiti dela.

aplikacije, ki si jo je potrebno namestiti. Navodilom so dodani Vključevanje e-knjig v šolsko prakso pa se vidimo kot

tudi e-vodiči za posamezne korake. koristno tudi za možnost večjega poudarka na lastni izbiri

E-knjige lahko beremo tudi na osebnih računalnikih ter literarnega dela v obliki e-knjige za branje, saj lahko učenec

bralnikih e-knjig in bralnikih inkBOOK. Preprost je tudi način zlahka dostopa do druge e-knjige, če ga tematika ene ne pritegne.

iskanja želene e-knjige, saj je brskalnik enak iskalniku tiskanih To zagotovo deluje zelo motivacijsko pri današnjih generacijah,

knjig v knjižnicah. To je zagotovo velika prednost, saj ni ki so vešče dela z digitalnimi viri.

potrebnega “novega učenja” iskanja želene e-knjige.

V bazi e-knjig najdemo večino sodobnih slovenskih SKLEP



nagovarjajo z njim aktualnimi vsebinami/temami. E-knjigo si negativnega, vendar je potrebno pristopiti v smeri iskanja lahko izposodimo na dveh različnih napravah, na voljo pa nam je priložnosti in pozitivnih vidikov. Šole in učiteljev računalniki (in mladinskih romanov, ki najbolj motivirajo mlade bralce, saj jih Digitalizacija v šolstvu je pogosto videna kot nekaj

21 dni od dneva izposoje. Naenkrat si lahko uporabnik izposodi naraščajoča uporaba umetne ingteligence) ne bodo mogli

5 e-knjig. To je lahko uporabno za učence, saj lahko e-knjige nadomestiti, lahko pa učitelji tehnologijo izkoristimo sebi in

berejo na dveh različnih napravah, na eni npr. v šoli in na eni namenom pouka v prid. Glede na to, da so naši učenci in dijaki

doma. Prav tako lahko hkrati berejo več različnih knjig, jih med “digitalin pionirji” in “digitalin domorodci”, je potrebno

seboj primerjajo (iščejo podobnosti in razlike, analizirajo didaktične pristope prilagajati njim in značilnostim njihovega

posamezne motive ipd.). učenja. S tem vsekakor ne mislimo na popolno opustitev

3.2.2 sBiblos “tradicionalnih” pristopov in načinov dela, ampak tehnologijo

Beletrinin sBiblos prav tako deluje na mobilnih napravah z izkoristiti za motivacijo tistih področij/dejavnosti, ki jih

operacijskima sistemoma Android in iOS. Na spletni strani [9] zaznavamo kot šibkejše oz. manj učinkovite.

so podrobna navodila, kako po korakih namestiti potrebne V prispevku smo v skladu s tem namenom želeli prikazati,

aplikacije za dostop do e-knjig. Zelo uporabni so tudi e-vodiči in kako en segment pri pouku slovenščine, tj. motivacija za branje

slikovno gradivo, ki pomaga pri nameščanju in uporabi bralnika. (morda tudi domače ali vsaj skupno dolgo branje v razredu)

E-knjige lahko uporabniki berejo na osebnih računalnikih, skušati razvijati skladno z značilnostmi šolajočih generacij, da bo

mobilnih napravah in e-bralnikih. čim učinkoviteje in da bodo uspeli uresničiti (vsaj del) ciljev

Do storitev sBiblos lahko učenci dostopajo z AAI pouka.

uporabniškimi podatki, v kolikor ima šola urejen dostop do

storitve Moja Knjižnica in podpisano pogodbo o obdelavi VIRI IN LITERATURA

osebnih podatkov z IZUM-om. [1] Alenka Kepic Mohar. 2021. Nevidna moč knjig: branje in učenje v

V zbirki e-knjig najdemo večino literarnih del, ki se digitalni dobi. Zbirka Bralna znamenja. Cankarjeva založba, Ljubljana.

[2] Alenka Žbogar. 2022. Branje kot potovanje. Branje v slovenskem jeziku,

obravnavajo v šoli in tudi novejša dela z učencem in dijakom literaturi in kulturi. 58. seminar slovenskega jezika, literature in kulture.

privlačnimi temami. Tudi sBiblosovi zbirki pedstavljajo navečji Založba Univerze v Ljubljani, Ljubljana. 5–6. DOI:

https://doi.org/10.4312/SSJLK.58.2386-058X.

del literarnega opusa romani in kratka proza. Učenci in dijaki si [3] Alojzija Zupan Sosič. 2014. Literarno branje. Jezik in slovstvo 59, 4. 47–

lahko naenkrat izposodijo 3 e-knjige, na mesec 10 e-knjig, vsaka 65.

knjiga pa jim je na voljo 30 dni od dneva izposoje. V primerjavi [4] Anne Mangen idr. 2013. Reading linear texts on paper versus computer

screen: Effects on reading comprehension. International Journal od

s COBISS Elo je čas izposoje e-knjig uporabnikom prijaznejši, Educational Research 58. 61–68. DOI:

je pa na voljo manj hkrati izposojenih knjig. https://doi.org/10.1016/j.ijer.2012.12.002.

[5] Boža Krakar Vogel. 2020. Didaktika književnosti pri pouku slovenščine.

Zbirka Učiteljeva orodja. Založba Rokus Klett, Ljubljana.



[6] Britannica: e-book. DOI: https://www.britannica.com/technology/e-book.

[7] COBISS Ela. DOI: https://www.cobiss.si/bralci/ela/#2.

[8] Gradniki bralne pismenosti: teoretična izhodišča. 2020. Univerzitetna

založba Univerze v Mariboru, Maribor. DOI:

https://press.um.si/index.php/ump/catalog/book/515.

[9] Kako deluje sBiblos? DOI: https://www.sbiblos.si/kako-deluje.

[10] Kindle ali Kobo? Kateri e-bralnik izbrati? DOI: https://racunalniske-

novice.com/e-bralniki-amazon-kindle-e-bralniki-kobo-e-knjige/.

[11] Marjan Kodelja. 2023. Zakaj in kako izbrati e-bralnik? DOI:

https://vsebovredu.triglav.si/tehnologija/zakaj-in-kako-izbrati-e-bralnik.

[12] Marjan Povše. 2022. Zvočne knjige v slovenskem jeziku. DOI:

https://blog.cobiss.si/2022/07/04/zvocne-knjige-v-slovenskem-jeziku/.

[13] Matic Razpet. 2017. Biblos. E-gradiva kot bližnjica do uspeha? E-zbornik

Bralnega društva Slovenije od 12. strokovnem posvetovanju 8. septembra

2017. Bralno društvo Slovenije, Ljubljana. 71–76. DOI:

https://www.bralno-drustvo.si/wp-content/uploads/2013/06/E-

GRADIVA_KOT_BLIZNJICA_DO_USPEHA_BDS20171.pdf.

[14] Miran Hladnik. 2009. Za elektronsko knjigo. Šolska knjižnica 19, 2–3.

126–133. DOI: https://lit.ijs.si/zaeknjigo.html.

[15] Petra Javrh. 2017. V digitalnem svetu beremo drugače. E-gradiva kot

bližnjica do uspeha? E-zbornik Bralnega društva Slovenije od 12.

strokovnem posvetovanju 8. septembra 2017. Bralno društvo Slovenije,

Ljubljana. 43–48. DOI: https://www.bralno-drustvo.si/wp-

content/uploads/2013/06/E-

GRADIVA_KOT_BLIZNJICA_DO_USPEHA_BDS20171.pdf.

[16] Simona Tancig. 2016. Od Prousta do Twitterja – nevroedukacijske

raziskave bralne pismenosti v digitalni dobi. Bralna pismenost kot izziv in

odgovornost. Pedagoška fakulteta Univerze v Ljubljani, Ljubljana, 9–26.

DOI: https://www.pef.uni-lj.si/wp-content/uploads/2022/07/Bralna-

pismenost_Posvet-PeF-2016.pdf.

[17] Slovar slovenskega knjižnega jezika. DOI: https://fran.si/.

[18] Sonja Pečjak in Ana Gradišar. 2002. Bralne učne strategije. Zavod

Republike Slovenije za šolstvo, Ljubljana.

[19] Sonja Pečjak. 2022. Psihološka perspektiva učinkov digitalizacije na

procese branja in učenja pri ljudeh. Branje v slovenskem jeziku, literaturi

in kulturi. 58. seminar slovenskega jezika, literature in kulture. Založba

Univerze v Ljubljani, Ljubljana. 61–68. DOI:

https://doi.org/10.4312/SSJLK.58.2386-058X.

[20] Tjaša Pucko in Sonja Pečjak. 2016. Primerjava učinkovitosti e-branja in

branja s papirja pri učencih v srednji šoli. Pedagoška obzorja 31, 3–4, 55–

69. DOI: https://www.dlib.si/listalnik/URN_NBN_SI_doc-

4Q2O9TK8/index.html.

[21] Učni načrt. Program osnovna šola. Slovenščina. 2011. Ministrstvo za

šolstvo in šport, Zavod Republike Slovenije za šolstvo, Ljubljana. DOI:

https://www.gov.si/assets/ministrstva/MVI/Dokumenti/Osnovna-

sola/Ucni-nacrti/obvezni/UN_slovenscina.pdf.

[22] Učni načrt. Slovenščina. Gimnazija: splošna, klasična, strokovna

gimnazija. 2008. Ministrstvo za šolstvo in šport, Zavod Republike

Slovenije za šolstvo, Ljubljana. DOI:

https://eportal.mss.edus.si/msswww/programi2018/programi/media/pdf/u

n_gimnazija/un_slovenscina_gimn.pdf.

[23] Wikivir. DOI: https://sl.wikisource.org/wiki/Glavna_stran.

[24] Wikipedija, prosta enciklopedija: elektronska knjiga. DOI:

https://sl.wikipedia.org/wiki/Elektronska_knjiga.

[25] Zbirka slovenskih leposlovnih besedil. DOI: https://lit.ijs.si/leposl.html.



Umetnointeligenčna pismenost in šolske knjižnice



Artificial Inteligence Literacy and School Libraries



Gregor Škrlj

Narodna in univerzitetna knjižnica

Ljubljana, Slovenija

gregor.skrlj@nuk.uni-lj.si



dele družbe, hkrati pa je poskočila tudi uporaba različnih orodij,

POVZETEK ki so narejena s pomočjo UI. Izjema niso niti knjižnice, šole in

V prispevku je podana razlaga termina umetnointeligenčna šolske knjižnice.

pismenost, predstavljen je vidik uporabnosti umetne inteligence Šolski knjižničarji doma in po svetu se vse bolj izobražujejo

v šolskih knjižnicah doma in po svetu. na to temo ter uporabljajo različna orodja in pripravljajo navodila,

Ob nastanku ter hitrem razvoju umetne inteligence je smernice in načrte za izobraževanje svojih uporabnikov

pomembno, da se tudi v šolskih knjižnicah o tej temi govori, se (različnih deležnikov).

uči ter uporablja pravilno in etično. Šolske knjižnice imajo tudi Narodna in univerzitetna knjižnica (dalje NUK) je nacionalna

izobraževalno vlogo na tem področju, saj so že do sedaj skrbele knjižnica Republike Slovenije, katere temeljno poslanstvo je

za ustrezne vire in literaturo ter izobraževanje svojih zbiranje in varovanje ter zagotavljanje uporabe nacionalne zbirke

uporabnikov. knjižničnega gradiva, strokovna podpora knjižnicam pri

Zelo hitro je z izobraževanji na področju nove pismenosti izvajanju javne službe in nacionalnemu bibliografskemu sistemu

pričela tudi Narodna in univerzitetna knjižnica, ki je pripravila ter vključevanje v mednarodne knjižnične povezave [1]. Skladno

izobraževanja ter soorganizirala konference na temo umetne z zakonodajo izvaja knjižnično dejavnost kot javno službo [2] in

inteligence. skrbi za dediščino ter sodeluje v nacionalnem vzajemnem

bibliografskem sistemu, opravlja pa tudi druge dejavnosti in



KLJUČNE BESEDE naloge. Med drugim je za knjižničarsko stroko izjemno pomembna Umetna inteligenca, knjižnice, šolske knjižnice, izobraževalna dejavnost, za katero v NUK skrbi Oddelek za umetnointeligenčna pismenost, Narodna in univerzitetna izobraževanje, razvoj in svetovanje. Izobraževalne vsebine so knjižnica razdeljene po različnih sklopih, med katerimi so vključene tudi



ABSTRACT izobraževanje je za vsakega posameznika nepogrešljivo ter vsebine o umetni inteligenci in posameznih aplikacijah. Stalno

The paper explains the term AI literacy, and presents an pomembno, česar se NUK zaveda in temu vsako leto posveča

aspect of the use of AI in school libraries at home and abroad. veliko pozornosti. [3]

With the emergence and rapid development of AI, it is V NUK smo izvajali izobraževanja (Uporaba orodij umetne

important that it is also discussed, taught and used in school inteligence v knjižnicah) ter soorganizirali konference na temo

libraries in a correct and ethical way. School libraries also have umetne inteligence (Artificial Intelligence and the Future of

an educational role to play in this area, as they have already been Digital Preservation), katerih so se udeležili tudi šolski

providing relevant resources and literature and educating their knjižničarji. users.

The National and University Library has been very quick to

take up the new literacy education agenda, providing training and 2 ŠOLSKE KNJIŽNICE IN UMETNA

co-organising conferences on artificial intelligence. INTELIGENCA



KEYWORDS Šolsko knjižnico v Sloveniji opredeljuje Zakon o organizaciji

in financiranju vzgoje in izobraževanja (2007), v 68. členu je

Artificial Intelligence, libraries, school libraries, Artificial zapisano, da ima šola knjižnico, v kateri se zbira, strokovno

Intelligence Literacy, National and University Library obdeluje, hrani, predstavlja in izposoja knjižnično gradivo ter ki

opravlja informacijsko-dokumentacijsko delo kot sestavino



1 vzgojno-izobraževalnega dela v šoli. Del naše šolske knjižnice je UVOD

tudi učbeniški sklad. [4]

Razvoj digitalne tehnologije je spodbudil globalizacijo in Šolske knjižnice že dolgo predstavljajo vir informacij in

modernizacijo življenja ter ustvarja vedno bolj medsebojno znanja za svoje deležnike (učence, dijake, učitelje in profesorje).

povezan svet. V letih po epidemiji (2020) so se pojavile mnoge Poleg ustaljenih knjižničnih storitev (izposoja knjig, dostop do

prosto dostopne oblike umetne inteligence (dalje UI), ki se strokovne literature in organizacija različnih dejavnosti), šolski

pospešeno razvijajo ter razširjajo in prodirajo v vse sisteme ter knjižničarji spodbujajo razvoj kritičnega mišljenja in



informacijskih spretnosti. Z nenehnim napredkom tehnologije in Osebe, ki so umetnointeligenčno pismene, razumejo temeljne

digitalizacije se vloga šolskih knjižnic hitro spreminja in koncepte UI, kot so strojno učenje, obdelava naravnega jezika in

prilagaja novim izzivom ter potrebam. Z razvojem ter nevronske mreže. Da bi svoje deležnike uporabnike opremili s

prisotnostjo UI orodji se je potrebno prilagoditi tudi tem potrebnimi veščinami za krmarjenje po tem zapletenem področju,

novostim. morajo knjižnice prednostno razvijati vire za usposabljanje, ki

Eden od najpomembnejših tehnoloških trendov, ki že danes posameznikom omogočajo natančno preverjanje informacij v

vpliva na šolske knjižnice (predvsem v ZDA), je UI, saj tudi že zvezi z aplikacijami UI. [6]

spreminja način, kako se zbira, obdeluje in uporablja informacije. Šolske knjižnice, poleg osnovnih knjižničnih dejavnosti,

V šolskih knjižnicah po ZDA se UI uporablja za izboljšanje izvajajo tudi različne učne oblike (medpredmetne povezave,

iskalnih algoritmov, personalizacijo priporočilnih seznamov pouk knjižnično-informacijskega znanja, interesne dejavnosti,

gradiv, avtomatizacijo administrativnih nalog in celo za pomoč izbirne predmete in mnoge druge dejavnosti), ki so predpisane,

pri učenju, kar so različni avtorji opisali v svojih prispevkih (An ter tako lahko svojim uporabnikom ponudijo še tisto nekaj več.

AI Toolbox for Librarians; ChatGPT and Generative Artificial [7]

Intelligence in drugi). Napredni sistemi UI omogočajo učencem Šolske knjižnice, kot središča znanja in informacij, so idealne

lažji dostop do informacij in prilagojene učne izkušnje, hkrati pa ustanove za spodbujanje učenja o UI. V ZDA je bilo veliko

šolskim knjižničarjem omogočajo učinkovitejše delo, tudi primerov predstavljenih na različnih konferencah, seminarjih in

upravljanje knjižničnih virov. Mnogo prispevkov je dostopnih drugih dogodkih. Več zanimivih primerov je bilo omenjenih na

tudi na portalu Knjižničarske novice, kjer se za lažje iskanje spletni konferenci Teaching and Learning with AI, kjer so

uporabljajo ključniki: #umetnainteligenca, predavatelji med drugim izpostavili ter predstavili primere dobih

#umetnointeligenčnapismenost, in #digitalnapismenost. praks. Spregovorili so o primerih uporabe UI: kako so ustvarili

Izobraževanje o umetni inteligenci v šolskih knjižnicah ne tutorja s pomočjo UI, klepetalnega robota za pomoč v knjižnici

samo da pripravi učence na tehnološko prihodnost, ampak tudi ter se učili uporabljati različne programe. [8]

spodbuja inovacije, povečuje različne pismenosti (digitalno, Šolski knjižničarji v ZDA uporabljajo UI za pripravo

bralno, medijsko …) in zagotavlja, da bodo mladi opremljeni z različnih navodil, plakatov, oznak za knjižnično gradivo, za

znanji in veščinami, potrebnimi za uspeh v sodobni družbi. Ob izdelavo bralnih seznamov, pisanje zgodb, izdelavo nalog in

vsem tem pa je potrebno razmisliti tudi o umetnointeligenčni različnih didaktičnih iger ter še mnogo drugega. Med drugim

pismenosti ter kritičnem mišljenju. uporabljajo aplikacije kot so: Canva, Claude, Gemini, Bard,

Midjourney, Dall-E, Copilot, Perplexity, Ideogram in druge.

2.1 Umetnointeligenčna pismenost Ameriške šolske knjižničarke so mnogo svojih primerov

Terminološka sekcija Inštituta za slovenski jezik Frana opisale tudi v spletnem podkastu Leading from the Library, kjer

Ramovša - Znanstvenoraziskovalni center Slovenske akademije so omenile klepetalne robote, pomen podatkov in virov ter

znanosti in umetnosti je predlagala, da se za prevod Artificial upoštevanje avtorskih pravic. [9]

Intelligence Literacy uporabi pridevnik umetnointeligenčni, ki je Precej primerov uporabe UI so uredniki zbrali v publikaciji z

jezikovno sistemsko ustrezen in je že del terminov, kot so npr. naslovom 101 creative ideas to use AI in education, A

umetnointeligenčni sistem, umetnointeligenčna slikovna crowdsourced collection, kjer ponujajo zbrane zamisli za prenos

analitika. S pridevnikom v levem prilastku (umetnointeligenčni) v prakso ob kritičnih in ustvarjalnih možnostih ter omejitvami

ustrezno določimo samostalniško jedro (pismenost). Tako uporabe UI pri učenju, poučevanju, podpori učencem pri učenju

govorimo o umetnointeligenčni pismenosti. [5] in drugih dejavnostih. [10]

V današnjem hitro razvijajočem se digitalnem svetu postaja V Sloveniji so šolske knjižnice začele s pospešeno rabo

umetnointeligenčna pismenost ključnega pomena za vse starosti, aplikacij in dela prek spleta v letu, ko je cel svet spremenila

tudi za mlade. Razumevanje in uporaba UI omogoča mladim, da pandemija. Kot so to zapisale posamezne avtorice v svojih

postanejo kompetentni uporabniki sodobnih tehnologij, kar jim prispevkih z naslovi: Virtualna šolska knjižnica OŠ Komenda

odpira vrata do številnih priložnosti v izobraževanju in Moste med poukom na daljavo (2020), Delo na daljavo v naši

kasnejšem poklicnem življenju. šolski knjižnici (2020), Šolska knjižnica kot popestritev pouka

Šolske knjižnice lahko in v prihodnje morajo igrati ključno na daljavo (2021), Šolska knjižnica na daljavo (2021), Šolska

vlogo pri poučevanju in spodbujanju umetnointeligenčne knjižnica in učenje na daljavo (2021), so šolske knjižnice

pismenosti. Knjižničarji in učitelji lahko skupaj ustvarjajo pogumneje začele uporabljati digitalne poti in gradiva, a še ne

pedagoške ure in dejavnosti, kjer učenci spoznavajo osnove posameznih orodji UI. To je prišlo z leti, ko nekateri šolski

umetne inteligence, njeno delovanje in praktične aplikacije. Tako knjižničarji uporabljajo in učijo uporabo UI ter poudarjajo

se učenci naučijo kritičnega razmišljanja, reševanja problemov pomen umentointeligenčne pismenosti. V šolskem letu

in etičnih vprašanj, povezanih z uporabo UI. 2023/2024 je kar nekaj slovenskih šolskih knjižničarjev začelo z

uporabo in učenjem pomena UI. Prispevke v uredništvu še

pričakujemo in bodo objavljeni na portalu Knjižničarske novice.

3 PRIMERI UPORABE UMETNE Če omenimo le nekatere, je bila med zanimivejšimi idejami

INTELIGENCE V ŠOLSKIH KNJIŽNICAH spletna soba pobega, ki so jo pripravili v eni izmed

osnovnošolskih knjižnic, kjer so se morali ujetniki vsakič znova

Da bi uporabniki sami uporabljali UI, jo morajo najprej

rešiti iz druge umetno generirane situacije (potrebno je bilo

razumeti in postati vešči na tem področju. Umetnointeligenčna

poznavanje določenih pravljic). Ravno tako je potrebno

pismenost obsega poznavanje, razumevanje, uporabo in

izpostaviti tudi šolsko knjižnico, ki je za učence pripravila lov za

vrednotenje umetne inteligence ter upoštevanje etičnih vprašanj.

knjižnim zakladom s pomočjo ChatGPT, slikovno gradivo pa je

bilo ustvarjeno z ustreznimi pozivi v aplikaciji Canva (slika 1) in vprašanjih, povezanih z UI, kot so zasebnost, pristranskost in

Copilot. vpliv na življenje. Tako učence pripravljajo na kritično

Poudariti je potrebno tudi, da so se v šolskih knjižnicah lotili ocenjevanje in premišljeno uporabo tehnologij v vsakdanjem

tudi priprave navodil, kako pravilno navajati vire (v besedilu in življenju.

v virih), ki so ustvarjeni s pomočjo UI.

VIRI IN LITERATURA



[1] Naloge, vizija, poslanstvo in vrednote NUK. https://www.nuk.uni-

lj.si/nuk/poslanstvo#

[2] Zakon o knjižničarstvu. 2015.

http://www.pisrs.si/Pis.web/pregledPredpisa?id=ZAKO2442

[3] Etični kodeks slovenskih knjižničarjev. 1995. https://www.zbds-

zveza.si/daljsa-6/

[4] Zakon o organizaciji in financiranju vzgoje in izobraževanja (ZOFVI).

2023. https://www.racunovodstvo.net/zakonodaja/zofvi

[5] Bon, M. 30. 8. 2023. Kako razvijati umetnointeligenčno pismenost.

https://knjiznicarske-novice.si/kako-graditi-pismenost-na-podrocju-

umetne-inteligence-ui/

[6] Bon, M. (20. 3. 2024). Ključni elementi razvijanja umetnointeligenčne

pismenosti. https://knjiznicarske-novice.si/kljucni-elementi-razvijanja-

umetnointeligencne-pismenosti/

[7] Fekonja, R. (ur.). (2014). Posodobitve pouka v osnovnošolski praksi:

Knjižnično informacijsko znanje. Zavod RS za šolstvo.

http://www.dlib.si/stream/URN:NBN:SI:DOC-ABAUB0ON/e946cb25-

a2e9-4845-a248-cd7ec5f22e8c/PDF

[8] Škrlj, G., Bon, M. (5. 7. 2024). Spletna konferenca Teaching and Learning

with AI. https://knjiznicarske-novice.si/spletna-konferenca-teaching-and-

learning-with-ai/

[9] AI in the School Library! (S7, E11). Leading from the Library Podcast.

https://www.youtube.com/watch?v=W4idSCuoqpI

[10] Nerantzi, C., Abegglen, S., Karatsiori, M. in Martínez-Arboleda, A. 2023.

101 creative ideas to use AI in education, A crowdsourced collection.

https://zenodo.org/records/8355454



Slika 1: S pomočjo aplikacije Canva ustvarjen primer

napotkov, kaj nudi šolska knjižnica svojim uporabnikom



4 ZAKLJUČEK

V današnji dobi digitalizacije in hitrega tehnološkega

napredka postaja umetnointeligenčna pismenost zelo pomembna.

Pri poučevanju o temi naj bi imeli tudi šolski knjižničarji vlogo,

saj delujejo kot most med tradicionalnim učenjem in sodobnimi

tehnologijami. Šolski knjižničarji naj bi bili ključni pri

spodbujanju radovednosti in kritičnega mišljenja med uporabniki.

Z organiziranjem dejavnosti in pouka o osnovah umetne

inteligence, ob sodelovanju strokovnih sodelavcev, bi učencem

omogočili uvid v delovanje UI, kje se uporablja in kakšne so

njene prednosti ter nevarnosti.

Poleg tega šolski knjižničarji zagotavljajo dostop do različnih

virov, ki učencem omogočajo poglobljeno raziskovanje umetne

inteligence. Knjižničarji skrbno izbirajo knjige, članke, e-knjige

in druga izobraževalna gradiva, ki so relevantni in aktualni. S tem

učencem omogočajo, da pridobijo celovito razumevanje

področja in ostanejo na tekočem z najnovejšimi trendi in

razvojem v svetu UI.

Šolski knjižničarji naj v bodoče tudi sodelujejo z učitelji pri

integraciji umetnointeligenčne pismenosti v učni načrt. S

skupnim načrtovanjem učnih enot in projektov, ki vključujejo

uporabo UI orodij, knjižničarji pomagajo učencem pridobiti

praktične izkušnje in veščine, ki jim bodo koristile v prihodnosti.

Poleg tega knjižničarji spodbujajo etično razmišljanje in

ozaveščanje o odgovorni uporabi umetne inteligence.

Organizirajo razprave in debate, kjer učenci razmišljajo o etičnih



E-igrifikacija kot metoda dela z učenci s posebnimi



potrebami v osnovni šoli



E-gamification as a Method of Working with Students with



Special Needs in Primary School



Urška Šuštaršič

šolska svetovalna služba

OŠ Danile Kumar

Ljubljana, Slovenija

sustarsicu@os-danilekumar.si



POVZETEK KEYWORDS

Igrifikacija kot metoda dela v osnovnošolskem izobraževanju, Gamification, primary school, students with special needs,

zlasti za delo z učenci s posebnimi potrebami, ima že stalno digitalization

prakso. Z uvedbo iger podobnih elementov, kot so točke in

nagrade, želijo učitelji pomagati učencem z motnja pozornosti s

hiperaktivnostjo (ADHD), avtizmom in drugimi učnimi izzivi, 1 UPORABA IKT V SLOVENSKEM

da ostanejo osredotočeni in motivirani dlje, kot bi bili drugače. Z ŠOLSTVU

dobo digitalizacije se je tudi področje igrifikacije razširilo in Slovenski šolski sistem je že dlje časa podvržen kritikam, da

ponudilo številna nova digitalna orodja, ki jih bomo raziskali v je staromoden, neprilagodljiv in da nujno potrebuje spremembe.

nadaljevanju. Kot odgovor je bila ustanovljena delovna skupina za pripravo

Dotaknili se bomo tudi pomislekov glede uporabe igrifikacije prenovljenega Nacionalnega programa vzgoje in izobraževanja

in se vprašali kaj je naš končni cilj v izobraževanju – ali samo 2023 – 2033. Eden izmed pomembnih strateških ciljev in

želimo, da so učenci v danem trenutku aktivni ali si prizadevamo ukrepov je tudi razvoj digitalnih kompetenc tako učiteljev kot

pri učencih spodbujati notranjo motivacijo in veselje do znanja? učencev [2].

Poleg tega niso zanemarljive raziskave, ki ugotavljajo, da se Med digitalne kompetence spada tudi uporaba IKT tehnologij

pozornost učencev zaradi uporabe digitalnih medijev skrajšuje kot metod in strategij učenja. Na podlagi raziskave Pedagoškega

[1]. Kako najti pravo ravnovesje? inštituta [3] slovenske šole zaostajajo pri uporabi IKT tehnologij

za poučevanje v osnovnih in srednjih šolah. Eden izmed

KLJUČNE BESEDE pomembnih razlogov za to navajajo, da je sama uporaba v veliki

Igrifikacija, osnovana šola, učenci s posebnimi potrebami, meri vezana na lastno zanimanje in motivacijo posameznih

digitalizacija učiteljev, saj je bilo v preteklosti izobraževanj vezanega na

uporabo IKT tehnologij za učitelje relativno malo. To ugotovitev

ABSTRACT lahko potrdim tudi sama, saj je angažiranost za vpeljevanje t. i.

Gamification as a method of work in primary school novosti v šolsko delo v veliki meri odvisno od angažiranosti tako

education, especially for work with students with special needs, vodstva kot strokovnih delavcev. Vlada RS je že leta 2021

already has a permanent practice. By introducing game-like sprejela in potrdila Nacionalni program spodbujanja razvoja in

elements like points and rewards, teachers aim to help students uporabe tudi umetne inteligence v Republiki Sloveniji do leta

with ADHD, autism and other learning challenges to stay 2025 [4], vendar konkretnih rezultatov v šolstvu še ni zaznati.

focused and motivated longer than they would otherwise. With Z desetletnimi izkušnjami z delom kot izvajalka dodatne

the era of digitalization, the field of gamification has also strokovne pomoči učencem s posebnimi potrebami sem se tudi

expanded and offered many new digital tools, which we will sama lotevala iskanja možnosti v informacijski tehnologiji,

explore in the article. predvsem na področju igrifikacijskih vsebin za učence s

We also have some concerns regarding the use of posebnimi potrebami – kar v nadaljevanju poimenujem kot e-

gamification especially regarding our ultimate goal in education igrifikacija.

- do we just want students to be active at a given moment or do



we strive to encourage their internal motivation and joy of 2 IGRIFIKACIJA knowledge? In addition, researches show that the attention span

of students is shortened true the years due to the usage of digital Pojem igrifikacije (angl. gamification), kot ga poznamo danes,

media [1]. How to find a perfect balance? je leta 2002 uvedel Nick Pelling. Izraz igrifikacija se je sicer

začel pogosteje pojavljati šele v drugi polovici leta 2010 [4]. Pri

igrifikaciji gre za uporabo pristopov, elementov in načinov 3 VKLJUČEVANJE E-IGRIFIKACIJE PRI

razmišljanja, ki se uporabljajo pri igranju iger, vendar so le-ti DELU Z UČENCI S POSEBNIMI

uporabljeni izven iger, npr. sistem, ki beleži število oddanih POTREBAMI nalog in ki nagrajuje dosežke z virtualnimi nagradami, kot so

pokali, zvezdice (Slika 1). Pomemben dejavnik doseganja Po konceptu dela z učenci z učnimi težavami v osnovni šoli

določenih ciljev predstavlja motivacija. Ko posameznik uspešno mora šola učencem ustrezno prilagoditi metode in oblike dela, da

reši zastavljeno nalogo oz. problem, doživi uspeh, kar v so primerne za posamezno skupino učencev. E-igrifikacija le to

posamezniku sproži motivacijo. Ko imajo posamezniki relativno preprosto omogoča, hkrati pa omogoča lažje

priložnost spopasti se z individualiziranimi učnimi izzivi, so uresničevanje temeljnih načel pomoči učencem kot so načelo

pripravljeni vložiti več energije v nalogo, poleg tega pa so tudi interdisciplinarnosti, odkrivanje močnih področij, načelo akcije

bolj dovzetni za povratne informacije. Vključitev igrifikacije se in samozagovorništva, predvsem pa načelo opravljanja

kaže v uporabi kvizov, odzivov, sodelovanj v skupinah in drugih optimalnih izzivov [7].

učnih aktivnostih [6]. V poglavju bomo pregledali glavne značilnosti uporabe e-

igrifikacije pri učencih s posebnimi potrebami.





Aktivnost in zavzetost

Ena izmed največjih prednosti uporabe igrifikacije je njena

zmožnost povečanja motivacije učencev. S preoblikovanjem

vsakdanjih nalog v interaktivne in tekmovalne dejavnosti lahko

naredi učenje bolj privlačno. Dr. Karl Kapp, znani strokovnjak

za izobraževalno igrifikacijo, trdi, da le-ta posega v osnovne

človeške želje po dosežkih, statusu in tekmovanju, zaradi česar

je močno orodje za motivacijo [8]. Za osnovnošolce, ki jih

pogosto vodijo takojšnje nagrade in povratne informacije, lahko

uporaba točk, značk in lestvic zagotovi občutek uspeha in jih

spodbudi, da bolj aktivno sodelujejo v učnem procesu.



Vključevanje sprememb



Slika 1: Prikaz delitve v igro usmerjenega poučevanja [6] izobraževanju je zmožnost individualizacije in prilagajanja Ena najpomembnejših prednosti igrifikacije v inkluzivnem



osnovnošolskem izobraževanju kot nov pristop za spodbujanje in učence s posebnimi potrebami, ki imajo pogosto težave z motiviranje učencev in posledično izboljševanje učnih rezultatov. osredotočenostjo in organizacijo. E-igrifikacija omogoča bolj Obljuba o spreminjanju učenja v bolj privlačno in zabavno prilagodljivo učno izkušnjo, kjer lahko učenci delajo s svojim V zadnjih letih je e-igrifikacija vedno bolj prisotna v metodami dela in določenim tempom dela je lahko izziv za učenja vsakemu posamezniku. Tradicionalna učilnica s svojimi

izkušnjo se zdi idealna predvsem za učence s posebnimi



potrebami, ki težje sledijo bolj tradicionalnim metodam učenja njihovim individualnim potrebam kot tudi različnim učnim tempom in se ukvarjajo z materiali na načine, ki ustrezajo

ali imajo drugačen tempo osvajanja znanja.

stilom. Nekaterim učencem bodo morda koristili vizualni

S pomočjo igrifikacije lahko povečamo motivacijo, aktivnost

pripomočki, medtem ko se bo drugim zdelo bolj učinkovito

in zavzetost udeležencev za učni proces, spodbujamo

avditivni ali kinestetični stil učenja.

sodelovanje, omogočamo sprotno povratno informacijo,

Dr. Thomas E. Brown, klinični psiholog poudarja, da

vključujemo spremembe, omogočamo varno učno okolje za

učencem z ADHD prilagojena učna okolja zelo koristijo, saj jim

udeležence, ki imajo težave pri deljenju mnenj ter omogočamo

omogočajo, da se lažje osredotočajo na svoje prednosti in

sprotne nagrade za opravljeno delo (Slika 2). Če pri tem



uporabljamo še IKT-tehnologijo, je uporaba navadno bolj ponudijo prilagodljive učne načrte, kjer je vsebina prilagojena interese [9]. Izobraževalne aplikacije in spletne platforme lahko

preprosta tako za učitelja, kot tudi bolj zanimiva za učenca. vsakemu učencu posebej. E-igrificirane učne aplikacije lahko

naredijo lekcije bolj privlačne in s tem pomagajo ohraniti



osredotočenost in motivacijo dlje, kot brez njih.



Varno učno okolje

Pomemben aspekt e-igrifikacije na inkluzivno izobraževanje

je tudi izboljšanje dostopnosti. Učenci s telesnimi oviranostmi,

senzoričnimi okvarami ali učnimi težavami lahko tako

uporabljajo tehnologijo, ki jim omogoča učinkovitejše

sodelovanje z izobraževalno vsebino kot tudi sodelovanje v

skupini. Orodja, kot so bralniki zaslona, programska oprema za

pretvorbo govora v besedilo in predvsem aplikacije, ki

spodbujajo sodelovanje, omogočajo učencem s telesnimi

okvarami ali socialnimi zadržki lažji dostop do izobraževalnih

vsebin kot tudi sodelovanja in deljenja svojih mnenj z ostalimi v

Slika 2: Prednosti vključevanja igrifikacije v učenje [6] skupini.



Po mnenju dr. Davida Rosa, soustanovitelja organizacije

CAST (Center for Applied Special Technology) in vodilnega

strokovnjaka za oblikovanje univerzalnega pristopa k učenju

(Universal design for learning- UDL) pravi, da tehnologija lahko

služi kot most za učence s posebnimi potrebami, saj jim omogoča

sodelovanje v izobraževalnih dejavnostih, ki bi sicer bile za njih

težje dostopne [10].

Slika 4: Aplikacija Kobi [12]



Sodelovanje in povratna informacija Kahoot – https://kahoot.com/ Uporaba e-igrifikacije ima lahko pomembno vlogo tudi pri Kahoot je norveška spletna učna platforma, ki temelji na omogočanju sodelovanja in komunikacije med učenci, učitelji in učenju preko iger. Vsebuje učne igre, znane tudi kot starši. Oblikovanje igrificiranih vsebin na spletnih učnih "kahoots"(Slika 5), ki so uporabniško ustvarjeni kvizi z več platformah, forumih za razprave in uporaba različnih orodij za možnimi odgovori, do katerih se lahko dostopa preko spletnega sodelovanje, udeležencem omogočajo skupno delo pri projektih, brskalnika ali aplikacije [13]. Primerna je predvsem za ne glede na njihovo lokacijo ali sposobnosti. To je še posebej spodbujanje motivacije, učence s težavami s pozornostjo in za pomembno v inkluzivnih učilnicah, kjer lahko učenci preko boljše vključevanje učencev s težavami na socialnem področju, zabavnih vsebin na različne načine pristopajo k učenju in saj učenci lahko sodelujejo preko pametni telefonov ali tablic. deljenju svojih znanj, mnenj.

Druga pomembne prednost e-igrifikacije je možnost hitre



povratne informacije o uspešnosti učenja. Povratna informacija

je eden od dejavnikov, ki najbolj vplivajo na učenčevo učenje.

Poglavitni cilj vzgojno-izobraževalnega procesa je namreč

pomoč pri identifikaciji stopnje osvojenega znanja v primerjavi

s končnim ciljem (»Kako mi gre?« sorazmerno z »Kam grem?«)

in ukrepanje v smislu predlogov nadaljnjih korakov (»Kam

naprej?«) [11]. Digitalizirane vsebine pa to omogočajo zelo

natančno in sprotno, kar daje povratni informaciji največjo Slika 5: Primer kviza v orodju Kahoot [13] učinkovitost in težo še posebno, če so le-te vezane tudi na

pridobivanje morebitnih nagrad (zvezdice, značke, pokali, CŠOD Misija - mobilna aplikacija za učenje na prostem

napredovanje …). Mobilni vodič CŠOD Misija (Slika 6) je inovativen

didaktičen pristop, ki nadgrajuje samostojno doživljajsko učenje



4 z doživetjem zunaj učilnice. Poleg tega, da spodbuja uporabnike PREDLOGI ORODIJ ZA OBLIKOVANJE



E-IGER ZA UČENCE S POSEBNIMI spoznanja koncepta igrifikacije, ki pedagoškemu procesu dodaja k učenju na terenu, upošteva tudi najnovejša pedagoška

POTREBAMI elemente iger in motivira uporabnike k doseganju boljših učnih

Pregled orodij za pomoč pri oblikovanju e-igrificiranih rezultatov. Gibanje spodbuja pomnjenje in razmišljanje, kar je

vsebin pri učenju glede na posebne potrebe učencev. tudi ena od prednosti orodja [14]. Primerna je predvsem za

učence s težavami na področju motivacije, koncentracije in za

Orodja za pripravo prezentacij učence z večjo potrebo po gibanju. Primerna je tudi za

Igrificirane vsebine se lahko oblikujejo tudi v orodjih za spodbujanje skupinskega dela.

pripravo prezentacij kot npr. Microsoft Powepoint (Slika 3) in so

primerne za vse vrste učencev.





Slika 6: Aplikacija CSOD Misija [14]



Slika 3: Primer kviza v orodju Powepoint



Baamboozle - https://www.baamboozle.com/

Kobi - https://kobiapp.si/ Baamboozle je učna platforma, ki temelji na igrah in se

Slovenska aplikacija, ki je namenjena branju, predvsem za uporablja predvsem za oblikovanje učnih kvizov (Slika 7).

učence z bralno-napisovalnimi težavami, težavami s pozornostjo Glavna prednost je možnost igranja v skupinah prav tako pa

ali težavami z razumevanjem prebranega (Slika 4). Prednost je omejuje tekmovalnost preko vključevanja naključnih dogodkov,

nalaganje vsebin preko fotografiranja gradiva. Podprt je tudi s kot je npr. menjava točk skupin, zmanjšanje točk za 20 in

strani Ministrstva za šolstvo. podobno. Primerna je predvsem za učence s težavami s

pozornostjo, učence, ki potrebujejo več-senzorno učenje in

učence s težavami na socialnem področju.

5 POMISLEKI



Nevarnost razpršene pozornosti



Medtem ko je IKT tehnologija lahko močno orodje za učenje,

predstavlja tudi veliko tveganje za večanje razpršene pozornosti,

zlasti za učence z ADHD. Prav naprave, ki se uporabljajo za

podporo učenju – računalniki, tablice in pametni telefoni – lahko

nudijo tudi dostop do iger, družbenih medijev in drugih motenj,

ki lahko iztirijo učenčevo osredotočenost. Dr. Edward Hallowell,

psihiater in strokovnjak za ADHD, opozarja, da je ista

Slika 7: Primer vprašanja Baamboozle kviza[15] tehnologija, ki pomaga učencem z ADHD, lahko tudi vir motenj,

če je ne uporabljajo previdno [18]. Pri učencih z ADHD, ki se že

Educaplay - https://www.educaplay.com/ borijo z ohranjanjem pozornosti, je lahko velika skušnjava, da bi

Educaplay je platforma za oblikovanje e-igrificiranih vsebin. zamenjali naloge ali se vključili v neizobraževalne dejavnosti.

S pomočjo že pripravljenih predlog se lahko enostavno oblikuje Zaradi tega je potrebno, da učitelji izvajajo nadzor nad uporabo

različne individualizirane izobraževalne igre kot so križanke, zaslonov, lahko tudi uporabo aplikacij, ki blokirajo druga spletna

igre dialogov, igre narekovanja, igre izpolnjevanja praznih mest, mesta med učenjem, ali določanje jasnih pravil o uporabi naprave.

kvize, igre ujemanja, igre spomina in podobno (Slika 8).

Primerna je predvsem za učence s težavami s pozornostjo, Zmanjševanje razvoja kompetenc

učence, ki potrebujejo več senzorno učenje, večjo slikovno Pretirano zanašanja na IKT tehnologijo lahko privede do

podporo in učence, ki se učijo s svojim tempom učenja. zmanjšanja sposobnosti kritičnega mišljenja in reševanja

problemov. Čeprav e-igrificirano učenje lahko podpira učenje,



ne sme nadomestiti temeljnih veščin, ki jih morajo učenci razviti,

kot so bralno razumevanje, matematične spretnosti in neodvisno

in kritično razmišljanje. Za učence s posebnimi potrebami

obstaja še večje tveganje, da bi jim zanašanje na tehnologijo za

organizacijo in upravljanje nalog preprečilo, da bi sami kasneje

razvili te bistvene spretnosti.

Pedagoška psihologinja dr. Larry Rosen trdi, da pretirano

zanašanje na tehnologijo lahko ovira razvoj pomembnih

kognitivnih veščin, zlasti pri učencih z ADHD, ki lahko

postanejo preveč odvisni od digitalnih pripomočkov [19].

Slika 8: Primeri aktivnosti na platformi Educaplay[16] Raziskave na slovenskih mladostnikih tudi kažejo, da se pri tistih,



Scratch - https://scratch.mit.edu/ rezultati pri samoregulativnem učenju. Izpostavljajo tudi ki so bolj zgodaj začeli igrati računalniške igrice, kažejo slabši

Scratch je orodje, ki je prvotno namenjen učenju pomembno ugotovitev, da so negativne povezave med rabo

programiranja za otroke. Zaradi enostavnega vmesnika pa je interneta za zabavo in samoregulativnim učenjem pri učencih s

primerno tudi za oblikovanje različnih igrificiranih vsebin. posebnimi potrebami močnejše [20].

Omogoča ustvarjanje digitalne zgodbe, različne igre in animacije. Za učitelje je pomembno, da najdejo ravnotežje med uporabo

Scratch je ustvarjen, razvit in vzdrževan pri Združenju Scratch, e-igrificiranih vsebin kot orodja za učenje in spodbujanjem

neprofitni organizaciji in je v celoti brezplačen. Primeren je za učencev v omejeni meri, da ne ovirajo razvoja lastne kognitivne

oblikovanje vseh vrst vsebin predvsem za učence, ki potrebujejo strategije za učenje. jasna in v naprej znana in strukturirana pravila igre.

Neenakopraven dostop in poznavanje tehnologij



Dostop do tehnologij ni enak za vse učence in ta digitalni

razkorak lahko še poveča obstoječe neenakosti v izobraževanju.

Učenci nimajo enakega dostopa do naprav ali internetnih

povezav, ki so potrebne, da bi v celoti izkoristili uporabo e-orodij

v izobraževanju. To je lahko še posebej problematično za učence

s posebnimi potrebami, ki se morda bolj zanašajo na tehnologijo

za podporo pri učenju. Dr. Mimi Ito, strokovnjakinja za digitalno

učenje, poudarja, da t. i. digitalni razkorak ustvarja velike izzive

za inkluzivno izobraževanje, zlasti za učence s posebnimi

potrebami, ki potrebujejo dostop do specializirane tehnologije

[21]. Brez dostopa do potrebnih orodij se lahko določeni učenci

znajdejo v slabšem položaju v primerjavi s svojimi vrstniki,

zaradi česar jim je težje uspeti v vse bolj digitalnem učnem

Slika 9: Primer igre v orodju Scratch[17] okolju.



Izzivi pri izvajanju in usposabljanju učiteljev [10] Rose, D. H., & Meyer, A. (2015). Universal Design for Learning: Theory

Učinkovita uporaba vseh vrst IKT tehnologij v inkluzivnem and Practice. CAST Professional Publishing.

[11] Hattie, J. A. C., & Timperley, H. (2020). Moč povratne informacije = The

izobraževanju zahteva, da so učitelji dobro usposobljeni tako za power of feedback. Vzgoja in izobraževanje, 51(6), 6–21.

samo tehnologijo kot za strategije za podporo učencem s [12] Kobi. https://kobiapp.io/sl/ (pridobljeno 10. 8. 2024).

[13] Kahoot. https://kahoot.com/ (pridobljeno 10. 8. 2024).

posebnimi potrebami. Vendar pa mnogi učitelji poročajo, da se [14] Mobilna aplikacija CSOD Misija. https://www.csod.si/stran/csod-misija.



počutijo premalo pripravljeni za integracijo tehnologij v svoje (pridobljeno 10. 8. 2024). [15] Baamboozle. https://www.baamboozle.com/smallclassic/1535141. poučevanje, zlasti ko gre za njeno uporabo pri poučevanju (pridobljeno 10. 8. 2024).

učencev s posebnimi potrebami. [16] Educaplay. https://www.educaplay.com/types-of-activities/. (pridobljeno

Na podlagi raziskave Pedagoškega inštituta [3] tudi slovenski 10. 8. 2024).

[17] Sctratch. https://scratch.mit.edu/. (pridobljeno 10. 8. 2024).

učitelji poročajo, da bi za uspešno vključevanje IKT tehnologij v [18] Hallowell, E. (2015). Managing Distractions: Technology's Double-Edged

poučevanju potrebovali več izobraževalnih vsebin. Sword for Students with ADHD. Harvard Business Review.

[19] Rosen, L. (2018). The Risks of Over-reliance on Technology for Students

with ADHD. Psychology Today.

[20] Opaka, M., & Tancig, S. (2022). Povezave med rabo interneta in

6 samoregulativnim učenjem pri mlajših mladostnikih s posebnimi RAZPRAVA

potrebami ali učnimi težavami v redni osnovni šoli. 227–238.

Uporaba igrificiranih vsebin kot didaktični pripomoček pri https://www.drustvo-bravo.si/2022/10/06/zbornik-6-bravo-konference/

učenju učencev s posebnimi potrebami je lahko učinkovit [21] Ito, M. (2016). The Digital Divide in Education: Implications for Students

with ADHD. Digital Media and Learning Research Hub.

pripomoček za spodbujanje motivacije, pozornosti in tudi kot [22] Barle Lakota, A., & Goriup, J. (2006). Poglavja iz sociologije vzgoje in

krepitev pozitivnega vedenja. Medtem ko e-igrifikacija ponuja izobraževanja: pregled sodobnih socioloških študij, perspektiv in

konceptov. UP Fakulteta za management.

dragocena orodja za obravnavo nekaterih edinstvenih izzivov, s

katerimi se soočajo otroci s posebnimi potrebami, je nujno najti

ravnovesje. Poraja se namreč velika skrb, da tovrsten način

učenja vse preveč poudarja pomen zunanjih nagrad, spodkopava

notranjo motivacijo in dolgoročne izobraževalne cilje. Poleg tega

večja osredotočenost na tekmovanje in zunanje nagrade

zmanjšuje vrednost učenja zaradi njega samega, kar bi lahko

vodilo k površnemu razumevanju znanja namesto k

poglobljenemu učenju [22].

Strokovni delavci morajo igrifikacijo uporabljati premišljeno,

selektivno in jasno ciljno usmerjeno s sprotnimi evalvacijami

glede učinkovitosti in primernosti uporabe. Končni cilj bi moral

biti spodbujanje motivacijo do učenja, ki presega takojšnje

zadovoljstvo, ki ga ponujajo igrificirane naloge, in učence

pripravi ne le na pomen rezultatov, ampak tudi na razvoj notranje

motivacije in spretnosti kritičnega mišljenja, potrebnih za

vseživljenjsko učenje.



REFERENCE

[1] Mark G., Mills K. (2023). Why our attention spans are shrinking.

American. Psychological Association (APA). DOI:

https://www.apa.org/news/podcasts/speaking-of-psychology/attention-

spans

[2] Ministrstvo za vzgojo in izobraževanje. 30. 6. 2024. Nacionalni program

vzgoje in izobraževanja za obdobje 2023–2033. DOI:

https://www.gov.si/assets/ministrstva/MVI/Dokumenti/Nacionalni-

program-vzgoje-in-izobrazevanja-za-obdobje-2023-2033-predlog.pdf

[3] Japelj Pavešić B., Peršolja A., Špegel Razbojnik A. 2020.

Zaostajanje uporabe IKT za poučevanjev v slovenskih osnovnih in

srednjih šolah.Pedagoški inštitut. DOI: https://www.pei.si/wp-

content/uploads/2021/01/Zaostajanje-Slovenije-v-uporabi-IKT-za-

poucevanje.pdf

[4] Republika Slovenija. 2021. Nacionalni program spodbujanja razvoja in

uporabe umetne intelligence v Republiki Sloveniji do leta 2025. DOI:

https://www.gov.si/teme/digitalizacija-druzbe/

[5] Deterding, S., Dixon, D., Khaled, R., in Nacke, L. (2011). From game

design elements to gamefulness: Defining “gamification”. V: A. Lugmayr,

H. Franssila, C. Safran, in I. Hammouda (ur.), MindTrek 2011, (str. 9–15).

New York, NY: ACM.David Kosiur. 2001. Understanding Policy-Based .

[6] Didakt.um. 2020. Igrifikacija, strokovna podlaga.. DOI:

https://didakt.um.si/oprojektu/projektneaktivnosti/Documents/Igrifikacija

_januar2020_final.pdf

[7] Magajna, L., Čačinovič Vogrinčič, G., Kavkler, M., Pečjak, S., & Bregar-

Golobič, K. (2008). Učne težave v osnovni šoli: koncept dela. Zavod

Republike Slovenije za šolstvo.

[8] Kapp, K. M. (2012). The Gamification of Learning and Instruction: Game-

Based Methods and Strategies for Training and Education. John Wiley &

Sons.

[9] Brown, T. E. (2017). A new understanding of ADHD in children and

adults: Executive function impairments. Routledge.



Malo po nemško z IKT



A Little German with ICT



Lea Tori

Osnovna šola Antona Martina

Slomška Vrhnika

Ljubljana, Slovenija

lea.tori1@gmail.com



svetovnem spletu. Učenci razvijajo zmožnosti uporabe

POVZETEK informacijsko-komunikacijske tehnologije tako, da:



Senekova misel, da se ne učimo za šolo, marveč za  spoznajo in uporabljajo elektronske in spletne slovarje,

življenje, še kako drži. V šolah danes ni predmeta na urniku, kjer  pridobivajo podatke s spleta (spletne strani, video

se ne bi pri pouku posluževali IKT, bodisi zaradi dodatnih razlag, portali, družbena omrežja) in uporabljajo brskalnike in

demonstracije, popestritve pouka ali pa samo kot prikaz tega, da iskalnike v nemščini,



pa čeprav so v realnosti naši koraki pogosto krajši kot tisti naših sodelovanje na daljavo (npr. s pomočjo e-pošte, e- učencev. Navsezadnje je to tudi naša dolžnost, da se tudi mi smo tudi učitelji v koraku s časom razvoja vseh vrst tehnologij,  jo uporabljajo za komuniciranje v nemščini in



naučimo in spremljamo novosti, ki so nam na razpolago, video konference, bloga, foruma ipd.), telefona, spletnega socialnega omrežja, klepetalnice,

zanimivo pa je tudi, da nas naši učenci v znanju uporabe IKT  pridobivajo gradivo o državah, v katerih govorijo marsikje že prekašajo, kar pa je za nas lahko tudi pozitivna

nemško, in o Sloveniji iz tujih virov,

izkušnja, saj se iz tega tudi sami lahko veliko naučimo. Za  izdelajo in objavljajo svoje izdelke in sporočila, •

življenje.

predstavljajo svoje izdelke v nemščini (grafično,



KLJUČNE BESEDE slikovno, pisno, zvočno, večpredstavno),

 se vključujejo v mednarodne mrežne projekte in šolska

Učenje za življenje, v koraku s časom, pozitivne izkušnje partnerstva. Uporabo IKT lahko vključimo tudi v

razvijanje sporazumevalne zmožnosti v nemščini.

ABSTRACT Pouk drugega tujega jezika nemščine kot neobveznega

Seneca's idea that we don't learn for school, but for life, is izbirnega predmeta v drugi triadi ter kot obveznega izbirnega

still true. In schools today, there is no subject on the schedule predmeta v tretji triadi na šoli poteka pred poukom ali po pouku.

where ICT is not used in lessons, either for additional Večinoma se za drugi tuj jezik odločajo tisti učenci, ki imajo

explanations, demonstrations, to enrich lessons, or just to show željo po učenju novega jezika, tisti, ki vedo, da bodo ta jezik

that teachers are also keeping up with the times of development imeli kot obvezni jezik v srednji šoli ter tudi tisti učenci, čigar

of all kinds of technologies, even though in in reality, our steps starši jih na nek način usmerijo k učenju dodatnega jezika, saj

are shorter than those of our students. After all, it is also our duty vedo, da jim bo v življenju to znanje prišlo prav. Na žalost pa so

to learn and follow the innovations that are available to us, and it v skupine nemščine vpisani tudi učenci, ki jih starši na nek način

is also interesting that our students already surpass us in the »prislijo« k učenju, morda zaradi tega, ker si želijo, da bi njihov

knowledge of using ICT in many places, which can also be a otrok znal čim več jezikov ali pa zato, ker sami nikoli niso imeli

positive experience for us, because we can learn a lot about this možnost učiti dodatnega tujega jezika. Toko so v vsaki skupini

ourselves. For life. učenci z različni predznanji in interesi, kar pomeni, da je

potrebno med poukom izvajati učno diferenciacijo in to sploh ne

KEYWORDS bi smelo predstavljati nikakršne težave – prav nasprotno!

Learning for life, keeping up with the times, positive Učitelju uspešno izpeljati strukturirano ter kvalitetno učno uro v

experience mešani učni skupini je res neke vrste izziv, a vendar je potrebno

ta izziv pogledati iz Takšnega zornega kota, da se tudi učitelj sam

kaj nauči in pozitivne izkušnje nadgrajuje, negativne pa vzame

1 FERTIG…LOS! kot nekaj, kar je možno vedno izboljšati. Torej »Fertig…los!«

V učnem načrtu za nemščino kot prvi tuji jezik v osnovni šoli (Pripravljeni, pozor, zdaj!, op.a.) Ena izmed stvari, ki učitelju pri

je uporaba IKT vključena v razvijanje digitalne pismenosti. tem ogromno pomagajo, pa je prav uporaba IKT.

Učenci pri pouku nemščine kritično uporabljajo informacijsko- Kljub vse večji težnji in potrebi po uporabi IKT pri pouku, je

komunikacijsko tehnologijo za pridobivanje, vrednotenje in potrebno pri izvedbi učne ureimeti tudi neko zmernost, saj

shranjevanje informacij, za njihovo tvorjenje, predstavitev in prepogosta uporaba pogosto zasenči prvotni cilj pouka, torej

izmenjavo ter za sporazumevanje in sodelovanje v mrežah na pridobivanje in utrjevanje znanja, ob prekomerni uporabi IKT pa

so učenci na koncu pogosto preveč pod vplivov nekih animacij,

distraktorjev in kasneje nekako ne znajo potem izluščiti bistva ob zabavi pri reševanju utrjuje tako besedišče, izgovorjava ter

učne snovi. Še vedno je pri pouku smiselno vpeljati tudi druge pravopis. Največkrat je bilo učencem med ponujenimi

učne metode in pripomočke, predvsem ob razlagi zapletenejših možnostmi ponujen tudi zapis, ki je enak izgovorjavi, kjer so

nemških slovničnih struktur, ko mora biti učenčev fokus učenci morali dvakrat premisliti, ali se beseda dejansko tako

usmerjen v razumevanje in logično povezovanje že pridobljene z napiše, kot se izgovori.

novo osvojeno učno snovjo. Popoldanske ure drugega tujega

jezika znajo biti za učence včasih naporne, koncentracija jim



pade hitreje kot pri rednem pouku, misli pa jim uhajajo na igrišče



ter k počitku doma. Kot učitelj je potrebno najti prav pristop, da

se pritegne pozornost učencev, hkrati pa imeti v mislih tudi to,

da se v teh urah fizične prisotnosti pri pouku učenci že samo s

poslušanjem naučijo največ. In prišlo je leto 2020. Pred zaprtjem

šol se je učence pogosto nagovarjalo o tem, kako je lepo

popoldne preživeti ob športu, druženju s prijatelji, sprehodih v

naravi in ne ob celovečernem buljenju v ekrane računalnikov,

tablic ali mobilnih telefonov. Kar naenkrat pa se je začelo

drugačno leto – leto izzivov, tako za učence, kot tudi za nas

učitelje, vodstvo šole in starše. Pa vendarle se je to leto v smislu



pridobivanja znanja IKT na koncu izkazalo kot zelo produktivno. Slika 1: Posnetek zaslona spletne strani E-učbeniki za

nemščino



2 WIR, DEUTSCH UND IKT



Postopoma se je po epidemiji pri vseh predmetih začelo več



uporabljati IKT, sploh na začetku, ko je bil pouk deloma še na

daljavo. Tudi pri nemščini so učenci najprej morali osvojiti

znanje uporabe ZOOM-a in okolja TEAMS- ter spletnih učilnic.

Popoldanski čas učiteljev je bil namenjen iskanju novih idej z

vključitvijo IKT, veliko primerov dobre prakse pa smo si učitelji

nemščine izmenjali na študijskih srečanjih v okviru Zavoda za

zaposlovanje ter na večernih »online« srečanjih v lastni

organizaciji, kjer smo imeli na razpolago čas za predstavitev

novih idej, predlogov za pouk ter seveda pogovoru ter refleksiji.

Vse aktualno in v okviru zmožnosti smo učitelji postopoma

prenesli tudi na pouk predmeta, ki ga poučujemo na svojem

delovnem mestu. To se ni zgodilo kar čez noč, saj smo se morali

učitelji najprej sami spopasti s programi, aplikacijami ter

spletnimi mesti, ki ponujajo pri poučevanju nekaj več, predvsem Slika 2: Posnetek zaslona vaje v E-učbenikih

pa imeti v mislih, da je ponujeno primerno starosti in znanju

učencev.





2.1 E-učbeniki



Kot dodatek klasičnemu učbeniku so nam na voljo tudi e-

učbeniki (sliki 1 in 2), ki so nastali v sodelovanju profesorjev

nemščine z namenom, da ponujajo dodatne vaje različnih vrst in

različnih stopenj težavnosti, dodatna besedila za branje in vaje

ob branju ter vaje za slušno razumevanje. Odlična izbira so

predvsem zato, ker imajo na spletu do njih dostop vsi, tako da

lahko učenci tudi sami doma samostojno delajo z njihovo

pomočjo ter si tudi sami pregledajo rešene vaje s pomočjo rešitev.



2.2 Kahoot!

Ta aplikacija je pri nemščini zelo popularna, saj se jo učenci



zelo hitro naučijo uporabljati. Večinoma se ga poslužujejo ob Slika 3: Primer vprašanja pri Kahoot kvizu koncu učnega sklopa za utrditev učne snovi, pogosto pa učence

kar presenetim z nenapovedanim kvizom za dvig adrenalina

(sliki 3 in 4). Z učenci je potrebno vzpostaviti zaupanje v njihovo

uporabo mobilnih telefonov, zato so jih po predhodnem

dogovoru z njimi in vodstvom šole tudi lahko uporabljali pri teh

kvizih. Pomembno pri sestavi vprašanj za kvize je bilo to, da se

pregovorov, posameznih besednih zvez, ki jih v slovenščini



uporabljamo v drugačnem pomenu oz. z drugačnimi izrazi.





Slika 6: Primer nesmiselnega prevoda v Google

Prevajalniku



Slika 4: Reševanje Kahoot kviza pri pouku nemščine



2.3 Duolingo

Ker imamo v šoli dobro opremljeno računalniško učilnico,

večkrat izkoristimo priložnost za pouk v njej, hkrati pa imajo

učenci možnost samostojnega dela z računalnikom pod

nadzorom. Na šoli pa imamo tudi petnajst tabličnih računalnikov,

ki jih lahko uporabljamo po dogovoru tudi v matičnih učilnicah,

kar je seveda zelo praktično. Priljubljena aplikacija za učenje

jezika je seveda zelo znan Duolingo (slika 5), ki omogoča učenje

več jezikov, pri nemščini pa izvajamo tudi medpredmetno

Slika 7: Primer napačnega prevoda v Google Prevajalniku

povezovanje, zato je Duolingo odlična pomoč pri primerjavi



besedišča več jezikov. Ker je tudi vizualno zelo privlačno Učencem je vedno na voljo brezplačni PONS slovar, trenutno

oblikovan, služi hkrati tudi kot dodatna motivacija. zelo kvaliteten in pregleden slovar v elektronski obliki, ki

zadostuje potrebam osnovnih in srednjih šol. Uporaba ni težka,



največ težav imajo učenci predvsem s kraticami ali okrajšavami,



zato se jim uporabo tega e-slovarja pojasni večkrat, za pomoč pa



imajo tudi zapiske v zvezku, kaj posamezni znaki predstavljajo.



Ker je pogosto uporaba e-slovarja PONS časovno zamudna, se



učenci premalo odločajo za samostojno delo z njegovo uporabo.



Pri nemščini imajo nekatere črke tudi preglase, zato se morajo

učenci naučiti tudi uporabe določenih simbolov na tipkovnici

(slike 8, 9 in 10).





Slika 5: Primer vprašanja v aplikaciji Duolingo





2.4 Google vs. PONS

Učenci se pri prevodih, še posebno daljših povedi z več

slovničnimi strukturami, pogosto zadovoljijo s hitro rešitvijo –

Googlovim prevajalnikom. Na pogoste napake v prevodu,

nesmiselne besedne zveze ter na splošno izraze brez pravega

pomena je učence potrebno opozoriti večkrat. Še vedno opažam,



nekaj narediti čez noč. Slika 8: Tipke za pomoč pri zapisovanju preglasov da se ga učenci prevečkrat poslužujejo, sploh takrat, ko je treba

Na slikah 6 in 7 sta prikazana primera, ko Google Prevajalnik

ne prevede pravilno in pogosto se zgodi, da so ti prevodi

nesmiselni. Največ napak je še vedno pri prevodih frazemov ali

 Padlet,

 Simple Mind,



 Movie Maker,



 www.iVerbs.de

 Termania,

 PowerPoint,

 http://www.wdrmaus.de,

 https://www.liveworksheets.com,

 https://www.youtube.com

 https://interaktivne-vaje.si

Izzivi pri nemščini so bili s pomočjo IKT poleg dejavnosti pri

pouku še naslednji:



 prevodi v sklopu tematskih dni (januar 2024),

 nastop v nemščini na Slomškovem tednu kulture (marec

2024) (slika 11),

 organizacija ogledov na ekskurziji v Avstrijo z

nadarjenimi učenci, ki obiskujejo izbirni predmet

Slika 9: Primer prevoda s slovarjem PONS nemščina (april 2024)

 sodelovanje na projektu Spodbujajmo prijateljstvo

(učenci so posneli kratke filme o prijateljstvu, maj

2024),

 dopisovanje s šolo v Nemčiji preko elektronske pošte

(maj 2024),

 organizacija poti in ogledov na ekskurziji v Celovec

(junij 2024) (slika 12).





Slika 11: Nastop učencev nemščine na Slomškovem tednu



Slika 10: Primer delnega prevoda pesmi s pomočjo PONS kulture



slovarja





2.5 Selbstgemacht (Narejeno samostojno)



Eden izmed ciljev uporabe IKT pri nemščini je torej



spodbujanje uporabe IKT naprav, ki jih imajo učenci v šoli na

razpolago v namene učenja jezika, hkrati pa tudi uporaba novih

ponujenih možnosti na spletu. S tem postanejo učenci vedno bolj

samostojni pri učenju in odločni v smislu, da se sami (tudi doma)

lotijo reševanja nalog brez učiteljeve pomoči. Učenci s tem, ko

naredijo nalogo sami (torej »selbstgemacht«), čutijo

zadovoljstvo, tudi ponosni so nase, naloga učitelja pa je, da

učencem poda kvalitetno povratno informacijo ter seveda

pohvalo za trud, ki ga je posamezen učenec vložil v delo.

Poleg že omenjenih možnosti, se učenci poslužujejo tudi

naslednjih spletnih orodij oziroma spletnih virov: Slika 12: Izlet v Celovec v samoorganizaciji učencev



3 ZAKLJUČEK

Z vsakimi generacijami mlajših učiteljev se v pouk vključuje

več IKT. Prav je tako, na nas ostalih pa je, da se prilagajamo

novim smernicam v digitalnem svetu ter jih na pozitiven način

izkoristimo otrokom in nam v prid. Posebna zahvala gre

ravnateljici OŠ Antona Martina Slomška Vrhnika ge. Sonji

Strgar za spodbudo, navdih, omogočanje uporabe IKT in

priložnost za naše vseživljensko učenje.



LITERATURA IN VIRI

[1] Kač, Lilijana: Smernice za uporabo IKT NEMŠČINA Ljubljana (2016).

Ministrstvo za šolstvo Zavod Republike Slovenije za šolstvo. Dostopno na

naslovu https://www.zrss.si/pdf/smernice-ikt-nem.pdf (5.8.2024)

[2] Lipovec A., M. Krašnja in I. Pesek, 2019. Izzivi in dileme osmišljene

uporabe IKT pri pouku. Univerzitetna založba Univerze v Mariboru.

[3] Digitalne kompetence za državljane. Osem ravni doseganja kompetenc in

primeri rabe. Dostopno na naslovu https://www.zrss.si/pdf/digcomp-2-1-

okvir-digitalnih-kompetenc.pdf (5.8.2024)

[4] Sio (n.d.) E-učbeniki. Dostopno na naslovu https://eucbeniki.sio.si

[5] Varno in spodbudno učno okolje. Zavod RS za šolstvo. Dostopno na

naslovu https://www.zrss.si/stiki-s-prakso/aktualno/varno-

spodbudnoucno-okolje (5.8.2024)

[6] ZRSS: Dvig digitalne kompetentnosti. Dostopno na naslovu

https://www.zrss.si/projekti/dvig-digitalne-kompetentnosti (5.8.2024)

[7] Kahoot. Dostopno na naslovu https://kahoot.com

[8] Pons. Dostopno na naslovu https://sl.pons.com/prevod

[9] Duolingo. Dostopno na naslovu https://www.duolingo.com/learn

[10] Preglasi kako jih zapisujemo. Dostopno na naslovu

https://www.adriatiqa.com/sl/preglasi-kako-jih-zapisujemo

[11] Kako napisati prijeglas (Umlaut) na našoj tipkovnici? - Pišite ä,ö,ü

Dostopno na naslovu https://www.youtube.com/watch?v=NyAsfVls3Yc





Understanding Factors Affecting Artificial Intelligence




Literacy among Higher Education Students



Silvia Torre Cambón Ines Kožuh

1 1 Faculty of Electrical Engineering and Computer Faculty of Electrical Engineering and Computer

Science Science

University of Maribor, Maribor, Slovenia University of Maribor, Maribor, Slovenia

2 2 University of A Coruña Faculty of Social Sciences

A Coruña, Spain University of Ljubljana

silvia.torre@student.um.si / Ljubljana, Slovenia

silvia.tcambon@udc.es ines.kozuh@um.si /

Ines.Kozuh@fdv.uni-lj.si



just education being transformed; AI is transforming the entire

ABSTRACT education sector. Gradually, intelligent systems are

In an era dominated by technological advancements, revolutionizing educational institutions at every level, including

understanding how Artificial Intelligence influences the digital primary and higher education, as well as adult and professional

literacy skills of the younger generation is crucial for educational learning. These systems aim to enhance human learning

and social development. The study aims to examine the effects experiences and facilitate the achievement of learning goals

of sociodemographic characteristics, digital literacy and patterns (Koravuna et al., 2020). Artificial Intelligence has also been

of using artificial intelligence tools on artificial intelligence referred to as the ability of a digital machine to carry out tasks

literacy. This study aspires to provide additional insights for commonly associated with intelligent beings. Its related

educators and artificial intelligence developers, while providing technologies are categorized into various branches, such as

awareness of implementing artificial intelligence technologies to Computer Vision, Speech, Machine Learning, Big Data, and

empower youth to navigate the digital landscape with proficiency, Natural Language Processing (Chiu, 2021; Chiu et al., 2022; Xia

responsibility, and ethical discernment. We employed an online et al., 2022, as cited in Xia et al., 2023). For this study, we will

survey questionnaire among higher education students in Spain. address AI in a broad sense, as a field of Computer Science that

The results revealed that individuals with higher self-reported solves different kinds of problems inspired by human

artificial intelligence literacy showed increased engagement with intelligence, and that may employ different technologies, such as

artificial intelligence tools in leisure and exam preparation. Neural Networks and Machine Learning, as the main ones. In

However, concluding about whether using these tools improves 1956, AI was defined as “the science and engineering of creating

artificial intelligence literacy requires further investigation. intelligent machines” (McCarthy, 2007, as cited in Su et al.,

2023). The primary purpose of these intelligent machines is to



KEYWORDS solve different kinds of problems via Natural Language Processing, Neural Networks, and Machine Learning (Mondal, Artificial Intelligence, higher education, artificial intelligence 2020, as cited in Su et al., 2023). literacy, frequency of usage In its initial conception, "literacy" referred to the capability to

articulate oneself and communicate through written language.



1 Recently, the concept of literacy has been extended to delineate INTRODUCTION skill sets across diverse disciplines, each possessing the capacity Over the past few years, various studies have been conducted to facilitate expression, communication, and access to in parallel with the ongoing development of Artificial knowledge. One example encompasses digital literacy. AI Intelligence (AI) (Xia et al., 2023). As this is a recent subject of literacy is a set of competencies that enable individuals to study, numerous aspects remain to be measured regarding these evaluate AI technologies critically, communicate and collaborate emerging technologies. The existing literature studies often effectively with AI, and use AI as a tool online, at home, and in overlook digital literacy as a crucial factor in AI literacy. In our the workplace. AI literacy is clearly related to other, previously study, we thus sought to examine how users' demographic defined literacies in related fields. These relationships manifest characteristics and habits of using contemporary technologies themselves in several ways; for instance, digital literacy is a affect AI literacy in university students. prerequisite for AI literacy, as individuals need to understand

how to use computers to make further use of AI (Long et al.,



2 2020). In addition to simply adopting the role of AI tool end users, THEORETICAL FRAMEWORK

AI literacy is a set of competencies that enables people to

AI is a field of Computer Science that focuses on creating evaluate, communicate, and collaborate effectively with AI

intelligent systems inspired by human intelligence. The diverse critically (Druga et al., 2021, as cited in Su et al., 2023). For this

applications of AI encompass a range of sectors, including study, we will address AI literacy as the knowledge and

medicine, manufacturing, and, notably, education. Thus, it is not



understanding of concepts related to Artificial Intelligence, and appropriately to topics they care about personally. They also had

the practical ability to use it to obtain specific desired goals. an opportunity to report whether they can leverage skills like

ethical thinking, creativity, empathy, and idea synthesis as they

create AI artefacts.

3 METHODOLOGY The frequency of using AI tools was measured with Likert-

A quantitative research approach was adopted in this study. type answer options, ranging from never to always (Never;

An online survey questionnaire was employed for data collection. Rarely, <11 times a year; Occasionally, 1-3 times a month;

Four research questions led our research: Frequently, 2-6 times a week; Always, >1 a day).

 RQ1: What is the effect of sociodemographic

characteristics (age, gender, and level of studies) on AI 3.2 Sample

literacy? Our study focused on university students aged between 18

 RQ2: What is the effect of digital literacy on AI and 25 years, aiming mainly at undergraduate and Master's

literacy? students (N=28).

 RQ3: What is the effect of the level of interest in AI on Most respondents, constituting a significant 78.6% of the

AI literacy? surveyed group, fell within the age bracket of 19 to 25,

 RQ4: What is the effect of the frequency of using AI representing a dominant age group in our study. The remaining

tools on AI literacy? 21.4% of respondents consisted of individuals aged 26 to 31.

Regarding gender, 46.4% of the participants were identified as

3.1 Measures and the measuring instrument female and 53.6% as male.

The measuring instrument was composed of three parts. The About the maximum level of studies achieved by the

first part was dedicated to demographic questions, the second to participants, the largest representation came from individuals

questions regarding digital literacy, and the third to questions with a Bachelor's degree, constituting 28.6% of the participants.

regarding AI. Following closely, 28.6% of the respondents had completed

Prior to developing the measuring instrument, we prepared a upper secondary education. Additionally, 21.4% of our sample



list of variables as follows: hold a Master's degree. There were also 21.4% of the participants  who came from post-secondary non-tertiary education and Variables related to sociodemographic characteristics college or vocational training. (age, gender, level of studies)



 Digital literacy 3.3 Procedure

 Interest in AI

 The data were collected online, mainly among students of the AI literacy



We used closed-ended questions with clear answer choices homework, exam preparation). Likewise, they had different backgrounds from different countries; this research may have also reached students from  University of Maribor of which some were Erasmus students. Frequency of using AI tools (during free time,

for each category. For age, we provided ranges focusing on other universities, especially Spanish universities, and other

younger generations (less than 18 years,18– types of study centers, such as vocational training students. 25 years, 26 –

The online questionnaire was held from 9th January 2024

30 years, >31 years or more). We measured gender in 3 until 12th January 2024, and it was distributed by WhatsApp

categories (male, female, and other). We followed the links to different chats and groups following a convenience

international ISCED standards (UNESCO Institute for Statistics, sampling method. 2011) to measure the level of studies, allowing for comparisons

across regions. Digital literacy and interest in AI were self-

reported through seven statements with 5-point Likert-type 4 RESULTS

answer options ranging from 1 (Strongly Disagree) to 5 (Strongly

Agree). 4.1 RQ1: Effect of the sociodemographic

AI literacy was measured according to Williams et al (2023) characteristics on AI literacy



through the indicators as follows. Technical AI knowledge where To find an answer to the first research question, we analyzed students had an opportunity to express whether they can define the effects of gender, age, and level of studies on AI literacy (see AI, identify AI systems daily, and articulate what makes them AI. Table 1). Accordingly, we used a Mann-Whitney U Test. The The students also reported to what extent they had a practical results revealed a statistically significant difference between understanding of how AI algorithms (pertaining to each males and females ( U = 30, p = .002, Z = -3.14). Specifically, the curriculum) function, and what is the humans’ role in creating AI males had a higher mean rank (19.00) than females (9.31), systems. suggesting that they scored higher on AI literacy measures on Ability to think critically about the implications of AI allowed average. The results revealed no statistically significant students to report whether they can think critically about the difference in AI literacy between the 18–25 and 26–30 age potential benefits or harms of AI systems and their impact on groups ( p > .05). Likewise, the p-value for level of studies ( p = stakeholders. The students had also possibilities to report to what 0.32) indicated that there was no significant difference in AI extent they can reflect on and discuss ethical issues (e.g., fairness, literacy between people with pre-university studies and those privacy, and responsible design) as they relate to AI. with higher studies. Ability to apply AI knowledge allowed the students to report

to what extent they can employ their AI knowledge and skills



Table 1: Sociodemographic characteristics two groups (U = 26, p = .024, Z = -2.26). The students with a

high interest in AI had a higher mean rank (16.32) than those

Variable with a low interest in AI (7.83), suggesting they scored higher on Value N Mean Ranks

Gender AI literacy measures on average (see Table 4). Male 15 19

Female 13 9.31



Age 18-25 years 22 14.27 Table 4: AI literacy according to AI interest

26-30 years 6 15.33

Level of studies Pre-university 8 12.06

Variable Value N Mean Ranks



Higher studies AI interest Low interest 6 7.83 studies

20 15.48

High interest 22 16.32



4.2 RQ2: Effect of digital literacy on AI

4.4 RQ4: Effect of the frequency of using AI

Literacy

tools on AI literacy



(85.8%, A Mann-Whitney U test was conducted, to investigate the Analyzing the collected results, most of the participants

(see Table 2). relationship between the frequency of using AI tools during free n=24) had self-reported a high level of digital literacy

time or for exam preparation and participants' AI literacy levels.

The participants were categorized into two groups based on their

Table 2: AI literacy according to the digital literacy reported usage: "at least once a week" and "less than once a

week."

Variable Regarding the frequency of use of AI during free time, the Value N Mean Ranks

Digital literacy results revealed a statistically significant difference between both Low level 4 9.75

High level groups (U = 53, p = .04, Z = -2.07). Individuals who reported 24 15.29

using AI tools at least once a week had a higher mean rank

Following the Mann-Whitney U Test, the results did not (17.92) compared to those who used them less often (11.53),

reveal a statistically significant difference between students with suggesting higher AI literacy on average (see Table 5).

low and high levels of digital literacy (U = 29, p = .207, Z = -



1.261). Specifically, the students with high digital literacy had a Table 5: AI literacy according to the frequency of using AI

higher mean rank (15.29) than those with low digital literacy tools during free time (9.75), suggesting that they scored higher on AI literacy



measures on average. However, this difference was not Frequency N Mean Ranks statistically significant ( p > .05), indicating that other factors Less than once a week 15 11.53 might be influencing AI literacy that are not related directly to At least once a week 13 17.92 the levels of digital literacy assessed in this study.

The results suggest that participants exhibited relatively high

Regarding the frequency of use of AI for exam preparation,

levels of both digital and artificial literacy. The mean digital

the results revealed a statistically significant difference between

literacy score of 1.86 out of a maximum of 2 indicates a strong

the groups (U = 38, p = .031, Z = -2.16). The results revealed a

grasp of digital technologies. However, the mean artificial

higher mean rank in the group that reported the highest frequency

literacy score of 21.54 appears notably high considering the

of using AI tools for exam preparation (19.75), compared to

maximum possible value of 30 (see Table 3).

those who used them less often (12.40), suggesting higher AI

literacy on average (see Table 6).

Table 3: digital literacy and artificial literacy



Table 6: AI literacy according to the frequency of using AI

Variable Mean Standard Min. Max. tools for exam preparation

Deviation value value



literacy Less than once a week 20 12.40 Artificial 21.54 4.53 13.00 30.00 At least once a week Digital 4.50 0.84 2 5 Frequency N Mean Ranks 8 19.75

literacy



4.3 RQ3: Effect of the level of AI interest on AI 5 CONCLUSIONS

literacy This study aimed to examine the relationship between various

The descriptive data revealed an interest in AI among factors and AI literacy among higher education students.

respondents, with 78.6% expressing a strong interest. Specifically, it explored the impact of demographic factors,

Meanwhile, 21.5% of respondents expressed disinterest. digital literacy, interest in AI, and frequency of AI usage.

A Mann-Whitney U test was conducted to compare AI After assessing the sociodemographic variables pertinent to

literacy between students with high and low interest in AI. The RQ1, the results revealed a statistically significant difference

results indicated a statistically significant difference between the between males and females in AI literacy, suggesting that male



participants scored higher in AI literacy than women. According ACKNOWLEDGMENTS

to Long and Magerko (2020), diversity in the Computer Science This work would not have been possible without all the

workforce, including AI, remains a significant issue. In 2018, students who participated in the survey, and we extend our

80% of AI Professors and 71% of applicants for AI-related jobs gratitude for their contribution to the study. were men. This lack of diversity can influence the target users of



developed systems, a critical concern in AI, where biased REFERENCES

algorithms can harm marginalized groups disproportionately. [1] Xia, Q., Chiu, T. K. F., Zhou, X., Chai, C. S., & Cheng, M. (2023). The specific field of study of the students was not assessed in this Systematic literature review on opportunities, challenges, and future

study, which could influence the results significantly, especially research recommendations of artificial intelligence in education.

if there was a concentration of Computer Science students. It is Computers & Education: Artificial Intelligence, 4, 100118.

https://doi.org/10.1016/j.caeai.2022.100118

essential to incorporate this variable in future research to obtain [2] Koravuna, S., & Surepally, U. K. (2020). Educational gamification and

more accurate results. artificial intelligence for promoting digital literacy. Proceedings of the 2nd

International Conference on Intelligent and Innovative Computing

While our findings do not show a significant gap in AI Applications, 1-6. https://doi.org/10.1145/3415088.3415107

literacy between the age groups analyzed, or between pre- [3] Su, J., & Ng, D. T. K. (2023). Artificial intelligence (AI) literacy in Early

university and university students, however, other studies Childhood education: the challenges and opportunities. Computers &

Education: Artificial Intelligence, 4, 100124.

suggest that low socioeconomic status (SES) children perform on https://doi.org/10.1016/j.caeai.2023.100124

average worse on AI abilities and AI literacy education than [4] Long, D., & Magerko, B. (2020). What is AI Literacy? Competencies and

Design Considerations. Proceedings of the 2020 CHI Conference on

children with higher SES backgrounds (Druga et al., 2019). Human Factors in Computing Systems. 1-16.

Digital literacy was examined to address RQ2. Our results did https://doi.org/10.1145/3313831.3376727

not show a significant correlation between digital literacy and AI Classification [5] UNESCO Institute for Statistics. (2011). International Standard

of Education (ISCED). Page 75-77.

literacy. However, it is worth noting that most participants https://uis.unesco.org/sites/default/files/documents/international-

expressed a high level of confidence in their abilities in both standard-classification-of-education-isced-2011-en.pdf

[6] Williams, R., Ali, S., Devasia, N. et al. (2023) AI + Ethics Curricula for

areas. Given that participants self-reported their AI literacy, there Middle School Youth: Lessons Learned from Three ProjectBased

is a likelihood of overestimation bias due to the Dunning-Kruger Curricula. International Journal of Artificial Intelligence in Education, 33,

325–383. https://doi.org/10.1007/s40593-022-00298-y

effect. Consistent with this, other research indicates that children [7] Druga, S., Vu, S. T., Likhith, E., & Qiu, T. (2019). Inclusive AI literacy

often overvalue the capabilities of AI agents, possibly reflecting for kids around the world. FL2019: Proceedings Of FabLearn 2019.

a limited understanding of underlying AI technologies (Long and https://doi.org/10.1145/3311890.3311904

[8] Kožuh, I., Blekač, E., Čakš, P., & Lovrenčič Držanič, I. (2023). Exploring

Magerko, 2020). To enhance the reliability and validity of the digital literacy and the use of ChatGPT among students with disabilities.

data, future research should consider utilizing a combination of In Information Society: Proceedings of the 26th International

Multiconference Volume H (pp. 25-28) https://is.ijs.si/wp-

measurement techniques, to obtain a more accurate and objective content/uploads/2023/11/IS2023_Volume-H.pdf

measurement of the participants' actual AI knowledge and

genuine understanding.

Our findings indicate that students with a higher interest in AI

(RQ3) and those who frequently use AI tools (RQ4), especially

for exam preparation, may possess a higher level of AI literacy.

Regarding homework, the AI literacy levels was not statistically

significant as a factor in relation to the frequency of use of

Artificial Intelligence as an assistant with homework. Drawing

conclusions about whether using AI tools improves literacy or

individuals with higher literacy simply use them more requires

further investigation.

Our results are also limited by the small sample size. Further

research is needed to confirm causality (between digital literacy,

interest in AI, frequency of use and AI literacy), and consider

another factor as the area of knowledge of the participants in

which they are specialized.

AI can be a powerful educational tool, offering more

personalized learning experiences and improved accessibility.

Some studies support the importance of the inclusivity of

chatbots to facilitate the learning process, but AI tools also

present ethical dilemmas, such as biased results and an uncritical

use of information (Kožuh et al., 2023).

This research shows a small insight into the relationship of

young students with these new technologies. For further

research, it is also suggested that the uses and gratifications of

these technologies be explored in more depth, exploring which

type of AI is the most used by the population and students.



Pogovorno programerstvo:



pot do boljšega razumevanja izotopov



Conversational Programming:



A Better Understanding of Isotopes



Marina Trost

Gimnazija Vič

Ljubljana, Slovenija

marina.trost@gimvic.org



ostalimi (pretežno) naravoslovnimi predmeti, torej MINUT -

POVZETEK M(atematika) I(nformatika) N(aravoslovje) U(metnost)

V članku je opisujemo primer medpredmetne povezave T(ehnologija) predmete.

kemije in informatike, pri čemer razvijamo temeljna znanja obeh V okviru projekta:

predmetov. Pri tem izpostavimo dejstvo, da je zelo pomembno, • Se redno sestajamo na spletnih (ZOOM) sestankih in

da se kemik zna pogovarjati z informatikom. Predstavljena je izmenjujemo izkušnje

izvedba učnih ur v 1. letniku gimnazijskega programa. Dijaki so • Razvijamo gradiva, kjer se prepletata RIN in še en predmet

z izdelovanjem programa, s katerim so narisali • Se srečujemo na delavnicah

submikroskopsko predstavitev izotopske sestave elementa, bolje • Sodelujemo na konferencah: NAK, STEAMColab,

spoznali funkcije s parametri, tabele in uporabo modula. Z DigiDakta, VIVID

izdelavo lastne slike se jim je izboljšal uvid v zgradbo snovi. • Obiskujemo kolege iz drugih držav, ki poučujejo na

podoben način



KJUČNE BESEDE V sklopu projekta je nastalo že več kot 10 obravnav tem, določene pa so še v razvoju. Za več informacij o projektu se algoritmično razmišljanje, MINT, medpredmetno lahko obrnete na napoj@404.si povezovanje



ABSTRACT kemija 1.1 Povezovanje predmetov informatika in

The article describes an example of interdisciplinary

integration between chemistry and computer science, where Dijaki si težko predstavljajo submikroskopsko zgradbo snovi,



lessons in the first year of the gymnasium program is presented. dostopni na spletu. S povezovanjem teh dveh predmetov želimo pri dijakih utrditi By creating a program that draws a submicroscopic znanje obeh predmetov in v dijakih povzročiti vsaj majhen representation of an element's isotopic composition, students premik v zavedanju, da bodo v poklicu, ki si ga bodo izbrali, pa gained a better understanding of functions with parameters, naj bo kakršen koli, potrebovali sposobnost, da informatiku tables, and module usage. Creating their own image also razložijo kakšen program potrebujejo za rešitev nekega , recimo improved their insight into the structure of matter.. highlights the importance of a chemist being able to da bo vizualizacija še učinkovitejša, če si prikaze dijaki ustvarijo sami s pomočjo želvje grafike. Pri tem uporabijo podatke, ki so communicate with a computer scientist. The implementation of fundamental knowledge of both subjects is developed. It pri tem jim lahko pomagajo grafični prikazi. Predvidevali smo,

kemijskega problema. Za čim bolj učinkovito razlago



KEYWORDS potrebujejo tudi nekaj temeljnega znanja informatike.

Druga želja pri povezovanju je, da bi dijaki videli, da je

Computational thinking, STEM, multi-disciplinary teaching znanje enega predmeta uporabno in potrebno tudi pri drugem.



1.1.1 Ideja projekta

1 UVOD Dijaki že na osnovni šoli spoznajo izotope, na gimnaziji pa to

Učitelji informatike na gimnazijah pogosto iščemo še nadgradijo z računanjem relativne atomske mase elementov s

konkretnega primere iz vsakdanjega življenja ali vsaj iz vsebine pomočjo relativnih atomskih mas posameznih izotopov in

nekega drugega gimnazijskega predmeta, kjer si lahko njihove pojavnosti v naravi. Izkaže se, da ima veliko dijakov

pomagamo z znanji računalništva in informatike. Ker vstopamo težave s submikroskopsko predstavo izotopske sestave, zato jim

v vedno bolj razvijajočo se digitalno družbo, je za šole postalo pri tem lahko pomaga grafičen prikaz za vsak element posebej,

nujno, da se prilagodijo in opremijo prihodnje generacije s ki ga lahko s pomočjo želvje grafike naredijo sami.

potrebnimi znanji in spretnostmi. [2] V projektu NAPOJ MINUT Pri poučevanju informatike na gimnaziji želimo dijakom

se povezujemo učitelji računalništva in informatike (RIN) z omogočiti pridobivanje temeljnih znanj računalništva in

informatike. Zato jih učim programiranja, pri čemer je primer v program. Vsak par dijakov je izbral svoje barvne sheme in,

risanja grafične predstavitve odličen za utrjevanje številnih glede na pridobljene podatke, zapisali zanko, ki je generirala

konceptov programiranja. Naloga vključuje uporabo funkcij s tabelo z ustreznim številom barv. Za naključno premešanje barv

parametri, tabel in modula random. so uporabili funkcijo shuffle iz modula random.



Na slikah 2 in 3 sta primera izdelkov dveh dijakov. Večina

dijakov se je odločila za izris v obliki piramide, nekateri pa so

izbrali drugačne geometrijske oblike, na primer kvadrat.





Slika 1: Zn



2 O PROJEKTU



2.1 Operativni učni cilji

V sklopu projekta smo zastavili naslednje učne cilje.

Dijaki bodo:

• sposobni razviti algoritem v psevdo kodi,

• sposobni razmišljati o potrebnih iteracijah zanke in

ustrezno prilagoditi algoritem,

• sposobni zapisat funkcijo v programskem jeziku,

• sposobni zapisat tabelo in elemente z uporabo modula

random naključno premešat, Slika 2: primer kode programa (Mg)

• spoznali osnovne ukaze grafične knjižnice,



• znali bodo pojasniti, kako se izotopi istega elementa

razlikujejo glede na atomsko maso, kljub enakim

kemijskim lastnostim,

• znali bodo interpretirati grafični prikaz izotopske sestave

elementov.



2.2 Načrtovanje projekta

Pred samo izvedbo projekta medpredmetnega sodelovanja je

pomembno, da se učitelji uskladimo, kakšno mora biti

predznanje dijakov pri vseh sodelujočih predmetih. S tem



Pogosto se pokaže, da so potrebne manjše prilagoditve v Slika 3: primer izdelka (Ti) določimo tudi časovne okvire v katerih lahko izvedemo projekt.

programih posameznih predmetov, kot tudi morebitno

Za izvedbo projekta pri kemiji ni bilo potrebnih dodatnih ur,

spremembo operativnih učnih ciljev projekta.

saj je bila vsebina že vključena v redni učni načrt. Prav tako so



elementa enaki oziroma, da ima večina elementov vsaj dva ali bile aktivnosti pri informatiki v skladu z učnim programom. Pri pouku kemije so se dijaki naučili, da niso vsi atomi istega

Celoten projekt je obsegal približno 5 učnih ur, vendar niso bile

več izotopov. V okviru pouka informatike smo najprej

izvedene v neprekinjenem zaporedju, saj smo vmes potrebovali

obravnavali temeljne koncepte programiranja, vključno z

dodatne vaje za utrjevanje osnov funkcij in tabel. Projekt je bil

zaporedjem ukazov, uporabo spremenljivk in zankami.

uspešno izveden v dveh različnih razredih.

Nadaljevali smo z osnovami želvje grafike in funkcijo za izris

Pri utrjevanju tabel in funkcij smo uporabljali pripravljene

kroga.

naloge iz projekta Tomo [3]. Dodali smo tudi nekaj novih nalog,



2.3 ki smo jih potrebovali med projektom. Izvedba

Skupaj smo zasnovali algoritem za izris prve vrstice krogcev

brez barv. V naslednji učni uri so dijaki v parih ta algoritem 3 REFLEKSIJA



implementirali v programskem jeziku Python. Nekateri Med izvajanjem aktivnosti se je pokazalo, da dijaki zelo naprednejši dijaki so program razširili na več vrstic, vendar sprva dobro sledijo procesu izdelave algoritma, kadar ga razvijamo nihče ni uporabil funkcij. Funkcijo smo definirali v kasnejši fazi skupaj. V drugem razredu sem preizkusila drugačen pristop: z in jo uporabili za izris celotne slike, pri čemer so dijaki spoznali dijaki smo sprva ustno načrtovali izris ene vrstice krogcev, nato uporabnost funkcij in pomen smiselne uporabe parametrov. pa so dijaki v parih samostojno zapisali algoritem. V naslednji V nadaljevanju smo se posvetili barvanju krogcev, kar je učni uri smo zapise algoritmov zamenjali med pari, tako da je zahtevalo definiranje tabel in postopek za izračun deleža vsak par dobil algoritem od sošolcev, njihova naloga pa je bila, izotopov v ustrezno število krogcev. da ta algoritem implementirajo v programskem jeziku Python. Dijaki so na ustrezni spletni strani pridobili podatke o Izkušnja je pokazala, da je bila ta naloga za mnoge dijake izotopski sestavi izbranega elementa [1], ki so jih nato vključili

pretežka. Nekateri so zapisali svojo različico algoritma, medtem kaže na uspešnost medpredmetnega pristopa pri izboljšanju

ko so bili drugi zmedeni in niso uspeli sestaviti niti svojega niti njihovega dojemanja predmeta.

sošolčevega algoritma. Na koncu ure smo zato skupaj zapisali

eno izmed možnih rešitev.

Po izvedbi aktivnosti smo se temeljito posvetili analizi anket,

ki so jih dijaki izpolnili pred in po aktivnosti. Rezultati

predhodnih anket so pokazali precejšen skepticizem in

nezainteresiranost dijakov glede povezovanja različnih

predmetov, kar je izpostavilo izziv pri vzpostavljanju

pozitivnega odnosa do interdisciplinarnega pristopa. Dijaki so

sprva izražali dvome o smiselnosti in uporabnosti tovrstnega

načina dela. To je potrdilo naše prepričanje o potrebnosti

dodatnih prizadevanjih za izboljšanje njihovega dojemanja

koristi takšnega poučevanja.





Slika 6: Oceni svoj odnos do predmeta, kjer boste uporabili

računalništvo (kemija, fizika ...)



Po aktivnosti je sicer odgovarjalo malo manj dijakov, vendar

jim je predmet bolj zanimiv, manj težak ipd.

Podrobnosti se vidijo na sliki 7.



Slika 4: Bi rad(a), da se tak pristop k snovi, ki ga boste

izvedli (prepletanje "druge" snovi in računalništva) še

večkrat izvaja



Kljub temu so povratne informacije po izvedeni aktivnosti

pokazale premik v razmišljanju dijakov, saj so nekateri

prepoznali vrednost v povezovanju različnih disciplin in izrazili

večje zanimanje za tovrstne projekte v prihodnje.





Slika 7: Oceni svoj odnos do predmeta, kjer ste uporabili



računalništvo (kemija, fizika ...)





Z veseljem opažamo, da so odgovori dijakov na vprašanja o

pridobljenih izkušnjah iz te aktivnosti v večini primerov

spodbudni, čeprav je na to vprašanje odgovorilo le 15 dijakov.

Le dva dijaka sta izrazila negativno mnenje, navajajoč, da jim je



Slika 5: Bi rad(a), da se tak pristop k snovi, ki ste ga izvedli tovrstna aktivnost povzročala več zmede in stresa. Nasprotno pa je 12 dijakov prepoznalo več prednosti, kot so: pridobitev (prepletanje "druge" snovi in računalništva) še večkrat dodatnega znanja, globlje razumevanje snovi, nov pristop k izvaja?

reševanju problemov in bolj zanimiv pouk. Posebej zanimiv je

Dijaki so ocenili, da je uporaba računalništva pri kemiji bil komentar enega izmed dijakov, ki je predlagal, da bi bilo

zapletena in težka, pri čemer so navedli, da je tak pristop koristno prilagoditi zahtevnost programov glede na predznanje

nerazumljiv, nesmiseln in včasih celo nemogoč. Le 10 od 51 dijakov, saj je to med njimi precej različno.

dijakov je izrazilo pričakovanja, da bodo pridobili novo znanje,

poglobljeno razumevanje obeh predmetov in da bo delo tudi

zabavno. Po izvedbi projekta se je mnenje, da je računalništvo 4 ZAKLJUČEK

zahtevno, še okrepilo, kar je bilo v nasprotju z našimi začetnimi Medpredmetno povezovanje kemije in informatike se je

predvidevanji. Po premisleku pa smo ugotovili, da je to izkazalo kot koristna strategija za poglabljanje razumevanja tako

pričakovano, saj so dijaki poleg osnov programiranja, s katerimi kemijskih kot tudi računalniških konceptov. Z aktivnostmi, kjer

so se srečali pred projektom, morali obvladati tudi naprednejše so dijaki ustvarjali grafične predstavitve izotopske sestave

koncepte, kot so tabele in funkcije. elementov, smo uspeli doseči več ciljev. Ne samo, da so dijaki

Pokazalo se je, da se je odnos dijakov do kemije po izvedbi poglobili svoje razumevanje kemije in informatike, temveč so se

aktivnosti pozitivno spremenil. Medtem ko so bili odgovori pred tudi spoprijeli z izzivom uporabe programiranja kot orodja za

aktivnostjo, prikazani na sliki 6, bolj zadržani in negativni, so reševanje kemijskih problemov.

rezultati po aktivnosti, prikazani na sliki 7, pokazali izboljšanje. Rezultati ankete so pokazali, da kljub začetnemu skepticizmu,

Dijaki so kemijo dojemali kot bolj zanimivo in dostopno, kar ki so ga dijaki izražali pred izvedbo projekta, večina po izvedbi



ceni prednosti takega integriranega pristopa. Pozitiven premik v

odnosu dijakov do kemije in informatike nakazuje na

pomembnost nadaljevanja in razvijanja podobnih projektov, pri

čemer je ključno prilagajanje zahtevnosti nalog glede na različno

predznanje dijakov.

Kljub izzivom, kot je občutek povečane težavnosti

informatike, je medpredmetno povezovanje smiselno, saj

izboljša njihovo razumevanje učnih vsebin, kar je tudi eden

izmed dolgoročnih ciljev projekta NAPOJ MINUT.

Pomemben poudarek projekta ostaja na sposobnosti dijakov,

da s pomočjo temeljnih znanj informatike in naravoslovja

prepoznajo uporabnost enega predmeta pri drugem, kar je

bistveno za njihovo prihodnjo poklicno pot, kjer bodo morali

znati komunicirati z različnimi strokovnjaki, vključno z

informatiki. Med načrtovanjem in izvedbo projekta je bilo

potrebno veliko pogovorov med učiteljem informatike in

učiteljem kemije, da sva se pravilno razumeli. Zadnjo uro pouka

v tem projektu sva bili pri pouku informatike prisotni obe

učiteljici, da smo zadevo zaključili z obeh zornih kotov.

Nadaljnje raziskave in izboljšave pri medpredmetnem

povezovanju lahko še naprej prispevajo k boljšemu

izobraževalnemu procesu in pripravijo dijake na izzive digitalne

dobe.

Vse navedeno poudarja potrebo po poglobljenem in

kontinuiranem razvoju medpredmetnih povezav, ki dijake ne le

učijo vsebin, ampak tudi razvijajo pomembne veščine, kot so

algoritmično razmišljanje, kritično mišljenje in

interdisciplinarno sodelovanje.



VIRI

[1] Atomic Weights and Isotopic Compositions for All Element (avgust,

2024) DOI: https://physics.nist.gov/cgi-

bin/Compositions/stand_alone.pl?ele=&all=all

[2] Bridging the Gap: Infusing Natural Science Classes with Computer

Science Concepts and Skills - na koncu Springer proceedings

(doi:10.1007/978-3-031-44900-0)

[3] M. Pretnar, „Spletna storitev za poučevanje programiranja,“ v Vzgoja in

izobraževanje v informacijski družbi – VIVID 2014 , Kranj: Fakulteta za

organizacijske vede, 2014



Z umetno inteligenco nad domače branje



Artificial Intelligence on Top of Home Reading



Eva Traven

OŠ Predoslje Kranj, Slovenija

eva.traven@guest.arnes.si



izkoristimo potencial umetne inteligence (UI) za izboljšanje učne

POVZETEK izkušnje. Umetna inteligenca je namreč že močno prisotna v

V prispevku je predstavljen primer iz prakse, in sicer življenju učencev, zato je ključno, da prepoznamo njen potencial

obravnava domačega branja pri pouku slovenščine s pomočjo za ustvarjalno in učinkovito poučevanje. V članku je predstavljen

umetne inteligence. Cilj dejavnosti je, poleg predpisanih ciljev primer iz prakse, in sicer uporaba umetne inteligence in

učnega načrta, razvijati tudi različne spretnosti in sposobnosti pri digitalnih platform za spodbujanje aktivnega branja in povečanje

učencih, npr. komunikacijske, informacijske in raziskovalne, literarne ter digitalne pismenosti med učenci. Obravnavali bomo

hkrati pa vzpostavljati pozitivno izkušnjo z literaturo in s tem pri primer praktične uporabe teh tehnologij in njihov potencial za

učencih vzbujati zanimanje za branje. S sodobnim pristopom ter izboljšanje učne izkušnje pri pouku.

približevanjem zanimanjem in trendom mladih lahko učitelji



pouk književnosti naredijo učencem privlačen. Učenci so za rabo 2 RABA UI PRI SLOVENŠČINI različnih tehnologij v zasebni rabi, tako za komunikacijo kot

zabavo, zelo motivirani. Vse te nove tehnologije imajo lahko tudi Marija Žveglič v svojem raziskovanju rabe IKT pri pouku

v šoli veliko moč, če so kakovostno osmišljene in povezane s cilji slovenščine ugotavlja, da so učenci v prostem času zelo

in vsebinami učenja. motivirani za rabo različnih tehnologij, tako na ravni

komunikacije kot tudi zabave; zastavlja pa se vprašanje, ali

KLJUČNE BESEDE zmoremo njihovo zasebno navdušenost na šolskem polju

Keywords in the language of the paper separated by commas kakovostno osmisliti ter povezati s cilji in z vsebinami učenja oz.,

ali smo glede na dejstvo, da je tehnološki razvoj v velikem



ABSTRACT zagonu, pripravljeni premisliti in prevrednotiti obstoječo



artificial intelligence into Slovenian language classes to support This article presents a practical example of integrating tehnologij [1]. Kot poudarja učni načrt za slovenščino (2019), se pedagoško prakso v luči ustvarjalne in učinkovite rabe novih

razvijanje digitalne zmožnosti povezuje z razvijanjem

home reading assignments. In addition to the prescribed sporazumevalne zmožnosti v slovenskem jeziku. Poleg spletnih

curriculum objectives, the aim of this activity is to develop učnih okolij, kot so forum, klepetalnica, uporaba anket ali drugih

various skills and abilities in students, such as communication, odzivnih sistemov za razne oblike preverjanja, videokonferenca

information literacy, and research skills. Furthermore, it seeks to za sodelovalno učenje itd., učni načrt omenja tudi uporabo

establish a positive experience with literature, thus fostering an družabnih omrežij, ob čemer naj učitelj učence tudi ozavešča o

interest in reading among students. By adopting a modern odgovornem poseganju v splet in o odgovorni rabi informacijske



people, teachers can make literature lessons more appealing. Uvajanje umetne inteligence v izobraževanje je prineslo Students are highly motivated to use different technologies in številne spremembe in priložnosti za izboljšanje izobraževalnega approach that aligns with the interests and trends of young tehnologije ter jih uči, kako lahko sami prispevajo za splet [2].



their private lives for both communication and entertainment. procesa. Z umetno inteligenco lahko razvijemo bolj prilagojene When thoughtfully integrated and aligned with learning in personalizirane učne dejavnosti, ki ustrezajo individualnim objectives and content, these new technologies can also hold potrebam učencev [4]. Kljub hitremu vdoru umetne inteligence great potential in the classroom. v naše učilnice, priložnosti, ki jih bo le-ta ustvarila za naše šole,

daleč presegajo tveganja, ki jih te tehnologije prinašajo.

KEYWORDS Prilagajanje šolstva UI je velik izziv, vendar lahko z ustrezno

Slovenian language, literature, home reading, artificial znanjem in veščinami učitelji pristopijo k temu novemu obdobju

inteligence, Chat GPT pripravljeni, da se osredotočijo na to, kar najbolje delajo:

poučevanje naših učencev. Kljub temu pa moramo, tako kot pri



1 UVOD prilagajamo našo prakso in pedagogiko v odzivu na umetno vsakem hitrem razvoju, skrbno premisliti, kako, kdaj in zakaj

inteligenco. V mnogih primerih namreč lahko umetna inteligenca

V sodobnem izobraževalnem okolju se učitelji nenehno izboljša naše obstoječe prakse in nam tako pomaga učinkoviteje

soočamo z izzivom motiviranja učencev za branje. V kontekstu doseči naše izobraževalne cilje [3].

poučevanja slovenščine se pojavlja vprašanje, kako lahko



V začetku poletja 2024 je Ministrstvo za šolstvo RS objavilo ga kot orodje za pomoč učencem pri oblikovanju obnov

predlog nacionalnega programa vzgoje in izobraževanja 2023- prebranih zgodb. Učenci so najprej prebrali knjigo in nato s

2033. Nacionalni program digitalni pismenosti namenja nekaj pomočjo Chat GPT ustvarili obnove posameznih zgodb.

pozornosti, vendar pri tem vsebine umetne inteligence umešča Predhodno branje izhodiščnih besedil je bilo pri tem pomembna

predvsem v srednje šole [5]. Ker pa se otroci z njo srečajo že osnova, saj so učenci le ob natančnem poznavanju vsebine lahko

dosti prej, vsaj z razvpitim Chatom GPT, bo dobrodošel kritično presojali rezultate, ki so jih dobili z uporabo UI. Med

razmislek, kako jim predstaviti delček teh vsebin tudi že v samo dejavnostjo v računalniški učilnici so bili zato razdeljeni v

osnovni šoli. Primer takšne aktivnosti podajamo v nadaljevanju. manjše skupine, da so lahko med sabo primerjali razumevanje

Domače branje je sestavni del pouka slovenščine, ki ga vsebine knjige in dobljenih rezultatov. Vsaka skupina je imela

številni učenci dojemajo kot nepotrebno breme. Kljub velikim nalogo v Chat GPT kopirati celotno zgodbo posameznih enot

prizadevanjem učiteljev, da bi literaturo približali učencem in jih izbrane knjige (obe književni deli sta v celoti objavljeni na

navdušili za branje, se pogosto zgodi, da učenci ne izpolnijo platformi Wikivir) ter od Chata zahtevati njeno obnovo. Dobljeni

pričakovanj. Namesto da bi se poglobili v knjige in jih prebirali rezultati so se od skupine do skupine precej razlikovali, nekatere

z zanimanjem, se mnogi odločijo za prepisovanje obnov, kar vodi obnove so bile precej dobre, druge bistveno manj, kar nam je

do slabih rezultatov na testih preverjanja znanja in občutka služilo tudi kot iztočnica za pogovor o napakah in pasteh uporabe

neuspeha tako pri učitelju kot pri učencih. Učitelji se pogosto Chat GPT. Ta pristop je učencem omogočil, da so se osredotočili

soočamo z izzivom, kako učence motivirati za branje, da bi na ključne elemente zgodbe in razvijali svoje sposobnosti analize

dosegli boljše razumevanje literature in bogatejšo izkušnjo in sinteze informacij. Po ustvarjanju obnov so učenci nato svoje

branja. delo predstavili s pomočjo spletne platforme Book Creator, ki

Pomanjkanje motivacije za branje med učenci je velik izziv, omogoča enostavno oblikovanje digitalnih knjig. Aplikacija

zato je še toliko bolj pomembno, da se učitelji slovenščine omogoča učiteljem in učencem, da na preprost način ustvarjajo

trudimo domače branje narediti čim bolj privlačno. Veliko truda in delijo interaktivne knjige. Deluje na različnih napravah,

in energije je potrebno, da učencem poleg obveznih kanonskih vključno s tablicami, prenosnimi računalniki in mobilnimi

del ponudimo tudi literarna dela, ki so jim po vsebini bližja in telefoni. Uporabniki lahko v svoje knjige vključijo različne

zanimivejša, ter jih obogatimo z dejavnostmi, ki jih naredijo multimedijske vsebine, kot so besedilo, slike, video posnetki,

privlačnejše in je tako obravnava literarnih del čim bolj raznolika zvok in risbe.

in inovativna. V razredu uporabljamo različne metode in pristope, Platforma podpira vlečenje in spuščanje, kar olajša urejanje

ki vključujejo interaktivne dejavnosti, diskusije, igranje vlog in strani in dodajanje vsebin. Ena od pomembnih funkcij je možnost

kreativno pisanje. Pri obravnavi knjig pogosto posegamo tudi po sodelovanja, kar pomeni, da lahko več učencev hkrati dela na isti

sodobnih orodjih in tehnologijah, kar omogoča bolj dinamično in knjigi, kar spodbuja timsko delo. Book Creator ponuja tudi

zanimivo učno izkušnjo. Na primer, uporabljamo digitalne številne predloge, ki uporabnikom pomagajo začeti ustvarjanje

platforme za skupinsko delo, spletne kvize za preverjanje znanja knjig (slika 1). Učitelji lahko nadzorujejo in upravljajo delo

in multimedijske vire za poglobljeno razumevanje konteksta svojih učencev, dodeljujejo naloge ter ocenjujejo njihovo delo

literarnih del. neposredno na platformi. Platforma podpira tudi izvoz knjig v

S takim pristopom želimo učencem približati literaturo na formate PDF in ePub, kar omogoča enostavno deljenje in tiskanje.

način, ki je zanje bolj dostopen in zanimiv. Verjamemo, da je Brezplačna različica omogoča osnovno uporabo platforme z

ključ do uspeha v prilagodljivosti in inovativnosti pri poučevanju, določenimi omejitvami, medtem ko plačljive različice ponujajo

saj lahko s tem pripomoremo, da bodo učenci razvili ljubezen do dodatne funkcionalnosti in zmogljivosti.

branja in literarne kulture. Na ta način dosežemo, da domače Vsak učenec je s pomočjo omenjenega programa oblikoval

branje postane manj obremenjujoče in prijetnejše doživetje za svojo digitalno knjigo, ki je vključevala obnovo zgodbe in

vsakega učenca, kar bo dolgoročno prispevalo k njihovemu ilustracije. Pri tem smo šli še korak dlje, saj smo slike ustvarjali

celostnemu razvoju in izobraževanju. s pomočjo programa Microsoft Copilot Designer (prej Bing

Image Creator). Gre za orodje, ki uporablja umetno inteligenco

za ustvarjanje slik na podlagi besedilnih opisov. Temelji na

3 PRIMER IZ PRAKSE tehnologiji DALL-E, ki jo je razvilo podjetje OpenAI. Designer



3.1 Domače branje malo drugače Microsoftovim računom. Ustvarjanje slik je učencem omogočilo, je brezplačno dostopen, vendar je potrebna prijava z

Med deli, ki se jih učenci za domače branje lotijo z manj da so vizualno predstavili svoje razumevanje zgodb. V program

navdušenja, sta deli iz obdobja slovenskega realizma, zbirki slik so vnesli opis slike, ki so jo želeli ustvariti. Opis je lahko

Med gorami Ivana Tavčarja ter Kmetske slike Janka Kersnika. vključeval podrobnosti o predmetih, okoljih, barvah in slogih. Za

Ker gre za starejša literarna dela, jim pri branju dela težave že dejavnost je bilo potrebno nekoliko več časa, kot smo načrtovali,

sam jezik, prav tako pa jim zgodbe niso blizu zaradi povsem saj so učenci potrebovali veliko pomoči pri ustvarjanju

drugačnega izkušenjskega sveta. V iskanju možnosti, kako kvalitetnih ukazov za generiranje slik. Nekaj dela, predvsem

spodbuditi učence k aktivnemu branju in sodelovanju, smo se uvodno seznanjanje z obema aplikacijama, smo zato opravili

osredotočili na inovativen pristop s pomočjo umetne inteligence. skupaj, v računalniški učilnici, učenci pa so svoje izdelke nato

dokončali doma. Ob zaključni uri smo si nato skupaj ogledali

3.2 Potek dejavnosti predstavitve izdelkov ter si v pogovoru izmenjali mnenja ter vtise

Chat GPT je napreden jezikovni model, ki omogoča o izvedeni dejavnosti in uporabljenih AI pripomočkih.

generiranje koherentnih in informativnih besedil. Uporabili smo





Slika 1: »Knjižnica« v programu Book Creator



4 Zaključek

Uporaba umetne inteligence v okvirih programov Chat GPT,

Book Creator in Copilot Designer je učencem omogočila, da so

se na kreativen in interaktiven način poglobili v literarno delo ter

hkrati razvijali veščine digitalne pismenosti in kreativnega

izražanja. Tak pristop spodbuja aktivno učenje in povečuje

motivacijo učencev za sodelovanje pri domačem branju, saj jim

omogoča, da literaturo doživijo na sodoben in zanimiv način.

Dodano didaktično vrednost opisane dejavnosti vidimo

predvsem v drugačnem pristopu, ustvarjanje je bilo učencem

zabavno, saj se jim je pri tem utrnilo ogromno idej, hkrati pa smo

ure pouka izkoristili tudi za osveščanje ter pogovor o etičnih

vidikih in pasteh uporabe umetne inteligence pri šolskem delu.



LITERATURA IN VIRI

[1] Žveglič, Marija, 2012: Umeščanje informacijsko-komunikacijskih

tehnologij v osnovni šoli in gimnaziji, Slovenščina v šoli 15/2, 13–24.

[2] Slovenščina, učni načrt: Ministrstvo za izobraževanje, znanost in šport,

Zavod RS za šolstvo, 2018, str. 76–77. Dostopno na naslovu:

http://mizs.arhiv-

spletisc.gov.si/fileadmin/mizs.gov.si/pageuploads/podrocje/os/prenovljen

i_UN/UN_slovenscina.pdf.

[3] Shah, P.: AI and the Future of Education: Teaching in the Age of Artificial

Intelligence. 2023. Dostopno na naslovu:

https://books.google.si/books?id=QkrUEAAAQBAJ

[4] Limna, Pongsakorn et al: A Review of Artificial Intelligence (AI) in

Education during the Digital Era (July 2022). Dostopno na

naslovu: https://ssrn.com/abstract=4160798

[5] Nacionalni program vzgoje in izobraževanja za obdobje 2023-2033:

predlog. Dostopno na:

https://www.gov.si/assets/ministrstva/MVI/Dokumenti/Nacionalni-

program-vzgoje-in-izobrazevanja-za-obdobje-2023-2033-predlog.pdf

[6] Obogateno s pomočjo umetne inteligence (slikovno gradivo).



Uporaba orodja OneNote pri delu z učencem z več



motnjami: študija primera



Using OneNote to Work with a Student with Multiple



Disabilities: Case Study



Lara Valentinčič

Osnovna Šola Kozara Nova Gorica

Nova Gorica, Slovenija

lara.valentincic@os-kozara.si



potrebami in omogočajo njihov optimalen razvoj, kot to

POVZETEK izpostavljata Navodila za izobraževalne programe s prilagojenim

Učenci s posebnimi potrebami se v šolskem prostoru pogosto izvajanjem in dodatno strokovno pomočjo za devetletno osnovno

srečujejo z ovirami, zato potrebujejo prilagoditve časa, prostora, šolo ter Zakon o usmerjanju otrok s posebnimi potrebami [1, 2].

pripomočkov, načina poučevanja in učenja ter preverjanja in Pri poučevanju učencev z več motnjami je pomembno, da učni

ocenjevanja znanja. Še posebno pozornost moramo temu proces prilagodimo vsakemu posamezniku, njegovim težavam

namenjati pri delu z učenci z več motnjami. V prispevku ter močnim področjem. Z upoštevanjem načel personalizacije in

predstavljamo primer uporabe digitalnega zvezka znotraj orodja diferenciacije učno snov učencu posredujemo na način, ki mu

OneNote pri poučevanju učenca z več motnjami v razredu, pri ustreza [3]. Eden izmed načinov je tudi uporaba informacijsko-

urah dodatne strokovne pomoči in doma. OneNote omogoča komunikacijske tehnologije. Ta si vedno bolj vtira pot v šolo in

hkratno sodelovanje strokovnih delavcev ter ponuja učinkovita v pouk. Učitelji so v zadnjih letih bolje seznanjeni z različnimi

orodja za učenčevo lažje spremljanje pouka ter šolsko delo. aplikacijami in računalniškimi orodji ter jih uporabljajo pri

svojem delu. Premisliti pa je potrebno še, kako ta orodja narediti



KLJUČNE BESEDE bolj dostopna za učence z različnimi posebnimi potrebami in kakšne možnosti ponujajo, da se lahko tudi ti enakovredno OneNote, digitalni zvezek, posebne potrebe, učenec z več vključujejo v pouk. motnjami V prispevku bomo predstavili poučevanje učenca z več



ABSTRACT znotraj orodja OneNote. motnjami, ki pri pouku uporablja računalnik in digitalni zvezek,

Students with special needs often encounter obstacles in the

school environment, so they need adaptations of time, space,

learning aids, teaching and learning methods, and knowledge 2 PREDSTAVITEV UČENCA TER UVEDBA

assesment. We have to pay special attention to this when working RAČUNALNIKA with students with multiple disabilities. In this paper, we present

Učenec zaradi kombinacije gibalne oviranosti, okvare vidne

an example of the use of a digital notebook within the OneNote

funkcije ter govorno-jezikovnih motenj izkazuje težave na

tool when teaching a student with multiple disabilities in the

področju:

classroom, during additional professional help, and at home.

OneNote enables the simultaneous collaboration of professionals - motorike, grafomotorike, koordinacije oko-roka:

and offers effective tools for the student's easier following of učenčeva pisava je slabše čitljiva, pri zapisu je

lessons and school work. počasnejši, potrebuje veliko prostora, težje prikliče

oblikovanost posameznih črk, za pisanje potrebuje

KEYWORDS veliko energije, njegova drža je pri tem pogosto

OneNote, digital notebook, special needs, student with neustrezna;

multiple disabilities - vidnega zaznavanja: več časa potrebuje za

osredotočanje pogleda na detajle, vidno polje je

zooženo, več napora vlaga v branje, težje preklaplja

1 UVOD pogled med zapisom na šolski tabli ter zapisom v

Učenci s posebnimi potrebami so raznolika skupina otrok in zvezku ali učbeniku;



ki jih ovirajo pri doseganju optimalnih rezultatov v šolskem opravilo, zato velikokrat težje sledi učiteljevi razlagi; - govora in jezika: več časa potrebuje za priklic besed, okolju. Eno izmed pomembnih načel za vzgojo in izobraževanje mladostnikov, pri katerih je pogosto prepletanje različnih težav, - pozornosti: pozornost zmore osredotočati na eno

učencev s posebnimi potrebami je tudi zagotavljanje ustreznih prisotna so mašila, upočasnjeno je jezikovno

pogojev, ki so osnova za uspešno učenje otrok s posebnimi procesiranje.



Učenčeva močna področja predstavljajo splošna razgledanost,

matematične spretnosti, ima bogat besedni zaklad in dobro

razvito slušno pozornost in pomnjenje. Učenec z ustreznimi

prilagoditvami dosega dobre učne rezultate. Z odločbo o

usmeritvi mu pripada spremljevalec za nudenje fizične pomoči.

V četrtem razredu smo pričeli z uvajanjem računalnika v

učenčevo delo. Pri poučevanju in delu z učenci z gibalno

oviranostjo se računalnik pogosto omenja kot pripomoček, ki

omogoča samostojnejše funkcioniranje, uspešnejše učenje in

enakovredno vključevanje v izobraževalne programe [1, 4].

Učencu smo z uporabo računalnika želeli olajšati izdelovanje

zapiskov, saj se pri uporabi tipkovnice izognemo priklicu oblike

črke in motorični izvedbi zapisa. Posledično smo želeli vplivati

na učenčevo držo za mizo, ki je bila pogosto pretirano sključena.

S projeciranjem vizualnega gradiva, ki ga je pripravil učitelj, na Slika 1: Odseki in strani v digitalnem zvezku



besedila poslušal namesto bral. Spoznaval je elektronske delovne domačem delu in učenju. Učenec je možnosti OneNote preizkusil pri individualnih urah dodatne strokovne pomoči, v zvezke, ki jih je izpolnjeval s pomočjo urejevalnika PDF razredu pa jih je uporabljal ob pomoči spremljevalca. dokumentov in s tem lažje povečeval vizualna gradiva ter imel Učitelji so pred učno uro v zvezek dodajali povzetke snovi, dovolj prostora za zapis na za to predvidena mesta. večpredstavnostne vsebine, kot so slike, videoposnetki, zvočni Učenec je z računalnik najprej začel delati v individualni posnetki (slika 2), učne liste, navodila za delo v času učenčeve situaciji, v okviru ur dodatne strokovne pomoči ob vodenju daljše odsotnosti (slika 3) in druge vsebine. Manjkajočo snov je specialnega in rehabilitacijskega pedagoga, kjer je vodeno dopolnjeval učenec sam ob podpori spremljevalca, saj sta lahko reševal šolske naloge s pomočjo računalniških orodij. Kasneje zaradi skupne rabe, zapiske na strani soustvarjala hkrati. smo uvedli računalnik pri pouku v razredu. Pri učencu je bil bili naloženi v učenčev digitalni zvezek, je učenec pozornost pouka seznanili strokovne delavce, ki so učenca poučevali. Digitalni zvezek in naložene vsebine smo predstavili tudi staršem lahko namenil učiteljevi razlagi, praktičnim prikazom ipd. in podali usmeritve za uporabo, da so učencu nudili oporo pri namesto zapisovanju. Z uporabo bralnika je učenec daljša tablo in delovno površino. Z vnaprej pripravljenimi zapiski, ki so Z načinom uporabe in dodajanjem vsebin smo pred začetkom računalniški zaslon, smo se izognili preklapljanju pogleda med

prisoten strah, da bo zaradi tega v razredu izstopal, zato smo



uvedli še dodaten, razredni računalnik, pri katerem so se ostali

učenci izmenjevali pri posameznih urah pouka in na njem

opravljali načrtovane aktivnosti.

V naslednjem šolskem letu se je zaradi večje kompleksnosti

posredovanih vsebin pokazala potreba po uporabi digitalnih

delovnih zvezkov in zvezka za zapiske. Odločili smo se za

uporabo storitev Microsoft 365, ki jih je šola pred tem

uporabljala. V ustvarjanje primernega digitalnega okolja so se

vključili vsi strokovni delavci, ki so delali z učencem, in njegov

spremljevalec.



3 Slika 2: Nalaganje večpredstavnostnih vsebin OneNote

OneNote je digitalni zvezek znotraj različnih storitev

Microsoft 365, ki samodejno shranjuje ustvarjene zapise tekom



našega dela. Njegova prednost je, da z njim lahko kreiramo več

različnih zvezkov, ki jim dodajamo poljubne strani in odseke,

znotraj katerih se lahko hitro in preprosto orientiramo, vanje pa

lahko poleg zapiskov vnašamo tudi druge vsebine. Omogoča

skupno rabo in sodelovanje z drugimi uporabniki, dostopen je s

katere koli naprave in se povezuje tudi z ostalimi storitvami

Microsoft 365. Mogoče ga je uporabljati tako v spletni kot v

namizni različici [5].

Za učenca smo ustvarili digitalni zvezek in znotraj tega dodali

posamezne odseke za posamezen predmet, pri posameznem

predmetu pa smo dodajali strani glede na obravnavano vsebino

(slika 1). Zvezek je bil shranjen v oblaku OneDrive poleg ostalih

datotek, kot so delovni zvezki ali učni listi, in je bil za učenca Slika 3: Učiteljeva navodila za delo v času učenčeve daljše

tako hitro dostopen. odsotnosti

Učenec je orodje OneNote hitro usvojil, poznal je njegove urejevalnikih besedil, kot je npr. Word, in z njegovo pomočjo

najpomembnejše funkcije, dodajal je posamezne strani v težje oblikujemo sheme ali druge grafike. Z dodatki lahko orodju

ustrezne odseke in jih naslovil. Prenašal je naložene datoteke na dodamo nekatere funkcije, a ti niso del storitev Microsoft 365.

računalnik za izpolnjevanje teh, uporabljal potopni bralnik, ki se V zvezek naloženih PDF ali drugih dokumentov ne moremo

ga je posluževal predvsem pri branju daljših besedil ter pri učenju urejati neposredno v OneNote ampak jih moramo prenesti na

angleščine za namene poslušanja izgovorjave ter razumevanja želeno mesto na računalniku, razen če v zvezek namesto

pomena besed. Potopni bralnik se je izkazal kot najučinkovitejše dokumenta vstavimo povezavo do mesta v oblaku OneDrive,

orodje za zagotavljanje dostopnosti vsebin kamor smo datoteko predhodno naložili.

Kljub precejšnjemu napredku pri uporabi slovenščine in

3.1 Orodja za zagotavljanje dostopnosti možnostim njene uporabe (branje, narek v slovenščini), je še

Potopni bralnik je orodje, ki ga je učenec poleg ustvarjanja vedno veliko več funkcij na voljo v angleškem jeziku (npr.

zapiskov najpogosteje uporabljal. V poštev je prišel predvsem pri slikovni slovar, prepoznavanje besednih vrst).

daljših besedilih. Bralnik poleg poslušanja besedila omogoča Kljub naštetim omejitvam, ki so se pojavile pri našem delu,

tudi prilagajanje velikosti pisave, razmika med vrsticami in pa OneNote še vedno predstavlja učinkovit način oblikovanja,

besedami, barve ozadja itd. za lažje sledenje besedilu. Poleg organizacije in shranjevanja zapiskov na enem mestu.

poslušanja besedil pa je potopni bralnik uporabljal pri učenju



angleščine (slika 4). Z njegovo pomočjo je preveril pomen 4 ZAKLJUČEK posamezne besede, besedo je slišal ter s pomočjo trenerja za

branje (»Reading Coach«) utrjeval izgovorjavo. Predstavljen primer nam pokaže, kako lahko z

zagotavljanjem ustreznih prilagoditev in uporabo digitalnih

orodij, kot je OneNote, učencem s posebnimi potrebami



olajšamo vključevanje v pouk. Uporaba orodja OneNote je

olajšala sodelovanje med strokovnimi delavci, omogočila

posredovanje vsebin na način, ki učencu najbolj ustreza, in s tem

pripomogla k zagotavljanju vključujočega in podpornega

šolskega okolja.



VIRI IN LITERATURA

[1] Navodila za izobraževalne programe s prilagojenim izvajanjem in

Slika 4: Pomoč potopnega bralnika pri razumevanju dodatno strokovno pomočjo za devetletno osnovno šolo, 2003.

besed v angleščini https://www.gov.si/assets/ministrstva/MVI/Dokumenti/Izobrazevanje-

otrok-s-posebnimi-potrebami/OS/Navodila_9-letna_OS.pdf

[2] "Zakon o usmerjanju otrok s posebnimi potrebami (ZUOPP-1)," Uradni

Ker učenec uporabe tipkovnice še ni avtomatiziral, je občasno list RS, št. 58/11, 40/12 – ZUJF, 90/12, 41/17 – ZOPOPP, in 200/20 –

posamezna besedila (predvsem v individualni situaciji ali doma) ZOOMTVI, 2011.

http://pisrs.si/Pis.web/pregledPredpisa?id=ZAKO5896

tvoril s pomočjo orodja narek (»Dictate«), ki je učenčev govor [3] J. T. Nganji in M. Brayshaw, »Disability-Aware Adaptive and

pretvoril v zapisano besedilo. Najbolje se je to obneslo pri Personalised Learning for Students with Multiple Disabilities,« Int. J. of

Information and Learning Technology, vol. 34, št. 4, str. 307–321, 2017.

tvorjenju odgovorov na vprašanja npr. o prebranem besedilu. Pri DOI: 10.1108/IJILT-08-2016-0027

tem je učenec moral biti pozoren, da je imel odgovor v mislih v [4] S. Rener, »Računalniške prilagoditve za učence s posebnimi potrebami«,

celoti oblikovan, preden ga je izgovoril in posnel, saj je za to v Gibalno ovirani otrok gre v šolo: priročnik za učitelje, ur. (Dušan

Rutar). Kamnik: CIRIUS Kamnik, 2020, 213–234.

zaradi slabšega priklica besed potreboval več časa. Pri tvorjenju https://cirius.splet.arnes.si/files/2023/03/Gibalno-ovirani-otrok-FINALE-

daljših besedil se je narekovanje slabše obneslo, saj je učenec 2.pdf

[5] »Kaj je OneNote, Ustvarjanje zvezka v OneNotu - Microsoftova

težje oblikoval misli, velikokrat se je želel med narekom podpora,« Microsoft.com, 2016. https://support.microsoft.com/sl-



popraviti in je prihajalo do manj smiselnega zapisa na ekranu. si/office/ustvarjanje-zvezka-v-onenotu-6be33cf9-f7c3-4421-9d74- 469a259952d3?wt.mc_id=otc_onenote Omogočeno je tudi vizualno prilagajanje posameznih strani z

dodajanjem črtovja (pri tem vnašanje besedila ni vezano na

postavitev črt ali mreže, ampak te predstavljajo zgolj ozadje),

obarvanjem strani in ustvarjanjem različnih kontrastov med

besedilom in ozadjem, povečevanjem pogleda, česar se je učenec

posluževal v manjši meri.



3.2 Možne izboljšave pri uporabi OneNote

Zaradi svoje preprostosti in enostavnih funkcij je OneNote

prijazen za uporabnika. Kljub temu pa smo se pri njegovi uporabi

srečevali z nekaterimi omejitvami.

V primeru sočasnega urejanja določene strani je občasno

prihajalo do težav v sinhronizaciji opravljenih sprememb med

različnimi napravami, kar je velikokrat povzročilo podvajanje

vsebin.

Glavni namen orodja OneNote je izdelovanje zapiskov, zato

ne omogoča tolikšnih funkcij, kot jih najdemo v drugih



Uporaba spletnega okolja MS Teams za izboljšanje učnih



dosežkov mladinskega pevskega zbora



Using the MS Teams Web Environment to Improve the



Learning Achievements of a Youth Choir



Mitja Vaupotič

OŠ Polzela

Polzela, Slovenija

mitjav.si@gmail.com



POVZETEK

Učenci s sodelovanjem v pevskem zboru pridobijo veščine, 1 UVOD



ki so dragocene za njihov razvoj in ki jim nezavestno pomagajo Pevski zbor je interesna dejavnost, katere namen je razvijati

pri oblikovanju osebnosti. Pevski zbori pogosto nastopajo na interesna področja učencev s poudarkom na kakovosti izvedbe,

prireditvah in se morajo na ta način neprestano dokazovati. Da ki se udejanja ob aktivnostih vseh, ki so vključeni v proces [1].

so nastopi kvalitetni, je potrebno ogromno število vaj in tudi Ker gre za dejavnost, ki je dostopna vsem otrokom, ne glede na

nekaj samostojnega dela doma. Ker je postala informacijsko- posameznikovo predznanje, je še toliko bolj pomembno, da zna

komunikacijska tehnologija dostopnejša kot kadarkoli prej in je zborovodja s pravilnimi metodami dela učencem pomagati k

pomemben pripomoček v vzgojno izobraževalnem procesu, sem pevskemu napredku. Sam sem se pri delu z mladinskim zborom

se pri delu z mladinskim pevskim zborom posluževal spletnega posluževal spletnega okolja MS Teams, ki nam je omogočil, da

okolja MS Teams. Z omenjeno aplikacijo so pevci samostojno so lahko otroci poslušali skladbe ob prostem času in ob njih tudi

opravili dodatne vaje, hkrati pa so se spoznavali s sodobno prepevali. To se je izkazalo za zelo učinkovito metodo, katera je

tehnologijo. Takšen način dela s tehnologijo je odličen, saj je za pripomogla k razvoju tako večglasja, kot tudi razvoju zbora kot

otroke zanimiv in predstavlja dodano vrednost pri delu s pevskim celote.

zborom. Dodatno dokazuje tudi, da je potrebno v vzgojno

izobraževalni proces uvajati sodobno tehnologijo, ker ima le-ta

pozitivne učinke. 2 VEČGLASEN PEVSKI ZBOR IN NAČIN

DELA



KLJUČNE BESEDE Zborovodja naj bo pevcem zgled, naj razvija svoj glas, išče

Mladinski pevski zbor, MS Teams, večglasno petje nove ideje in načine doseganja ciljev, se dodatno izobražuje in se

do glasu (pevca) vede odgovorno [2]. Zborovodja je odgovoren



them form their personalities. Choirs often take part in events usposablja za skupinsko glasbeno umetniško poustvarjanje [3], mora izbrati primeren program, ki vsebuje enoglasne in where they have to constantly keep proving themselves. In order večglasne skladbe. Večglasne skladbe zahtevajo več utrjevanja to deliver quality performances, many rehearsals are needed, in in poglobljenega razumevanja glasbe in si je za njih potrebno addition to some independent practice at home. Seeing that today, vzeti več časa. Ob tem ima zborovodja na voljo celo paleto information and communication technology is more easily načinov, kako pevcem pomagati pri hitrejšem osvajanju accessible than ever before, and presents an indispensable tool in valuable for their development, and which unconsciously help Students who participate in a choir gain skills which are pevskega zbora. Ker je pevski zbor interesna dejavnost, pri kateri zborovodja razvija smisel za večglasno muziciranje in pevce ABSTRACT tudi za sistematsko urejene vaje, ki sledijo viziji razvoja

the educational process, I used the MS Teams software while večglasnih skladb.

working with the youth choir. With the help of this app, singers



did additional exercises on their own, while at the same time 3 MS TEAMS getting familiar with modern technology. This way of working

with technology is perfect because it is interesting for students Microsoft Teams je aplikacija za sodelovanje, ustvarjena za

and presents added value when working with the choir. It also hibridno delo [4]. V času šolanja na daljavo je bila omenjena

shows that it is necessary to introduce modern technology into aplikacija ena izmed spletnih orodij, ki so omogočale učiteljem

the educational process because of its positive effects. sistematično posredovanje snovi, ki so jo otroci utrjevali v času

in načinu kot jim je najbolj ustrezalo. Čeprav je šolanje na



KEYWORDS daljavo stvar preteklosti, sem za mladinski pevski zbor ohranil elemente tega načina poučevanja, vendar sem mu spremenil Youth choir, MS Teams, polyphonic singing namen. Po novem smo MS Teamse uporabljali za dodatno

utrjevanje večglasnih skladb, ki smo jih prepevali na vajah. svoje naprave, povezave na koreografije določenih skladb in

Osnovni način dela z mladinskim zborom na vajah sem še vedno izgovorjava besedila (Slika 3). Pri delu sem skušal biti čim bolj

ohranil, MS Teamsi so nam služili le kot dodatek in pripomoček raznolik, da so učenci spoznavali skladbo na več različnih

na poti k izboljšanju rezultatov. načinov.



3.1 MS Teams in mladinski pevski zbor

Preden smo z učenci začeli uporabljati omenjeno spletno

okolje, smo morali obnoviti znanje o uporabi Teamsov, saj so

nekateri to že pozabili. Ko sem se prepričal, da delo z

računalnikom in Teamsi ne predstavlja večjih ovir, sem ustvaril

skupino in vključil vse člane mladinskega pevskega zbora. V

objavah sem otrokom posredoval informacije o novem gradivu

in navodila za samostojno delo (Slika 1).





Slika 3: Različni dokumenti za utrjevanje skladb



Glavni namen samostojnega dela otrok je bilo utrjevanje, zato

sem posnetke posameznih glasov skladb razdelil na manjše enote,

ki sem jih različno označeval (Slika 4). Da ne bi prišlo do

nezaželene zmede, sem se držal pravila, da so deli skladb v notah

enako označeni kot na posnetkih. Za ta način sem se odločil, ker

so nekatere skladbe obsežne in jih je bolje utrjevati po

Slika 1: Objavljanje navodil posameznih delih.



Pevci so do gradiva prosto dostopali, ob morebitnih težavah

sem jim svetoval in jih pomagal odpraviti. Večjih težav z

dostopom in delom z gradivom pri učencih nisem zaznal.



3.2 Urejanje gradiva pevskega zbora v MS

Teams

Gradivo sem urejal sistematsko, kar je pevcem omogočalo

lažjo orientacijo pri delu in s tem doseganje boljših rezultatov na

vajah. Skladbe sem dodajal sproti, in tako je nastala cela zbirka

not, do katerih so otroci imeli dostop in so jih lahko utrjevali

(Slika 2). Slika 4: Označevanje posnetkov skladb s posebnimi

oznakami





4 ZAKLJUČEK



S pomočjo sodobne tehnologije so učenci na inovativen način



utrjevali skladbe, ki smo jih spoznavali pri pevskem zboru. Pevci

so zaradi samostojnega domačega dela prihajali na vajo bolje

pripravljeni in smo na ta način lahko hitreje osvojili

kompleksnejše skladbe. Pred pomembnimi nastopi so te skladbe

ponovno poslušali doma in tako še bolj utrdili že pridobljeno

znanje. Lahko trdim, da smo tudi zaradi uporabe aplikacije MS



Slika 2: Delovno gradivo za pevski zbor razporejeno po priznanje na mednarodnem tekmovanju. Teams, prejeli zlato priznanje z odliko na državnem in zlato

mapah

Ta način dela predstavlja za zborovodjo izziv, ki zahteva



način so imeli učenci pregleden dostop do gradiva posameznih Note sem v mapah razdelil po posameznih glasovih in na ta nekoliko več ustvarjalnosti in iznajdljivosti kot sicer, zagotovo pa pripomore k izboljšanju rezultatov. Iz tega razloga je smiselno in pravilno vnašati sodobno tehnologijo ter aplikacije kot je MS glasov. Otrokom sem posredoval gradiva v različnih formatih in Teams v pedagoški proces. oblikah, in sicer: notno gradivo v pdf formatu, posnetke celotnih

skladb, posnetke posameznih glasov, ki smo jih posneli na vaji,

mp3 posnetki posameznih glasov, ki so si jih lahko prenesli na





CITIRANA DELA


[1] K. Metoda, „Interesne dejavnosti za 9-letno osnovno šolo,“ Ministrstvo za

šolstvo in šport, Zavod RS za šolstvo, Ljubljana, 2008.

[2] M. Burger, „Naši zbori,“ JSKD, 5. 6. 2014. [Elektronski]. Available:

https://www.nasizbori.si/zborovodja-vokalna-tehnika/. [Poskus dostopa

11. 8. 2024].

[3] P. k. k. z. g. vzgojo, „Zavod republike Slovenije z ašolstvo,“ 2003.

[Elektronski]. Available:

https://www.gov.si/assets/ministrstva/MVI/Dokumenti/Osnovna-

sola/Ucni-nacrti/Drugi-konceptualni-dokumenti/Pevski_zbor.pdf.

[Poskus dostopa 11. 8. 2024].

[4] Microsoft, „Microsoft Podpora,“ [Elektronski]. Available:

https://support.microsoft.com/sl-si/office/uvod-v-microsoft-teams-

b98d533f-118e-4bae-bf44-3df2470c2b12. [Poskus dostopa 11. 8. 2024].



Analiza projektnega dela za pridobitev praktičnih izkušenj



in znanj študentov v delovnem okolju



Analysis of Project Work to Gain Practical Experience and



Knowledge of Students in the Working Environment



izr. prof. dr. Borut Werber

Fakulteta za organizacijske vede

Univerza v Mariboru

Kranj, Slovenia

borut.werber@um.si



Po klasifikaciji je ta projekt spadal pod izvajanje projektnih

POVZETEK aktivnosti sodelovanja z gospodarstvom (v nadaljnjem besedilu:

Prispevek predstavlja nastanek, izvedbo in rezultate Sklop A V projektu so sodelovali študent Univerze v Ljubljani iz

projektnega dela za pridobitev praktičnih izkušenj in znanj Medicinske fakultete, študent Univerze v Mariboru, Fakultete za

študentov z naslovom Uporaba umetne inteligence za organizacijske vede (FOV) ter 5 študentov Univerze v Mariboru,

prepoznavanje bazalnoceličnega karcinoma. Gledano v celoti je Fakultete za elektrotehniko, računalništvo in informatiko (FERI)

bil projekt uspešen, kljub temu pa so predstavljene omejitve, ki smeri Informatika in tehnologije komuniciranja (2 študenta),

lahko vplivajo na izvedbo takih projektov. Med omejitve so Informatika in podatkovne tehnologije, Računalništvo in

navedene omejeno financiranje, vrednotenje takih projektov, informacijske tehnologije ter Medijske komunikacije. Trije

omejeno število prijav iz fakultet, čas trajanja ter predpisane študenti so bili iz 1. stopnje UNI in 4 iz 2. stopnje mag. Vključena

vrste sodelujočih študentov, sva bila dva pedagoška mentorja iz FOV in FERI, ter dva delovna



Bazalnocelični karcinom KEYWORDS / KLJUČNE BESEDE mentorja, eden iz Dermatologije Bartenjev [2], drugi iz Doma za upokojence Idrija [3]. Na projektu so sodelovali še trije Študentski projekt, praktične izkušnje, umetna inteligenca, koordinatorji za izvedbo projektov iz FOV in UM. Projekt je trajal od 1.2.2023 do 30.6.2023.



ABSTRACT

The paper presents the origin, implementation and results of 2 NAČRT PROJEKTA



the project work to gain practical experience of students entitled Ideja za projekt izhaja iz lastnih potreb avtorja, ki je sam Use of artificial intelligence to identify basal cell carcinoma. onkološki pacient z več posegi kirurške odstranitve kožnega raka Taken as a whole, the project was successful, however, - bazalno celičnega karcinoma (BCC). Sam BCC sicer ni limitations that may affect the implementation of such projects življenjsko ogrožajoč, njegovo ne zdravljenje pa lahko pripelje are presented. Limitations include limited funding, evaluation of do poškodb kože ali pretvorbo v druge bolj ogrožajoče vrste such projects, limited number of applications from faculties, kožnega raka kot sta ploščato celični karcinom ter melanom, ki duration and prescribed types of participating students. je najnevarnejši in lahko od razširitvijo v telo povzroči smrt.



KEYWORDS slutnjo, da se bo katero od znamenj spremenilo v kožnega raka. Slednje je razlog, da imamo onkološki pacienti vedno prisotno

Student project, practical experience, artificial intelligence, Običajno so dermatološki pregledi na onkologiji enkrat letno. Da

Basal cell carcinoma bi si lahko sami diagnosticirali morebitne sumljive spremembe v

vmesnem času, smo si zadali cilj narediti programsko rešitev za



1 UVOD karcinoma. Cilj je bil torej le prepoznava BCC, ker smo menili, uporabo umetne inteligence za prepoznavo bazalnoceličnega

Prispevek obravnava analizo izvedbe projektnega dela za da je slednji najlažje prepoznan tudi laikom. Osnove projekta sva

pridobitev praktičnih izkušenj in znanj študentov v delovnem s so-mentorjem iz FERI napisala v treh srečanjih. S pomočjo

okolju z naslovom Uporaba umetne inteligence za prepoznavo socialnih mrež sva našla interes v Domu upokojencev Idrija, ki

bazalnoceličnega karcinoma, ki je bil del razpisa Ministrstva za so bili pripravljeni ponuditi brezplačen pregled njihovim

izobraževanje znanost in šport v okviru Operativnega programa upokojencem, hkrati pa je direktorica doma nastopala kot

za izvajanje Evropske kohezijske politike v obdobju 2014 -2020 mentorica študentu medicine, ki je te preglede izvedel. Da bi

[1]. dobili strokovne informacije o BCC smo iskali sodelavce med

Javni razpis za izbor operacij delno financira Evropska unija, javnimi onkološkimi inštitucijami, a je bil odziv negativen saj je

in sicer iz evropskega socialnega sklada. Razpis je bil objavljen financiranje delovnih mentorjev skromno. Ponovno smo s

4.7.2022 in je obsegal izvedbo projektov v obdobju 2022-2023. pomočjo socialne mreže našli privatno podjetje Dermatologija

Bartenjev, ki so bili pripravljeni usposobiti študenta medicine in ponoviti. Na tečaju smo se naučili, da računalnik ne primerja

pomagati pri diagnostiki – potrditvi ali je študent zaznal ustrezna slike kot celote, temveč jo razreže v dolgo vrsto bitov in iz te

spremenjena kožna znamenja ali so to le starostne pege. vrste s pomočjo algoritmov ugotavlja, kaj je na sliki. Da lahko

Fotografije pacientov smo nameravali uporabiti kot test za našo izvede primerjavo, mora nevronska mreža skozi postopek učenja.

mobilno rešitev, ki bi naj s pomočjo umetne inteligence na osnovi To poteka tako, da se ji omogoči obdelava čim večjega vzorca

analize fotografij v odstotkih ocenila znamenje na fotografiji kot fotografij (vsaj 10.000 slik). Večji kot je vzorec, večja je

sumljivo ali nesumljivo. Kot dodaten cilj je bil razviti verjetnost, da bo sistem bolje deloval. Med najdenimi javno

spletno/namizno rešitev za zdravnika, da bi lahko opažanja dostopnimi zbirkami slik, smo se odločili za učno množico

zapisal v bazo pacienta. znamenj, ki poleg samih slik in kategorije kožnega znamenja,

prikazuje še nekatere druge atribute, ki bi lahko pripomogli k

natančnejši klasifikaciji znamenj.

3 ANALIZA Pred izvedbo kliničnih pregledov starostnikov in za

Glavni cilj projekta je bil razviti prototipno mobilno rešitev, morebitno objavo v revijah višjega ranga se zahtevajo potrdila

ki uporabniku omogoča, da sam na sebi fotografira sumljiva etičnih komisij. Odločili smo se, da damo vlogi na etično

kožna znamenja, sistem pa s pomočjo uporabe umetne komisijo UL MF in na FOV. Med postopkom smo ugotovili, da

inteligence preko naučenih nevronskih mrež obdela fotografijo je potrebno vlogo za UL MF plačati. Ker teh sredstev nismo

in oceni verjetnost, da je to BCC. Delo je potekalo v parih ali planirali, smo oddali vlogo le na Etično komisijo FOV, ki nam je

posamezno, glede na znanja študentov. Študent FOV je skrbel za delo dovolila z izjemo, da se mora mobilna rešitev preden se da

koordinacijo, organiziral tedenske sestanke preko MS Teamsov v javno uporabo, klinično preizkusiti. Iz te izkušnje sta se največ

in pisal zapisnike sestankov. Sestanke je vodil eden od naučila študent medicine in študent FOV, ki sta vlogi pripravljala.

pedagoških mentorjev, ki je tudi usmerjal in dodeljeval nadaljnje Vsi ostali deležniki pa so bili seznanjeni s pomenom takih potrdil

naloge glede na zastavljene cilje. Študent FOV je bil spoznan z pri izvedbi raziskav, ki vsebujejo sodelovanje ljudi (pacientov)

vodenjem tima in pomenom in izzivi, ki ga takšno delo povzroča. tudi, če se ne uporabljajo invazivni postopki.

Prvi korak projekta je bila študija literature s področja Da bi se naučili na napakah drugih, smo analizirali obstoječe

kožnega raka in s področja umetne inteligence ter nevronskih brezplačne programe, ki omogočajo fotografiranje znamenj in

mrež. Delovni mentor iz Dermatologije Bartenjev je v tem delu njihovo oceno. Med testom ene od teh rešitev smo spoznali

priskrbel literaturo, ki jo sam uporablja za strokovne namene in podobno kot EurekaAlert [11], da nekateri program ne loči med

nam je predstavljala vir osnovnih podatkov, kako BCC posnetkom rdeče rože in dejanskim kožnim znamenjem (slika 1

prepoznati med ostalimi kožnimi raki in benignimi kožnimi in 2), zato smo se odločili za dvostopenjski sistem. Študenta

znamenji. Ta literatura bo koristna predvsem za študenta FERI iz skupine za nevronske mreže sta vsak zase razvijala in

medicine in za nas, ko bomo nadaljevali to temo v naslednjem testirala nevronske mreže. Prvi je preverjal, če je na sliki sploh

projektu. Kot je pokazala študija strokovnih spletnih strani in kožno znamenje ali pa slika ni dovolj ostra (slika 5). Drugi je

pisnih objav se BCC pojavlja v različnih oblikah in v različnih nato ugotavljal ali znamenje na fotografiji ustreza opisu ene od

stadijih [4][5][6][7][8][9][10]. oblik BCC.



Pridobljena literatura s področja umetne inteligence in

nevronskih mrež je uporabna pri več predmetih tako na FERI,

kot na FOV na smeri informacijskih sistemov. Preučili smo tudi

nekaj obstoječih mobilnih aplikacij za klasifikacijo kožnih

znamenj in potencialne pripomočke, ki na bi olajšali zajem

nekaterih dodatnih podatkov o kožnih znamenjih, ki bi izboljšale

natančnost napovedovanja.

Sama izkušnja projekta z njegovimi težavami in omejitvami

se lahko uporabi kot študijski primer pri predmetih na FERI in

na FOV. Kot primer navajamo predmet Informacijske in

komunikacijske tehnologije, ki obravnavajo superračunalnike,

med drugim tudi Vego, s katero imamo sedaj praktične izkušnje.

Izkazalo se je, da uporaba superračunalnika kljub dovoljenjem za

uporabo sploh ni enostavna, saj je potrebno nastaviti veliko

parametrov in se postaviti v čakalno vrsto, da se vaš proces

požene in izvede.

Za člane projekta in sodelavce smo izvedli izobraževanje z

mikrokrediti v sistemu NVIDIA s področja umetne inteligence in

uporabe nevronskih mrež za prepoznavanje objektov. To se je

pokazalo kot zelo koristno in lahko prispeva k širjenju znanja

tako med akademsko kot poslovno sfero. Iz pridobljene izkušnje

smo ugotovili, da je zanimanje za take vrste znanj tako med

študenti, kot med zaposlenimi, zato bomo tečaj periodično

ponavljali za širšo javnost. Smo se pa na tečaju tudi naučili, da

delo na oddaljenem namizju zahteva dobro omrežje, če tega ni, Slika 1: Prikaz napake programa Skin Check, ki ne

se lahko izvedba procesa na daljavo zaustavi in je potrebno vajo prepozna, da je na sliki roža in ne BCC.





Slika 2: Prikaz napake programa Skin Check, ki ne Slika 3: Prijavno okno mobilne aplikacije.



prepozna, da je na sliki roža in ne BCC.





Za klasifikacijo slik smo uporabili globoke nevronske mreže.

Učenje nevronskih mrež je potekalo na osebnih računalnikih v

programskem jeziku Python in z uporabo programskih knjižnic

za strojno učenje TensorFlow (https://www.tensorflow.org/) in

Keras (https://keras.io/). TensorFlow je odprtokodna platforma

za strojno učenje in simbolična matematična knjižnica, ki se

uporablja za aplikacije strojnega učenja, vendar je nekoliko

kompleksnejša za uporabo[12]. Keras je odprtokodna knjižnica

za nevronske mreže, ki deluje na osnovi knjižnice Tensorflow.

Zasnovana je tako, da jo lahko uporabnik uporablja hitro in

enostavno, ter omogoča izdelavo katerega koli algoritma

globokega učenja na abstrakten način [13].

Kasnejša klasifikacija slik z naučenimi nevronskimi mrežami

poteka na oblačni storitvi Firebase ML

(https://firebase.google.com/docs/ml). S tem je omogočeno, da

se klasifikacija kožnih znamenj izvaja tudi na manj zmogljivih

mobilnih napravah.

Seveda pa zaledni programi niso primerni za uporabo oseb, ki

niso programerji zato smo morali izdelati uporabnikom prijazno

mobilno aplikacijo. Ta del je dela drugi par študentov iz FERI

zadolžen za ospredje »front end«.

Ob zagonu aplikacije se mora uporabnik najprej prijaviti

(Slika 3), s čemer je zagotovljena zasebnost podatkov.

Sama slika BCC brez podatkov ni primerna za nadaljnjo

uporabo zato smo program dopolnili z nekaterimi pomembnimi Slika 4: Prikaz sofinancerjev projekta.

podatki, ki jih je potreboval študent medicine in programerja za

mobilno aplikacijo (slika 5 in 6).





Slika 6: Prikaz sheme človeškega telesa, na kateri



Slika 5: Prikaz vmesnika mobilne aplikacije, ki beleži uporabnik izbere želeno lokacijo, kjer se nahaja sumljivo

dodatne podatke o posnetem kožnem znamenju kožno znamenje.



Sama rešitev deluje na mobilni platformi za Android telefone.



Študenta FERI za ospredje sta pripravila navodila za namestitev

in uporabo.

Da je aplikacija skladna in barvno usklajena, ima svoj logotip

in ustrezno obarvana ozadja in gradnike je poskrbela študentka

iz FERI na študijski smeri Medijske komunikacije. Ena od

funkcionalnosti spletne aplikacije je bila tudi beleženje mesta na

telesu kje se to znamenje nahaja. Tudi tukaj smo potrebovali

risbo telesa ( slika 8), ki jo je poiskala in prilagodila študentka

zadolžena za dizajn.

Za potrebe komuniciranja med študentom medicine in

delovnim mentorjem oziroma za namene naknadne obdelave

fotografij sta študenta za ospredje programirala še spletno rešitev Slika 7: Prikaz prijave v spletno rešitev.

(Slika 7), ki je vezana na isto bazo podatkov in fotografij. Ta

rešitev omogoča prikaz večjih fotografij kot na telefonu in tako Ena od funkcionalnosti spletne rešitve je tudi, da lahko

omogoči lažjo komunikacijo in svetovanje med izvajalcem pregledamo vse slike za nazaj. Izbiro omogočajo različni

pregleda in specialistom. Slednji tako lahko na daljavo oceni ali pogoji( Slika 8), ki omejijo izpis na določen datum.

je znamenje sumljivo in ali so potrebni dodatni postopki.





Slika 8: Prikaz možnosti izbire prikazov posnetih fotografij

na spletni različici programa.



Dodatno smo želeli povečati natančnost prepoznave malignih

znamenj s pomočjo pripomočka, na katerem bi namestili merila

in barvne lestvice za kalibracijo slik. Študent FOV je najprej

predlagal nekaj tipov vmesnikov. Zaradi enostavnosti uporabe

smo se odločili za slednjega (Slika 9), ki ima odprtino 2x2 cm in

že s tem omogoča umerjanje fotografija. Z dodanim merilom in Slika 10: Prikaz logotipa in barvne lestvice mobilne rešitve

barvno lestvico pa omogoča kalibracijo. Pred tiskanjem na 3D v zeleni in modri barvi.

tiskalnik je potrebno v računalniškem 3D programu narisati načrt.

Pri tiskanju se določijo parametri, ki opredelijo strukturo in Ena od nalog študentke iz FERI smer medijske komunikacije

gostoto natisnjenega materiala. je bila tudi, da se izdela reklamni plakat, ki smo ga uporabili za

Merila so se natisnila na debelejši papir in nalepila na 3D promocijo na dogodkih v prihodnosti. Sledila je še izdelava

pripomoček (Slika 9). V praksi so se spet pokazale težave, saj spletne ankete uporabnikov, ki se bo uporabil, ko bo rešitev na

tiskalnik ni natisnil lestvic v pravilnih dimenzijah. Sliko smo na razpolago uporabnikom.

koncu spremenili v tako obliko, da jo je v ustrezni velikosti lahko Na koncu projekta je sledilo testiranje in dopolnjevanje

natisnil tudi študentu iz FOV v Kranju. Pridobljeno znanje iz tega rešitve. Pri tem sta poleg programerjev sodelovala še študent

področja bo prispevalo k izkušnjam obeh udeleženih v tej dvojici, FOV in študentka FERI smeri Medijske komunikacije ter študent

študentu FOV in študentki FERI smer Medijske komunikacije. medicine, ki je programerjem pripravil seznam zahtevanih meta

Za grafično podobo, logotip in dizajn je bila zadolžena podatkov in vsebine za spustne menije.

študentka FERI iz smeri Medijske komunikacije. Pripravila je



več verzij logotipov in barvnih shem. Skupina je na tedenskem 4 OCENA USPEŠNOSTI PROJEKTA sestanku izbrala končno verzijo logotipa in barvnih shem. Te

barvne sheme je posredovala programerjema, ki sta na osnovi Sam cilj projekta je bil sicer dosežen, a se je v praksi pokazalo

tega izdelala mobilno in spletno rešitev. več omejitev, ki jih nismo pričakovali pri planu projekta..

Kot prvo se je po študiji literature izkazalo, da kljub temu, da

smo kot cilj določili obravnavati le najpogostejšo obliko kožnega



raka BCC, ki naj bi se najenostavneje prepoznal, smo ugotovili,

da obstajajo 4 podvrste BCC, ki si niso podobne ne po obliki in

ne po barvi. Zadevo smo rešili z združitvijo vzorcev po meta

podatkih le na nivoju BCC in ne na podnivojih.



Kot drugo se je izkazalo, da ni bilo mogoče uporabiti



superačunalnik Vega zaradi specifike poganjanja programja in

dostopov do uporabniških računov. Zaradi tega smo uporabili

močnejši računalnik z več grafičnimi karticami pedagoškega

mentorja iz FERI in na njem izvedli učenje nevronskih mrež

seveda v daljših časovnih obdobjih.

Kot tretje se je izkazalo, da je potrebno za vlogo na Etično

komisijo UL MF plačati prijavnino, ki ni bila planirana v

projektu. Zadevo smo rešili tako, da smo projekt prijavili na

Etično komisijo FOV, kjer plačilo ni potrebno.

Potrdilo Etične komisije FOV smo dobili pod omejitvijo, da

Slika 9: Prikaz meril in 3D natisnjenega pripomočka. program ne sme v javno uporabo dokler ne bo klinično testiran.

Zaradi časovne omejitve in malega vzorca pozitivnih primerov



bo prototip dostopen, a le raziskovalni in strokovni sferi, ne pa 5.3 Omejeno število možnih prijav



posameznim uporabnikom. Kadar se razpis izvede se omeji število možnih prijav iz Predvsem smo se naučili, da ena učna baza ni dovolj, če posamezne univerze in posledično fakultete. Ker ne obstaja način vsebuje 10.000 enot, saj se z meta podatki vzorec razbija na izbire, se lahko na posamezne projekte prijavi le eden iz iste podskupine s tem pa se zaželena meja 10.000 učnih enot za fakultete neglede na to, da imajo tudi drugi izdelane ideje za nek nevronsko mrežo ne doseže. Še posebej je pomembno, da je projekt. vzorec testnih fotografij dovolj velik, ki predstavlja testno enoto



in s katerimi se lahko preveri, če sistem pravilno deluje. Vmesnik 5.4 Predpisano število študentov različnih

za izboljšanje rezultatov prepoznave BCC z merskimi in stopenj študija barvnimi kalibracijami, je bil uspešno izdelan, a zaradi časovne

omejitve ni bila prilagojena nevronska mreža tako, da bi te Projekt zahteva, da se v projekte vključuje študente iz vseh

podatke uporabila pri prepoznavi. Da bi to delovalo je potrebno stopenj študija, to pomeni, da pridejo v projekt študenti 1 letnika

posebno programiranje, ki iz fotografije pobere mere in barvne VS, ki niso opravili niti vseh izpitov prvega letnika in še nimajo

odtenke, ne upošteva pa fotografijo in vmesnik pri obdelavi niti teoretičnih znaj, kaj šele prakse, da bi lahko konkretno

posnetka. prispevali k rezultatom projekta. To se je še posebej prikazalo v

Študent medicine je moral počakati na odobritev etične našem primeru, ker je projekt kljub izvedenem tečaju iz umetne

skupine preden je lahko začel z delom. Da bi povečal število inteligence zahteval veliko predznanj s področja mobilnih in

pregledov je obiskal poleg planiranega doma upokojencev še spletnih rešitev, nevronskih mrež, računalniškega vida, … Pet

Dom upokojencev Črni vrh. Kljub temu, da je pregledal 112 ljudi mesecev pa je premalo, da se nekoga iz nič priuči v strokovnjaka.

in se z dermatologom posvetoval o 51 fotografiranih ljudeh, od Pričakuje se, da imajo študenti neka teoretična znanja in znanja

teh je bilo napotenih v nadaljnjo obravnavo 9. Tak vzorec programiranja, ki jih lahko nadgradijo na problemu, ki ga

fotografij je občutno premajhen za testiranje tako kompleksnega obravnava projekt.

sistema zato nam ta del, testiranje rešitve, ostaja za naslednje

projekte kjer bomo lahko pridobljeno znanje in rešitve 5.5 Čas trajanja projekta

implementirali v praksi. Med tem ko se običajni projekti izvajajo med 3-5 let ali več,

so ti projekti med 3-5 mesecev. Ta čas je razmeroma kratek, ker



5 ANALIZA DEJAVNIKOV USPEH mesecu. Ta omejitev je nesmiselna, ker imajo različni člani ekipe so omejitve pri številu ur, ki jih študent lahko opravi v enem

Kljub temu, da je bil projekt uspešno zaključen na tem mestu različne dobe intenzivnosti, Na primer, analitiki in programerji

predstavljamo omejitve oziroma negativne dejavnike pri takih imajo največ dela na začetki, testerji na koncu.

projektih.

ZAHVALA

5.1 Financiranje projekta Projekt je bil financiran s strani Republike Slovenije in

Vedno se vse začne pri denarju in tudi tokrat je tako. Kljub Evropske unije iz evropskega socialnega sklada. Projekt sta

temu, da projekte sofinancira EU so to nizko proračunski projekti omogočili Fakulteta za organizacijske vede in Fakulteta za

(do 20600€). Najprej je težko dobiti pedagoškega mentorja, ki bo elektrotehniko, računalništvo in informatiko Univerze v

pripravljen pet mesecev skrbeti za študente in partnerje za cca Mariboru. Iskreno se zahvaljujemo sodelujočim iz

1500€ bruto, 300€ bruto na mesec, torej cca. 180€ neto, 6€ na Dermatologije Bartenjev ter Doma upokojencev Idrija in Črni

dan. S tem, da se prijava projekta ne šteje v te ure. Mi smo imeli Vrh.

tedenske sestanke, ki so trajali vsaj eno uro. Sam tečaj iz umetne

inteligence nam je vzel en dan, kje je testiranje rešitev, iskanje REFERENCES

literature, iskanje baz vzorčnih fotografij, svetovanje študentom, [1] Ministrstvo za izobraževanje znanost in šport, “Javni razpis Projektno delo

pisanje mesečni poročil in končnih poročil, kar se vsakič zavleče, za pridobitev praktičnih izkušenj in znanj študentov v delovnem okolju,”

ker vedno zahtevajo neka dopolnila in milijon obrazcev. Javni razpis, 2022. https://www.gov.si/novice/2022-07-04-javni-razpis-

projektno-delo-za-pridobitev-prakticnih-izkusenj-in-znanj-studentov-v-

Podobno vrednost honorarja so dobili delovni mentorji. delovnem-okolju/

Verjetno zato ni bilo odziva iz javnih inštitucij, ker je vsak [2] “Dermatologija Bartenjev.” https://www.bartenjev.si/# (accessed Sep. 12,

2024).

izračunal, da bo več dela kot plačila. [3] Dom upokojencev Idrija, “Dom upokojencev Idrija - Življenje z nami je

Plačilo študentov je druga ovira. Večina naših študentov iz kvalitetnejše,” Predstavitev, 2023. https://www.duidrija.si/ (accessed Sep.

12, 2024).

smeri informatike in računalništva že honorarno programira, zato [4] DermNet, “Basal Cell Carcinoma: Symptoms, Causes, and Treatment —

je bil odziv slab, v bistvu smo vse sprejeli, Da bi slednjim lahko DermNet,” 2024. https://dermnetnz.org/topics/basal-cell-carcinoma

plačali več ur nismo zapolnili celotno kvoto študentov(10) Za (accessed Sep. 12, 2024).

[5] C. Bichakjian et al., “Guidelines of care for the management of basal cell

študente je bilo namenjeno 80% vseh sredstev. carcinoma,” J. Am. Acad. Dermatol., vol. 78, no. 3, pp. 540–559, Mar.

2018, doi: 10.1016/J.JAAD.2017.10.006.



5.2 [6] N. R. Telfer, G. B. Colver, and C. A. Morton, “Guidelines for the Vrednotenj projekta za pedagoške mentorje

management of basal cell carcinoma,” Br. J. Dermatol., vol. 159, no. 1,

Pedagoški mentor mora za vsako habilitacijo v petih letih pp. 35–48, Jul. 2008, doi: 10.1111/J.1365-2133.2008.08666.X.

[7] K. Freeman et al., “Algorithm based smartphone apps to assess risk of skin

zbrati določeno število točk. Največ točk jim prinesejo članki iz cancer in adults: systematic review of diagnostic accuracy studies,” BMJ,

JCR skupine. Pomembni so mednarodni projekti ali večji vol. 368, 2020, doi: 10.1136/bmj.m127.

[8] A. Marka, J. B. Carter, E. Toto, and S. Hassanpour, “Automated detection

projekti znotraj Slovenije, med tem, ko se ti študentski projekti of nonmelanoma skin cancer using digital images: A systematic review,”



točkujejo zelo nizko ali celo nič. BMC Med. Imaging, vol. 19, no. 1, Feb. 2019, doi: 10.1186/S12880-019- 0307-7.



[9] H. W. Rogers, M. A. Weinstock, S. R. Feldman, and B. M. Coldiron,

“Incidence estimate of nonmelanoma skin cancer (keratinocyte

carcinomas) in the us population, 2012,” JAMA Dermatology, vol. 151,

no. 10, pp. 1081–1086, Oct. 2015, doi:

10.1001/JAMADERMATOL.2015.1187.

[10] O. T. Jones et al., “Artificial intelligence and machine learning algorithms

for early detection of skin cancer in community and primary care settings:

a systematic review,” Lancet Digit. Heal., vol. 4, no. 6, pp. e466–e476,

Jun. 2022, doi: 10.1016/S2589-7500(22)00023-1.

[11] B. Kermenova and C. Martin, “Direct-to-consumer skin cancer detection

apps | EurekAlert!,” EurekaAlert, 2021. https://www.eurekalert.org/news-

releases/929922 (accessed Sep. 12, 2024).

[12] A. Davies et al., “Advancing mathematics by guiding human intuition with

AI,” Nature, vol. 600, no. 7887, pp. 70–74, Dec. 2021, doi:

10.1038/S41586-021-04086-X.

[13] “Keras: Deep Learning for humans.” https://keras.io/ (accessed Sep. 12,

2024).



Nekaj izkušenj pri uporabi ChatGPT pri poučevanju



študentov na FUDŠ – smer Mediji in novinarstvo



Some Experiences with the Use of ChatGPT in Teaching



Students at FUDŠ – Media and Journalism Programme



Srečo Zakrajšek Miha Zakrajšek

IAM Visoka šola za multimedije, Ljubljana IAM Visoka šola za multimedije, Ljubljana

Ljubljana, Slovenia Ljubljana, Slovenia

sreco.zakrajsek@iam.si miha.zakrajsek@iam.si



POVZETEK KEYWORDS

Umetna inteligenca je s pojavom brezplačne platforme ChatGPT, artificial intelligence, media, media production,

ChatGPT postala zelo aktualna na področju medijev, saj so higher education

strokovnjaki s področja filmske industrije in drugih umetniških

poklicev zelo hitro spoznali pomen in tudi nevarnosti umetne

inteligence. Zaradi trendov v stroki in zanimanja študentov smo 1 UVOD

se odločili, da pripravimo raziskavo, s katero smo želeli ugotoviti, Umetna inteligenca je s pojavom brezplačne platforme

kako lahko smiselno vključimo ChatGPT v študijski proces in to ChatGPT v letu 2022 postala splošno dostopna in se je zelo hitro

posebej v del, kjer študenti v okviru projektnega dela pripravljajo razširila tudi med mladimi in s tem tudi v izobraževalnem

multimedijske izdelke. sistemu in ustanovah po vsem svetu. Med njimi so bile med

Raziskava je pokazala, da študenti podpirajo uporabo umetne prvimi visokošolskimi ustanovami tiste, ki se ukvarjajo z

inteligence pri študiju in produkciji izdelkov ter da jo praktično medijsko produkcijo, saj so strokovnjaki s področja filmske

vsi tudi uporabljajo, vendar hkrati poudarjajo potrebo po industrije in drugih umetniških poklicev zelo hitro spoznali

kritičnem ovrednotenju podatkov ter zavedanju omejitev tega pomen in tudi nevarnosti umetne inteligence ter v letu 2023

orodja. pripravili stavko, s katero so želeli čim bolj zmanjšati uporabo

Raziskavo smo opravili v času od 1. 10. 2023 do 25. 12. 2023 UI v filmski industriji oziroma predvsem preprečiti, da bi se

na FUDŠ, Ljubljana, pri predmetu Vodenje in organizacija zmanjšali honorarji in število zaposlenih [1].

medijske produkcije, s skupino 36 študentk in študentov. Zaradi zanimanja študentov, ki večinoma delajo v praksi in so

v vedno večji meri soočeni z uporabo UI, smo se odločili, da

KLJUČNE BESEDE pripravimo raziskavo, s katero smo želeli ugotoviti, kako lahko

ChatGPT, umetna inteligenca, mediji, medijska produkcija, ChatGPT smiselno vključimo v študijski proces in to posebej v

izobraževanje, visokošolsko izobraževanje del, kjer študenti v okviru projektnega dela pripravljajo

multimedijske produkte. Želje študentov, da čim prej dobijo

ABSTRACT koristne napotke glede uporabe UI v praksi, so postavile pred nas

Artificial intelligence has become highly relevant in the field poseben izziv, saj na tem področju ni bilo poglobljenih raziskav,

of media with the advent of the free ChatGPT platform, as the večina dokumentov pa je svarila (in to počne še danes) pred

significance and potential dangers of AI have been quickly uporabo UI.

recognised by experts in the film industry and other artistic UNESCO je konec leta 2023 izdal publikacijo Education in

professions. Due to trends in the field and the interest of students, the age of artificial intelligence (okt., dec. 2023), v kateri



products as a segment of their project work. The study dezinformacije in sovražni govor, spletni viri pa ne upoštevajo revealed 95 odstotkov živečih človeških jezikov. Zato je treba šole that students support the use of artificial intelligence in their zaščititi pred škodo, ki bi jo lahko povzročila umetna inteligenca studies and product production, and practically all of them use it. [2]. However, they also emphasise the need for critical evaluation UNESCO je opozoril, da se uporaba ChatGPT v šolah and awareness of the limitations of this tool. uveljavlja prehitro, brez zadostnih preverjanj, pravil ali The research was conducted from October 1, 2023, to meaningfully incorporate ChatGPT into the educational process, kolikor to ne bo dostopno vsem učencem. Z vsako novo tehnologijo se veča digitalna revščina. Obenem se z UI lahko širi specifically the part in which students prepare multimedia we decided to conduct a study to determine how we can opozarja na morebitne negativne učinke UI v izobraževanju, v

December 25, 2023, at FUDŠ, Ljubljana, in the course predpisov. Izobraževalni sektor je v veliki meri nepripravljen na



Management and Organisation of Media Production, etično in pedagoško integracijo teh hitro razvijajočih se orodij. involving V publikaciji so objavljena tudi navodila za regulacijo umetne

a group of 36 students. inteligence v izobraževanju (str. 8), ki jih je UNESCO objavil 7.



septembra 2023. Opozarja tudi na problem varstva podatkov in Raziskovalno vprašanje: Kako vključitev umetne

zasebnosti in nujnost izobraževanja učiteljev. Posebno težavo inteligence (ChatGPT) v predavanja in vaje vpliva na učno

predstavlja dejstvo, da so trenutni modeli ChatGPT izdelani na izkušnjo študentov pri predmetu Vodenje in organiziranje

podatkih spletnih uporabnikov, ki odražajo vrednote in medijske produkcije?

prevladujoče družbene norme globalnega severa [3]. Hipoteza 1: Vključitev umetne inteligence (ChatGPT) v

Raziskave smo se lotili s pregledom virov, pri čemer smo predavanja in vaje pomeni večjo zanimivost in učinkovitost.

ugotovili, da je ChatGPT izredno interaktiven, odziven, hiter in Hipoteza 2: Vključitev umetne inteligence (ChatGPT) v

zelo priljubljen, saj je sposoben praktično v trenutku odgovoriti predavanja in vaje omogoča bolj učinkovit študij, ker je uporaba

na različna vprašanja z različnih področij, kar omogoča hitro personalizirana

pridobivanje podatkov in napotil, razjasnjevanje določenih

pojmov in dilem ipd. Strokovnjaki za akademsko integriteto [5]

trdijo, da namesto da bi preprečevali uporabo orodij, kot je 3 METODE

ChatGPT, moramo kritično ovrednotiti, kako izvajamo uporabo Uporabili smo kvalitativno raziskovalno metodo, pri kateri

in ocenjevanje. Klasične metode, kot so izpiti na daljavo, pisanje smo podatke dobili s pregledom in analizo virov, anketo med

domačih nalog in spletni izpiti, se ponovno vrednotijo in študenti ter pogovori in intervjuji s študenti in strokovnjaki o

nadomeščajo z netekstovnimi izdelki, z razvojem veščin in z rezultatih raziskave. Pri anketah smo uporabili štiri tipe vprašanj

multimodalnimi preizkusi, izpiti in nalogami v razredu. – odprtega in zaprtega tipa, Likertovo lestvico in vprašanja z

Multimodalni preizkusi se nanašajo na preizkuse in predstavitve, rangiranjem.

ki vključujejo več različnih modalnosti ali načinov izražanja in Študenti so se seznanili z osnovnimi principi in možnostmi

komunikacije. Multimodalni preizkusi združujejo več oblik uporabe UI, opravili so precej vaj, pripravili pa tudi tri projekte:

izražanja, na primer besedilo, slike, zvok, video in drugo. Ti  program za izboljšanje organizacije svojega življenja

preizkusi so zasnovani tako, da študentom omogočajo uporabo je pripravil vsak študent sam;

različnih veščin in izraznih sredstev ter spodbujajo raznolikost  strokovni priročnik s področja vodenja in organizacije pri izražanju znanja.

medijske produkcije na izbranem področju, glede na

Uporaba umetne inteligence ima precejšen vpliv na izvedbo

študijskega procesa in študente. Članek [4] prikazuje težave pri interes študentov, so pripravili študenti sami ali pa po

zanesljivem odkrivanju vsebine, ki jo ustvari umetna inteligenca, dva ali trije skupaj, glede na željo študentov;

in vzbuja zaskrbljenost glede morebitnih lažnih obtožb proti  1,5-minutni film, posnet v enem kadru na določeno

študentom. V članku poudarjajo, da je nujno, da damo prednost temo, so študenti pripravili v skupinah po 5 študentov

etični uporabi umetne inteligence, negujemo pismenost z umetno in tako spoznavali tudi način projektnega dela v

inteligenco in razvijemo okvire, ki študentom in učiteljem skupinah.

omogočajo varno izkoriščanje celotnega potenciala teh

tehnologij (str. 11).

Uporaba umetne inteligence (ChatGPT) pomeni za vsakega 4 REZULTATI IN DISKUSIJA

učitelja, pa tudi za študenta, velik izziv, ker so učinki zelo hitri, Raziskovalno vprašanje: Kako vključitev umetne

veliki in raznovrstni, obenem pa je še veliko nedorečenega in inteligence (ChatGPT) v predavanja in vaje vpliva na učno

nepreizkušenega, kar zahteva še dodatno razmišljanje in skrbno izkušnjo pri predmetu Vodenje in organiziranje medijske

pripravo vključevanja možnosti, ki jih omogoča UI, v produkcije?

poučevanje. Večina študentov (86,1 %) meni, da je uporaba umetne

Z raziskavo želimo dati skromen prispevek k poznavanju in inteligence pozitivno vplivala na jasnost razumevanja

razumevanju uvajanja ChatGPT v študijski proces, obenem pa predstavljenih vsebin v predavanjih in vajah. 63,9 % študentov

koristne predloge uporabiti pri načrtovanju predmeta v verjame, da je uporaba umetne inteligence naredila predmet bolj

naslednjem študijskem letu in splošno tudi pri dejavnostih uporaben za njihovo prihodnje delo v medijski produkciji. Kar

fakultete in njene vizije ter strategije uvajanja UI. 80,5 % študentov je bilo zadovoljnih z uporabo umetne

inteligence v izobraževalnem procesu, pri čemer je 44,4 %



2 izrazilo precejšnje zadovoljstvo. Študenti najbolj cenijo večjo NAMEN IN CILJI RAZISKAVE

učinkovitost pri izvajanju nalog in razvoj analitičnih veščin,

Raziskavo smo pripravili in izvedli z željo, da spoznamo, ali medtem ko menijo, da umetna inteligenca ne prispeva k boljšemu

in kako je mogoče ChatGPT smiselno vključiti v poučevanje pri sodelovanju med študenti.

predmetu Vodenje in organizacija medijske produkcije, v 2. Najpomembnejši vidiki uporabe umetne inteligence za

letniku visokošolskega strokovnega programa Mediji in študente so hiter in enostaven dostop do gradiv in podatkov, 24/7

novinarstvo na FUDŠ, Ljubljana. Obenem smo želeli ugotoviti, dosegljivost, interaktivnost in hitra odzivnost.

katere so prednosti in težave, ki se pri tem pojavijo, in jih Kljub pozitivnim izkušnjam so nekateri izrazili previdnost pri

primerjati z ugotovitvami raziskav iz drugih izobraževalnih zanesljivosti umetne inteligence, zlasti glede natančnosti

sistemov. podatkov in virov in etičnih vprašanj.

Z raziskavo smo želeli tudi ugotoviti, kako študenti Študenti so izrazili večjo motivacijo zaradi lažjega

ocenjujejo vpliv uporabe ChatGPT na zanimivost in kakovost premagovanja izzivov, vendar so hkrati opozorili na potrebo po

izvedbe predmeta, učno izkušnjo in ocenjevanje znanja pri jasnih mejah uporabe, da se prepreči zmanjšanje motivacije, ker

predmetu. UI prevzame delo posameznika.



Raziskava kaže, da ima uporaba umetne inteligence večinoma  Manj pomembne prednosti

pozitiven vpliv na učno izkušnjo študentov, vendar je potrebna Manj pomembne prednosti vključujejo spoznavanje

previdnost glede zanesljivosti, meja uporabe in ohranjanja učinkovitejšega učenja, razvoj praktičnih znanj, hitrejše



motivacije in angažiranosti. reševanje nalog, več truda in znanja pa tudi večjo Navajamo še odgovore na dve odprti anketni vprašanji: individualnost med učenci. a.) Kaj vidite kot največjo oviro pri uporabi umetne 

Najmanj pomembne prednosti

inteligence v svojem študiju? Od 1 (najmanjša ovira) do

Kreativnost z UI je ocenjena kot najmanj pomembna

5 (največja ovira).

prednost v primerjavi s tradicionalnimi učnimi metodami.

Študenti so pripravili prek 150 predlogov, ki smo jih tudi s

Na podlagi rezultatov in diskusije lahko ocenimo ustreznosti

pomočjo UI razvrstili v pet kategorij.

 hipotez:

Nizka ovira: uporabniki prepoznajo le prednosti

Hipoteza 1 je potrjena: Vključitev umetne inteligence

umetne inteligence in ne zaznavajo ovir.

(ChatGPT) v predavanja in vaje pomeni večjo zanimivost in

 Srednje nizka ovira: različna dolžina odgovorov, učinkovitost.

slovnične napake, omejitve v originalnosti besedila, Hipoteza 2 je delno potrjena: Vključitev umetne inteligence

podobnosti besed med uporabniki in jezikovne (ChatGPT) v predavanja in vaje omogoča bolj učinkovit študij,

karakteristike uporabniškega vmesnika. Dodaten čas ker je uporaba personalizirana. Še večji pozitivni učinek pa

za uporabo računalnika in omejitve pri dostopu do študenti namenjajo nadgradnji razlage profesorja, visoko

nekaterih spletnih strani in podatkov. prilagodljivost, 24/7 dosegljivost, možnost sodelovanja pri

 nalogah ter hitro dostopanje do informacij. Srednje visoka ovira: omejena uporaba, skrbi glede

varstva podatkov in zasebnosti ter etičnih vprašanj in Rezultati celotne raziskave so dostopni na [6].

strahu pred nadomestitvijo ljudi z UI. Problem so tudi



zastareli podatki, verjetno prekomerna raba umetne 5 ZAKLJUČEK inteligence na račun zmanjšanja samostojnega

V letu 2023 smo se odločili opraviti raziskavo, s katero bi

razmišljanja. ugotovili nekatere osnovne značilnosti, možnosti in tudi

 Visoka ovira: umetna inteligenca vedno ne ustreže potencialne težave uporabe UI (konkretno ChatGPT) pri

zahtevam, določenih vsebin ne pojasni zadosti, načrtovanju in izvedbi študijskega procesa. Uvodno raziskavo

spodbuja lenobo pri pripravi nalog, saj lahko smo opravili pri predmetu Vodenje in organiziranje medijske

prevzame reševanje vaj in pripravo gradiv, pogosto produkcije v visokošolskem strokovnem programu Mediji in

poda napačne informacije in deluje nepredvidljivo. novinarstvo na FUDŠ, Ljubljana, pri načrtovanju in izvedbi

 študijskega procesa, ko smo poskusili dobiti še mnenje študentov Najvišja ovira: netočnost in nenatančnost ter

zastarelost podatkov, pomanjkanje transparentnosti, o tem, kako je vključitev UI v poučevanje predmeta vplivala na

njihovo učno izkušnjo, na zadovoljstvo s predmetom in tudi na

etična vprašanja in odgovornosti ter v določeni meri



regulativa, ki je še v pripravi. V raziskavi smo se pri uporabi ChatGPT osredotočili ocene, ki so jih pridobili pri predmetu.

b.) Katere prednosti vidite pri uporabi tradicionalnih učnih predvsem na tiste funkcije, ki so ključne za pridobivanje in

metod v primerjavi z umetno inteligenco? Od 1 (najmanj obdelovanje podatkov in jih študenti tudi največ uporabljajo. UI

pomembno) do 5 (najbolj pomembno). omogoča predvsem hiter pregled in zapisan prikaz določenega

obstoječega znanja, kar pomeni hiter dostop do ustrezno

Študenti so pripravili prek 170 predlogov, ki smo jih tudi s kakovostnega in urejenega materiala, v katerem so praviloma

pomočjo UI razvrstili v pet rangiranih kategorij. rešitve in predlogi, do katerih bo posameznik težje prišel sam ali

 Najpomembnejše prednosti pa le z dolgotrajnim iskanjem po raznih virih, ki pogosto niti niso

Ključne prednosti UI so: nadgradi razlago profesorja, dostopni. UI jim lahko veliko pomaga tudi pri delu kot pomoč pri

visoka prilagodljivost, 24/7 dosegljivost, možnost brainstormingu, analizi, sintezi in vrednotenju rezultatov ipd.

sodelovanja pri nalogah in hitro dostopanje do informacij. Za ljudi, ki niso navajeni in preveč vešči pisanja in branja, kar

 Pomembne prednosti je precej značilno za današnjo mladino, je tak pristop precej

Pomembne prednosti so: hitro iskanje informacij, motivacijski, obenem pa tudi edukacijski, saj so primorani in tudi



spodbujanje kreativnega in kritičnega razmišljanja, praviloma morajo še preveriti, dodelati in nadgraditi z raznimi viri, pa tudi svojimi raziskavami, razmišljanji, dognanji in personalizacija učenja ter razvoj socialnih veščin, medtem vzpostavljanje motivirani za branje in študij z UI pridobljenim gradivom, ki ga medčloveških stikov in odnosov,



ko je umetna inteligenca manj zahtevna glede znanja in strokovnjaki ob načrtovanju in izvedbi konkretnih nalog. predvsem s poglobljeno razpravo z mentorji ali drugimi

napora. Ugotovili smo, da večina študentov meni, da je uporaba

 Srednje pomembne prednosti umetne inteligence pozitivno vplivala na jasnost razumevanja

Enostavno iskanje informacij, ustvarjanje zgodb, socialna predstavljenih vsebin v predavanjih in da je uporaba umetne

interakcija in prilagodljivost delovanja profesorja so inteligence naredila predmet bolj uporaben za njihovo prihodnje

ocenjeni kot srednje pomembni, saj lahko učenci prihranijo delo v medijski produkciji. Študenti najbolj cenijo večjo

čas, se lažje spomnijo gradiva in se bolje prilagajajo svojim učinkovitost pri izvajanju nalog in razvoj analitičnih veščin,

potrebam.



medtem ko menijo, da umetna inteligenca ne prispeva k boljšemu ZAHVALA

sodelovanju med študenti. Za pomoč in zelo tvorno sodelovanje se zahvaljujemo

Za spodbudo učiteljem pri uvajanju ChatGPT v študijski študentom 2. letnika programa Mediji in novinarstvo (v š.l.

proces navajamo misel študentke: 2023/24) na FUDŠ, Ljubljana.

 Z uporabo ChatGPT sem za študij in delo veliko bolj

motivirana, saj vem, kje lahko dobim potrebne VIRI

informacije za katerokoli vsebino. Vem, da bom s [1] Anguiano Dani , Beckett Lois, 1. 10. 2023, How Hollywood writers

pomočjo UI naredila boljšo nalogo, zato tudi z večjim triumphed over AI – and why it matters, How Hollywood writers

triumphed over AI – and why it matters | US writers' strike 2023 | The

veseljem začnem z delom. Guardian (pridobljeno 15.12.2023).

Raziskava kaže, da ima uporaba umetne inteligence večinoma [2] Education in the age of artificial intelligence (okt., dec. 2023), Education

in the age of artificial intelligence | The UNESCO Courier (pridobljeno

pozitiven vpliv na učno izkušnjo študentov, vendar je potrebna 15. 12. 2023).

previdnost glede zanesljivosti, meja uporabe in ohranjanja [3] Artificial intelligence and the Futures of Learning, 2023, Artificial

intelligence and the Futures of Learning | UNESCO (pridobljeno

motivacije in angažiranosti. Pri nadaljnjem uvajanju UI v 15. 12. 2023).

izobraževanje bo treba upoštevati napredek v teoretičnih in [4] Tom Farrelly Tom, Baker Nick, 2023, Generative Artificial Intelligence:

Implications and Considerations for Higher Education Practice, Educ.

praktičnih dognanjih in praktične izkušnje pri načrtovanju in Sci. 2023, 13(11), 1109; https://doi.org/10.3390/educsci13111109

izvedbi predmeta. [5] Shir Ali, 2023, Information Integrity, Academic Integrity, and Generative

AI, Information Integrity, Academic Integrity, and Generative AI -

Information Matters (pridobljeno 15. 12. 2023).

[6] ZAKRAJŠEK, Srečo. 2024, Uporaba umetne inteligence (ChatGPT) pri

predmetu Vodenje in organiziranje medijske produkcije na FUDŠ,

Ljubljana: poročilo o raziskovalni nalogi. Ljubljana: IAM Visoka šola za

multimedije: FUDŠ, 2024. 60 str., ilustr. Digitalna knjižnica Slovenije -

dLib.si. [COBISS.SI-ID 180637955] (pridobljeno 1. 6. 2024).



Zbornik 27. mednarodne multikonference



INFORMACIJSKA DRUŽBA – IS 2024

Zvezek H



Proceedings of the 27th International Multiconference



INFORMATION SOCIETY – IS 2024

Volume H



Digitalna vključenost v informacijski družbi



Digital Inclusion in Information Society



Urednika / Editors



Matjaž Debevc, Ines Kožuh



http://is.ijs.si



9. oktober 2024 / 9 October 2024

Ljubljana, Slovenia





PREDGOVOR




Danes, v času hitrega razvoja digitalnih tehnologij, digitalna vključenost ostaja ključen steber

naše informacijske družbe. Širjenje dostopa do tehnologij in digitalnih storitev na vedno večji

del prebivalstva, vključno z osebami z različnimi oblikami oviranosti, zahteva nenehno

izboljševanje dostopnosti, razumljivosti in uporabnosti digitalnih rešitev. Ob naraščajoči

potrebi po prilagojenih rešitvah je naš cilj zagotoviti, da vsi posamezniki – ne glede na

njihove zmožnosti – lahko v celoti sodelujejo v digitalni dobi in izkoristijo njen polni

potencial.



Osrednja tematika konference "Digitalna vključenost v informacijski družbi – DIGIN 2024"

je, kako lahko sodelujemo za ustvarjanje digitalno dostopnega okolja, ki omogoča

enakopravno vključitev vseh. Naša druga hibridna konferenca, ki se ponovno izvaja v

sodelovanju z evropskim centrom virov o dostopnosti – AccessibleEU, bo združila vodilne

slovenske in mednarodne raziskovalce ter strokovnjake, ki bodo predstavili svoje rešitve,

spoznanja in metode za izboljšanje digitalne vključenosti. Skupaj bomo raziskali, kako lahko

s podporno tehnologijo ter prilagojenimi digitalnimi pristopi premagamo ovire, s katerimi se

srečujejo osebe z različnimi oblikami oviranosti, in ustvarimo digitalni svet, ki je resnično

dostopen vsem.



Prispevki, zbrani v letošnjem zborniku, predstavljajo bogat vir znanja in inspiracije za vse

udeležence. Preko njih bomo bolje razumeli, kako tehnologijo uporabiti za izboljšanje

kakovosti življenja in omogočiti enakopravno vključevanje v digitalno okolje. Verjamemo, da

bomo skupaj s tem dogodkom pomembno prispevali k bolj vključujoči in dostopni digitalni

prihodnosti.



Uredniški odbor





FOREWORD




Today, in an era of rapidly advancing digital technologies, digital inclusion remains a

fundamental pillar of our information society. Expanding access to digital technologies and

online services to an increasingly diverse population, including persons with disabilities,

requires ongoing improvements in accessibility, clarity, and usability. As the need for tailored

solutions grows, our goal is to ensure that all individuals – regardless of their abilities – can

fully engage with the digital world and leverage its full potential.



The central theme of the conference “Digital Inclusion in the Information Society – DIGIN

2024” is how we can collaborate to create a digitally accessible environment that enables

equal participation for all. Our second hybrid conference, organized once again in cooperation

with the European Resource Centre on Accessibility – AccessibleEU, will bring together

leading Slovenian and international researchers and professionals to share their solutions,

insights, and methods for advancing digital inclusion. Together, we will explore how

supportive technologies and adapted digital approaches can help overcome the barriers faced

by persons with disabilities, thereby building a digital world that is truly accessible to

everyone.



The contributions compiled in this year’s proceedings provide a rich source of knowledge and

inspiration for all participants. Through these insights, we will gain a deeper understanding of

how technology can enhance quality of life and enable equal participation in the digital

environment. We believe that, through this event, we will make a significant contribution

towards a more inclusive and accessible digital future.



The Editorial Board



PROGRAMSKI SVET / STEERING COMMITTEE



Matjaž Gams (Predsdnik / Chair) – Institut Jožef Stefan



Matjaž Debevc – AccessibleEU



Zoran Ren – Univerza v Mariboru, Fakulteta za strojništvo



Marta Licardo – Univerza v Mariboru, Pedagoška fakulteta



Boštjan Šumak – Univerza v Mariboru, Fakulteta za elektrotehniko, računalništvo in informatiko



Damjana Kogovšek – Univerza v Ljubljani, Pedagoška fakulteta



Matevž Pogačnik – Univerza v Ljubljani, Fakulteta za elektrotehniko



Dušan Caf – Alma Mater Europaea



PROGRAMSKI ODBOR / PROGRAMME COMMITTEE



Ines Kožuh (Predsednik) – Univerza v Mariboru, Fakulteta za elektrotehniko, računalništvo in

informatiko



Andraž Petrovčič (Podpredsednik) – Univerza v Ljubljani, Fakulteta za družbene vede



Maja Pušnik – Univerza v Mariboru, Fakulteta za elektrotehniko, računalništvo in informatiko



Katja Kous – Univerza v Mariboru, Fakulteta za elektrotehniko, računalništvo in informatiko



Darinka Verdonik – Univerza v Mariboru, Fakulteta za elektrotehniko, računalništvo in informatiko



Mirjam Sepesy Maučec – Univerza v Mariboru, Fakulteta za elektrotehniko, računalništvo in

informatiko



Patricija Sedminek – Univerza v Mariboru, Pedagoška fakulteta



Natalija Špur – Univerza v Mariboru



Narvika Bovcon – Univerza v Ljubljani, Fakulteta za računalništvo in informatiko



Jaka Čibej – Univerza v Ljubljani, Fakulteta za računalništvo in informatiko



Dejan Lavbič – Univerza v Ljubljani, Fakulteta za računalništvo in informatiko



Aleš Smrdel – Univerza v Ljubljani, Fakulteta za računalništvo in informatiko



Klemen Pečnik – Univerza v Ljubljani, Fakulteta za elektrotehniko



Žana Juvan – Univerza v Ljubljani, Fakulteta za elektrotehniko



Meta Furlan – Univerza v Ljubljani, Filozofska fakulteta



Nives Ličen – Univerza v Ljubljani, Filozofska fakulteta



Jelena Ficzko – Univerza v Ljubljani, Zdravstvena fakulteta



Anamarija Kejžar – Univerza v Ljubljani, Fakulteta za socialno delo



Jana Mali – Univerza v Ljubljani, Fakulteta za socialno delo



Simona Hvalič Touzery – Univerza v Ljubljani, Fakulteta za družbene vede



Exploring Digital Media Literacy: A Case Study of Elderly



Deaf Users in Slovenia



Laura Horvat† Peter Čakš Irena Lovrenčič Ines Kožuh

Faculty of Electrical Faculty of Electrical Držanič Faculty of Electrical

Engineering and Computer Engineering and Computer Engineering and Computer Faculty of Electrical

Science Science Science (University of

Engineering and Computer

University of Maribor Slovenia University of Maribor Slovenia Maribor), Faculty of Social Science

laura.horvat2@student.um.si peter.caks@um.si Sciences (University of University of Maribor Slovenia

irena.lovrencic@um.si Ljubljana)

ines.kozuh@um.si,

ines.kozuh@fdv.uni-lj.si



Abstract 2 Digital media literacy



This study explores digital media literacy and usage patterns 2.1 Digital Media Literacy among elderly deaf individuals in Slovenia. Interviews with two

Digital media literacy refers to the ability to access, understand



media literacy, particularly in critical media consumption. The and create content using digital media [3]. These skills are participants revealed moderate digital literacy but limited digital

participants use social media mainly for communication and essential in the information age, where rapid technological

information. The findings highlight a potential relationship advancements demand constant adaptation and learning [4].

between digital media literacy and usage frequency, Digital media literacy involves not only technical skills, but also

underscoring the need for tailored digital literacy training for the includes emotional response, consumption and cultural

deaf community. evaluation of media content [3, 5].

With social media becoming a key source of information for

Keywords many, the importance of social media literacy increased. These

Digital media literacy, social media literacy, deaf, social media, platforms became a popular way of spending free time, places for

digital media, new media, online media platforms. socialisation and communication with others in a personal or

business environment [6]. Digital media usage has also shifted

from desktops to mobile devices and tablets, changing the user

1 experience, frequency and mode of use [7, 8]. Introduction



In the digital age digital media play a central role in daily life, 2.2 Indicators of Digital Media Literacy

shaping how individuals access, understand and create digital Digital media literacy involves four main areas: critical

content. This shift presents unique challenges for D/deaf and consuming, critical prosuming, functional consuming and

hard of hearing (DHH) individuals in navigating the digital functional prosuming [9]. Critical consuming includes the ability

landscape [1, 2]. This paper explores the digital literacy and to analyse, synthesise and evaluate media content [5]. Critical

digital media literacy of DHH individuals in Slovenia, focusing prosuming involves interactive participation and critical content



of digital media’s accessibility. The objectives were to evaluate creation in new media environments, such as chatrooms. The on their abilities to navigate digital media and their perceptions

the level of (digital) media literacy among DHH individuals in individuals’ critical abilities are exhibited in their understanding

Slovenia, and examine their experience with social media and of social and cultural values [10, 11]. Functional consuming

online media platforms. By examining these aspects, the study refers to technical skills in using media content, such as

seeks to contribute to a better understanding of how DHH navigating the internet and using various digital devices [9, 12].

individuals interact with digital media. Since the study is Functional prosuming includes the technical skills needed for

ongoing, this paper presents only the preliminary findings from content production and distribution, such as creating and sharing

interviews with two deaf individuals. photos and videos on social media [12].



3 DHH in the digital media



3.1 Statistics on DHH Individuals

Permission to make digital or hard copies of part or all of this work for personal or In Slovenia there are approximately 1,500 deaf people and

classroom use is granted without fee provided that copies are not made or distributed around 50,000 hard of hearing individuals, according to the

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must Association of the Deaf and Hard of Hearing of Slovenia [13].

be honored. For all other uses, contact the owner/author(s). The European Union of the Deaf (EUD) reports further that

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s). around 1,850 individuals in Slovenia identify as deaf, and 1,000

https://doi.org/10.70314/is.2024.DIGIN.1 of them use Slovenian Sign Language [14].



3.2 Communication Habits of DHH 4.2 Research Questions



Individuals The study focused on two main research questions:

Communication among DHH individuals involves primarily the (1) What is digital media literacy among elderly deaf

use of sign language, with Slovenian Sign Language being individuals?

recognised officially and protected by law [15]. The right to use (2) How do elderly deaf individuals use digital media?

sign language includes access to interpreters and information

through adapted techniques, ensuring inclusion in social and 4.3 Participants

professional environments [16, 17]. In addition to sign language, The interviewees were two deaf individuals residing in Slovenia,

DHH individuals use written text and visual aids to facilitate who belong to the group of elderly people aged 60 years or older.

communication, particularly in digital contexts. Video calls via They were selected randomly from a larger sample of 12

platforms (e.g., Skype, Teams and Zoom) have become essential, participants, to present the preliminary results of the ongoing

especially during the COVID-19 pandemic, enabling real-time research study. Participant 1 (P1) is a 63 year old female with a

visual communication [18]. hearing loss level of 96 dB. Her educational level is lower or

vocational secondary. Participant 2 (P2) is a 60 year old female

3.3 Habits of Using Digital Media with a hearing loss level of 100 dB, and has primary school

DHH individuals use digital media increasingly for various education. Both participants lost their hearing early in life and

purposes, from social interaction, information access to their primary language is Slovenian Sign Language.

entertainment. Research indicates that DHH users use social

media platforms frequently to connect with their community and 4.4 Measuring Instrument

share experiences [19, 20]. DHH users particularly value video The measuring instrument consisted of several parts. Prior to

content with subtitles, text transcriptions or sign language for participation, the participants were questioned on their

improved understanding. The shift to digital media has also seen demographic data (gender, age, education level), followed by

a rise in the use of mobile applications tailored to DHH needs, data on their hearing loss (level of hearing loss, primary

such as those for real-time captioning and sign language language, hearing loss history). Next, the participants were

interpretation [21, 22, 23]. questioned on their use of digital media (social media and online

media platforms). This part consisted of questions on their

3.4 Challenges in Using Digital Media purpose of use, content creation, frequently encountered content

Despite many benefits, DHH individuals also encounter and frequency of use.

challenges in the digital environment. Accessibility is a common The next section examined digital literacy [29, 30], digital

issue, as many digital platforms lack subtitles or sign language media literacy [31, 32], and literacy about social media [33].

features [24, 25]. This can lead to social isolation and exclusion Lastly, the participants were also asked about their perception of

from mainstream digital interactions and content [26, 27]. the accessibility of digital media for DHH, as proposed by

Similar constraints arise in everyday occasions, such as public Maiorana-Basas & Pagliaro [34] and Kožuh & Debevc [35].

speeches, often conducted without an interpreter, and official

information delivered to individuals via phone calls, posing

5 Results

significant barriers for DHH individuals. Another significant

challenge is the lack of digital literacy training tailored to the 5.1 Digital Media Literacy Among Elderly

DHH community, such as those in sign language [28]. Deaf Individuals (RQ1)

Both interviewees demonstrated a satisfactory level of digital



4 literacy. P1 uses only a phone, while P2 uses a computer and Methodology

tablet, with a clear preference for the phone. Both have been

4.1 Procedure using digital devices for years, learning about them informally.

This case study examines two DHH individuals in Slovenia, who P1 learned to use them about 10 years ago with a partner’s

were recruited through Associations of DHH people. The assistance, and P2 15 years ago with her son’s help, and, later, at

inclusion criteria were (a) Identification as a DHH individual, (b) work. Neither expressed a desire to improve their digital literacy,

Membership in associations of DHH people, (c) Active use of at believing their current skills were sufficient for their needs. Both

least one social media platform, (d) Use of online media rely on others for help when encountering difficulties (e.g.,

platforms, e.g., dostopno.si. We collected data using semi- suspected money fraud), with P1 turning to her partner and P2

structured joint interviews, providing a holistic view and direct consulting an interpreter.

observation of the participants` interactions [36, 37]. A sign Both interviewees also follow traditional media, either

language interpreter was present to ensure accurate television or newspapers. While critical thinking skills are a

communication and provide additional support and information. crucial part of media literacy, it is concerning that neither verified

The interview took place at the Association of the Deaf and Hard the veracity of the online content. P1 typically relies on

of Hearing of the Podravje, Maribor unit, and was audio-recorded acquaintances to check veracity, and has occasionally clicked on

to facilitate the analysis of the interpreter's translation. The data fake news due to its attractive content, though she did not share

were later transcribed and analysed according to Roblek [39]. it. She mentioned, "I can tell when it is fake news". P2 is more

suspicious of online content and fears internet scams, as



indicated by her concern about the dangers of Facebook: "There to their lower media literacy. This leads us to conclude that the

are more and more of these scams". Neither participant interviewees might not think critically, and we might therefore

demonstrated an understanding of media bias and political classify their digital media literacy as lower. Their lack of

influence on media impartiality. understanding of it does not bother them, nor do they wish to

Social media literacy was examined in regard to the improve their digital media literacy. We believe that they may be

proficiency in navigating the Facebook application. P2 showed a compelled to use social media platforms (e.g., WhatsApp,

significantly greater proficiency, as she could navigate the main Facebook Messenger) and to follow the online media portal

features (post, comment, share) with more ease and dostopno.si, as they cannot communicate or inform themselves

independence, but also caution. In contrast, P1 required more otherwise, no matter how well they understand digital media.

time, and showed some confusion, asking P2 for help in learning

to post stories. Both participants acknowledged the positive

impact of social media platforms on maintaining relationships 6 Discussion

and staying informed. To assess the interviewees’ digital media literacy, we focused on

answering two Research Questions. Both participants have

5.2 Use of Digital Media by Elderly Deaf exhibited a moderate level of digital literacy (RQ1). Despite their

Individuals (RQ2) age and long-term hearing loss, they adapted to using digital

Both interviewees use social media platforms frequently, mainly devices and social media platforms, learning informally through

in the morning. Occasionally, they access them multiple times a family members and work experiences. However, they lacked

day. They both use Facebook, Facebook Messenger and critical thinking skills, a key component of media literacy.

WhatsApp, while P2 also uses Instagram and Imo. They use Neither of the interviewees verified the accuracy of online

social media to browse content, communicate and interact with content, indicating a gap in their digital literacy abilities. It is

posts from other users, particularly those featuring gestures in important to note that the primary language of both interviewees

photos or videos. One example is the travelling videos from is Slovene sign language, while social media platforms and most

fellow deaf individuals, which P2 finds interesting, and online media portals are based on text-based communication. An

exception is the online media portal dostopno.si, which the

understands due to international gestures. Their primary reasons interviewees know, and use mainly because the video content has

for using social media platforms are entertainment and staying an interpreter subtitling and transcripts.

updated with current events in their hometown. P2 also uses Both interviewees are active users of social media platforms

social media platforms for direct communication and occasional and digital media platforms (RQ2). They use social media

posting, such as photos and videos/stories of interesting things. platforms frequently (e.g., Facebook, WhatsApp) and mobile

Contrarily, P1 does not publish her own content. apps for communication (e.g., Imo). They use it for browsing

The interviewees follow Slovenian online media portals, content, communication, and interacting with posts, especially

including rtvslo.si, dostopno.si, and 24ur.com, with P1 also using those that include images and videos with gestures for better

zurnal24.si. P2 also follows foreign, mainly Croatian, portals understanding. There was a notable difference in content creation

displayed in her Facebook feed. They access these portals once a habits, as one participant frequently published her own content

day, usually in the morning, and seek mainly for information, but and the other avoided such behaviour.

also entertainment. Both access online media content through Following these findings, the connection between the use of

social media platforms rather than direct searches. The content digital media and digital media literacy is unclear. Although both

they typically encounter is genre-specific, including news, interviewees were cautious in using social media, they exhibited

culture, entertainment, sports and arts. When asked about the different levels of proficiency. P2, who shows higher digital

media portal dostopno.si, a portal for people with disabilities, literacy, engages more actively and confidently with digital

both expressed familiarity with it and past use, as they require media, creating and sharing content regularly. She is proficient

the presence of an interpreter to understand the content. in using multiple devices and platforms, indicating a higher level

The interviewees were also asked about the accessibility and of digital literacy compared to P1. In contrast, P1's lower

adaptation of the said social media platforms and online media engagement and more passive consumption of digital media

portals for the needs of deaf people, which they both agreed were reflect her more limited digital literacy. Both interviewees follow

sufficient. Nevertheless, P2 has had difficulties with verbal various online media portals for information and entertainment,

communication, as she finds it difficult or impossible to accessing content primarily through social media platforms

understand speech without gestures. rather than searching for it.

We observed that neither participant understood the Based on Lin et al`s (2013) indicators of digital media

importance of media impartiality. They also failed to avoid fake literacy, both interviewees fall into the functional category.

news, often clicking on links and reading the content (for P2 Neither can be classified as critically consuming or critically

there is a possibility that she would believe the content). When prosuming, although P1 exhibits some skepticism towards fake

asked about information, neither of them checks information news. Both possess the technical skills necessary for functional

from different online media sources. They do ask acquaintances consuming, with P1 proficient in using a phone, the internet and

about the veracity, but even they cannot know for sure if a certain social media platforms, and P2 is also competent with a computer

piece of information is really accurate. The fact that the and tablet. P2 additionally fits into the functional prosuming

interviewees do not seek out media content intentionally, but category, due to her active content creation on social media

only access it via social media platforms, might also contribute platforms. Despite initial help from family members, she now



shares photos and videos independently, underscoring her [15] Zakon o uporabi slovenskega znakovnega jezika (ZUSZJ). 2002. Uradni

functional literacy. list RS, 96/02. Accessed on:

https://pisrs.si/pregledPredpisa?id=ZAKO1713

Apart from the small sample size, an important limitation of [16] Jasna Bauman, Veronika Ciglar, Darja Holec, Matjaž Juhart, Damjana

this study are potential misunderstandings due to the presence of Kogovšek, Stanislav Košir, Marjetka Kulovec, Martina Ozbič, Petra Rezar

an interpreter during data collection. This may result in and Andreja Žele. 2009. Stanje slovenskega znakovnega jezika:

ekspertiza. Zavod Združenje tolmačev za slovenski znakovni jezik.

misinterpretations of expressions between the interviewees and [17] Laura Horvat. 2021. Slovenski znakovni jezik in njegova uporaba na

the interviewer. javnem radiotelevizijskem servisu v Sloveniji (RTV SLO) Bachelor’s

thesis. Univerza v Mariboru, Fakulteta za elektrotehniko, računalništvo in

informatiko. Accessed on:

https://dk.um.si/Dokument.php?id=152687&lang=slv

7 [18] Maartje De Meulder and Nienke Sijm. 2024. “I feel a bit more of a conduit Conclusions

now”: Sign language interpreters coping and adapting during the COVID-

Our study reveals that digital media literacy is crucial in using 19 pandemic and beyond. Interpreting and Society 4, 1 (2024), 3–25.

digital media effectively. The interviewees' limited digital media [19] Ines Kožuh, Manfred Hintermair, Andreas Holzinger, Zala Volčič and

Matjaž Debevc. 2015. Enhancing universal access: deaf and hard of

literacy led to less critical and cautious use of digital media. This hearing people on social media platforming sites. Universal Access in the

study identified the social media platforms and online media Information Society 14 (2015), 537–545.

portals they use, as well as their reasons and ways of use. The [20] Muhammad Ali and Anza Nasir. 2023. Investigating the Social Media

Practices among Aurally Challenged Students (ACS): The challenges and

most important reasons for using them were communication, issues. Human Nature Journal of Social Sciences 4, 1 (2023), 518–532.

entertainment, and information. We found that higher digital [21] Ehsan Toofaninejad, Esmaeil Zaraii Zavaraki, Shane Dawson, Oleksandra

Poquet and Parviz Sharifi Daramadi. 2017. Social media use for deaf and

media literacy may increase their presence in the digital hard of hearing students in educational settings: A systematic review of

environment, as well as enhance their understanding of digital literature. Deafness & Education International 19, 3–4 (2017), 144–161.

[22] Jess A. Cuculick. 2014. Facebook among deaf college students: Deaf-gain

devices and media. and funds of knowledge. Ph.D. Dissertation. University of Rochester.

[23] Emem P. Udofia, Daniel A. Aloysius and Victoria David Jimmy. 2017.



Acknowledgments Internet Resources And Information Literacy Of Hearing And Speech

Impaired Students In Nigerian Academic Libraries. Computing &

We extend our sincere gratitude to the two interviewees for their Information Systems 21, 1 (2017).

[24] Andrew Solomon. 2012. Far from the Tree: Parents, Children and the

insights and participation in the interview. We also thank the Search for Identity. Simon and Schuster. Accessed on:

Association of the Deaf and Hard of Hearing of the Podravje http://archive.wilsonquarterly.com/sites/default/files/articles/24_CBFarfr

(Društvo gluhih in naglušnih Podravja) for their invaluable omtheTree.pdf

[25] David M. Frost 2011. Social stigma and its consequences for the socially

support and contribution to this study. stigmatized. Social and Personality Psychology Compass 5, 11 (2011),

824–839.



References [26] Margaret P. Brown and Andrew Cornes 2015. Mental health of deaf and

hard-of-hearing adolescents: What the students say. Journal of Deaf

Studies and Deaf Education 20, 1 (2015), 75–81.

[1] Julio Abascal, Simone D. J. Barbosa, Colette Nicolle and Panayiotis

[27] Sylvia Olsson. 2021. Longing to belong: Deaf and hard of hearing young

Zaphiris. 2016. Rethinking universal accessibility: A broader approach adults’ social interaction, social relationships, and identity. Doctoral

considering the digital gap. Universal Access in the Information Society dissertation. Mälardalen University. Accessed on: https://www.diva-

15, 2 (2016), 179–182.

portal.org/smash/get/diva2:1566410/FULLTEXT02

[2] Nathan Shelena Soosay, Hussain Azham and Hashim Nor Laily. 2016. [28] Mary Rudner, Josefine Andin, Jerker Rönnberg, Mikael Heimann, Anders

Studies on deaf mobile application: Need for functionalities and

Hermansson, Keith Nelson and Tomas Tjus. 2015. Training literacy skills

requirements. Journal of Telecommunication, Electronic and Computer

through sign language. Deafness & Education International 17, 1 (2015),

Engineering 8, 8 (2016), 47–50.

8–18.

[3] Sora Park. 2011. Access to digital devices and its relationship to digital

[29] Sarah Soyeon Oh, Kyoung-A Kim, Minsu Kim, Jaeuk Oh, Sang Hui Chu

media literacy. In Proceedings of the Communications Policy and

and JiYeon Choi. 2021. Measurement of digital literacy among older

Research Forum (Sydney, Australia, Nov. 7–8, 2011).

adults: systematic review. Journal of Medical Internet Research 23, 2

[4] Eric M. Meyers, Ingrid Erickson and Ruth V. Small. 2013. Digital literacy

(2021), e26145.

and informal learning environments: an introduction. Learning, Media and [30] Kristīne Mackare and Anita Jansone. 2018. Habits of Using internet and

Technology 38, 4 (2013), 355–367.

digital devices in education. In SOCIETY. INTEGRATION.

[5] David Buckingham, Shaku Banaji, Andrew Burn, Diane Carr, Sue

EDUCATION: Proceedings of the International Scientific Conference,

Cranmer and Rebekah Willett. 2005. The media literacy of children and

vol. 5, 348–356.

young people. London, UK: Ofcom.

[31] Katherine N. Hayles. 2012. How we think: Digital media and

[6] Andrey Glukhov and Yuliya Stakhovskaya. 2021. Social media platform

contemporary technogenesis. University of Chicago Press.

digital literacy.

[32] Dirk Jahn and Alessandra Kenner. 2018. Critical thinking in higher

[7] Laurel J. Felt and Michael B. Robb. 2016. Technology addiction:

education: how to foster it using digital media. In The Digital Turn in

Concern, controversy, and finding balance. San Francisco, CA: Common

Higher Education: International Perspectives on Learning and Teaching

Sense Media.

in a Changing World, 81–109.

[8] Yolanda R. Chassiakos, Jenny Radesky, Dimitri Christakis, Megan A.

[33] Andreas M. Kaplan and Michael Haenlein. 2010. Users of the world,

Moreno, and Corinn Cross. 2016. Children and adolescents and digital

unite! The challenges and opportunities of Social Media. Business

media. Pediatrics 138, 5 (2016).

Horizons 53, 1 (2010), 59–68.

[9] Tzu-Bin Lin, Jen-Yi Li, Feng Deng and Ling Lee 2013. Understanding

[34] Michella Maiorana-Basas and Claudia M. Pagliaro. 2014. Technology use

new media literacy: An explorative theoretical framework. Journal of

among adults who are deaf and hard of hearing: A national survey. Journal

Educational Technology & Society 16, 4 (2013), 160–170.

of Deaf Studies and Deaf Education 19, 4 (2014), 400–410.

[10] Wasson Haidee. 2009. Convergence culture: Where old and new media [35] Ines Kožuh and Matjaž Debevc. 2018. Challenges in social media use

collide. Film Quarterly 62, 4 (2009), 84

among deaf and hard of hearing people: In Social Media Platforms

[11] Henry Jenkins and Mark Deuze. 2008. Convergence culture. Convergence

Science: Design, Implementation, Security, and Challenges: From Social

14, 1 (2008), 5–12.

Media Platforms Analysis to Social Media Platforms Intelligence, 151–

[12] Der-Thanq “Victor” Chen and Jing Wu. 2011. New media literacy in the

171.

21th century society: Key findings, gaps and recommendations. National

[36] Graham Allan. 1980. A Note on Interviewing Spouses Together. Journal

Association for Media Literacy Education, Philadelphia, PA, USA.

[13] Zveza društev gluhih in naglušnih Slovenije (ZDGNS). n.d. Gluhost. of Marriage & Family 42, 1 (1980).

[37] Hilary Arksey. 1996. Collecting data through joint interviews. Social

Accessed on: https://zveza-gns.si/o-zvezi/o-gluhoti/

Research Update (15).

[14] Lorraine Leeson, Jemina Napier, Tobias Haug, Teresa Lynch and Haaris

[38] Julie Seymour, Gill Dix, and Tony Eardley. 1995. Joint Accounts:

Sheikh. 2021. Access to justice for deaf signers: The Justisigns project. In

Methodology and Practice in Research Interviews with Couples. Social

UNCRPD Implementation in Europe—a deaf perspective: Article 9:

Policy Research Unit, University of York.

Access to information and communication, 161–175. European Union of

[39] Vasja Roblek. 2009. An Example of Performing a Text Analysis in a

the Deaf.

Qualitative Research). Management 4, 1 (2009), 53–69.





Innovative Technological Solutions for Cities




as a Response to the Challenges



of Multimodal Mobility for All Citizens



Inovativne tehnološke rešitve mest kot odgovor na izzive

multimodalne mobilnosti za vse prebivalce



Jani Demšar Marina Lovrić Roman Rener

Cartography Spatial Management Cartography

Geodetic Institute of Slovenia Geodetic Institute of Slovenia Geodetic Institute of Slovenia

Jamova cesta 2 Jamova cesta 2 Jamova cesta 2

Ljubljana, Slovenia Ljubljana, Slovenia Ljubljana, Slovenia

jani.demsar@gis.si marina.lovric@gis.si roman.rener@gis.si



Abstract often face significant barriers when navigating public spaces and utilizing public transportation. Mobility is a fundamental right of all people, crucial for social Digital data play a crucial role in supporting efficient participation, as recognized also by the UN Convention on the multimodal mobility, especially when it comes to understanding Rights of Persons with Disabilities. Despite this, persons with the entirety of the system, from standards to implementation. disabilities often face challenges when using public spaces and Real-time route planning apps that consider accessibility needs, transport. This paper emphasizes the importance of physical designated accessible parking information, and integrated accessibility, digital tools, and stakeholder collaboration to foster ticketing systems across various transportation modes facilitate inclusivity. Using mixed-methods approach, including data independent mobility. The primary aim of this paper is to explore cataloguing, field verification, surveys, and workshops, the study innovative technological solutions that can bridge the gap highlights the need for comprehensive data integration and between current challenges and the vision of inclusive urban accessible infrastructure to support independent mobility for mobility. persons with disabilities. It is important to approach and develop Furthermore, this paper presents research based on key mobility solutions for people with disabilities. These include projects including, "Enabling Multimodal Mobility for Persons methodology implementation, standards, pilot testing, with Various Disabilities" [3], "Spatial Data Support for Public adaptation, and integration of the entire system in the country, Passenger Transport Management – Persons with Disabilities in and the knowledge transfer to all stakeholders. Public Passenger Transport" [4] and "Analysis of Conceptual



Keywords Designs of Information and Technical System Support for

Persons with Disabilities to Increase Social Inclusion" [4]. These

multimodal mobility, spatial data, persons with disabilities, projects provide insights into the current state of accessibility of

public transport, accessibility public spaces and public transport infrastructure, addressing the

specific experiences of PwD.



1 Introduction

Mobility is a fundamental aspect of human life, enabling 2 Methods

individuals to navigate their environments and actively Understanding the travel needs of PwD is crucial for planning

participate in society. This right to independent movement cities with effective public transport, accessible public spaces,

extends to everyone, regardless of personal circumstances. The and diverse options for multimodal mobility. We employed a

United Nations Convention on the Rights of Persons with mixed-methods approach in our research to address this topic.

Disabilities guarantees this fundamental right to independent We developed various methodologies and object catalogues for

mobility for persons with disabilities (PwD) [1]. However, PwD data capture in the office. We developed methodologies for the

four groups focused on in the project, such as persons with

mobility impairments, persons with visual impairments, the deaf

Permission to make digital or hard copies of part or all of this work for personal or and those with hearing impairment, and the elderly. We have

classroom use is granted without fee provided that copies are not made or distributed covered the elements related to the mobility and accessibility of

for profit or commercial advantage and that copies bear this notice and the full the space for these people in the object catalogue. In the

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s). catalogue, elements are described as attributes with

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia corresponding values. All elements are depicted in real-life

https://doi.org/10.70314/is.2024.DIGIN.2 scenarios with maps and photos, providing precise locations and © 2024 Copyright held by the owner/author(s).

real-world contexts. With the help of catalogues, we captured

data in the office on a desktop computer with an Open-Source

Geographic Information System (GIS) program tool. We used tests that have already been carried out, such as research

the latest orthophoto in Slovenian national coordinate system conducted by Triglav Čekada and Radovan [6]. Participatory

D96/TM with an accuracy of 0.25 m as the basis for captured data maintenance is the editing and maintenance of data after the

data as background. completion of data capture. Emphasis is placed on the

For this purpose, we created libraries composed of thematic participation of various PwD, who are often ignored in wider

characters, which we took from the object catalogues of the research or data capture. Many research projects confirm that

project. Verification of captured field data was carried out using people respond positively to changes when they are offered the

a digital approach; with open-source mobile data capture opportunity to actively participate [7] in decision-making or

application for Android, iOS, and macOS. Field data verification spatial planning. The approach to data capture was based on

was supported by persons with mobility impairments as a Volunteer Geographic Information (VGI), as a new source of

participatory method. We created data capture layers that include geospatial data. Term Volunteer Geographic Information is

areas for data capture, physical barriers (both permanent and relatively new according to Goodchild, and it is understood as

temporary), routes, public facilities, parking spaces for PwD, bus geographic data created and contributed voluntarily by

stops and railway stations, as well as points of interest. Each layer individuals, often through online platforms and mobile

contains different types of data. For example, for paths, we technologies [8]. Numerous studies show that GIS, through the

captured information on sidewalks, roads without sidewalks, use of smart mobile phones, has established itself as an effective

bicycle paths, pedestrian crossings, tactile pavement elements, tool for ensuring participation [9][10][11][12]. This technology

building edges, green spaces, fences, and stairways. Data on allows residents to actively contribute to shaping their living

physical barriers includes obstacles such as steps, high curbs, environment through mobile GIS applications, by capturing and

hazardous objects on paths, unsuitable surfaces, traffic signs sharing spatial data in real-time [13][14].

blocking sidewalks, low pegs, and movable barriers like parked



cars, flower beds, advertising boards, and outdoor seating from

restaurants. Temporary obstacles, such as construction sites,

were also documented, including details on the duration of

closures for excavation works. Additionally, we captured

location data of accessible toilets for PwD, and accessibility of

public transport (e.g. accessibility of bus stops and vehicles) to

support route planning.

To better understand the travel habits, experiences with

different modes of transport and the willingness to use the “on-

call transport model among people with disabilities, we

conducted a survey. The survey was conducted using the online

platform - 1ka, chosen for its user-friendly interface, open-source

nature, and accessibility features that cater to users with visual

impairments. To ensure comprehensive data capture and increase

survey participation, we employed two supplementary methods. Figure 1: Demonstration of a data capture workshop.

The first method involved collecting data at social events



organized by organizations for people with disabilities, allowing 3 Results us to engage closely with respondents and gather opinions and

information in a more informal setting. The second method As highlighted in the previous chapter, we captured data for four

involved collecting data through employment centres that cater groups of PwD included in the project. This data was then

to PwD. In both cases, respondents were given the option to entered into the database, organized topologically, and verified

complete the questionnaire either online or in person. for suitability in navigation applications. To visualize the

Finally, we conducted various workshops with persons with captured data, we developed the Web Accessibility Viewer - a

different types of disabilities, including those with mobility web-based GIS tool equipped with basic functionalities for

impairments, the blind and visually impaired, and the deaf and displaying and defining the properties of the various layers.

hard of hearing. These workshops were organized as focus Our database currently contains data for 79 municipalities,

groups, where we facilitated guided conversations through covering all groups of PwD considered in the project. The

specific questions to gather as much useful information as captured data includes 3,170 parking spaces designated for PwD

possible. These sessions provided us with in-depth insights into and 5,100 physical barriers, such as dropped kerbs, inadequate

how PwD navigate through different environments, how they use ramps, unsuitable pavement surfaces, and obstructions on paths.

assistive devices, and how they prepare for travel and use public Additionally, we have documented 1,650 accessible bus stops

transportation. We also gained new knowledge about various and 59 railway stations. In terms of public facilities, there are

technological aids and applications. 4,750 identified, of which 288 have accessible toilets for PwD.

For individuals with sensory impairments, 508 audible traffic

2.1 Volunteer Graphic Information for Data signals and 921 tactile pavement elements have been captured.

maintenance Furthermore, we have captured data on 712 points of interest

For data maintenance, we used the participatory data across the municipalities.

maintenance method. We relied on similar voluntary data capture

urban areas tend to be more accessible and the overall situation



is gradually and consistently improving.



4 Discussion

As we previously established, digital data on the accessibility of

public spaces and infrastructure can be captured both

methodically and in a participatory way, when shared by PwD

via smart mobile phones using a GIS app, thus becoming part of

big data. We define big data as large-scale datasets originating

from heterogeneous sources and collected in the urban

environment using sensors integrated in an Internet of Things

Figure 2: Web Accessibility Viewer displaying parking (IoT) system [17]. An IoT system combines devices such as

place from data bases. smartphones and different sensors into a single network

connected to a common server. As central elements of the IoT,

Analysis of the survey responses highlighted current sensors enable cities to collect real-time data from public

challenges in public transport, the experiences and travel habits transport usage and traffic flow analysis to energy consumption

of the respondents, as well as their actual intention to use the "on- and weather conditions [18]. Advanced analytics tools can

call transport" service. A total of 921 respondents participated in process and correlate this data to extract useful information to

the survey. The majority of respondents are from urban areas of improve city planning and management decisions [19].

Slovenia, more specifically Ljubljana, Maribor, Celje, Kranj, and The importance of big data in the context of accessibility is

Novo mesto. multi-layered. These data enable the analysis and understanding

We understand the "on-call transport" service as a service that of mobility patterns, use and accessibility of public spaces and

provides transportation from the starting point to the endpoint of infrastructure. With the help of this data, administrations or urban

a journey, with both points being within the operational area of planning and design professionals can identify areas where

the call center. If a passenger travels outside this area, the service improvements of accessibility are needed. Moreover, big data

offers transport to the nearest accessible public transport station. allows for a dynamic response to the needs of the population, as

This allows passengers to continue their journey using other it can detect and predict changes in mobility and space use

forms of public transport, enhancing their mobility and access to patterns. Providing accurate, real-time information on

a broader area. It is essential to mention that the "on-call accessibility of public transport using big data improves travel

transport" model acts as a bridge between individual transport planning options for PwD. This not only reduces the risk of

and complements existing public transport systems. encountering unexpected barriers, but also highlights the

The "on-call transport" service not only supports the importance of big data in the design and planning of more

independence and improves mobility of vulnerable groups but accessible and inclusive public spaces. The Ljubljana Passenger

also strengthens their social inclusion and independence. Transport's Urbana mobile app is a good example of the use of

Although originally designed for the transportation of PwD, the big data and sensor technologies in practice, as it allows users in

"on-call transport" service has the potential to benefit the wider need of an accessible bus to order transport for a specific time

community. It presents a solution that could be adapted for other and location, while the location of the bus can be monitored in

vulnerable groups, such as the elderly, children, or those without real time directly through the app.

private transportation in remote areas. The integration of different databases which include data on

The majority of respondents expressed interest in using "on- accessibility and data on public passenger transport is essential

call transport" if the advance notice period was 12 hours or less. for improving and enabling independent mobility of PwD.

A small number of respondents indicated that they would Combining data on accessible bus locations, accessible station

consider replacing their personal car with "on-call transport" if locations, accessible sidewalks and physical barriers, and

given the option. timetables allows people with different disabilities to access key

Our data analysis showed that most of respondents are information for route planning. All data are in standard format

employed, which is probably the consequence of distributing the for GIS and can be converted for various application use.

survey in employment centers to capture a wide range of To successfully implement knowledge into practice and

individuals with diverse experiences and travel habits. Retirees ensure accessible public spaces, efficient public transport, and

represent the second largest group, while students and self- other mobility options within the context of multimodal mobility,

employed individuals constitute the smallest portion of cooperation among stakeholders at local, regional, and national

respondents. levels, as well as experts in spatial planning, is essential. For

The analysis of field data capture, survey findings and digital data on accessibility of public spaces and infrastructure to

different workshops reveals that public transport accessibility for be useful and effectively integrated into various systems,

PwD is often inadequate. Respondents cited unreliability, collaboration with urban planners, public transport operators,

inaccessibility, and poor organization as some of the most organizers of on-demand transport services, and representatives

significant issues, leading many to prefer using personal cars of PwD is necessary. This collaborative approach will help

over public transport. Most stations and intercity public transport address the shortcomings and contribute to the improvement of

vehicles are inaccessible, as highlighted also by the Advocate of data in the future.

the Principle of Equality in Special Reports [15][16]. However,



5 Conclusion References

In an era of rapid technological transformation, urban population [1] United Nations. 2006. Convention on the Rights of Persons with

growth, and ageing population, urban design faces increasing Disabilities. Treaty Series, 2515, 3. Retrieve 12 Jul. 2024 from

https://www.un.org/en/development/desa/population/migration/generalas



challenges to create spaces that are accessible and inclusive. The sembly/docs/globalcompact/A_RES_61_106.pdf [2] Luis G. Natera Orozco, Laura Alessandretti, Meead Saberi, Michael Szell, rise in the number of individuals with disabilities necessitates and Federico Battiston. 2022. Multimodal Urban Mobility and Multilayer barrier-free access to public spaces. Over the past decade, we Transport Networks. Environment and Planning B: Urban Analytics and City Science . 50(8), 2038-2070 DOI: have witnessed a transition from traditional city concepts to the https://doi.org/10.1177/23998083221108190 . design of dynamic and adaptable urban environments. This shift [3] Roman Rener, Jani Demšar, Marina Lovrić, Miran Janežič, in Helena



is not merely physical but represents a profound change in our Žnidaršič. 2022. Omogočanje multimodalne mobilnosti oseb z različnimi oviranostmi: končno poročilo. Ljubljana: Geodetski inštitut Slovenije approach to urban design. [4] Marina Lovrić, Maruša Vertačnik, in Roman Rener. 2023. Prostorska By centering the design process around people and leveraging podatkovna podpora upravljanju javnega potniškega prometa – Invalidi v javnem potniškem prometu: končno poročilo. Ljubljana: Geodetski innovative technological solutions, cities are becoming not only inštitut Slovenije. technologically advanced but also inclusive and responsive to the [5] Katja Oven, Roman Rener, Dalibor Radovan, Marina Lovrić . 2022. Analiza konceptualnih zasnov informacijske in tehnične sistemske needs of their inhabitants. This new paradigm prioritizes podpore invalidom za povečanje socialne vključenosti. Vmesno poročilo accessibility and inclusiveness, ensuring that urban spaces ciljnega raziskovalnega projekt št. V2-2252.



accommodate everyone, regardless of their physical abilities. [6] Mihaela Triglav Čekada, Dalibor Radovan. 2019. Comparison of the volunteered geographic information usage to monitor floods and Through this approach, we are crafting cities that embody the earthquakes. Geografski vestnik . 127-135(2). DOI: principles of universal design, fostering environments where all https://doi.org/10.3986/GV91207 [7] Rachael Luck. 2018: Inclusive design and making in practice: Bringing individuals can thrive and participate fully in society. bodily experience into closer contact with making. Design Studies . The article describes the approach to the development of the Volume 54, 96–119. DOI: https://doi.org/10.1016/j.destud.2017.11.003



entire system from methodology, standards, the use of new [8] Michael F. Goodchild. 2007. Citizens as sensors: the world of volunteered geography. GeoJournal. 69, 211–221. DOI: technologies and building a base for the purpose of mobility of https://doi.org/10.1007/s10708-007-9111-y persons with disabilities. It is especially worth emphasizing the [9] Renee Sieber. 2006. Public Participation Geographic Information Systems: A Literature Review and Framework. Annals of the Association involvement, sharing knowledge and collaboration between of American Geographers. 96:3, 491-507, DOI: system developers and users. https://doi.org/10.1111/j.1467-8306.2006.00702.x [10] Maria Antonia Brovelli, Marco Minghini, in Giorgio Zamboni. 2016. In this context, technological solutions play a pivotal role in Public participation in GIS via mobile applications. ISPRS Journal of advancing multimodal mobility and enhancing urban Photogrammetry and Remote Sensing. 114. 306-315. DOI:



accessibility. The integration of various transportation options https://doi.org/10.1016/j.isprsjprs.2015.04.002 [11] Raquel Pérez-delHoyo, Maria Dolores Andújar Montoya, Higinio Mora such as walking, cycling, and public transit requires seamless and Virgilio Gilart Iglesias, 2018. Citizen Participation in Urban Planning-coordination facilitated by cutting-edge technologies. Real-time Management Processes - Assessing Urban Accessibility in Smart Cities. In Proceedings of the 7th International Conference on Smart Cities and data, smart infrastructure, and advanced analytics are Green ICT Systems – SMARTGREENS. 206-213. DOI: transforming how cities manage and optimize their https://doi.org/10.5220/0006704202060213



transportation networks. [12] Katarzyna Roszewska. 2021. Geographic information systems and accessibility for persons with disabilities. GIS Odyssey Journal . 1(2), 21–



Acknowledgments 29. DOI: https://doi.org/10.57599/gisoj.2021.1.2.21

[13] Aimi Hamraie. 2018. Mapping Access: Digital Humanities, Disability



We would like to extend our sincere thanks to the Ministry of Justice, and Sociospatial Practice. American Quarterly. 70. 455-482. DOI:

https://doi.org/10.1353/aq.2018.0031



Environment, Climate, and Energy for funding the project [14] Sultan Kocaman and Nadire Ozdemir. 2020. Improvement of Disability Rights via Geographic Information Science. Sustainability 12, no. 14: "Enabling Multimodal Mobility for Persons with Various 5807.DOI: https://doi.org/10.3390/su12145807 Disabilities" for eight consecutive years, allowing us to improve [15] The Advocate of the Principle of Equality. 2023. Inaccessibility of Public Inter-Urban Bus service for Persons with Sensory Disabilities. Retrieve 17 and raise awareness about the importance of accessible July 2024 from https://zagovornik.si/wp- environments. content/uploads/2024/02/Special-Report_Inaccessibility-of-Public-Inter- Urban-Bus-Service-for-Persons-with-Sensory-Disabilities.pdf We are also grateful to The National Council of Disability [16] The Advocate of the Principle of Equality. 2023. Inaccessibility of Public Organizations of Slovenia for their extensive support in Inter-Urban Bus service for Persons with reduces mobility. Retrieve 17

implementing the project "Spatial Data Support for Public July 2024 from https://zagovornik.si/wp-content/uploads/2024/01/2023-

Passenger Transport Management – Persons with Disabilities in 11-23-Posebno-porocilo_Nedostopnost-javnega-medkrajevnega-

avtobusnega-prevoza-za-gibalno-ovirane_EN1.pdf



The authors of the article would also like to thank the Slovenian Oana-Matilda Sabie. 2023. Digital Twin Simulation Tools, Spatial Cognition Algorithms, and Multi-Sensor Fusion Technology in Research and Innovation Agency (ARIS) and the Ministry of Public Passenger Transport." [17] Elvira Nica, Gheorghe H. Popescu, Milos Poliak, Tomas Kliestik, and

Sustainable Urban Governance Networks. Mathematics 11, no. 9: 1981.



Labour, Family, Social Affairs and Equal Opportunities DOI: https://doi.org/10.3390/math11091981 [18] Amardeep Das, Sumanta Mishra Sharma, and Bikram Ratha. 2019. The (MDDSZ), who financing the research project entitled "Analiza New Era of Smart Cities, From the Perspective of the Internet of Things. konceptualnih zasnov informacijske in tehnične sistemske Smart Cities Cybersecurity and Privacy . 1-9. DOI: 10.1016/B978-0-12-815032-0.00001-9. podpore invalidom za povečanje socialne vključenosti " No. V2-[19] Michael Batty. 2016. Big Data and the City. Built Environment . 42. 321-2252, targeted research program "CRP-2022", made it possible 337(17). DOI: https://doi.org/10.2148/benv.42.3.321 .

to carry out part of the research and results of this article.

Note that there is a section break at the end of references to

balance the columns (and this text is a part of the new section).

If you have no space left at the end of your paper, you can delete

it.





The role of communication skills in the development of




digital literacy competences using the "DigInGreen"



model



Mateja Forte † Matevž Obrecht Bojana Slomšek Šlamberger

Faculty of logistics Faculty of logistics Faculty of logistics

University of Maribor University of Maribor University of Maribor

Celje, Slovenia Celje, Slovenia Celje, Slovenia

mateja.forte@um.si matevz.obrecht@um.si bojana.slomsek@um.si



Abstract / Povzetek sensory needs and the dilemmas, associated with them. One of

them being the so-called digital literacy, meaning the ability of

The paper deals with the field of developing communication and individuals to effectively use digital technologies and the internet

digital literacy with the help of the lifelong learning model to find, create, evaluate and communicate information. This

"DigInGreen" for all layers of modern society, including people includes understanding and using digital tools, such as

with special needs. The aim of the authors is a targeted search for computers, smartphones, applications and social networks as

new solutions in the field of using digital learning and well as awareness of privacy, security and ethical issues in the

communication tools for the development of digital competences digital environment. Digital literacy and digital tools therefore

in the light of "Society 5.0". We focus on a unique way of play a key role in achieving the goal of equal opportunities for

connecting paradigms, such as: inclusive society, competences, all, as they can be completed by users in different periods of time,

digital, special needs, literacy and lifelong learning. We want to which are different for each individual and they help themselves

connect the thinking about the communicative ability, an in solving challenges with specialized learning or work tools

indispensable part of which is an individual's digital literacy and (letter enlargement, volume, translation into another language,

point out, that the circumstances of living in a modern, but still font transcription). In relation to the mentioned challenges, the

partially exclusive society, limit the development of an overall aim of this paper is to connect very different research

individual's communicative and digital literacy, due to a fields with the focus to find new solutions in the field of using

demonstrated special need or disability. We believe, that the digital learning and communication tools for the development of

activities we carry out, are a unique example of good practice, digital competences in the light of "Society 5.0". because they connect many scientific disciplines and fields into

a modern concept of competence development without excluding 1.1 Language and communication ability or conditioning the end user. We combine the paradigms of

logistics, informatics, statistics, social studies, linguistics and We cannot talk about the concept of digital literacy without first

didactics. All the above and more represent the concept of talking about language and the communicative ability, because

"Society 5.0", which cannot exist as a scientific construct, if it the latter represents the dividing line between human and other

does not consider the key aspect of one's own identification - communicative codes. Language represents our means of

social equality and the provision of equal opportunities. communication, it is a code, that humans have developed,

perfected, thereby enabling communication. Language is an



Keywords / Ključne besede agreed system of signs, which means, that it is a standardized

communication code and leads to the realization, that language

communication ability, digital literacy, competences, inclusion, is not just an individual property, it is a social property. With this

special needs, lifelong learning finding, ensuring the possibility of literacy development for the

individual in all its forms and variations becomes a social

responsibility. The latter is implemented not only in the home

1 Digital literacy is an exclusive component of environment, but also in the process of education at the primary,

developed communication skills secondary or tertiary level and in all forms of social interaction,

including models of lifelong learning. When communicating,

In the last few years, we have witnessed the rapid and general

two processes always take place; encoding, creating a message

progress of the digitization process everywhere, including in

and decoding or understanding and responding, to what is

education, which brings many opportunities, but also challenges,

communicated. At the same time, we must assume, that both the

especially when it comes to the inclusion of people with special

communicator and the addressee have developed communication

needs in the educational process. In the past, the field was more

skills. These are the knowledge and abilities of both the

related to physical and intellectual challenges, but today it is

addressee and the communicator to create any number of texts in

increasingly focused on emotional, behavioral, psychosocial and

a wide variety of speaking positions. The ability to communicate

is an important part of a person's general ability, as it enables

Permission to make digital or hard copies of part or all of this work for personal or him/her to creatively adapt to new circumstances and needs. The

classroom use is granted without fee provided that copies are not made or distributed ability to communicate therefore defines us through expression,

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must communication with others and our social involvement. Above

be honored. For all other uses, contact the owner/author(s). all, it determines our oral and written communication, both

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia classic and digital. We can all be creators/communicators as well

© 2024 Copyright held by the owner/author(s). as addressees, but we can only cooperate with each other with

https://doi.org/10.70314/is.2024.DIGIN.3



developed communication skills. So, we learn the means, easier management of digital content. We also know about the

through which we enter the communication process. implementation of individual distance educations, that enable

less mobile individuals to participate and do it more easily. Good

practices include collaboration with various stakeholders,

2 Accessibility for developed communication including non-governmental organizations, that liaise with

and digital skills people with special needs and developers of digital technologies

Universal digitization enables adaptation of learning content and to ensure comprehensive support and adaptations of learning

teaching methods, which can significantly improve accessibility and ensure constant updating of competences, which contributes content. In addition, it is important to promote lifelong learning

to educational opportunities for people with special needs. We to greater employability and inclusion of everyone, including

can highlight the example of online platforms and digital people with special needs in the labor market. Digitization and

materials; these allow learners to access learning content in an the development of green and digital competences through

adapted format, such as texts in larger fonts, audio recordings or lifelong learning are important steps towards ensuring equal

video content with subtitles. In addition, interactive tools and opportunities for all, including people with special needs. With

applications can enable individual adaptation of the learning legal frameworks and good practices we can create an inclusive

process, which increases the effectiveness of learning and the environment, that promotes sustainable development and digital

engagement of learners. The legal regulation in the field of literacy for everyone. But for this it would be necessary to

education of people with special needs in Slovenia is based on new approaches, advantages and tools for increasing an inclusive actively increase financial resources and educate personnel about

the Act on Equalizing the Opportunities of the Disabled and the educational environment. Act on the Guidance of Children with Special Needs. This

legislation determines the rights and obligations of educational

institutions regarding the provision of adaptations for people 4 Acquired competences are the basis for

with special needs. In practice this means that institutions must developed communication skills with an

ensure accessibility to all educational resources and services, emphasis on digital literacy



including digital platforms and technologies, but sometimes legal Digitization represents a major challenge, as the appropriate commitments are a burden to institutions, as there is no knowledge of users, including people with special needs, guaranteed systemic funding for their implementation. With the represents an essential pool of skills for the efficient operation of successful implementation of education, we condition the many processes. At the Faculty of Logistics of the University of

development of competences and the latter can be applied to all Maribor, we prepare, develop and implement short workshops,

final stakeholders, both people with special needs and people, that cover a wide variety of topics, such as: research into the

who do not have demonstrated special needs. As far as we management of sustainability and resilience in supply chains,

discuss the developed competences for digital literacy, of course digitization, interdisciplinary and applied knowledge. The use of

we cannot bypass the models, that take care of the development skills in the field of digitization enables efficient data collection

of these competences. We have developed such a model. and analysis, as well as high-quality tracking and optimization of

various processes. With the appropriate knowledge for all users,

we want to get closer to up-to-date and reliable information. In



3 lifelong education and learning within the "DigInGreen" model, A unique model of lifelong learning, that



develops both green and digital competences workshops in the field of digitization, such as: planning and we focus on offering all interested participants educational

We start from the concept of a special example of good practice implementation of digitization, use of simulation tools as

of lifelong education and learning of green and digital skills for decision support, information and computer security, use of

all social groups. They were designed and developed at the artificial intelligence tools, information and digital literacy,

Faculty of Logistics of the University of Maribor as part of the methods of communication, creation of video content. We enable

NOO project "DigInGreen". The acronym NOO defines the participants to study selected topics in depth and participate in

financial mechanism, namely the Recovery and Resilience concrete tasks and practical examples. This type of method of

Mechanism, from which operations, addressing the transition to implementing the lifelong learning model has proven to be very

a green, digital and resilient society are financed. As part of this successful in imparting and acquiring new knowledge and skills.

project, the Faculty of Logistics of the University of Maribor has As an educational institution, we see the development of lifelong

determined a research and development niche, that is aimed at learning according to the "DigInGreen" model in using digital

the development of green and digital competences through tools with the end user in mind, including users with special

lifelong learning and focus on various areas, which include both needs.

short-term and long-term goals for the development of social

studies, information sciences and social skills. The short-term

goals include training individuals to use digital literacy tools and 4.1 Model implementation

practices, that improve the efficiency of many processes, while Our most effective and also the most innovative form of work or

the long-term goals aim for the sustainable development and activity is the framework of the workshops, that we offer as part

strengthening of digital literacy at all levels of the population, of the project. All of them are innovative in terms of content, as

especially with an emphasis on understanding complex they all address topics, that are strongly related to “Society 5.0”.

challenges and possible approaches for people with special The main ideas are: digital, sustainable and resilient. All

needs. To ensure inclusive education, it is necessary to adapt workshops are interconnected and together provide a

teaching materials and methods, that enable equal access to comprehensive insight into the topic of sustainable and

knowledge and skills. For example, courses and workshops can digitalized future for all walks of society. Their innovation lies

include adapted digital interfaces, that are also accessible to in the method of imparting knowledge, as it is not a lecture.

people with visual or hearing impairments and tools, that enable

These are experiential workshops, that encourage participants to Figure 1: Model performance analysis

come to certain insights on their own. Former concepts of

learning and teaching have been abandoned and learners need In order to determine the effectiveness of the proposed model of

experiential learning to quickly assimilate the elements, that then lifelong learning and education to improve sustainable and

make up the so-called “micro education”. digital literacy within different social groups, we conducted an

analysis of satisfaction at the level of the participant. In the

4.2 Following the principles of an egalitarian sample, we analyzed the following parameters: content of the

society – “Society 5.0” workshop, length of the workshop, method of delivering content,

The “DigInGreen” lifelong learning model is defined by the place of performance and its accessibility, ease of registration,

principles of the “Society 5.0”, therefore an egalitarian society in accessibility of material, relationship between theoretical and

practical elements, dynamics and compliance with individual

its entirety. We follow: equal opportunities; all stakeholders of

expectations. 66 % of the participants expressed their intention

our project - participants in the planning of activities,

to deepen and upgrade the knowledge they acquired at the

implementation of workshops, participation in workshops - have

workshop. 95 % of them will also recommend participation in

the same opportunities to participate. Our basic criteria are the

the workshop to their colleagues, acquaintances and business

demonstrated area competence of the individual and potential

partners. Most of the participants pointed out that at the attended

gaps in knowledge. We offer all interested parties’ equal

workshop they upgraded their existing or acquired new

opportunities to participate in the project. Non-discrimination;

knowledge to improve their digital and sustainable skills.

we treat individuals equally. Individuals are not compared to



someone else because of gender, nationality, race or ethnic Acknowledgments / Zahvala origin, religion or belief, disability, age, sexual orientation or

other personal circumstance. Accessibility for the disabled and The article is funded from the European Union – Next

people with special needs; for all participants in the activities of Generation EU and the Ministry of higher education, science and

the “DigInGreen” lifelong learning model we provide suitable innovation, Slovenia. The research was carried out within the

materials, that contribute to the development of their digital and project, entitled: »Establishing an environment for green and

communication skills, using appropriate teaching methods and digital logistics and supply chain education«.

ensuring the correct guidance.

References / Literatura

4.3 Model performance analysis [1] Mateja Forte. 2015. Slovene language infrastructure for blind/impaired

students in regular primary school conceptual design of a primary school



textbook of the Slovene language for the inclusion program. University of

Primorska, Koper

[2] Maria Golden. 2001. About language and linguistics. Department of

Comparative and General Linguistics, Faculty of Arts, University of

Ljubljana, Ljubljana

[3] Olga Kunst Gnamuš. 1994. Theory of Communication. Center for

Discursive Studies, Ljubljana

[4] Matevž Obrecht, Tina Cvahte Ojsteršek, Mateja Čuček, Martin Fale,

Milena Kajba, Polona Vičič, Bojana Slomšek Šlamberger, Lazar Pavić.

2024. Knowledge supply chains - Guidelines for green and digital

transition st , (1. ed.), University of Maribor, University Press, Maribor

[5] Bojana Slomšek Šlamberger. 2024. ESG – Environment, company,

management no. 188, DigInGreen – an innovative concept of lifelong

education



1



The Moore’s Law for Education and the Need for



Inclusion



Marius M. Balas∗ ∗ † † ‡ , Valentina Balas , Dana Rad , Gavril Rad and S´ergio D. Correia

∗ Engineering Faculty, Aurel Vlaicu University of Arad, Arad, Romania

† Center of Research Development and Innovation in Psychology, Faculty of Educational Sciences Psychology and

Social Work, Aurel Vlaicu University of Arad, Arad, Romania

‡ Portalegre Polytechnic University, Portalegre, Portugal

Correspondance: {marius, balas, dana}@drbalas.ro, radgavrilarad@gmail.com, scorreia@ipportalegre.pt



Abstract—From within the electronics industry, Gordon Moore We believe that the current situation requires a new fun-

observed an exponential rate of development of technology. Over damental approach, capable to sustain in the long term a

time, the same pace of development has been observed for proper Moore’s Law for Education, helping students to include

most human activities, resulting in an exponential growth in

the volume of accumulated knowledge. This phenomenon also into the Moore’s Law for Knowledge era. For this, we must

has adverse consequences, especially the increasing difficulty of increase the efficiency of learning, so that it keeps pace with

the educational process, mainly in the scientific and engineering the increase in the volume of knowledge. This approach is

fields, which ultimately results in an increase in school dropout. encouraged by the statement according to which learning

To cope, we will have to initiate the inclusion of students in the resources are practically unlimited [4]. new context, through measures that will support for long term

a Moore’s law for education. To begin with, we propose several

measures, such as increasing the share of visual representations II. PSYCHOLOGICAL THEORIES EXPLAINING THE

of knowledge or applying the recommendations of the Bologna PHENOMENON

Process more rigorously.

A. Cognitive Load Theory

Index Terms—Moore’s Law, Visual Representation of Knowl- Cognitive load theory (CLT), developed by John Sweller,

edge, Bologna Process, Inclusive Education, System Engineering

provides a valuable framework for addressing the challenges

posed by exponential growth in knowledge and its implica-

tions for education. CLT posits that working memory has a

I. THE MOORE’S LAWS AND EDUCATION limited capacity for processing information. As the volume

and complexity of educational content increase, the cognitive

In 1965, Gordon Moore, Intel CEO & Co-founder, made a

load on students also rises. This escalation in cognitive load

perspicacious empirical observation: the number of transistors

can lead to cognitive overload, where students struggle to

in integrated circuits doubles every two years. This observation

process or retain information effectively, ultimately resulting

became the well-known First Moore’s Law. In the following

in reduced learning efficiency and increased dropout rates [5],

decades, and until today, the law has been verified and even

[6]. To counteract these issues, strategies that manage cognitive

knowingly applied in the management of the electronics in-

load, such as simplifying complex information and utilizing

dustry. Moreover, the exponential development pattern noticed

multimedia tools to present knowledge visually, are essential.

by G. Moore has been observed in many more areas of human

These strategies help accommodate the rapid advancements

activity: (1) Moore’s Second Law (Arthur Rock’s Law): the

described by Moore’s Law in education, thereby supporting

cost of a semi-conductor chip fabrication plant doubles every

more effective learning experiences.

four years; (2) More than Moore: the Moore’s law is now

Van Merrienboer and Sweller emphasize that recent de-

beginning to be revalued in the sense of performance indexes

velopments in CLT highlight the importance of designing

per chip instead of sheer number of transistors per chip; (3)

instructional methods that manage cognitive load effectively,

Moore’s Law for Knowledge: the general human knowledge

especially when dealing with complex learning scenarios [7].

doubles every year [1]; and (4) Moore’s Law for Everything:

By integrating these principles into educational practice, ed-

addresses our entire society as a whole [2].

ucators can better support students in handling the growing

An undesirable consequence of the Moore’s Law for Knowl-

complexity of their studies.

edge appears in education: young students face more and more

difficulty the increased volume and complexity of the knowl-

edge they have to assimilate B. Self-Determination Theory . Therefore, we are witnessing

an alarming increase in school dropouts. In 2001 G. Moore Self-determination theory (SDT), developed by Deci and

complained: “It’s hard to come up with ways to increase Ryan, emphasizes the essential psychological needs for au-

productivity in education” [3]. Eventually one can observe tonomy, competence, and relatedness as central to motivation

that Moore’s Law is beginning to make its presence felt in and learning [8]. The exponential growth of knowledge, as

education too [3]. highlighted by Moore’s Law, may challenge students’ sense of



812



2



competence and autonomy, especially when the pace of learn- D. Complex Adaptive Systems Theory

ing feels overwhelming or unmanageable. This can adversely Complex adaptive systems theory (CAST) provides a con-

impact their motivation and academic performance. According temporary cognitive psychological perspective that integrates

to SDT, when students perceive their learning environment as concepts from Systems Engineering. CAST focuses on how

supportive of their need for autonomy, they are more likely to systems, including educational systems, adapt and evolve in

experience intrinsic motivation and engage more deeply with response to environmental changes [14]. This theory empha-

the material. Similarly, providing constructive feedback and sizes the dynamic interactions among various components of

fostering meaningful interactions with peers and educators can a system and the necessity for flexibility in adapting to rapid

enhance students’ sense of competence and relatedness, which changes.

are crucial for maintaining motivation and engagement [9]. In the context of Moore’s Law for education, CAST posits

Research supports the idea that educational practices align- that educational systems must be designed to be adaptable and

ing with SDT principles can lead to more effective learning resilient to the exponential growth in knowledge. As knowl-

outcomes. For instance, Guay, Ratelle, and Chanal argue edge expands at an accelerating rate, educational environments

that optimal learning environments, which cater to students’ must continuously evolve to meet diverse learning needs and

psychological needs, contribute to better educational outcomes integrate new technologies and methodologies. This adapt-

[10]. They found that when students experience a high degree ability involves creating learning systems that can effectively

of self-determination within their learning contexts, they are incorporate feedback and undergo iterative improvements to

more likely to engage actively and perform well academically. refine educational practices [15].

In light of these insights, incorporating measures that sup- CAST also highlights the importance of feedback loops

port the psychological needs identified by SDT - such as and iterative processes in educational settings. By applying

providing opportunities for self-directed learning, ensuring CAST principles, educators can develop more responsive

frequent and constructive feedback, and facilitating collabo- and inclusive educational systems. For instance, leveraging

rative and supportive peer interactions - can help mitigate the iterative feedback mechanisms and promoting flexibility in

negative effects of rapid knowledge expansion. This approach teaching approaches can help address the challenges associated

not only addresses the challenges posed by Moore’s Law with rapid knowledge expansion, ultimately supporting long-

for education but also promotes more inclusive and effective term student success [16].

educational practices that cater to diverse student needs. By implementing CAST, educational institutions may create

settings that are more suited to managing the complexity

of modern education and responding to the needs of fast

C. Information Processing Theory knowledge expansion, as outlined by Moore’s Law.

Understanding the psychological underpinnings of educa-

Information processing theory, as developed by Atkinson

tional challenges in the context of rapid knowledge expan-

and Shiffrin, provides a framework for understanding how

sion provides valuable insights into how to support students

information is encoded, stored, and retrieved [11]. This theory

effectively. By applying theories such as cognitive load the-

becomes increasingly relevant as the volume of information

ory, self-determination theory, information processing theory

grows exponentially, presenting challenges for processing and

and complex adaptive systems theory, educators can develop

organizing vast amounts of data effectively. According to the

strategies to enhance learning outcomes and address the needs

theory, cognitive strategies such as chunking and rehearsal are

of a diverse student population. In addition to supporting the

crucial for managing large data sets, which can help students

inclusion of all students in a changing educational environ-

process and retain information more efficiently.

ment, these strategies aid in managing the growing body of

Chunking involves breaking down complex information

information.

into smaller, manageable units, making it easier to encode

and recall [12]. Rehearsal, which includes techniques like

III. CONCRETE INCLUSION MEASURES

repetition and active engagement with the material, also plays

a significant role in strengthening memory retention. Integrat- A. Improving the Bologna Process Application

ing these cognitive strategies into educational practices can School dropout is all the more expensive the more advanced

aid students in navigating the challenges associated with the the student is on a higher level of preparation and the larger

rapid expansion of knowledge. For example, employing visual the target group. That is why we will focus with priority on

aids, interactive tools, and other multimedia resources can students in the undergraduate stage.

enhance students’ ability to process and retain information by The Bologna Declaration (19 June 1999) proposed a higher

presenting it in more digestible formats [13]. educational system with two main cycles, undergraduate and

Incorporating information processing techniques into teach- graduate, leading to the master and/or doctorate degree (Fig.

ing methodologies supports increased inclusion and provides 1). Access to the second cycle requires successful completion

effective support mechanisms for students facing the com- of first cycle of minimum three years. The first cycle degree

plexities of modern education. By applying these strategies, (Bachelor) is already relevant to the European labour market

educators can help students manage cognitive load more as an appropriate level of qualification.

effectively, fostering an environment where learners are better A key issue is the balance between mass and elite higher

equipped to handle the demands of rapid knowledge growth. education [17]. This balance is perfectly designed, but our



813

3



observations (direct, empirical, and at a restricted scale) point As previously mentioned, in fields like science, technology,

out that many professors apply elite standards for the mass engineering, and mathematics, visualizations such as graphs,

level, leading to scholar dropout. Besides personal attitudes, models, and simulations are crucial for understanding abstract

this might also be caused by a possible lack of attention and complex concepts. Visual tools can illustrate historical

paid to the mass level compared to the elite one. There are timelines, sociological theories, and literary analyses, making

important differences between mainstream and elite education. these subjects more accessible and engaging. To give an

The inclusiveness of mass education is meant to be broad, but example, we mention that visual reasoning is beginning to be

its depth remains shallow, as does its specialization and rigor. adopted even in areas where formal mathematical approaches

On the other hand, elite education is supposed to be deep, seemed immutable, such as automation. These approaches are

rigorous, and specialized, thus losing out on inclusiveness. specific to Artificial Intelligence methods, which, let’s not

That is why the next measures will point to possible technical forget, seek to emulate human reasoning. In the broadest sense,

solutions leading to effectiveness, notably in mass education. in automation, one can observe a revaluation of methods based

on quantitative or even qualitative time analysis, such as the



analysis of the phase trajectory of the control error, which has

appeared since the nineteenth century, comparing to methods

based on the precise frequency analysis (transfer functions,

pole placement, etc.). More specifically, we can mention: (1)

The sliding mode [20]; (2) The qualitative analysis [21]; (3)

The self-adaptive fuzzy-interpolative controllers [22]; and (4)

Sculpting the state space [23].

Systems engineering (SE) is an increasingly significant

scientific field that successfully manages multidisciplinary

Fig. 1. The higher education structure (Bologna Process) and very complex systems (including educational ones). SE

applications are based almost entirely on visual software pack-

ages: UML (Unified Modeling Language), IDEF (Integration

B. Visual Representation of Knowledge

Definition), QFD (Quality Function Deployment), etc.

Homo Sapiens is a diurnal species, essentially relying on

sight [18]: 90% of information transmitted to the brain is

visual; C. The Top-Down Approach 50% of the brain’s surface is used for the vision; and

visual information gets to brain 60.000 times faster than text. SE is proposing another way to increase education’s effi-

These facts lead us to the conclusion that we should look ciency, which is derived from its holistic feature: embracing

to favorize the visual representation of knowledge, especially the Top-Down approach [24]. The top-down approach means

when addressed to the undergraduate students [19]. Figure 2 essentially fewer details and more comprehension.

illustrates how visual adds enhance texts’ understanding [18]. Bottom-Up teaching starts with small details and broadens

the scope of the lessons step by step as students master the



skills. This way is rigorous, yet instruction-driven, cumber-

some, and time-consuming, eventually suited for elite educa-

tion and company specific training.

The strategy of top-down teaching begins with the big,

abstract concept and works down to the specific details,

Fig. 2. Better formulated questions lead to better results according to the available time. This motivates students to

learn through direct interaction and their own experience and

Visuals can break down complex concepts into more di- is fast and suited to mass education. The Top-Down Approach

gestible parts, making it easier for students to grasp intricate in education provides a structured and motivating way to learn,

ideas. Diagrams, flowcharts, and mind maps can illustrate promoting a deeper and more integrated understanding of

the relationships between different concepts, helping students subjects while aligning well with real-world problem-solving

understand how they interconnect. Moreover, visuals are often and diverse learning styles. It is structured and linear, ideal

more memorable than text alone. The dual-coding theory for subjects where a foundational understanding is crucial,

suggests that information presented both visually and verbally aiming for a comprehensive understanding by building from

is more likely to be remembered, as visual representations general to specific. Students gain an overall understanding

can reinforce learning by providing multiple ways to process of the subject, helping them see how individual parts fit

and recall information. When considering student motivation, into the bigger picture. This broad perspective can enhance

visuals can make learning more engaging and interesting, comprehension and retention of detailed information. By un-

potentially increasing student motivation and participation. Vi- derstanding the ultimate goals and applications of what they’re

sual tools like graphs and charts encourage students to analyze learning, students can be more motivated to engage with and

data and identify patterns, trends, and outliers, and they help master the detailed content. Knowing the relevance and end

visualize problems and brainstorm solutions, fostering critical goals of their studies can make the learning process more

thinking skills. meaningful and interesting. Moreover, starting with general



814

4



concepts allows students to quickly grasp the subject’s scope REFERENCES

and identify areas that need more focus. This can make the [1] R. B. Fuller, Critical path. St. Martin’s Press, 1981, (ISBN: learning process more efficient, as students can prioritize their 9780312174910).

efforts on more challenging aspects once they understand the [2] . Sam Altman · March 16, “Moore’s law for everything.” [Online].



overall framework. [3] G. Available: https://moores.samaltman.com/

Anders, “Moore’s law touches education at

Given the current complexity of education, illustrated in last – to techies’ delight,” Apr 2014. [On-

section 2, a mechanical application of any of the meth- line]. Available: https://www.forbes.com/sites/georgeanders/2014/04/16/

ods proposed above is virtually impossible. Hybrid solutions [4] J. W. Botkin, M. Elmandjra, and M. Malitza, moores-law-touches-education-at-last-to-techies-delight/ No limits to learning

have the highest chance of success. Let’s give an example bridging the human gap; a report to the club of Rome James W. Botkin;

from digital electronics: the Hardware Description Languages Mahdi Elmandjra; Mircea Malitza. Pergamon Press, 1979, (ISBN:

(HDL). HDLs admit two types of descriptions for the same [5] J. Sweller, J. Plass, R. Moreno, and R. Brunken, “Cognitive load theory: 0-08-024705-9).

circuit: Behavioral and architectural, which are interchange- Recent theoretical advances,” Cambridge Univerisity Press, p. 29–47,

able according to needs (Fig. 3). The behavioral description 2010.

is top-down functional, while the architectural is bottom-up tional design: Recent developments,” [6] F. Paas, A. Renkl, and J. Sweller, “Cognitive load theory and instruc-Educational psychologist, vol. 38,

structural. no. 1, pp. 1–4, 2003.

[7] J. J. G. Van Merrienboer and J. Sweller, “Cognitive load theory and com-

plex learning: Recent developments and future directions,” Educational



psychology review, vol. 17, pp. 147–177, 2005.

[8] E. L. Deci and R. M. Ryan, “Self-determination theory,” Handbook of

theories of social psychology , vol. 1, no. 20, pp. 416–436, 2012.

[9] J. Reeve, “A self-determination theory perspective on student engage-

ment,” in Handbook of research on student engagement. Springer, 2012,

pp. 149–172.

[10] F. Guay, C. F. Ratelle, and J. Chanal, “Optimal learning in optimal con-

texts: The role of self-determination in education.” Canadian Psychology

/ Psychologie Canadienne, vol. 49, no. 3, p. 233, 2008.

[11] R. C. Atkinson and R. M. Shiffrin, “Human memory: A proposed system

and its control processes,” in Psychology of learning and motivation.

Elsevier, 1968, vol. 2, pp. 89–195.

[12] J. T. Wixted, “Atkinson and shiffrin’s (1968) influential model overshad-

owed their contemporary theory of human memory,” Journal of Memory

Fig. 3. The HDL descriptions and Language, vol. 136, p. 104471, 2024.

[13] G. Sucharitha, A. Matta, K. Dwarakamai, and B. Tannmayee, “Theory

and implications of information processing,” Emotion and Information

Processing: A Practical approach, pp. 39–54, 2020.



IV. C [14] T. Fidan and A. Balcı, “Managing schools as complex adaptive systems: ONCLUSION International Electronic Journal of Elementary

A strategic perspective,”

The paper discusses the ever-widening gap between the Education , vol. 10, no. 1, pp. 11–26, 2017.

volume of general knowledge, which obeys Moore’s law of [15] P. McMillan and M. Jess, “Embracing complex adaptive practice:

the potential of lesson study,” in Non-Linear Perspectives on Teacher

exponential knowledge development, and the effectiveness of Development. Routledge, 2023, pp. 77–92.

education, which cannot keep up. The effect of this gap is [16] J. K. Garner and A. Kaplan, “A complex dynamic systems approach

critical for young people, especially for undergraduates, who to the design and evaluation of teacher professional development,” in

Non-Linear Perspectives on Teacher Development. Routledge, 2023,

have a high dropout rate. This is a peculiar case of inclusion, pp. 93–118.

because it is addressing a wide group of perfect valid persons: [17] K. Miklaviˇc, “Elite vs. mass higher education - a reflection on the

virtually any student. In order to promote their inclusion in the purpose of higher education,” in The GAPS Think Piece. GAPS -

Global Access to Postsecondary Education, 2016, no. 14, pp. 1–8.

context of Moore’s law for knowledge, several fundamental [18] E. McCoy, “How our brains are hardwired for visual content,” Killer

measures are proposed in order to give education a greater Visual Strategies, 2019.

dynamic: the more rigorous application of the Bologna Process [19] M. Chein, M.-L. Mugnier, and M. Croitoru, “Visual reasoning with

graph-based mechanisms: the good, the better and the best,” The

system, the prioritization of visual methods of representation knowledge engineering review, vol. 28, no. 3, pp. 249–271, 2013.

and manipulation of knowledge and the broader application [20] K. D. Young, V. I. Utkin, and U. Ozguner, “A control engineer’s guide to

in education of methods derived from Systems Engineering, sliding mode control,” IEEE transactions on control systems technology,

vol. 7, no. 3, pp. 328–342, 1999.

such as the top-down approach. [21] W. F. Clocksin and A. J. Morgan, “Qualitative control,” In Proceedings

Incorporating visual representation of knowledge in higher of the Seventh European Conference on Artificial Intelligence, vol. 350,

education not only caters to diverse learning styles but also no. 6, 1986.

[22] M. M. Balas, “The fuzzy-interpolative methodology,” Soft Computing

enhances comprehension, retention, engagement, and critical Based Modeling in Intelligent Systems, pp. 145–167, 2009.

thinking. By leveraging the power of visuals, educators can [23] J. M. Mendel, “Comparing the performance potentials of interval and

create a more effective and enriching learning environment that general type-2 rule-based fuzzy systems in terms of sculpting the state

space,” IEEE Transactions on Fuzzy Systems, vol. 27, no. 1, pp. 58–71,

prepares students for both academic and real-world challenges. 2018.

The application of such a set of measures is a large, lasting [24] S. Blakeley, “Bottom up & top down teaching

action and will require the involvement of all teachers, whose strategies.” [Online]. Available: https://study.com/academy/lesson/

bottom-up-top-down-teaching-strategies.html

task is difficult because they will have to start by changing

themselves.



815



Multimedia Based Sign Language Dictionaries:



How are Potential Users Involved?

∗ †

Žana Juvan Klemen Pečnik

zana.juvan@f e.uni- lj.si klemen.pecnik@fe.uni- lj.si

Faculty of Electrical Engineering Faculty of Electrical Engineering

University of Ljubljana University of Ljubljana

Ljubljana, Slovenia Ljubljana, Slovenia



Abstract from poor accessibility and usability. Modern multimedia tech-

nologies offer a promising solution for a better representation of

Sign languages are rich visual-gestural languages that serve as

the spatio-temporal content. These technologies can be used to

the primary means of communication for the d/Deaf commu-

create more accessible and effective dictionaries that meet the

nity. Despite their complexity and cultural significance, sign

target users’ needs [3]. Yet, there is a notable lack of testing and

language dictionaries face challenges in terms of documentation,

evaluation of these dictionaries with all target audience groups

accessibility, and usability. Despite their linguistic richness, the

[4]. This gap highlights a critical problem where effectiveness

development of comprehensive and user-friendly sign language

and practicality of these resources are often assumed rather than

dictionaries remains limited, especially in terms of adequate user

empirically tested.

testing, leading to a gap between technological advances and

This paper explores the development and key features of multi-

user satisfaction.

media Sign Language Dictionaries (SLD), focusing on appropriate

This paper examines the history and evolution of sign lan-

empirical testing with target groups. This could improve accessi-

guage dictionaries, from the early printed versions to interactive

bility and usability, as well as better meet the needs of the Deaf

digital formats. It highlights their key features and technologi-

community, their relatives, friends, and anyone who wishes to

cal advances and discusses the benefits of integrating modern

communicate in sign language.

technologies such as motion capture and artificial intelligence

into sign language dictionaries to improve the accuracy and ac-

cessibility of sign language resources. The paper emphasizes 2 Technical Aspects of Sign Language

user-centered design and calls for a thorough evaluation involv- Dictionaries ing diverse target groups, including d/Deaf, hard of hearing, and

In the interconnected world, access to information in one’s nat-

hearing users. By addressing the current lack of empirical test-

ural language is a fundamental right that is essential for equal

ing, this paper proposes a hybrid approach to the development

participation and involvement in society. Language is a carrier

of sign language dictionaries that are accessible, effective, and

of culture, identity and knowledge. When people have access to

culturally sensitive, ensuring equal access to communication and

information in their natural language, they can better engage

information for all.

and break down the barriers that exclude people from important



Keywords conversations and opportunities. If information is only available

in a few languages, most people become marginalised and their

online dictionary, sign language, sign language dictionary, user

voices go unheard.

testing, evaluation

Conventional methods of information dissemination fall short

when it comes to sign languages. General dictionaries rely on

1 Introduction words, phonetics, pronunciation, and text-based explanations

The natural language of the d/Deaf, where "deaf " refers to a that are inadequate for sign languages. Sign languages are visual

physiological state, while "Deaf " refers to a member of the Deaf and spatial, and include hand and body movements as well as

community [1], is sign language, characterised by its unique facial expressions which can modify sign meaning. To accurately

visual-gestural modality, with each sign functioning as a lex- represent these elements, SLDs must use additional means of

ical unit within a comprehensive grammatical system. These communication such as videos, or other alternatives like anima-

languages are fully-fledged and serve as important means of com- tions, 3D animations, 360 videos, spatial videos and in addition,

munication for people with varying degrees of deafness, whether specialised notations [5].

prelingual or postlingual. Developing effective SLDs requires a hybrid approach that

Despite their richness, sign languages face significant chal- combines features of both learner’s and explanatory dictionar-

lenges in terms of documentation and supporting materials, like ies. Learner’s dictionaries simplify definitions and provide clear

textbooks, grammar books and dictionaries [2]. High-quality sign examples to help beginners. Explanatory dictionaries provide de-

language dictionaries are rare, and those that exist often suffer tailed descriptions and a broader vocabulary for advanced users

[6]. An ideal SLD would include a learner’s section with videos

∗

ORCID: 0000-0001-8859-180X

† demonstrating basic signs, accompanied by 3D animations and

ORCID: 0009-0004-1430-8054

text descriptions, with the goal to focus on fundamental vocab-

Permission to make digital or hard copies of all or part of this work for personal

ulary and common phrases, making it accessible to beginners.

or classroom use is granted without fee provided that copies are not made or

The explanatory part would cover complex signs with multiple

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this examples in different contexts. This hybrid model would ensure

work must be honored. For all other uses, contact the owner /author(s).

that SLDs are versatile resources for all users, promoting equal

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

access to information and supporting the development of sign

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.DIGIN.5 language skills. By creating advanced hybrid SLDs that embrace



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Ž. Juvan and K. Pečnik



the visual and dynamic nature of sign languages, inclusion and of cultural sensitivity and user-centred design in the development

equality would be promoted in the best possible way. of these aids [15, 16].



2.1 History and Development

2.2 Advantages

The history and development of SLDs reflect the evolving under-

Sign language dictionaries offer numerous advantages that signif-

standing and appreciation of sign languages. The first sign lan-

icantly support the learning process and enhance communication

guage dictionaries were created in the 20th century [7] providing

and inclusion. They provide a standardised way of showing and

a valuable resource for both the Deaf and hearing communities.

understanding signs, which helps to standardise their meanings

These dictionaries were printed works that used static images,

for different users and contexts. This promotes the recognition

line drawings of signs and textual explanations to depict the

and use of the sign language, raises awareness and increases the

signs, which helped to somewhat bridge the gap in communica-

participation of the d/Deaf in society. In addition, these dictionar-

tion. These early dictionaries, while pioneering, had significant

ies enable anyone to learn new signs or refresh their knowledge,

limitations when it came to accurately capturing the dynamic and

ensuring accurate and effective communication, regardless of

spatial nature of sign languages. The static nature of used images

skill level - whether beginner or certified sign language inter-

often failed to convey the fluid and expressive aspects of sign

preter. As a comprehensive resource, sign language dictionaries

languages, often making it very hard for people to understand

play a crucial role in promoting inclusion and bridging commu-

and learn the signs correctly [3, 4].

nication gaps between d/Deaf and hearing people [3, 11].

A significant advancement in the representation of sign lan-

guage was achieved with the introduction of notation systems.

The first notation system was developed in 1960 by Stokoe, which

was closely followed by SignWriting (1974) and HamNoSys (1985) 2.3 Key Features

notation systems. They use pictures and abstract symbols to de- SLDs have evolved from simple printed resources to sophisticated,

scribe the elements of each sign, which makes analysis of sign interactive tools that are essential for learning, communication

language structure possible. To describe a sign accurately, five and inclusion. Initially, these dictionaries were based on static

parameters are needed: movement, handshape, location, palm images and textual explanations, which were basic but limited in

orientation and non-manual signals [4, 8]. Printed SLDs used four their ability to capture the full nature and details of sign language.

of the five parameters to describe included signs, which allowed Modern SLDs [3, 6, 11, 17, 18, 19] have evolved considerably,

for easier categorisation as well as more efficient searching, since incorporating a robust entry structure that typically includes

it was possible to organise signs by their characteristics instead the lemma, a video demonstration, and a detailed explanation.

of alphabetically [7]. Each entry can also provide usage examples and information

Video technology made a revolutionary advance for SLDs. on frequency of use, along with grammatical details, synonyms,

Early video dictionaries which began appearing in the late 20th antonyms and collocations. Visual aids and notation systems

century and used 2D video to demonstrate the signs [9, 10]. These such as Stokoe, HamNoSys, SignWriting further break down the

video dictionaries provide a more accurate representation of signs components of each sign, improving understanding and learning.

compared to static images, as they show body movements and The search function in these dictionaries is versatile and allows

facial expressions. However, 2D video still lacks spatial informa- users to find signs by entering a word or phrase, search by sign

tion, which is crucial for fully capturing the three-dimensional components, or by thematic groups. The search results are listed

nature of sign languages [11]. by relevance, ensuring the most accurate match.

The development of 3D avatar dictionaries has addressed some Interactivity is a key feature of modern SLDs. Video content

of the limitations of 2D videos by incorporating spatial informa- offers controls to play, pause, change playback speed and jump to

tion. These systems use computer-generated avatars to perform specific sections. Some dictionaries offer multiple synchronised

signs, offering users to depict signs from multiple viewing an- 2D videos filmed from different viewing angles for better under-

gles, which facilitates better understanding and learning. Despite standing, or even 3D avatars with 360-degree views, giving users

these advances, 3D avatars often struggle to adequately repro- the ability to freely change rotation or perspective.

duce natural facial expressions and body movements, which are As up-to-date and unlimited resources, online SLDs contin-

integral parts of signing, causing the Deaf community to not yet ually add new information, avoiding the limitations of printed

fully accept 3D avatars [12]. versions. This approach mitigates the pressure of deciding which

The most advanced SLDs today are exploring the use of sophis- entries to include and allows for quick updates, keeping the dic-

ticated interactive systems to improve inclusion and interaction. tionary current and comprehensive.

Researchers are utilising technologies like motion capture, aug- Additional features of modern SLDs increase their usefulness

mented reality and artificial intelligence to create more natural and educational value. Instructions on how to use the dictionary

and accurate representations of signs. These systems aim to cap- effectively, information on the history and context of sign lan-

ture the full complexity of sign languages, including accurate guage, and up-to-date information on the latest developments

facial expressions and subtle body movements [13, 14]. E.g. in- keep users informed and engaged. Features such as the "word

teractive platforms may allow users to view signs from different of the day" encourage regular learning and exploration of new

angles, slow down movements to study them in detail, and even signs. These interactive features lead to better learning outcomes.

converse with virtual assistants in sign language in real time. More advanced SLDs incorporate additional features, allowing

Scientific publications have documented these advances and users to participate in quizzes and games, practise with struc-

highlighted the associated technical and social challenges. Studies tured exercises, and save and print customised vocabulary lists.

have examined the effectiveness of various technologies and their In addition, SLDs can also support language rehabilitation by

acceptance in the Deaf community, emphasising the importance providing tools for continuous practise and improvement.



Multimedia Based Sign Language Dictionaries: How are Potential Users Involved? Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



3 User Centered Design and Evaluations 3.2 Evaluation Studies

The development of SLDs has mainly focused on its features, pro- The evaluation of SLDs can focus on both technical criteria and

gramming processes, and design. Numerous articles and papers user experience. Categorising features of SLDs and determining

meticulously describe the intricacies of these aspects, often em- what should be evaluated are crucial steps in this process.

phasising the technical and linguistic challenges. However, there From a technical perspective, the evaluations often look at how

is a glaring absence in the literature when it comes to user evalu- well the dictionary code or service works. However, this type of

ation of these dictionaries with the users. Since SLDs are used by evaluation is insufficient, if the user experience is not taken into

deaf, hard of hearing and hearing users, these solutions should be account. An effective evaluation should also include an assess-

tested and evaluated by all target groups with comparable tools. ment of functionality and usability, ideally involving the target

User testing is crucial to the development of any effective audience. For example, the evaluation of an Electronic Malaysian

educational resource, which is especially true for SLDs. Many Sign Language Dictionary [21] was conducted with actual users

studies have shown that involving end users in the testing phase and provided valuable insights into its effectiveness. In contrast,

is crucial to ensure effectiveness, efficiency and user satisfaction. most SLD evaluations often involve non-target groups of users,

One of the most widely used methods for this purpose is User typically in very small numbers, which limits the relevance of

Centered Design (UCD) [2]. This methodology emphasises effec- the obtained feedback.

tiveness, efficiency and user satisfaction. Effectiveness ensures Unfortunately, many SLD projects do not report on user testing

that the tools or services fulfil the intended purpose, efficiency at all. Instead, they rely on simple functional testing or use metrics

minimises the effort and time required for users to achieve their such as Google Analytics as a substitute for evaluation. This

goals, while user satisfaction creates a positive and engaging approach overlooks important aspects of user interaction and

user experience. UCD focuses on improving usability and aims satisfaction.

to develop tools that are intuitive and user-friendly. Ideally, SLD A few articles emphasise the importance of thorough user eval-

design should follow all UCD steps, including thorough user uation and show how user feedback can lead to improvements.

testing and evaluation. Unfortunately, in practise, many SLD de- A study on the users of an online dictionary of sign languages

velopment projects tend to skip the crucial fourth step of UCD — titled "Proposing an instrument for evaluation of online dictio-

evaluation with actual users. naries of sign languages" [4] provides a proposed framework for

This omission leads to a gap between the theoretical benefits of conducting comprehensive evaluations and further illustrates

SLDs and their practical usability. Without empirical evaluation, the benefits of involving users in the testing process.

developers miss out on important feedback that could lead to



improvements and adjustments. As a result, SLDs may not fully 3.3 Guidelines and Importance meet the needs and preferences of the d/Deaf and hard of hearing

The evaluation of sign language dictionaries is crucial to ensure

communities and hearing users, limiting their effectiveness as a

that they are effective, user-friendly, and meet the needs of their

learning or helping tool.

diverse users. Key aspects of evaluation must include testing

specific features, assessing usage and usability, ensuring ease of

use and reviewing the quality of videos and content. Feedback and

3.1 Target Users ideas from the community are invaluable for further development

The users of SLDs are diverse and they use these tools for various and improvement of any dictionary.

reasons, e.g. to learn a sign language, to interpret, to prepare for An evaluation of SLD should include all target user groups,

specific interpretations or to test their knowledge. Understanding including the d/Deaf and hard of hearing people, as well as profes-

these different user groups is essential for the creation of effective sional interpreters, teachers and other hearing learners, in order

and accessible dictionaries. to gain comprehensive insights. Methods that include interviews

Users can be categorised according to their purpose: Sign lan- are preferable to written surveys to obtain more in-depth, nu-

guage learners who are seeking to expand their vocabulary, in- anced feedback [22]. In a separate ongoing study, we were able to

terpreters who need accurate signs for communication, teachers demonstrate that it is important to use tools translated into sign

who use SLDs as teaching tools, and developers who are creating language for all users whose natural language is sign language,

or improving sign language resources. Another categorisation in addition to semi-structured interviews with participants. To

is by hearing status: d/Deaf people for whom sign language is summarise, it is important to tailor the evaluation methods to

a primary form of communication and hearing people who use the specific user group and allocate sufficient time for thorough

SLDs for learning, interpreting or teaching. testing and feedback collection. Although there are some differ-

Sign language competence also defines user categories: Begin- ences between countries in terms of certified interpreters and

ners who are just starting to learn, advanced beginners with basic legal obligations, each method should consider all possible op-

proficiency, proficient users who engage in complex interactions, tions to reduce the pressure on users by providing them with a

proficient users who often work in a professional context, and comfortable environment and relaxed user testing to minimise

experts, including native signers and experienced certified inter- the negative impact of testing methods on the final results.

preters or educators. It is important to understand all different The importance of these evaluations cannot be overempha-

user groups through user studies. They provide insight into how sised as they lead to a better end product or service. Systematic

the different groups interact with SLDs, their specific needs, chal- evaluations improve functionality, usability, and overall quality,

lenges and preferences. This knowledge is invaluable when it and ensure that the SLD or any other product or service actually

comes to customising features, increasing usability and improv- serves the target audience for which it is intended. This itera-

ing accuracy to ensure SLDs are effective and user-friendly for tive process ensures that the final product truly meets the needs

all experience levels within the sign language community [2, 3, and expectations of its users, resulting in a more effective and

20]. user-friendly sign language dictionary.



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Ž. Juvan and K. Pečnik



4 Conclusions Retrieved May 17, 2024 from https://ijeecs.iaescore.com/index.php/IJEECS/a

rticle/view/22518.

The evolution of sign language dictionaries from printed images

[6] Pavel Campr, Marek Hrúz, Jiří Langer, Jakub Kanis, Miloš Železný, and

to advanced interactive systems reflects significant technological Luděk Müller. 2010. Towards Czech on-line sign language dictionary - tech-

nological overview and data collection. In Proceedings of the LREC2010 4th

and linguistic advancements. While early efforts provided basic

Workshop on the Representation and Processing of Sign Languages: Corpora

resources, modern approaches strive to capture the complex- . European Language Resources Association and Sign Language Technologies

ity and richness of sign languages. As research and technology (ELRA), (May 2010), 41–44. isbn: 2-9517408-6-7. Retrieved May 19, 2024

from https://www.sign- lang.uni- hamburg.de/lrec/pub/10043.html.

advance, the goal is to develop inclusive, accurate, and widely

[7] William C. Stokoe, Dorothy C. Casterline, and Carl G. Croneberg. 1976. A

accepted resources that empower the Deaf community and im- . Linstok Press. Dictionary of American Sign Language on Linguistic Principles

isbn: 9780932130013. Retrieved July 31, 2024 from https://books.google.si/b

prove communication for all. Sign language dictionaries have

ooks?id=WjAFAQAAIAAJ.

thus become indispensable, continually updated resources that

[8] Joe Martin. 2000. A Linguistic Comparison: Two Notation Systems for Signed

support learning, foster inclusivity, and bridge communication Languages – Stokoe Notation and Sutton SignWriting, 1–33. doi: 10.1145/1

188913.1188915.

gaps between deaf and hearing individuals.

[9] Martin L. A. Sternberg. 1994. . American Sign Language Dictionary on Cd-Rom

To bridge this gap, it is essential that future SLD development Harpercollins Interactive. isbn: 978-0-06-279016-3. Retrieved July 31, 2024

incorporates comprehensive user testing into their methodolo- from https://search.library.wisc.edu/catalog/999897039602121.

[10] Aleš Jaklič, Dora Vodopivec, and Vito Komac. 1995. Learning Sign Lan-

gies. By doing so, developers can ensure that their dictionaries are Proceedings of the International Conference

guage through Multimedia. In

not only linguistically and technically sound, but also genuinely . IEEE, (May 1995), 282–285. isbn: on Multimedia Computing and Systems

useful and accessible for their intended users. 978-0-8186-7105-0. Retrieved July 31, 2024 from http://ieeexplore.ieee.org/d

ocument/484935/.

In conclusion, for SLDs to be truly effective, evaluations must

[11] Luka Cempre, Aleksander Bešir, and Franc Solina. 2013. Dictionary of the

extend beyond technical functionality and include thorough user Slovenian Sign Language on the WWW. In Human Factors in Computing testing. This involves categorising features appropriately, assess- and Informatics. Springer Berlin, Heidelberg, (July 2013), 240–259. isbn: 978-

3-642-39061-6. Retrieved May 17, 2024 from http://eprints.f ri.uni- lj.si/2099/.

ing content accuracy, and involving all targeted user groups to [12] Lorna Quandt, Athena Willis, Melody Schwenk, Kaitlyn Weeks, and Ruthie

ensure the dictionaries meet the actual needs of the Deaf and Ferster. 2022. Attitudes Toward Signing Avatars Vary Depending on Hearing

Status, Age of Signed Language Acquisition, and Avatar Type. Frontiers in

Hard of hearing communities, as well as all other user groups

Psychology, 13, (Feb. 2022), 1–14. Retrieved May 19, 2024 from https://pubm

who are using SLDs. ed.ncbi.nlm.nih.gov/35222173/.

[13] 2020. . IGI Recent Advances in 3D Imaging, Modeling, and Reconstruction

Finally, our research on SLDs and user testing methods has

Global, (Jan. 2020), 50–78. isbn: 9781522552956. Retrieved July 31, 2024 from

identified some exemplary approaches [20, 21, 22] that integrate

https://www.mdpi.com/journal/applsci/special_issues/MCV6B6MQRX.

users into the design, development, and evaluation stages. On [14] Ilias Papastratis, Christos Chatzikonstantinou, Dimitrios Konstantinidis,

Kosmas Dimitropoulos, and Petros Daras. 2021. Artificial Intelligence Tech-

the other hand, however, we could not find a suitable tool for

nologies for Sign Language. , 21, (Jan. 2021), 5843–5868. Retrieved Sensors

user testing that was adapted to or created in sign language. This July 31, 2024 from https://www.mdpi.com/1424- 8220/21/17/5843. highlights a significant area for improvement, particularly in [15] Vaishnavi Upadrasta, Astrid Oehme, and Sandra Böhm. 2020. User-Centred

Design for Accessibility in Media Content – Sign Language and Virtual

ensuring that solutions meet user needs without assuming that HCI International 2020 – Late Breaking Papers: Universal Access and

Signer. In

every solution and current evaluation methods are good enough. . Springer International Publishing, (Sept. 2020), 126–143. Inclusive Design

isbn: 978-3-030-60148-5. Retrieved July 30, 2024 from https://link.springer.c



Acknowledgements om/chapter/10.1007/978- 3- 030- 60149- 2_11. [16] Gregor Burger, Katja Oven, and Matevž Pogačnik. 2021. Uporabnost pamet-

This research was in part supported by Slovenian Research and nih telefonov in mobilnih aplikacij kot pripomočkov za osebe z invalidnostjo.

In Proceedings of the 30th International Electrotechnical and Computer Science

Innovation Agency (ARIS), P2-0425: Decentralized solutions for Conference ERK 2021. ERK 2021, (Sept. 2021), 105–108. Retrieved Aug. 1, 2024

the digitalisation of industry and smart cities and communities. from https://erk.f e.uni- lj.si/2021/erk21.pdf .

[17] Ines Kožuh, Primož Kosec, and Matjaž Debevc. 2016. Evaluating User Expe-

The results of the research were partially produced within the

rience of the Online Dictionary of the Slovenian Sign Language. In Proceed-

framework of the research project titled "Analysis of Concep- ings of the LREC2016 7th Workshop on the Representation and Processing of

tual Designs for Information and Technical System Support for . European Language Resources Association Sign Languages: Corpus Mining

(ELRA), (May 2016), 129–132. Retrieved July 10, 2024 from http://www.lrec-

the Disabled to Increase Social Inclusion" No. V2-2252, as part

conf .org/proceedings/lrec2016/index.html.

of the Target Research Program "CRP-2022," co-financed by the [18] Jette Hedegaard Kristoffersen and Thomas Troelsgård. 2010. The Danish

Sign Language Dictionary. In Proceedings of the XIV Euralex International

Slovenian Research and Innovation Agency (ARIS) and the Min-

Congres. Fryske Academy, (July 2010), 1549–1554. isbn: 978-90-6273-850-3.

istry of Labour, Family, Social Affairs, and Equal Opportunities

Retrieved July 6, 2024 from http://euralex.f ryske- akademy.eu/.

(MDDSZ). [19] Chiara Vettori and Mauro Felice. 2008. E-LIS: Electronic Bilingual Dictionary

Italian Sign Language-Italian. In Proceedings of the 13th EURALEX Interna-

References tional Congress. Institut Universitari de Linguistica Aplicada, Universitat

Pompeu Fabra, (July 2008), 791–796. isbn: 978-84-96742-67-3. Retrieved

[1] Gary Morgan and Bencie Woll. 2002. . July 6, 2024 from https://euralex.org/publications/e- lis- electronic- bilingual-Directions in Sign Language Acquisition

John Benjamins Publishing, (June 2002). isbn: 9789027297242. Retrieved dictionary- italian- sign- language- italian/.

Aug. 1, 2024 from https://api.semanticscholar.org/CorpusID:57766071. [20] Mireille Vale. 2015. A Study of the Users of an Online Sign Language Dic-

[2] Tania Di Mascio and Rosella Gennari. 2008. An Intelligent Visual Dictionary tionary. In . Trojina, Institute Electronic Lexicography in the 21st Century

for Italian Sign Language. , 7, (Jan. 2008), 318–338. for Applied Slovene Studies, (Aug. 2015), 281–303. isbn: 978-961-93594-3-3. Journal of Web Engineering

Retrieved July 10, 2024 from https://journals.riverpublishers.com/index.php Retrieved July 16, 2024 from https://api.semanticscholar.org/CorpusID:2183

/JWE/article/view/4081. 58484.

[3] Rachel Locker McKee and David McKee. 2013. Making an Online Dictionary [21] Siew Ow. 2009. User Evaluation of an Electronic Malaysian Sign Language

of New Zealand Sign Language. , 23, (Dec. 2013), 500–531. Retrieved Dictionary: e-Sign Dictionary. Lexikos Computer and Information Science, 2, (Apr.

July 6, 2024 from https://lexikos.journals.ac.za/pub/article/view/1227. 2009), 34–52. doi: 10.5539/cis.v2n2p34.

[4] Klara Majetić and Petra Bago. 2017. Proposing an Instrument for Evaluation [22] Amelie Unger, Dieter P. Wallach, and Nicole Jochems. 2021. Lost in Transla-

of Online Dictionaries of Sign Languages. In tion: Challenges and Barriers to Sign Language-Accessible User Research. INFuture2017: Integrating ICT

in Society. Department of Information, Communication Sciences, Faculty of In Proceedings of the 23rd International ACM SIGACCESS Conference on Com-Humanities, and Social Sciences, University of Zagreb, Croatia, (Jan. 2017), . Association for Computing Machinery, (Oct. 2021), puters and Accessibility

189–203. Retrieved July 10, 2024 from https://inf oz.f f zg.hr/inf uture/2017/i 1–5. isbn: 978-1-4503-8306-6. Retrieved July 31, 2024 from https://dl.acm.or

mages/papers/INFuture2017.pdf . g/doi/10.1145/3441852.3476473.

[5] Ahmed H. Aliwy and Ahmed A. Alethary. 2021. Development of arabic

sign language dictionary using 3D avatar technologies. Indonesian Journal

of Electrical Engineering and Computer Science, 21, (Jan. 2021), 609–616.



(In)Accessibility of Slovenian E-commerce the Year Before



the European Accessibility Act



Bogdan Cerovac †

Institute for Digital Accessibility

A11Y.si

Ljubljana, Slovenia

Bogdan.Cerovac@gmail.com



Abstract Unfortunately, automated accessibility testing is quite limited

and cannot cover all success criteria and accessibility

The European Accessibility Act (EAA, Directive (EU) requirements [10]. Furthermore, it can never definitively

2019/882) aims to harmonize accessibility requirements, with a determine conformance [3].

significant focus on e-commerce services due to their socio- Such tests are unable to address context-based accessibility



549 requirements, aligned with WCAG 2.1 success criteria at result, they often fall short in detecting real accessibility issues levels A and AA, to provide more relevant results for future [4]. economic impact. This study manually assessed 26 key EN 301 requirements, which can only be evaluated by a human. As a

accessibility trend comparisons and to advocate for necessary



awareness and education initiatives. We examined 20 major and Testing the accessibility of websites and native mobile

original, Slovenian e-commerce platforms expected to fall within applications using automated tools can therefore only provide a

the EAA's scope. None of the websites were fully accessible; superficial impression of the state of accessibility [5]. each failed at least 13 out of the 26 selected requirements, with



an average of 17.1 criteria unmet per website. Our findings The main objective of the European Accessibility Act (EAA)

highlight the urgent need for improved accessibility to ensure is to harmonize accessibility requirements for a wide range of

digital inclusivity in Slovenia and identify specific accessibility essential services and products, including e-commerce services,

and usability issues that must be addressed to achieve compliance starting from 28 June 2025. The EAA does not specify a

and bridge existing digital barriers. particular accessibility standard to be used. Instead, it



Keywords systematically removing and preventing barriers, preferably emphasizes that accessibility should be achieved by

Accessibility, E-commerce, European Accessibility Act, through a universal design or "design for all" approach. [6].

WCAG, EN 301 549, Design for all, EN 17161, Accessibility

culture Studies using manual testing for accessibility are sparse due

to the significant time investment and specialized knowledge



1 Introduction accessibility of 20 representative Slovenian e-commerce required. Our intention with this study was to manually audit the

With the increased digitalization of e-commerce and beyond, it websites, present an overview of our findings, highlight the most

is crucial that digital solutions are inclusive and accessible to significant barriers for people with disabilities, and briefly

everyone, including individuals with permanent, temporary, or suggest ways to prevent or address these accessibility issues.

situational disabilities. International studies show that 95.9% of

the top one million home pages had detected WCAG 2 failures Based on our experience from other audits, we expected to

[1]. find that all these websites would be inaccessible.



Recent studies in Europe reveal that 94% of European

websites fail to meet accessibility requirements [2]; however, 2 Methods

these large-scale studies primarily rely on automated Since the EAA does not define a specific technical standard, we

accessibility testing due to its speed, convenience, and low effort. used selected parts of EN 301 549 accessibility requirements to

evaluate functional performance statements.



† ORCID: 0009-0007-2254-2459 These requirements are mandated by the Web Accessibility

Permission to make digital or hard copies of part or all of this work for personal or Directive (Directive (EU) 2016/2102) [7], which applies to the

classroom use is granted without fee provided that copies are not made or distributed public sector in the European Union. for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s). Our study focused on a selection of 26 relevant accessibility

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia requirements from section 9 (check table 1) of the EN 301 549

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.DIGIN.6 (version 3.2.1) [8] standard during our manual evaluation,

supported by different tools (contrast checker, static code



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia B.Cerovac



validators (HTML, CSS and ARIA) and developer tools built in Table 1: Accessibility failures detected with manual

browsers). accessibility testing of selected EN 301 549 requirements

of 20 e-commerce sites

The selection of websites was based on three criteria: EN 301 549 Requirement Failed %

1. The company has at least 10 employees. sites of

2. The company has an annual turnover of at least 2 failed

million euros. sites

3. The website is original Slovenian e-commerce, rather 9.1.2.2 Captions (pre-recorded) 10/10 * 100%

than being a translation of content from international 9.1.1.1 Text alternatives 20/20 100%

companies. 9.1.3.1 Info and relationships 20/20 100%

9.1.4.3 Contrast (minimum) 20/20 100%

The sampling of webpages was conducted based on the following 9.4.1.3 Status messages 20/20 100%

essential end-user digital journeys: 9.2.4.1 Bypass blocks 18/20 90%

• Home page and navigation mechanisms. 9.2.4.4 Link purpose (in Context) 18/20 90%

• Product search, filtering, comparison. 9.2.4.7 Focus visible 18/20 90%

• Add to cart procedure. 9.3.3.2 Labels or instructions 17/20 85%

• Registration and login mechanisms. 9.1.4.5 Images of text 16/20 80%

• 9.2.2.2 Pause, stop, hide 16/20 80% Contact and customer support functions.

9.2.4.6 Headings and labels 15/20 75%

The manual accessibility audits were conducted from April to 9.2.4.3 Focus order 15/20 75%

May of 2024, by four auditors: two senior auditors with multiple 9.3.3.3 Error suggestion 15/20 75%

years of experience, one auditor with a couple of years of 9.2.1.1 Keyboard 15/20 75%

experience, and one junior auditor. The results were coordinated 9.1.4.11 Non-text contrast 15/20 75%

and verified to ensure the highest possible quality and reliability. 9.1.3.5 Identify input purpose 14/20 70%

9.1.4.1 Use of colour 12/20 60%

To ensure the integrity of our findings and maintain 9.3.3.1 Error identification 11/20 55%

confidentiality, we anonymized the e-commerce platforms under 9.3.1.1 Language of page 9/20 45%

study. This approach allowed us to objectively highlight the 9.1.4.10 Reflow 8/20 40%

barriers identified, aiming to raise awareness and mitigate 9.2.4.2 Page titled 8/20 40%

potential biases. 9.1.4.4 Resize text 5/20 25%

9.2.4.5 Multiple ways 4/20 20%



3 9.2.1.2 No keyboard trap 2/20 10% Results

9.1.3.4 Orientation 1/20 5%

All websites failed to conform to the EN 301 549 and WCAG 2.1



standards at both A and AA levels. * 9.1.2.2 Captions (pre-recorded) was only tested on 10 sites

as the other 10 did not have any videos.

On average, each website had 17.1 accessibility issues out of 26



selected criteria. The two worst-performing websites failed 20 The secondary goal of manual accessibility testing was to out of 26 criteria, while the best-performing website failed 13 out document specific issues that can be used for further research. of 26 criteria. We are only briefly reporting them here:

• Use of CAPTCHA was often conducted with

The specific accessibility requirements that all websites failed

inaccessible solutions that were totally blocking blind

include:

• 9.1.2.2 Captions (pre-recorded) – 10 sites of 20 used people relying on screen readers.

• Modal implementations, especially cookie consent

videos and all the videos were without captions of

modal windows, were often inaccessible for multiple

any kind.

• 9.1.1.1 Non-text content – mainly missing or wrong groups of users, especially blind screen reader users



• 9.1.3.1 Info and relationships – mainly wrong or data collection consent management. missing semantics of HTML code. • Unstoppable, auto-playable carousels loaded with • 9.1.4.3 Contrast (minimum) – mainly text contrasts information that is unusable for all people. alternative text on images and functional icons. keyboard). This also has further legal implications for and keyboard only (or assistive technologies based on

that were below 4.5:1 for normal size texts. • Pages coded as English with Slovene texts and

• 9.4.1.3 Status messages – mainly for features like add components that use English alternative texts for

to cart, filter/search result changes and some form screen readers on pages with Slovene language.

validations. • Inaccessible mobile (“hamburger”) menu buttons,

often totally unusable for keyboard-based assistive

technologies, screen readers and voice input.



(In)Accessibility of Slovenian E-commerce the Year Before the

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

European Accessibility Act



• 260 tab presses needed to bypass site navigation and We are aware of multiple limitations of this study, but would like

get to the content of the page. especially to point out the following:

• Background images coded as decoration including • We scaled down the scope of testing with the full

important information. EN 301 549 set of requirements to expedite

• When pages are zoomed in or in landscape mode on testing. Testing a full list of requirements would

mobile devices, the site framework like navigation, most likely produce even worse results, but our

footer, and sticky buttons cover almost 90% of the selection was based on the relevance of

screen, thus making content practically invisible. requirements for e-commerce.

• Keeping the selected 20 e-commerce websites

Results of manual accessibility testing of the selected EN 301 intentionally undisclosed makes comparison of

549 requirements / WCAG success criteria, show that even if all trends of the same websites impossible, but we

20 sites were to pass various types of automatic accessibility still believe that they are a well-chosen and

testing (which they did not), they would not conform to relevant representative sample for high-level

accessibility requirements and would be inaccessible to certain inaccessibility situation indication and may be

groups of users, especially screen reader users, keyboard-only compared with similar websites on a requirement

(and similar assistive technologies) users, users with visual basis.

impairments and deaf users or users with hearing impairments. • Our scope was limited to a set of the most vital

parts of the user journey, and testing beyond that

would most probably find more failures, but we

4 Discussion wanted to focus on the most important parts from

The study confirmed our expectation that all websites audited an end-user perspective instead of mapping the

were inaccessible to groups of users, especially people with situation of the technically wider but less realistic

disabilities. The pervasive neglect of web accessibility in scenarios.

Slovenian e-commerce not only excludes users with disabilities • We would like to involve people with different

but also represents a missed opportunity for businesses to reach disabilities to support the study with parallel

a broader audience. usability testing, using their own ways and

assistive technology, but that was not possible due

A year before the new accessibility legislation is enforced, we to limited resources.

would like to see better results – less inaccessibility, highlighting

the need for increased awareness and education among

stakeholders, project managers, web developers and designers, 5 Conclusion

content providers, and everyone else involved in the planning, This study's findings highlight the pervasive inaccessibility of

production, maintenance, and implementation of e-commerce. Slovenian e-commerce websites, with none of the 20 audited

sites fully conforming to EN 301 549 and WCAG 2.1 standards.

Besides people with permanent disabilities, we also need to Each e-commerce website failed an average of 17.1 out of 26

consider situational and temporary disabilities that are often left selected criteria, with issues ranging from missing captions and

out of demographics and statistics. It is important to be aware of alternative texts to inadequate contrast and problematic

the negative implications of inaccessibility on society as well as navigation mechanisms. These shortcomings exclude users with

its negative impact on the business sector. disabilities and represent missed opportunities for businesses to

engage a broader audience.

Automatic accessibility testing alone can never be enough to

test for conformance to accessibility standards, but it is As the European Accessibility Act's implementation approaches,

nevertheless a useful complementary tool, helping to make it is imperative to raise awareness and educate stakeholders,

manual auditing slightly faster and more efficient. There are including project managers, web developers, designers, and

numerous automatic accessibility testing tools that also have content providers. Improving accessibility is not only a legal

issues with false positives (reporting accessibility issues falsely) obligation and financial repercussion prevention, but also a

[10] and human interpretation will always be vital for quality of moral and business imperative. Automatic testing tools, while

the end results. useful, cannot replace the nuanced insights gained from manual

audits.

We would like to point out that it is obvious that there is Therefore, a combination of both methods, with a focus

ample evidence that accessibility needs to be integrated into on manual evaluation, is essential for meaningful progress

organizations from top to bottom, and it is also evident that toward digital inclusivity. When baseline conformance is

standards such as Design for All (EN 17161:2019) [9] and EN guaranteed and there are no obvious barriers, we highly

301 549 get insufficient attention. We encourage stakeholders recommend the involvement of people with disabilities, to

and all interested persons to study, implement, and share further improve the usability aspects beyond technical guidelines

knowledge to raise awareness, improve accessibility at scale, and and standards.

with that enable more people to use their services for common

benefits.



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia B.Cerovac



Addressing accessibility issues benefits everyone, including https://www.w3.org/WAI/standards-guidelines/act/rules/about/#act-rules-

those with temporary or situational disabilities, and enhances the are-partial-checks

[4] Adrian Roselli. Comparing Manual and Free Automated WCAG Reviews.

overall user experience. The study underscores the need for 2024. Retrieved from https://adrianroselli.com/2023/01/comparing-manual-

ongoing efforts to bridge digital barriers and ensure that e- and-free-automated-wcag-reviews.html#Highlights

[5] Kollotzek, G., Zimmermann, G., Ableitner, T., & Nebe, A. M.. 2021.

commerce platforms are accessible to all users, thereby fostering Comparison of Manual Evaluation Methods for Assessing the Accessibility

an inclusive digital environment in Slovenia. of Websites based on EN 301 549. In CHIRA (pp. 24-35).

https://doi.org/10.5220/0010647000003060

[6] Official Journal of the European Union. 2019. Directive (EU) 2019/882 of

the European Parliament and of the Council of 17 April 2019 on the

accessibility requirements for products and services (Text with EEA



Acknowledgments relevance). Retrieved from https://eur-lex.europa.eu/legal-

content/EN/TXT/?uri=CELEX%3A32019L0882

I would like to express my gratitude to the following individuals, [7] Official Journal of the European Union. 2016. Directive (EU) 2016/2102 of

the European Parliament and of the Council of 26 October 2016 on the

members of the Institute for Digital Accessibility A11Y.si, accessibility of the websites and mobile applications of public sector bodies

Slovenia, who made this research possible by auditing and (Text with EEA relevance). Retrieved from https://eur-lex.europa.eu/legal-

content/EN/TXT/?uri=celex:32016L2102

reviewing the audit results (listed alphabetically): [8] European Telecommunications Standards Institute. 2021. EN 301 549

V3.2.1: Accessibility Requirements for ICT Products and Services.



• CEN/CENELEC/ETSI Joint Technical Body. 2024. Retrieved from

Andreja Bevc MSc, https://www.etsi.org/deliver/etsi_en/301500_301599/301549/03.02.01_60/



• en_301549v030201p.pdf Grega Fajdiga [9] European Committee for Standardization. EN 17161:2019 Design for All - • Tilen Škraba Accessibility following a Design for All approach in products, goods and

services - Extending the range of users. 2019. Retrieved from

References https://standards.cencenelec.eu/dyn/www/f?p=CEN:110:0::::FSP_PROJEC

T,FSP_ORG_ID:62323,2301962&cs=1AECBCDFF18BED2C84BA2E5F

[1] WebAIM (Web Accessibility In Mind). The WebAIM Million. 2023. A7AF6E955

Retrieved from https://webaim.org/projects/million/#wcag [10] Markel Vigo, Justin Brown, and Vivienne Conway. 2013. Benchmarking

[2] Craftzing.com and Wieni.com. Digital Trust Index. digitaltrustindex.eu. web accessibility evaluation tools: measuring the harm of sole reliance on

2024. Retrieved from https://www.digitaltrustindex.eu/#summary automated tests. In Proceedings of the 10th International Cross-Disciplinary

[3] The ACT Task Force of the W3C Accessibility Guidelines Working Group Conference on Web Accessibility (W4A '13). Association for Computing

(AG WG). W3C Web Accessibility Initiative (WAI). 2024. Retrieved from Machinery, New York, NY, USA, Article 1, 1–10.

https://doi.org/10.1145/2461121.2461124





Digital Inclusion of Children with Special Needs in




Extracurricular Sports Activities



Nuša Maver † Matjaž Debevc Katja Košir

Pedagoška fakulteta AccessibleEU Filozofska fakulteta

Univerza v Mariboru Maribor, Slovenia University of Maribor

Maribor, Slovenia matjaz.debevc@guest.um.si Maribor, Slovenia

katja.kosir@um.si nusa.maver@gmail.com



to all, as they offer numerous benefits to everyone [2]. Including

children with special needs in sports activities requires properly

Abstract trained staff, suitable equipment and tools, and environmental

This article explores the inclusion of children with special needs and content adjustments. For successful inclusive sports training,

in extracurricular sports activities, focusing on necessary it is crucial that coaches and other professionals in sports

adaptations for their full participation. It analyzes the current understand the specific needs of different groups of children with

state of sports offerings in the Maribor region and identifies special needs, implement necessary adjustments during the

challenges faced by sports clubs. Key recommendations include training process, and remain open to exploring new possibilities

specific coach training, systematic program adjustments, and the and solutions [5].

use of digital tools to enhance engagement and track progress. A In recent years, digital tools have become essential in

handbook for coaches is presented, offering strategies and adapting sports activities for children with special needs. Modern

adaptations for various groups of children with special needs. technologies, such as exercise tracking apps, virtual reality, and

The aim is to ensure accessible sports activities for all children, interactive training programs, enable adaptations that facilitate

promoting their social, emotional, and physical development. inclusion in sports activities [7]. These tools not only ease the

implementation of training but also increase children's

Keywords motivation, allow monitoring of progress, and adapt activities to

Children with special needs, inclusion, sport clubs, inclusive their needs and capabilities.

sport activities, assistive technologies Through a review and analysis of the current situation,

conducted via survey research, we investigated the possibilities

for including children with special needs in organized sports

1 activities in the Maribor area and the surrounding regions, and Introduction

the knowledge of this field among the leaders of sports clubs and

The inclusion process is based on the mindset that children have

associations. Special emphasis was also placed on the use of

diverse abilities, expectations, and needs, but can learn and

digital tools, which can significantly contribute to improving the

develop together in a suitably designed environment. In a

quality of sports programs for children with special needs.

positive school climate, where diversity is valued, every child

Based on the analyzed research results, which will be

can realize their potential [3]. The Slovenian »Placement of

described in the following sections, and the review of

Children with Special Needs Act« facilitates the inclusion of

documented good practices from abroad, we have developed a

children with special needs into regular educational programs

handbook [4] providing guidelines for successfully

with adapted implementation and additional professional

implementing an inclusive training process. This handbook aims

assistance [6]. For effective work and active involvement of all

to enable coaches to effectively work with children with special

participants, teachers must understand the characteristics of the

needs within regular sports programs.

children and the necessary adjustments concerning content, tools,

environment, and other factors.



such as work, cultural life, leisure, recreation, and sports, Inclusion extends beyond education to other areas of life, 2 Empirical research

although it is rarely organized in these fields. This article focuses 2.1 Problem definition and purpose



by maintaining physical and mental health, boosting self-esteem, The issue of including children with special needs in organized on sports, which play a significant role in the lives of individuals



fostering friendships, and integrating into the broader social extracurricular sports activities is critical, as we observe in practice that this group of children has limited opportunities to environment [1]. Organized sports activities should be accessible participate in mainstream programs of sports clubs and

Permission to make digital or hard copies of part or all of this work for personal or associations.

classroom use is granted without fee provided that copies are not made or distributed The purpose of the research was to analyze the existing

for profit or commercial advantage and that copies bear this notice and the full situation and identify barriers preventing the inclusion of these

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s). children in sports activities. By doing so, we aimed to contribute

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia to a better understanding of the issue and develop guidelines for

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.DIGIN.7 working with each group of children with special needs in sports.



2.2 hindrance to the progress of other participants, and fear of Objective of the empirical research

responsibility.

The primary objectives of the empirical research were:

• The analysis of the inclusion of children with special needs in

To determine the opportunities available for children with sports clubs and associations showed a highly varied situation.

special needs to participate in extracurricular sports Key findings from the analysis can be linked to foreign practices,

• identifying the main reasons for the non-inclusion of children activities in the Maribor area and its surroundings.

To identify the barriers and reasons why sports

clubs/associations reject children with special needs. with special needs in additional sports activities as follows:

• To propose solutions and formulate guidelines to improve

the inclusion of children with special needs in sports Social environment:

activities. 1. Negative societal attitudes towards individuals with special

• To examine the role of digital tools in adapting sports needs.

activities for children with special needs. 2. Low level of trainer qualifications: sports clubs and

associations reported that their trainers have limited

2.3 knowledge and experience working with children with Methodology

The research was conducted in three parts. Initially, we special needs. Most surveyed trainers expressed a need for

performed a local environmental scan by contacting 13 disability additional education and training.

organizations via telephone to gather data on the availability of 3. Inappropriate attitudes and communication towards

sports programs for children with special needs and the children with special needs.

possibilities for their inclusion. In the second part of the research, 4. High costs: the groups with children with special needs are

we employed a survey method to collect data on the existing significantly smaller in number compared to others, as a

practices and challenges faced by sports clubs/associations in higher degree of individualization is necessary for the safe

including children with special needs in their regular sports conduct of training. It is also necessary to provide more

programs. coaches.

The survey questionnaire, answered by the leaders of the

Physical environment:

sports clubs/associations, consisted of ten closed-ended and two

1. Access issues and inadequate equipment (facilities, devices,

open-ended questions. A total of 48 local sports clubs and

tools).

associations, covering 18 different sports disciplines,

2. Lack of programs that enable children with special needs to

participated in the research.

engage in organized sports.

Regarding the use of digital tools in sports activities, we

3. Unadapted existing programs: although some clubs

conducted interviews and a brief review of studies published in

implement adjustments to their sports programs, these are

journals. The handbook written following the research is based

often not systematic or specific enough for individual

on the conceptual framework of foreign sports organizations that

groups of children with special needs. The most common

exemplarily implement inclusion in sports practice. To deepen

adjustments include changes in the intensity of training and

our understanding of the inclusive paradigm, we attended online

individually tailored tasks.

courses organized by their umbrella organizations, studied the

methods and principles essential for achieving inclusion in sports,

Personal reasons:

and supplemented these insights with our own experiences.

1. Lack of confidence among trainers in their abilities.

2. Lack of confidence among parents in the abilities of trainers.



3 3. Time constraints for parents. Analysis of the sports offer by disability

4. Financial constraints for parents.

organizations

5. Lack of parental knowledge about the importance of sports

The analysis of the sports offerings by disability organizations in activities for their child’s quality of life, available sports

in Slovenia (town Maribor area and its surroundings) revealed programs, and possible adaptive equipment and tools that

that while the range of sports activities is extensive, no would enable their child to safely engage in chosen sports

association allows participation by children, only by their adult activities. users.

From the analysis of the second part of the research, in which Use of digital tools:

the leaders of sports clubs/associations reported on their 1. Digital tools, such as exercise tracking apps, virtual reality,

experiences and challenges in including children with special and interactive training programs, are rarely used. However,

needs in their regular programs, we found that most leaders first the clubs that have introduced them reported positive effects,

inquire about the child’s disability, impediment, or deficiency including increased motivation among children, better

and implement certain adjustments in their work. Data on the progress tracking, and more tailored training programs.

number of children with special needs included in regular sports

programs was not provided by most sports clubs/associations. Based on these findings, we have written a handbook that

Those that did provide information reported having only one to formulates guidelines for improving the inclusion of children

six children with special needs in their sports clubs/associations. with special needs in sports activities.

The main reasons cited for non-inclusion were the lack of

trained personnel capable of making appropriate adjustments,



4 encourage positive interactions and praise achievements, thereby Handbook for inclusive coaches

increasing the child’s motivation and self-esteem.

The handbook, developed based on the findings of our research,

Relevant digital tools include applications for visualizing

aims to empower coaches to successfully implement inclusive

exercises (as videos or pictures), which help children understand

sports activities [4]. It includes guidelines and recommendations

and follow instructions. Additionally, coaches can use apps to

for working with children with special needs, with an emphasis

create structured training schedules, allowing children to better

on adaptations that enable these children to fully participate in

plan and execute activities.

organized sports activities.



inclusive paradigm, addressing the knowledge gaps among 5.2 Blind and Visually Impaired Children The first part of the handbook defines and describes the

sports professionals on how to approach, communicate, and For blind and visually impaired children, it is crucial to adapt the

establish successful relationships with children with special training environment to be safe and accessible. The use of

needs. auditory signals, tactile markers, and verbal descriptions of

A successful inclusive coach must possess specific exercises can significantly enhance understanding and safety.

knowledge and skills for working with children with special Coaches must pay attention to the orientation and mobility of

needs and create an encouraging environment where all these children, ensuring they feel safe and accepted.

participants can actively engage and realize their potential. Digital tools include auditory applications that guide children

The inclusive approach of a coach is based on empathy, through exercises and devices for auditory orientation. These

understanding, and respect. This requires careful planning of tools can help children navigate the space and participate in

sports activities and adapting methods, the environment, training without additional barriers. Additionally, the use of

equipment, tools, and content. Using respectful and positive braille displays for presenting information about exercises is

language is crucial for successful communication. Coaches must beneficial.

develop skills to adapt the training process, be innovative in

creating a motivational environment, and be adept and 5.3 Deaf and Hard-of-Hearing Children

understanding in resolving conflicts. Adaptations for deaf and hard-of-hearing children include using

The handbook describes the so-called inclusive spectrum, visual signals, gestures, and clear facial expressions. Coaches

which includes various methods and approaches to integrate must ensure that all visual representations of tasks are clear and

children with special needs into sports activities. Based on visible. Communication should be adapted so that children can

foreign practices, we derived inclusion models (TREE, STEP, follow instructions and participate in the training process without

and CHANGE IT), which are practical tools designed to support difficulty.

coaches in adapting and modifying sports activities. Digital tools include speech-to-text applications and video

Additionally, the handbook outlines an important friendship instructions with subtitles. Coaches can also use video calls with

system that promotes peer cooperation and creates a supportive sign language interpreters to facilitate communication.

network for children with special needs.

The second part of the handbook provides detailed guidelines

for adapting sports training for each group of children with 5.4 Children with Speech and Language

special needs. Disorders



can contribute to a more inclusive environment where every child disorders, it is important to use short and simple instructions and feels accepted and valued. visual aids. Coaches should encourage communication in various By following these recommendations and guidelines, coaches When working with children with speech and language

ways, including gestures, picture cards, and other visual tools to

facilitate understanding and participation.

5 Proposed adaptations for implementing



inclusive sports training interactive storytelling apps that assist in learning and Digital tools include visual communication apps and

When implementing inclusive sports training, it is crucial to communication. Applications for creating picture

understand the specific needs of each child, the limitations they communication boards are also useful, enabling children to

face, and the specific adaptations required to enable safe express their needs and desires more easily. participation in sports activities, along with the appropriate

digital tools. Each child has unique characteristics that affect 5.5 Children with Motor Disabilities



their abilities and needs during training. Adaptations for children with physical disabilities involve The following summarize key adaptations for each group to adjusting exercises according to their physical abilities and using ensure full participation and optimal development of every child. specific equipment, such as wheelchairs or orthopedic aids.

Coaches need to be creative in designing exercises that allow



5.1 Children with Intellectual Disabilities limitations. participation from all children, regardless of their physical

Adaptations for children with intellectual disabilities include Digital tools include motion control applications that aid in

using simple and clear instructions. Significant time should be rehabilitation and virtual reality for simulating various sports

dedicated to reinforcement, and it is essential to strive for a activities. Movement tracking devices can also be used to

predictable and structured training environment. Coaches should monitor progress and adjust exercises based on capabilities.



5.6 Adaptations for individual groups of children with special Children with Long-term Illness

needs are essential for the effective implementation of inclusive

For chronically ill children, it is crucial to consider their health

sports training. With the right approach and the use of digital

limitations and adjust the intensity of the training. Coaches must

tools, coaches can create an inclusive environment where each

be aware of the children’s health conditions and any

child feels accepted and valued, contributing to their optimal

contraindications for specific physical activities. Training should

development and full participation in sports activities.

be tailored to enhance the health and well-being of the children.

The use of digital tools is fundamental in facilitating these

Digital tools include health and fitness tracking apps that

adaptations and enabling the full participation of children with

allow for adjusting training programs based on health conditions.

special needs in sports. Applications for visualizing exercises,



enabling children to monitor their health status and adjust their auditory and visual signals, interactive stories, motion control, Apps for recording symptoms and medications are also useful,

and health monitoring are just some of the tools that can

activities as needed.

significantly improve the quality of training and increase

children's motivation.

5.7 Children with Learning Disabilities

It is also important to emphasize the need for continuous

Adaptations for children with learning disabilities involve using education and training for coaches working with children with

multisensory approaches and offering various learning methods. special needs. Only with appropriate knowledge and skills can

Coaches should use a combination of visual, auditory, and they create an inclusive and supportive environment.

kinesthetic methods, providing additional support and time for Future efforts should focus on continuing research and

understanding instructions. developing new approaches and technologies to support

Digital tools include interactive learning and brain training inclusion in sports. This will ensure that sports are accessible to

apps that help improve cognitive functions and motor skills. all children, regardless of their abilities or special needs,

Applications for creating visual schedules and monitoring providing them with equal opportunities for holistic development.

progress during training are also beneficial.



5.8 Children with Autism Spectrum Disorders Acknowledgments

Adaptations for children with autism spectrum disorders include The research was co-funded by the European Commission, in the

creating a structured and predictable environment and using framework of the project European Accessibility Resource

visual schedules and clear rules. It is important to reduce (or Centre - AccessibleEU (VT/2022/027). increase) sensory stimuli and adapt the training to allow focus

and sustained attention.



Digital tools include apps for creating visual schedules and References

social stories that help understand and adapt to changes. Sensory [1] Herman Berčič. 2001. Športna rekreacija v funkciji kakovosti življenja integration apps are also useful, helping children manage sensory prebivalcev Slovenije. V H. Berčič (ur.), Zbornik 2. Slovenskega kongresa športne

stimuli during training. rekreacije (9–20). Športna unija Slovenije.

[2] Anita Goltnik Urnaut. 2007. Šolske športne dejavnosti in samopodoba

5.9 Children with Emotional and Behavioral mladostnikov z ovirami v gibanju. [Doktorska disertacija, Univerza v Ljubljani,

Disorders Pedagoška fakulteta].

For children with emotional and behavioral disorders, it is [3] Kefallinou Anthoula, Simoni Symeonidou and Cor JW Meijer. 2020.

important to create a supportive and stable training environment. Understanding the value of inclusive education and its implementation: A review

Coaches should use positive reinforcements, set clear boundaries of the literature. PROSPECTS, 49(3–4), 135–152.

and rules, and be consistent in their training approach. [4] Nuša Maver. 2024. Vključevanje otrok s posebnimi potrebami v izvenšolske

Adaptations should enable children to manage their emotional športne dejavnosti. [Magistrsko delo, Univerza v Mariboru, Pedagoška fakulteta].

and behavioral challenges and participate in sports activities. [5] Rajko Vute. 1999. Izziv drugačnosti v športu. Narodna in univerzitetna knjižnica

Digital tools include apps for meditation and emotional v Ljubljani.

regulation that help children manage stress and anxiety. Apps for [6] Zakon o usmerjanju otrok s posebnimi potrebami (ZUOPP-1). 2011. Uradni list recording and monitoring emotional states are also useful, RS, št. 58/11, 40/12 – ZUJF, 90/12 in 41/17 –ZOPOPP. Accessed on:

enabling children and coaches to recognize behavior patterns and http://pisrs.si/Pis.web/pregledPredpisa?id=ZAKO5896#

adjust training accordingly. [7] Isabel Cristinna do Nascimento da Silva and Ana Cristina de Jesus Alves. 2020.

Identification of assistive technologies in paralympic sports: contributions and

barriers. Cadernos Brasileiros de Terapia Ocupacional, 28, 917-930.

6 Conclusions

The inclusion of children with special needs in organized sports

activities is crucial for their social, emotional, and physical well-

being. This article presents various adaptations necessary for the

successful integration of this group of children into sports

activities. We found that adaptations depend on the specific

needs of each group of children, making it essential for coaches

to understand and recognize these characteristics and needs,

maintain a positive attitude, be open to adjusting the training

process, and seek innovative solutions.





A micro-Learning Units Package for Improving Inclusive




Digital Education in HEI



Sérgio D. Correia Ana Cunha Maja Pušnik

Lusófona University Lusófona University Boštjan Šumak

Lisbon, Portugal Lisbon, Portugal maja.pusnik@um.si

Portalegre Polytechnic University ana.cunha@ulusof ona.pt bostjan.sumak@um.si

Portalegre, Portugal University of Maribor

sergio.correia@ulusofona.pt Maribor, Slovenia

scorreia@ipportalegre.pt



Abstract factors include leadership techniques, teamwork and network-

ing, infrastructure needs, professional development, pedagogical

In Higher Education Institutions (HEIs), digital inclusion plays a

support systems, classroom implementation, assessment proce-

crucial role in providing equitable and just access to educational

dures, and digital competency. By addressing these areas, the

materials for every student, irrespective of their varied origins,

learning units can be tailored to provide practical and actionable

capacities, or impairments. To leverage HEIs and their ability to

insights that support educators and administrators in fostering

implement an inclusive educational system, this work considers

an inclusive digital environment [4, 5].

identifying the factors that impact e-inclusion in HEIs to cre-

This paper presents the development of a micro-Learning Units

ate a new content package of micro-learning units. These will

cover categories such as leadership techniques, teamwork and Package

designed to enhance inclusive digital education in HEIs.

By incorporating the identified factors that impact e-inclusion,

networking, infrastructure needs, professional development, ped-

this package aims to provide a structured and flexible pedagogi-

agogical support systems, classroom implementation, assessment

cal content delivery system. The micro-learning units will cover

procedures, and digital competency. The structured pedagogical

various essential categories and will be supplied to users through

content is supplied to the user through a recommendation model

a recommendation model that includes an online self-evaluation

that constitutes an online self-evaluation tool.

tool. This approach ensures that the learning experience is person-



Keywords alized, responsive to individual needs, and conducive to fostering

a more inclusive digital education environment.

Inclusion, Learning Unit, Higher Education Institutions, Educa-

The remaining paper is organized as follows. Section 2 ad-

tion, Educational Content

dresses the key factors previously identified as enablers for dig-

ital inclusive education, and Section 3 describes the Learning

1 Introduction package, its structure and content. Finally, Section 4 concludes

Digital technologies have revolutionized the landscape of Higher the paper.

Education Institutions (HEIs), enhancing accessibility and facili-

tating innovative teaching and learning methods. However, this

digital transformation also underscores the imperative for digital 2 Factors for Inclusive Digital Education as

inclusion, ensuring that all students have equitable access to ed- Building Blocks ucational resources regardless of their diverse backgrounds and

Identifying the factors that affect digital accessibility is here con-

abilities. Digital inclusion in HEIs is not merely about provid-

sidered a previous starting point for constructing educational

ing access to digital tools but involves a comprehensive strategy

content. Policymakers, educators, and institutions can create

that encompasses leadership, infrastructure, professional devel-

plans and initiatives to support e-inclusion in HEI by recogniz-

opment, and pedagogical innovations [6, 2].

ing these elements and how they interact. In this case, a set of

Micro-learning units are small, self-contained learning mod-

micro-learning units was produced, making educational content

ules designed to deliver targeted educational content in a flexible

available that relays and is organized based on the structure of

and accessible manner. These units can be particularly effective

the identified factors [4, 5].

in promoting digital inclusion by allowing students, teachers, and

The overall methodology consisted of four different stages.

HEI staff to engage with material at their own pace and according

Firstly, a theoretical framework of inclusive digital education was

to their individual learning needs. Recent studies have shown

developed, considering technology, pedagogy, content, manage-

that micro-learning can enhance engagement and retention by

ment aspects, and different e-learning settings and modes. From

breaking down complex subjects into manageable segments [1,

this analysis, factors and indicators were obtained. Secondly, on-

3]. To create a comprehensive package of micro-learning units

line workshops for refinement and validation of the variables and

aimed at improving inclusive digital education, it is essential to

indicators were conducted in five countries (Italy, Portugal, Slove-

identify the key factors that impact e-inclusion in HEIs. These

nia, Spain, and Turkey). In the workshops, the audience discussed

and evaluated the framework, the factors, and the indicators).

Permission to make digital or hard copies of all or part of this work for personal

With the framework, the micro-learning package was created,

or classroom use is granted without fee provided that copies are not made or

consisting of a wide range of units as e-learning materials that

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this cover all the factors and indicators previously identified. Thirdly,

work must be honored. For all other uses, contact the owner /author(s).

a second set of workshops was conducted to refine and validate

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

the e-learning materials produced. Finally, a piloting stage was

© 2024 Copyright held by the owner/author(s).

implemented in five European Universities, where management

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Correia et al.





Figure 1: Micro-Learning Units Mapping with Categories, Factors, Indicators, and Estimated Duration



and teachers were invited to test the entire pipeline and evaluate For the first category, Leadership/School’s perspective, three

its performance. factors are considered: (1.1) Inclusive Digital Strategy and Policy

The model of factors enabling inclusive digital education was for Empowering Inclusive Digital Education: This factor reflects

categorized into 8 categories: (1) Leadership/ School’s perspec- the importance of an inclusive digital strategy, collaboration with

tive; (2) Collaboration and Networking; (3) Infrastructure and leaders and teachers, and the establishment of enabling policies

Equipment/ Tools and Technology; (4) Continuous Professional to reduce barriers to learning and participation; (1.2) Inclusive

Development; (5) Pedagogy: Supports and Resources; (6) Peda- Digital Pedagogy & Supportive Culture: This factor encompasses

gogy: Implementation in the classroom; (7) Assessment Practices/ supporting contemporary pedagogical approaches with inclusive

Inclusion Assessment; and (8) Student Digital Competence/ Stu- digital technologies, minimizing digital discrimination, and fos-

dent’s Perspective, where each category embraces 1 to 4 factors. tering collaboration and communication between HEI staff and

Besides each category, the theoretical model that identifies the the organization; (1.3) Professional Development for Inclusive

factors is operationalized by considering a set of indicators. By Digital Education where scheduled time for staff to explore in-

tracking these indicators, stakeholders can gain insights into clusive digital technologies and supporting their acquisition of

the effectiveness of digital inclusive education initiatives and digital literacy knowledge and skills are considered.

identify areas needing improvement. In the context of digital When considering the second category, Collaboration and

inclusive education, an indicator is a specific measure or met- Networking, a set of three factor are considered: (2.1) Collabo-

ric that provides information about the extent to which digital rative Evaluation and Planning for Inclusive Digital Teaching:

education resources and opportunities are accessible, equitable, highlighting the importance of reviewing progress in teaching

and effective for all learners. These indicators help assess and and learning with inclusive digital technologies, discussing the

monitor various factors such as access to technology, digital lit- advantages and disadvantages of inclusive teaching, and engag-

eracy, inclusivity, engagement, policy support, and educational ing in collaborative planning for inclusive courses and teaching

outcomes, allowing educators, policymakers, and stakeholders methods; (2.2) Shared Vision and Collaborative Culture for Inclu-

to identify strengths, gaps, and areas needing improvement in sive Digital Education: This factor emphasizes the importance

implementing digital inclusive education. of promoting a common vision of digital educational inclusion

A micro-Learning Units Package for Improving Inclusive Digital Education in HEI Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia





Figure 2: Recommendation Tool Model



among the main actors in the institution (management, pedagog- The sixth category, Pedagogy: Implementation in the class-

ical staff, students, and parents), as well as fostering inclusive room, involves three factors: (6.1) Personalized and Engaging Dig-

digital collaboration and communication between staff and stu- ital Learning: This factor emphasizes the importance of HEI staff

dents; and (2.3) Collaborative Engagement for Inclusive Digital using digital technologies to personalize and tailor their teaching

Partnerships: This factor highlights the importance of enabling to student’s individual needs, incorporating digital learning ac-

inclusive collaboration with local communities, caregivers, and tivities that foster creativity, and setting inclusive digital learning

parents through digital technologies, as well as actively collab- activities that actively engage and motivate all students; (6.2)

orating with other HEIs or organizations to support the use of Inclusive Classroom Collaboration and Respectful Environment:

inclusive digital technology. This factor focuses on promoting inclusive collaboration and

Regarding Infrastructure and Equipment/ Tools and Technol- creating a respectful classroom environment using digital tech-

ogy, the third category, four factors are considered: (3.1) Accessi- nologies; and (6.3) Enhanced Pedagogical Digital Inclusiveness:

ble Infrastructure for Inclusive Learning: This factor highlights This factor emphasizes the focus on enhancing digital inclusive-

the importance of both digital and physical infrastructure in cre- ness among pedagogical staff.

ating an inclusive learning environment; (3.2) Accessible Devices The seventh category, Assessment Practices/ Inclusion Assess-

for Inclusive Teaching and Learning: This factor emphasizes the ment, embraces three factors: (7.1) Inclusive Digital Assessment

importance of providing accessible digital devices and assistive Practices: This factor emphasizes the importance of HEI staff

products for inclusive teaching in HEI; (3.3) Digital Accessibility using inclusive and accessible digital technologies for assessing

Support for Inclusive Learning: This factor emphasizes the impor- students’ skills; (7.2) Digital Feedback and Self-Reflection: This

tance of HEI providing support for digital accessibility in order factor emphasizes the importance of HEI staff using inclusive

to facilitate inclusive learning; and (3.4) Inclusion and Equity digital technologies to provide timely feedback to students, en-

in Digital Learning: This factor emphasizes the importance of able students to self-reflect on their own learning and facilitate

HEI’s efforts to identify and address the digital divide and chal- peer feedback on other student’s work, and (7.3) Data-Driven

lenges related to students’ learning needs and socio-economic Improvement for Inclusive Digital Learning: This factor high-

backgrounds in the context of digital learning. lights the focus on leveraging digital data analysis and evaluation

Only one factor is identified for the fourth category, Contin- metrics to identify students’ needs and improve their inclusive

uous Professional Development: (4.1) Continuous Professional digital learning experience.

Development for Inclusive Digital Teaching. This factor empha- Finally, the last category, Student Digital Competence/ Stu-

sizes the importance of HEI leaders discussing and addressing dent’s Perspective, is considered with one factor: (8.1) Inclusive

continuing professional development needs for inclusive teach- Digital Communication and Skills Development: This factor em-

ing with digital technologies. It also highlights the importance of phasizes the importance of HEI ensuring that students learn to

providing staff with opportunities to participate in CPD and fos- communicate in an inclusive way using digital technologies and

tering the sharing of experiences within the school community. that they develop their digital skills and knowledge on inclusive-

Concerning the fifth category, Pedagogy: Supports and Re- ness across subjects. The factor reflects the focus on promoting

sources, two factors are considered: (5.1) Inclusive Digital Teach- inclusive communication practices and fostering the develop-

ing Practices: This factor emphasizes the importance of HEI teach- ment of digital skills in relation to inclusivity.

ers using virtual learning environments and digital technologies

in an inclusive way, as well as being trained and instructed on how

to address student diversity; and (5.2) Inclusive Digital Resources:

This factor emphasizes the importance of HEI staff searching for 3 The Micro-Learning Units Package

inclusive digital educational resources, creating digital resources Based on the factors previously identified and described in Sec-

to support inclusive teaching, using inclusive open educational tion 2, a newly package of micro-learning units was created. A

resources, and developing and updating resources that support micro-learning unit is a small, focused segment of educational

inclusive learning and participation. content designed to teach a specific skill or concept in a brief pe-

riod. These units are typically short, ranging from a few minutes

to about 15 minutes, and are meant to be easily digestible and

accessible. To organize the micro-learning units and to guarantee

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Correia et al.





Figure 3: Example of a Micro-Learning Unit Content



their integrity, the units are correlated to the indicators associ- nature of micro-learning units can improve retention by mini-

ated with each factor, also allowing directing the reader with the mizing cognitive overload and reinforcing key points through

help of a self-evaluation tool as described in Fig. 2. The figure repetition and varied presentation. Although the micro-Learning

presents a comprehensive framework for the micro-Learning units are integrated into a larger educational framework, they

Units Package aimed at improving inclusive digital education in are also effective as standalone lessons.

Higher Education Institutions (HEIs). It is organized into several

categories, each addressing specific factors contributing to digital 4 Conclusion

inclusion. Each category is further divided into factors, indicators, Using a self-evaluation tool, a set of micro-learning units previ-

and corresponding micro-learning units with specified durations.

ously mapped using factors and indicators of inclusive digital

The user is first invited to answer a questionnaire that relies

education is made available to a user through a recommendation

on the indicators. From this point, a scoring system evaluates the

model.

user’s perception of each factor, and globally, in its aptitude in

The global set of micro-training units covers a broad set of

relation to digital and inclusive education. Then, a decision-based

skills, considered based on the factors that involve inclusive

model advises the user on which micro-learning units it should

digital education. With this procedure, the personnel involved

be involved in. The estimated time frame for each unit varies

in higher education institutions, teachers or staff, acquire the

from 15 to 30 minutes, depending on how many indicators it

necessary skills to provide truly inclusive digital education in

covers, since some indicators are cross-correlated and are thus

their institution.

blended into one micro unit. A total of 32 micro-learning units

constitute the overall package, totaling seven hours, although the Acknowledgments purpose of the tool is for the user to only study the units related

This research was conducted within the activities of the Self-

to the topics they least mastered. Besides, the tool can be used

Evaluation Tools for e-Inclusion in HEI (SET4Inclusion) project

cyclically, meaning it can be used several times until the overall

(2022-1-SI01-KA220-HED000088368), co-funded by the Erasmus+

score obtained from the self-evaluation reaches a satisfactory

Programme of the European Union.

value, allowing the user to manage the learning path. Both the



self-evaluation tool and the micro-Learning unit contents are References

allocated online at www.set4inclusion.eu. Figure 3 exemplifies

[1] Mary Jo Dolasinski and Joel Reynolds. 2020. Microlearning: a new learning

how the unit , model. , 44, 3, 551–561. "Is my digital infrastructure accessible and inclusive?" Journal of Hospitality & Tourism Research

related to indicators , and "IET 3.2:

"IET 3.1: Digital infrastructure" [2] Paulo Ricardo dos Santos, Débora NF Barbosa, Eduardo G Azevedo Neto, Jorge

LV Barbosa, Sérgio D Correia, and Valderi RQ Leithardt. 2021. Learning and

Internet access", associated to the factor "Accessible Infrastructure well-being in educational practices with children and adolescents undergoing

for Inclusive Learning", under the category "Infrastructure and cancer treatment. Education Sciences, 11, 8, 442.

Equipment/Tools and Technology" [3] Karen Mossberger, Caroline J Tolbert, and Ramona S McNeal. 2007. Digital

is presented to the user.

citizenship: The Internet, society, and participation. MIt Press.

When considering the presented micro-Learning package, [4] Maja Pušnik, Katja Kous, Tatjana Welzer Družovec, and Boštjan Šumak.

wider content is broken down into small, manageable pieces, 2024. Identification and analysis of factors impacting e-inclusion in higher

education. In . IOS Press, Information Modelling and Knowledge Bases XXXV

making it easier for learners to consume and understand, where

308–317.

each unit targets a single, specific learning objective, ensuring [5] Boštjan Šumak, Sergio Duarte Correia, Ana Cunha, Tuncer Can, Irfan Simsek,

Katja Kous, and Maja Pušnik. 2024. Identification of factors that impact e-

that learners can quickly grasp the intended concept or skill.

inclusion in hei. In , 2024 47th MIPRO ICT and Electronics Convention (MIPRO)

Designed to be accessed on-demand through the self-evaluation

478–483. doi: 10.1109/MIPRO60963.2024.10569746.

tool, it allows learners to engage with the material at their own [6] UNESCO. 2020. Global Education Monitoring Report 2020: Inclusion and edu-

pace and on their own schedule. Also, the concise and focused cation: All means all. UNESCO, Paris. https://unesdoc.unesco.org/ark:/48223

/pf 0000373718.





Segmentation of students with special needs at UL




∗

Andrej Košir Janja Košir

Janez Zaletelj ∗ Pedagoška Fakulteta, Univerza v Ljubljani

andrej.kosir@f e.uni- lj.si Ljubljana, Slovenia

janez.zaletelj@f e.uni- lj.si janja.kosir@pef.uni- lj.si

Fakulteta za elektrotehniko, Univerza v Ljubljani

Ljubljana, Slovenia



ABSTRACT segmentation of students with disabilities at the University of

Ljubljana (UL). The initial goal of segmentation is to plan and

This paper deals with the identification and interpretation of seg-

implement effective and needs-based support for students with

ments of students with disabilities at the University of Ljubljana.

disabilities at UL.

The data on which this segmentation is based are the students’

We have concluded that this study does not require ethical re-

responses to a dedicated questionnaire created through an op-

view. The study is fully consistent with the purpose of collecting

erationalization process according to the selected aspects and

data from students with disabilities and its results will be used

objectives of the segmentation.

for the direct benefit of the population.

The segmentation aimed to identify homogeneous groups of

students with disabilities in order to 1. improve the understanding

of students’ needs, 2. prepare reliable data for the selection of 1.1 Aims of the student segmentation

technological support for students and 3. create the basis for The main aim of the segmentation is to learn about the main

a subsequent project on ICT-based support for special needs groups of students with disabilities in relation to the opportuni-

students. ties for technological support for their studies provided by the

Through the operationalization process, we identified five university. The sub-objectives are:

areas (aspects), namely 1. Technology and overcoming study

(1) Identify the groups of students with disabilities at UL along

barriers, 2. Technology and study outcomes, 3. ICT and Study

with their basic characteristics.

Obligations, 4. ICT and Study Skills, and 5. Opportunities to

(2) Identify meaningful groups of active students with disabil-

use ICT technology. Based on student responses, we identified

ities.

segments for the first three areas, with all three areas well covered

To achieve the goal ad 1. we designed a questionnaire. The

by three segments each. We identified three segments in each

design of the questionnaire resulted in 12 questions, with 5 main

area: 1. Technology Enthusiasts (accept everything), 2. Skeptical

sub-questions and a larger number of sub-questions.

Users (reject everything but the exceptions), and 3. Picky Users

To achieve objectives ad 2, we conducted an extensive data

(accept almost everything but the exceptions). From the second

analysis (see Sec. 4) and consulted domain experts.

and third segments, we extracted the main characteristics by

technology and by activity. The results are applicable for the next

steps in technology support for students with special needs. 2 STATE OF THE ART



2.1 User segmentation

KEYWORDS

User segmentation is the process of dividing users into different

user segmentation, special needs students, segmentation algo-

groups or segments based on common characteristics. It was

rithm

developed in the field of business and management. An organiza-

tion can segment users by language preference, product version,

1 INTRODUCTION geographic region or user personality [1]. A similar segmentation

Effective technological support for students with special needs is has been successfully applied to other areas, e.g. to the users of

crucial for modern teaching and learning at universities. Over the ICT and also to the area of ICT in special education [6].

last decade, the landscape of teaching and learning has changed Data mining techniques entered the field quite early on [21].

rapidly [17].On the other hand, the rapid development of in- Machine learning-based techniques are the most important ap-

formation and communication technologies [6] and studies on proach to user segmentation today [2].

technological support for students [4] has added a variety of Successful segmentation methods lead to homogeneous sub-

new support options. There is no effective technological support groups of users. A necessary next step is to understand these

without prior knowledge of the needs of the users - in our case, segments, i.e. to define and understand their typical represen-

the specific needs of our students. tatives. Such a representative is called a persona, and to clearly

In this paper, we present the operationalization of the segmen- understand the part of their behavior that is relevant to them,

tation instrument (aspects and questions) and the results of the they are described as a living person [20]. A persona is therefore

a fictional character whose characteristics and goals best fit the

∗

Both authors contributed equally to this research.

segment.

Permission to make digital or hard copies of part or all of this work for personal

or classroom use is granted without fee provided that copies are not made or 2.2 Operacionalisation distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this Operationalization is a process of 1. selecting relevant aspects

work must be honored. For all other uses, contact the owner /author(s).

of the designed instrument (set of questions) and 2. selecting an

Information Society 2024, 10–14 October 2022, Ljubljana, Slovenia

initial set of questions representing the selected aspects. There

© 2022 Copyright held by the owner/author(s).

https://doi.org/https://doi.org/10.70314/is.2024.DIGIN.9 is a long history of research on this approach [16]. It is based



Information Society 2024, 10–14 October 2022, Ljubljana, Slovenia Košir, Košir and Zaletelj, et al.



on human expert judgment. The aim of operationalization is to (5) Opportunities to use ICT technology

construct a measurement instrument that is used in data collec-

To cover these five aspects, we also constructed, selected and

tion [12]. The validity and reliability of the resulting instruments

modified questions that led to the final instrument (not listed

are of central importance and the research community has devel-

here for space reasons).

oped strict guidelines on how to achieve and ensure this [3]. In



our case, we have focused on relevant aspects of technological 3.2 Participants and data collection support for students with special needs.

The selected population is all UL students with disabilities. Given



2.3 the estimated number of 700 to 800 such students, the entire Unsupervised clustering for user

segmentation population was included in the sample.

The data collection was carried out by the University of Ljubl-

User segmentation is a subfield of the highly developed field

jana (UL) administration services. They provided us with a list

of customer segmentation. Machine learning techniques found

of all students with special needs at ULat UL. We then manually

their way into customer segmentation decades ago [19]. From

screened this list with the baseline descriptions and selected 723

various data mining approaches [21], the focus shifted to neural

respondents.

networks [18] and deep learning [15].

The inclusion criteria for selection were self-selected cate-

Unsupervised clustering with visualization of the cluster hier-

gories of special needs selected by the students at the time of

archy is a necessary step in customer segmentation [11]. In the

enrolment. They covered general special needs, and deficits from

case of this study, understanding the user segments obtained is

disabilities (hearing, vision, speech, physical, emotional, mental

very important. The technique of explainable customer segmen-

health). There were no exclusion criteria for selection into the

tation is discussed in [14].

sample.



3 MATERIALS AND METHODS 3.3 Unsuprvised clustering for user segment

3.1 Operacionalisation and existing determination

instruments We applied the unsupervised clustering technique K-Means with

dendrogram visualization [5]. The feature space was a space of

The operationalization procedure of this research focused on the

participant responses, no dimensionality reduction or location-

existing support for students with special needs at our and other

dependent transformation [10] or similar was used. The Eu-

universities. Theoretical knowledge and practical experiences

clidean distance is used for K-means clustering. Other distances

in supporting students with special needs were also taken into

lead to similar clusters. Since all responses were on the same Lik-

account. This is crucial to achieve good validity and reliability of

ert scale, no prior scaling was performed. We used elbow curves

the resulting instrument in fewer iterations.

to determine the optimal number of segments and basic statistics

To incorporate existing measurement instruments related to

to determine the most important characteristics of the selected

the use and benefit of assistive technologies, we also examined

segments. The initial number of clusters was set to 𝑘 12. =

available measurement instruments. As early as 1996, the authors

of [7] developed a scale ”Quebec User Evaluation of Satisfaction

with Assistive Technology (QUEST)”. The instrument contains 4 EXPERIMENTAL RESULTS



18 items and two dimensions. The construct was later reviewed 4.1 Student data

in [8], where 12 items were further selected.

The sampling method chosen was to send emails to the e-mail

A scale called the Psychosocial Impact of Assistive Devices

addresses provided by the students with disabilities at the time

Scale (PIADS) was developed by the authors of [13]. The PIADS

they obtained their special education student status. The inclu-

is a 26-item self-report questionnaire to assess the impact of

sion and exclusion criteria were specified in subsection 3.2. The

an assistive device on functional independence, well-being and

questionnaire was administered via the web-based system 1ka

quality of life. The construct is three-dimensional and includes 1.

(https://www.1ka.si/d/sl) and the questionnaire with response

competence (measures feelings of competence and effectiveness),

categories is available on request. Invitees received an email with

2. adaptability (indicates willingness to try new things and take

explanations and instructions.

risks) and 3. self-esteem (indicates feelings of emotional health

The invitations were sent out in three phases

and happiness).

The authors Dijcks et. al developed a scale (one-dimensional (1) First invitation: by e-mail from the university e-mail ad-

construct) to assess the quality of service delivery in the provi- dress;

sion of assistive technology (KWAZO) [9]. It aims to assess the (2) Second invitation: via the Disability representatives at the

quality of the provision of assistive devices from the customer’s faculties and academies;

perspective. The instrument consists of seven questions relating (3) Third invitation: via the Association of Students with Dis-

to accessibility, knowledge, coordination, efficiency, flexibility abilities

and user influence. The reliability of the scale is rated as very

The invitation was sent to 733 people. A total of 18 (3.4%)

good.

responses to the first invitation were received within one week,

As a result of all considerations, we decided to include the

next 7 to the second invitation and next 18 to the third invitation,

following aspects:

altogether 43 (5.9%). At the time of study design, we estimated

(1) Technology and overcoming study barriers that the response rate of respondents would be around 10%. We

(2) Technology and study outcomes did not formally identify the reasons for this low response rate,

(3) ICT and Study Obligations but we did gather some opinions. These can be grouped as fol-

(4) ICT and Study Skills lows:

Segmentation of students with special needs at the University of Ljubljana Information Society 2024, 10–14 October 2022, Ljubljana, Slovenia



(1) There is enough freely available technology that I can use overcoming study barriers, Q2: Technology and study outcomes

myself and I do not need any special support from the and Q3: ICT and Study Obligations.

university; The interpretation of the identified segments is based on man-

(2) In the responders’ mind, the university does not have ual inspection of identified clusters’ specifics such as frequency

enough technical support to help individuals; of answers etc. Due to lack of space, we cannot reproduce these

figures here.

We present the results of the segmentation according to the

criteria specified in Subsec. 3.1 identified aspects.

4.3 Aspect Q1: Technology and overcoming

4.2 study barriers Basic statistics

The main question was: "Please indicate how important each

The breakdown of the 43 respondents by gender can be found in

of the ICT assistive technologies listed is to you in overcoming

Fig. 1 (top left), where NA indicates those who did not wish to

the challenges you face in your studies due to your own specific

state their gender.

needs."





Seg. Q1a Q1b Q1c Q1d Q1e Q1f



1 4.3 4.0 5.6 6.2 5.9 6.2

2 4.4 3.9 2.9 2.5 2.3 2.9

3 4.1 3.6 5.0 5.2 4.8 6.2

Q1g Q1h Q1i Q1j Q1k Q1l



1 5.9 4.1 6.1 5.5 4.9 6.0

2 2.2 3.5 4.0 3.6 3.2 3.4

3 5.3 3.6 3.5 3.8 3.9 5.3



Table 1: Aspect Q1: Averages of answers per segment.



According to the inertia curve and the dendrogram of the

development of the segments, the number of identified segments

can be set to 3 or 5. In line with the segment structure, we have

opted for three segments.

The observed characteristics by segment are:



Seg. 1: The answers to all questions are the highest, i.e. all tech-

nologies are rated as very important. These are technology

Figure 1: Response of participants: by gender (top left), age

enthusiasts.

(second row left), special needs (second row right), study

Seg. 2: Most questions are answered with low values. These

style (third row left) and university (third row right), where

are . Little importance is attached to most technology sceptics

UL stands for the University of Ljubljana and UMB for the

technologies, with the exception of e-materials, e-environments

University of Maribor

and multimedia content. This segment therefore scores well

for e-materials, but not for content conversion tools, etc.

The histogram by age is shown in Fig. 1 (center left), where NA Importance varies considerably on average. These are Seg. 3:

denotes those who did not want to reveal their age. Respondents those who believe in and use some technologies but not others.

aged 44 and 55 were excluded from the analysis. The responses They rate most technologies well, with the exception of visual

by special needs are shown in figure 1 (middle right), where the and design customization tools.

meanings of the special needs are as follows:

P1: ’Deficits in certain areas of learning (dyslexia, dysgraphia, 4.4 Aspect Q2: Technology and study

dyscalculia ...)’, outcomes

P2: ’Physical disability’, The guiding question was: "Please indicate how important each

P3: ’Partial or total hearing loss’, of the ICT support services listed is to you in overcoming the

P4: ’Partial or total loss of vision’, challenges you face in your studies due to your own specific

P5: ’Speech-language difficulties’, needs."

P6: ’Emotional and behavioral disorders’, The number of segments was determined by combining the

P7: ’Long-term or chronic illness’, dendrogram and the "inertia" curve". We decided on 3 segments.

P8: ’Autism spectrum disorder’, The observed characteristics by segment are:



In terms of university, one student was from the University Technologies are of varying importance. These are Seg. 1: criti-

of Maribor and 18 students did not specify their university. We . They rate most technologies well, with the exception cal users

assume that most of the students were from the University of of audio-to-sketch, dictation, e-interpreting and audio-to-text

Ljubljana. tools.

As there are not enough complete responses for aspects Q4 They rate all technologies as very important. This is seg. 2:

and Q5 (less than 15) to draw more reliable conclusions, we only , the first segment from a segmenta- technology Enthusiasts

report the results for the first three aspects Q1: Technology and tion into two segments.



Information Society 2024, 10–14 October 2022, Ljubljana, Slovenia Košir, Košir and Zaletelj, et al.



Seg. Q2a Q2b Q2c Q2d Q2e Q2f ACKNOWLEDGMENTS

This research was supported by the project P2-0246 ICT4QoL -

1 3.4 3.0 3.4 2.6 3.1 4.3

Information and Communications Technologies for Quality of

2 4.4 6.0 4.9 6.2 6.3 5.8

Life.

3 1.5 1.5 1.0 1.0 1.0 1.0



Q2g Q2h Q2i Q2j Q2k REFERENCES

1 4.1 3.9 4.3 3.0 6.3 [1] Miguel Alves Gomes and Tobias Meisen. 2023. A review on customer seg-

mentation methods for personalized customer targeting in e-commerce use

2 5.0 5.0 5.3 5.7 6.2

cases. en. , 21, 3, (Sept. 2023), Information Systems and e-Business Management

3 2.5 1.5 1.5 2.5 2.0

527–570. doi: 10.1007/s10257- 023- 00640- 4.

Table 2: Aspect Q2: Averages of answers per segment. [2] R Amutha and Aazmaan Ahmed Khan. 2023. Customer Segmentation using

Machine Learning Techniques. en. 44, 3.

[3] APA. 2020. APA Guidelines for Psychological Assessment and Evaluation.

en. Institution: American Psychological Association. (2020). doi: 10.1037/e5

10142020- 001.

[4] Shu-Chen Cheng and Chiu-Lin Lai. 2020. Facilitating learning for students

with special needs: a review of technology-supported special education

Seg. 3: All technologies are classified as unimportant. These studies. en. Journal of Computers in Education, 7, 2, (June 2020), 131–153.

are . They classify most technologies as doi: 10.1007/s40692- 019- 00150- 8. technology sceptics

[5] Parag Dange et al. 2023. EXPLORING CUSTOMER SEGMENTATION WITH

unimportant, with the exception of electronic communication

K-MEANS CLUSTERING ALGORITHM. en. 11, 4.

and customized hardware. [6] Rabi Das. 2023. ICTs In Special Education: A Review. en. 11, 8.

[7] Louise Demers, Rhoda Weiss-Lambrou, and Bernadette Ska. 1996. Develop-

ment of the Quebec User Evaluation of Satisfaction with assistive Technol-

4.5 Q3: ICT and Study Obligations ogy (QUEST). en. Assistive Technology, 8, 1, (June 1996), 3–13. doi: 10.1080

/10400435.1996.10132268.

The guiding question was: "Please indicate to what extent you

[8] Louise Demers, Rhoda Weiss-Lambrou, and Bernadette Ska. 2000. Item

consider the use of ICT support important to fulfil the study Analysis of the Quebec User Evaluation of Satisfaction with Assistive Tech-

requirements listed below." nology (QUEST). en. , 12, 2, (Dec. 2000), 96–105. doi: Assistive Technology

10.1080/10400435.2000.10132015.

[9] Béatrice P. J. Dijcks, Roelof D. Wessels, Suzanne L. M. De Vlieger, and Marcel

W. M. Post. 2006. KWAZO, a new instrument to assess the quality of service

Seg. delivery in assistive technology provision. en. , Q3a Q3b Q3c Q3d Q3e Q3f Q3g Disability and Rehabilitation

28, 15, (Jan. 2006), 909–914. doi: 10.1080/09638280500301527.

1 4.2 3.6 2.4 2.2 4.4 2.2 5.4 [10] Xiaofei He and Partha Niyogi. 2003. Locality preserving projections. In Pro-

ceedings of the 16th International Conference on Neural Information Processing

2 4.2 4.5 5.0 4.9 5.0 4.9 5.9

Systems (NIPS’03). MIT Press, Whistler, British Columbia, Canada, 153–160.

3 1.0 1.0 1.0 1.0 1.0 1.0 1.0 [11] Nouri Hicham and Sabri Karim. 2022. Analysis of Unsupervised Machine

Learning Techniques for an Efficient Customer Segmentation using Cluster-

Table 3: Aspect Q3: Averages of answers per segment. ing Ensemble and Spectral Clustering. en. International Journal of Advanced

Computer Science and Applications, 13, 10. doi: 10.14569/IJACSA.2022.01310

16.

The number of segments was determined by combining the [12] Louise Norman Jespersen, Susan Ishøy Michelsen, Bjørn Evald Holstein,

dendrogram and the "inertia" curve". We estimate that a reason- Tine Tjørnhøj-Thomsen, and Pernille Due. 2018. Conceptualization, opera-

tionalization, and content validity of the EQOL-questionnaire measuring

able number of segments is again 3.

quality of life and participation for persons with disabilities. en. Health and

The observed characteristics by segment are: , 16, 1, (Dec. 2018), 199. doi: 10.1186/s12955- 018- 10 Quality of Life Outcomes

24- 6.

Seg. 1: The technologies are characterized by different appli- [13] Jeffrey Jutai and Hy Day. 2002. Psychosocial Impact of Assistive Devices

cability. These are Scale (PIADS). en. , 14, 3, (Sept. 2002), 107–111. Gert Critical Users Technology and Disability

, which are Segment 1 of a

Jan Gelderblom and Luc P. De Witte, editors. doi: 10.3233/TAD- 2002- 14305.

two-segment segmentation. All technologies are classified as

[14] Riyo Hayat Khan, Dibyo Fabian Dofadar, and Md. Golam Rabiul Alam. 2021.

useful, with the exception of ICT to support independent work, Explainable Customer Segmentation Using K-means Clustering. en. In 2021

to support group work and to support examination require- IEEE 12th Annual Ubiquitous Computing, Electronics & Mobile Communication

Conference (UEMCON). IEEE, New York, NY, USA, (Dec. 2021), 0639–0643.

ments.

isbn: 978-1-66540-690-1. doi: 10.1109/UEMCON53757.2021.9666609.

Seg. 2: In this segment, all technologies are rated as very useful. [15] Dr Chitralekha Navneet Kumar, Dr Ramchandra Vasant Mahadik, Dr Sangeeta This is , which is virtually identical

Technology Enthusiasts Paliwal, Dr Pallavi, and Chaitali B Kasar. 2024. Leveraging Deep Learning

for Customer Segmentation: Patterns and Preferences Unveiled. en. Interna-

to segment 2 of the two-segment segmentation. . tional Journal of Intelligent Systems and Applications in Engineering

Seg. 3: Here, the majority of respondents consider the technolo- [16] Theodoros A. Kyriazos and Anastasios Stalikas. 2018. Applied Psychometrics:

The Steps of Scale Development and Standardization Process. en. , Psychology

gies to be of little use. These are technology . They describe all

09, 11, 2531–2560. doi: 10.4236/psych.2018.911145.

technologies as not very useful, with the exception of support [17] 2018. . en. OCLC: Learning and teaching in the European higher education area

1142675866. European University Association asbl, Brussels ; Geneva. isbn:

for direct distance learning.

978-90-78997-65-8.

[18] Sukru Ozan and Leonardo O. Iheme. 2019. Artificial Neural Networks in

5 CONCLUSIONS AND DISCUSSIONS Customer Segmentation. en. In 2019 27th Signal Processing and Communi-

cations Applications Conference (SIU). IEEE, Sivas, Turkey, (Apr. 2019), 1–4.

The student response rate was relatively low (5.8%, total 𝑁 isbn: 978-1-72811-904-5. doi: 10.1109/SIU.2019.8806558. =

[19] Rakeshkumar Umanath Upadhyay and Prof. Nilesh Choudhary. 2023. A

43). For the first three aspects listed in subsection 3.1, we found

Review of Customer Segmentation Methods. en. International Journal of

meaningful segmentations (for the last two aspects, there was not Advanced Research in Science, Communication and Technology, (Mar. 2023),

enough data to create segments). We did not find any common 684–692. doi: 10.48175/IJARSCT- 8904.

[20] David A. Siegel. 2010. The mystique of numbers: belief in quantitative ap-

segments between the aspects. Obtained segments were expected

proaches to segmentation and persona development. en. In CHI ’10 Extended

and it was confirmed a useful grouping of students is doable. . ACM, Atlanta Georgia Abstracts on Human Factors in Computing Systems

Further investigation of the segments would require at least 10 USA, (Apr. 2010), 4721–4732. isbn: 978-1-60558-930-5. doi: 10.1145/1753846

.1754221.

new responses in the first three aspects. In the near future, we

[21] Hasan Ziafat and Majid Shakeri. 2014. Using Data Mining Techniques in

will study the usage patterns of these students based on carefully Customer Segmentation. en. 4, 9.

designed case studies.

Introducing a Solution:



The Self-Evaluation Tool in Practice*



Isabel Leandro García† Marta Mármol Muñoz

Project Department Innovation Department

INERCIA DIGITAL INERCIA DIGITAL

Huelva, Spain Huelva, Spain

isabelleandro@inerciadigital.com martamarmol@inerciadigital.com



Abstract / Povzetek



Permission to make digital or hard copies of part or all of this work for Self-Evaluation Tool has to be used as a practical guide

personal or classroom use is granted without fee provided that copies are not for the management to define the good and not-so-good



made or distributed for profit or commercial advantage and that copies bear things about digital education and communication about this notice and the full citation on the first page. Copyrights for third-party inclusion. On the other hand, a Self-Evaluation Tool for HEI components of this work must be honored. For all other uses, contact the

owner/author(s). teachers on inclusive digital teaching practice is used as a

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia practical guide for the teachers to define the good and not-

© 2024 Copyright held by the owner/author(s).

http://doi.org/DOI_RECEIVED_AFTER_REVIEW so-good things about their digital teaching practice

concerning inclusion. (Bennett, 2021)

Digitalization of education in Higher Education Institutions



can have many positive impacts on students and their The article focuses on the introduction of the Self-evaluation

learning success still, at the same time, it can also pose a Tool developed under the Erasmus+ project

significant barrier for some students, especially students “SET4Inclusion”. The SET4Inclusion project is a European

with different disabilities or special needs. Digital inclusion collaborative initiative designed to tackle these challenges by

is crucial in providing equitable and just access to developing and implementing a Self-evaluation Tool tailored

educational materials for every student, irrespective of their for HEIs. [3]. origins, capacities, or impairments. The structured

pedagogical content is supplied to the user through a

recommendation model that constitutes an online Self-

Evaluation Tool to determine the inclusiveness of their

digital education. [1]. 1





Keywords / Ključne besede

Digital Education ,Inclusiveness, equality, disabilities.



1 Introduction

In modern education, continuous improvement is key to

both student and organizational success. One effective

method to foster this growth is through self-evaluation. The

Self-Evaluation Tool is a practical solution designed to

empower institutions to assess digital inclusion, set goals,

and track progress over time.



Based on that, the theoretical framework of inclusive Figure 1:The tool on the website set4inclusion.eu

digital education including all 3 elements, technology,

pedagogy, content, and management aspects, considering

different e-learning settings and modes is the base

foundation to develop a Self-Evaluation Tool for HEI on

inclusive digital education in HEI. [2].

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia I. Leandro et al.





Development section contains five management-related

questions that support us in acquiring the knowledge and

pedagogical skills necessary for inclusive education.

From the Teacher's perspective, the Pedagogy, Supports, and

Resources section contains 7 questions about how it is

developed to provide inclusive digital resources to support

their inclusive teaching.¡



Pedagogy, Implementation in the classroom, consists of

seven questions that cover Digital technologies in the

classroom to facilitate the organization of teaching groups in

which students can collaborate by valuing the diversity of

each one. The student's perspective covers thirteen

Figure 2 :Self evaluation tool questions with multiple questions that are related to the proper use of the technology

choice answers. as an inclusion tool.





2 How does it works

The self-evaluation tool serves as a practical guide for the

management and teachers to define the good and not-so-

good things about their digital education and communication

concerning inclusion.



Our Self-Evaluation Tool is a practical guide for HEI

management to help them assess the capabilities of the

HEI regarding inclusive digital education. It consists of

many questions, representing influence factors, described

through specific indicators. This self-evaluation tool is also a Figure 3: Percentage of answers by categories.

practical guide for HEI teachers to help them assess the

capabilities of their teaching practice about inclusive digital Finally, the tool offers the results with a percentage of

education. It consists of many questions, representing answers by categories. Once you have done the self

influence factors, described through specific indicators. evaluation, the E-platform where the tool is located,

automatically creates a smart report for the users of Self-

Structured Reflection for users are prompted to review their Evaluation Tools with links to recommended e-learning

access to the digital inclusion practices, focusing on relevant content according to the result of the self-evaluation. You can

performance indicators: find the tool in multiple languages.



Twelve questions regarding Leadership/ School's

perspective. These questions are designed to evaluate the

support provided to teachers in digital literacy for inclusive 3 Conclusion

digital education. The responses and subsequent questions

are categorized into different variables, which will be used to In conclusion,the Self-Evaluation Tool is a powerful solution

determine the final results of the report. The categories for for those seeking to improve their performance and grow in

responses are: Not applicable, Totally agree, Agree, their careers. By providing a structured way to reflect, set

Somewhat agree, Disagree, and Totally disagree. goals, and track progress, this tool empowers individuals to

take charge of their development and contributes to the

The following section is Collaboration and Networking. It overall success of the organization. The pilot study of the

consists of eight questions that support the progress in Self-Evaluation Tool developed under the SET4Inclusion

inclusive teaching and learning with digital technologies. project has yielded significant insights into the tool itself, and

The third section involved Infrastructure and Equipment. It its effectiveness for the development of Digital Inclusion in

is developed with twelve questions that cover the digital HEIs.



infrastructure that supports inclusive and accessible 4. References / Literatura education with digital technologies.



Next, a section called Tools and Technology, and aims to [1] UNESCO, Global Education Monitoring Report 2020: Inclusion and

education: All means all. Paris: UNESCO, 2020. [Online]. Available:

investigate technology itself. Continuing Professional https://unesdoc.unesco.org/ark:/48223/pf0000373718



Introducing a Solution:

The Self-Evaluation Tool in Practice Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



[2] P. R. dos Santos, D. N. Barbosa, E. G. A. Neto, J. L. Barbosa, S. D.Correia,

and V. R. Leithardt, “Learning and well-being in educational practices

with children and adolescents undergoing cancer

treatment,”Education Sciences, vol. 11, no. 8, p. 442, 2021.

[3] Bennett, S. (2021). Self-Evaluation Tools in higher education: Bridging

the gap between assessment and improvement. Assessment &

Evaluation in Higher Education, 46(5), 723-

738. https://doi.org/10.1080/02602938.2020.1853698

ANALYSIS OF THE SELF-EVALUATION TOOL: A PILOT



STUDY*



Tuncer Can † Irfan Simsek Hüseyin Göksu

Department of Foreign Language Department of Computer Open and Distance Education

Education, Hasan Ali Yücel Education and Instructional Application and Research Center

Faculty of Education/ Istanbul Technology, Hasan Ali Yücel Istanbul University-Cerrahpasa

University-Cerrahpasa, Istanbul Faculty of Education/ Istanbul Istanbul Türkiye

Türkiye University-Cerrahpasa, Istanbul huseyin.goksu@iuc.edu.tr

tcan@iuc.edu.tr Türkiye

irfan@iuc.edu.tr



ABSTRACT Modern higher education systems are increasingly focused on acknowledging student diversity and creating educational This article focuses on the analysis of the self-evaluation tool environments that effectively respond to this diversity. developed under the SET4Inclusion project, with a particular The SET4Inclusion project is a European collaborative initiative emphasis on its application within the context of a pilot study. designed to tackle these challenges by developing and Pilot studies are crucial for testing the validity and reliability of implementing self-evaluation tools tailored for HEIs. These tools new tools or methods. The primary objective of this pilot study are intended to help institutions critically assess their current is to evaluate the effectiveness of the Self-Evaluation Tool (SET) practices, identify gaps, and take concrete steps toward creating developed within the SET4Inclusion project in facilitating e- a more inclusive educational environment. By enabling inclusion processes in higher education institutions (HEIs). The institutions to analyze their performance, self-evaluation tools pilot study began with the preparation of an invitation letter, create an internal cycle of continuous improvement [1]. They which was sent to potential participants. The pilot study involved provide not just a snapshot of the current state but also support participants from higher education institutions in five different the development of forward-looking improvement strategies. European countries: Turkey, Portugal, Slovenia, Italy, and Spain. When viewed through the lens of inclusivity, such tools The number of HEIs participating in the piloting was at least five, significantly enhance the capacity of learning environments to with at least fifty HEI teachers involved. cater to the needs of all students, irrespective of their



KEYWORDS This article focuses on the analysis of the self-evaluation tool backgrounds or abilities.

SET4Inclusion project, Self-Evaluation Tool, Micro Learing developed under the SET4Inclusion project, with a particular

Units emphasis on its application within the context of a pilot study.

Pilot studies are crucial for testing the validity and reliability of

new tools or methods. They serve as a foundational step before

1 INTRODUCTION wider implementation and can provide valuable insights into the

The rapidly evolving dynamics within higher education have practical challenges and opportunities associated with new

made inclusivity and accessibility more significant than ever initiatives [3]. In this study, key elements such as the

before. The increasing number of students with diverse learning functionality of the self-evaluation tool and the developed micro-

needs, including those with disabilities, necessitates a learning units in diverse contexts, areas requiring improvement,

comprehensive approach to ensure equitable access to education. and user feedback will be thoroughly evaluated. The results of

The rise of digital tools and technologies offers novel the pilot study have facilitated necessary adjustments to the tool

opportunities to foster e-inclusivity, particularly within Higher and units before their broader application in various HEIs across

Education Institutions (HEIs). This process not only contributes Europe.



to enhancing student success rates but also plays a vital role in

promoting fairness and justice in education. Inclusivity is closely

linked to student satisfaction and academic achievement, making

it a strategic priority for HEIs [2].



∗Article Title Footnote needs to be captured as Title Note

†Author Footnote to be captured as Author Note



Permission to make digital or hard copies of part or all of this work for personal or

classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s).

Information Society 2022, 10–14 October 2022, Ljubljana, Slovenia Figure 1: Home page of Self-Evaluation Tool © 2022 Copyright held by the owner/author(s).

identifying areas for improvement, and implementing strategies



to enhance inclusivity. The Micro Learning Units complemented

the tool by providing focused, easily accessible learning content

that could be used to support inclusive practices within the

institutions. Participants used these tools to evaluate and reflect

on the e-inclusion processes in their respective institutions.



2.3 Data Collection

The key instruments used in this research were the Self-

Evaluation Tool and the Micro Learning Units developed within

the SET4Inclusion project. The Self-Evaluation Tool was

designed to assist HEIs in assessing their e-inclusion processes,

identifying areas for improvement, and implementing strategies

to enhance inclusivity. The Micro Learning Units complemented

Figure 2: Home page of e- learning materials the tool by providing focused, easily accessible learning content

that could be used to support inclusive practices within the

institutions. Participants used these tools to evaluate and reflect

2 METHODOLOGY on the e-inclusion processes in their respective institutions.

The primary objective of this pilot study is to evaluate the

effectiveness of the Self-Evaluation Tool (SET) developed 2.4 Data Analysis

within the SET4Inclusion project in facilitating e-inclusion The data collected from the surveys were systematically

processes in higher education institutions (HEIs). The pilot study analyzed using graphical representation methods to visualize the

began with the preparation of an invitation letter, which was sent feedback and results. These graphs provided clear insights into

to potential participants. This letter provided detailed the effectiveness of the Self-Evaluation Tool and the micro-

information about the project, the scope of the pilot study, and learning units. The feedback from participants was thoroughly

the tools that participants would be evaluating. Links to the tools reviewed, with particular attention paid to their suggestions for

and a survey were included in the invitation, allowing improvements and their overall experience with the tool. The

participants to engage directly with the materials. analysis revealed several key areas where the tool could be

The tools developed for the project, including the Self- refined and optimized to better meet the needs of HEIs and their

Evaluation Tool and the associated micro-learning units, were diverse student populations.

distributed to participants via email. Participants were given a The pilot study's findings provided valuable information that

two-week period to complete their evaluations. At the end of this will inform the further development and dissemination of the

period, they were asked to complete a survey that provided Self-Evaluation Tool. The insights gained from this study will

feedback on their experiences with the micro-learning units and help ensure that the tool is effective, user-friendly, and capable

the self-evaluation process. of supporting HEIs in their efforts to create more inclusive

learning environments. As the tool is refined and prepared for

2.1 Participants broader application, these findings will serve as a crucial

The pilot study involved participants from higher education foundation for its successful implementation.

institutions in five different European countries: Turkey,

Portugal, Slovenia, Italy, and Spain. The number of HEIs In conclusion, the pilot study of the Self-Evaluation Tool

participating in the piloting was at least five, with at least fifty developed under the SET4Inclusion project has yielded

HEI teachers involved. Participants were selected based on significant insights into the tool's effectiveness and the user

voluntary participation, and efforts were made to ensure diversity experiences associated with it. The feedback gathered from

by including individuals from various departments and academic participants has highlighted both the strengths and areas for

levels. This approach aimed to gather a wide range of improvement, which will be addressed in the subsequent

perspectives and insights, thereby enriching the study's findings. development phases. By incorporating these insights, the tool

will be better equipped to support HEIs in fostering inclusivity

and accessibility in their educational offerings. The study also

Table 1: Number of participants underscores the importance of pilot testing in the development of

educational tools, as it provides a critical opportunity to refine

No Participants f(x) and enhance the tools before they are implemented on a larger

1 HEI participating 5 scale.

2 HEI teachers 50

REFERENCES

2.2 Instruments [1] Bennett, S. (2021). Self-evaluation tools in higher education: Bridging the

The key instruments used in this research were the Self- gap between assessment and improvement. Assessment & Evaluation in

Evaluation Tool and the Micro Learning Units developed within Higher Education, 46(5), 723-738.

https://doi.org/10.1080/02602938.2020.1853698

the SET4Inclusion project. The Self-Evaluation Tool was [2] Ehlers, U. D. (2020). Quality in e-learning from a learner's perspective.

designed to assist HEIs in assessing their e-inclusion processes, European Journal of Open, Distance and E-Learning, 23(1), 10-28.

https://doi.org/10.2478/eurodl-2020-0002



[3] Van der Kleij, F. M., Adie, L., & Cumming, J. J. (2019). The role of Review of Educational Research, 89(4), 635-676.

feedback in self-assessment: A systematic review of educational research. https://doi.org/10.3102/0034654



Definition of a Framework for Self-Evaluation Tool:



Optimizing Evaluation Practices for Enhanced



Performance in HEI



Maja Puš nik† Boštjan Šumak Katja Kous Florian Gallo Enrico Dolza

Faculty of Faculty of Faculty of Fondazione Ištituto Fondazione Ištituto

Electrical Electrical Electrical dei Sordi di Torino dei Sordi di Torino

Engineering and Engineering and Engineering and ONLUS ONLUS

Computer Science Computer Science Computer Science Torino, Italy Torino, Italy

Univeršity of Univeršity of Univeršity of erašmuš@ištitutoš direzione@ištitutoš

Maribor, Slovenia Maribor, Slovenia Maribor, Slovenia orditorino.org orditorino.org

maja.pušnik@um.ši boštjan.šumak@um.ši katja.kouš@um.ši



Abstract accelerated digitalization of teaching and learning in higher



Thiš paper propošeš a framework for a šelf-evaluation tool to SEND študentš. Recognizing that ševeral študentš štruggle education, which haš dišproportionately dišadvantaged

šupport the evaluation of inclušive digital education in with online education, it iš fundamental to promote



Special Educational Needš and Dišabilitieš (SEND). The higher education), helping them reach their full potential. We enhanced digitališation of education šhowed that many believe thiš endeavor would be šupported by a šelf- študentš and profeššionalš štruggled to engage with online evaluation tool, providing špecific feedback to HEIš and content due to rigid teaching methodš and inacceššible helping increaše inclušive digital education capabilitieš in rešourceš. By focušing on inclušivity, the propošed formal and non-formal šettingš [1]. framework aimš to addrešš theše challengeš by tackling Within the šcope of Erašmuš+ SET4Inclušion motivated by the COVID-19 pandemic, which emphašišed inclušive digital environment that šupportš študentš from inequalitieš in digital learning, particularly for študentš with the earliešt štageš of education (from primary šchool to Higher Education Inštitutionš (HEIš). The idea waš inclušivenešš aš a key šolution. HEIš aim to create an

digital technologieš to create adaptive, equitable learning project (2023-1-SI01-KA220-HED-000160853) the partnerš



šupport HEIš in evaluating and enhancing their digital education to addrešš the HEIš challengeš and prepare for a education practiceš, enšuring all študentš can fully šelf-evaluation tool. Thiš framework waš built upon the well- environmentš for diverše študent needš. The framework will developed a theoretical framework for inclušive digital

participate and šucceed in education. eštablišhed TPACK model (Technological Pedagogical



Keywords Content Knowledge), which guideš the integration of ICT in

education. TPACK outlineš the eššential knowledge areaš for

Inclušive Digital Education, Self-Evaluation Tool, Higher effectively integrating technology into teaching, šuch aš

Education Inštitutionš, Special Educational Needš and technological knowledge (focušing on acceššibility),

Dišabilitieš pedagogical knowledge (emphašizing Univeršal Dešign for

Learning - UDL), and content knowledge (innovative



1. applicationš of new technologieš to overcome barrierš for Introduction

SEND študentš). The TPACK framework alšo incorporateš

In the rapidly evolving higher education, integrating the latešt štandardš, šuch aš the Web Content Acceššibility

Information and Communication Technology (ICT) haš Guidelineš (WCAG 2.2), and inšightš from contemporary

become eššential for enhancing teaching and learning rešearch. Additionally, we aimed to include the SELFIE tool

proceššeš. However, enšuring that digital education iš for šelf-reflection on šchoolš' digital capabilitieš [2], an

inclušive—particularly for SEND študentš—remainš a evaluation šurvey promoted by the European Commiššion,

šignificant challenge [4]. Thiš paper addreššeš the challengeš by integrating a štronger focuš on inclušivity. Our framework

expošed by the COVID-19 pandemic, particularly the and developed šelf-evaluation tool aim to identify and

∗ dešcribe the key factorš and indicatorš that define inclušive

Article Title Footnote needs to be captured as Title Note

† Author Footnote to be captured as Author Note digital education, providing HEIš with an inclušive tool to

Permission to make digital or hard copies of part or all of this work for personal evaluate and enhance their digital education practiceš. The

or classroom use is granted without fee provided that copies are not made or development of the framework involved ševeral key taškš:



distributed for profit or commercial advantage and that copies bear this notice • A literature review, which included 131 paperš and and the full citation on the first page. Copyrights for third-party components of explored exišting rešearch, bešt practiceš, and the this work must be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia challengeš of implementing inclušive digital

© 2024 Copyright held by the owner/author(s). education (due to the length it iš not included in thiš

httpš://doi.org/10.70314/iš.2024.DIGIN.12

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia M. Pusnik et al.



paper, however, it iš available aš one of the The literature review provided an idea of levels of

deliverableš of SET4Inclušion project. inclusiveness within digital education, which are divided into

• Development of two šurveyš — one targeting HEI eight categories, further defined by factors and evaluated

štaff and another focušed on SEND študentš— to with indicators on a Likert scale from strongly agree to

gather inšightš on their experienceš and needš. The strongly disagree, presented in the following paragraphs [3].

šurveyš were performed with the help of Google The categories are the following: (1) Leadership/ School's

Formš and the reportš will be available aš one of the perspective, (2) Collaboration and Networking, (3)

deliverableš of the SET4Inclušion project. Infrastructure and Equipment/ Tools and Technology, (4)

• The co-development workšhopš involving Continuous Professional Development, (5) Pedagogy:

collaboration with HEIš partnerš, rešearcherš, and Supports and Resources, (6) Pedagogy: Implementation in

expertš. The report iš available aš one of the the classroom, (7) Assessment Practices/ Inclusion

deliverableš of the SET4Inclušion project. Assessment and (8) Student Digital Competence/ Student's

Perspective.

The rešultš from the lišted activitieš prešent a foundation for

the final framework that will enable HEIš to create and Category Leadership/ School's perspective includes the

maintain inclušive digital learning environmentš, enšuring following factors:

accešš to education for all študentš. It iš alšo the foundation • (1.1.) Inclusive digital strategy education strategy

for developing šelf-evaluation toolš and other project rešultš, • (1.2.) Collaborative digital strategy development

providing a štructured bašiš for inštitutionš to aššešš and • (1.3.) Contemporary pedagogical approaches

enhance their practiceš. The main rešultš, for which • (1.4.) Scheduled time to explore digital teaching

dešcribed framework iš neceššary, are the following: • (1.5.) Efforts to minimize discrimination,



• • (1.6.) Collaboration and communication Two Self-evaluation toolš for HE inštitutionš and HE teacherš to determine the inclušivenešš of their encouragement between school and teacher, • (1.7.) Inclusion policies,



• • (1.8.) Digital literacy. A šet of micro-learning unitš on Inclušive digital digital education,

education for HE leaderš and HE teacherš,

• The examples of indicators for each factor are Collection of ševeral good practiceš on Inclušive

digital education, transformed into statements in the self-evaluation tool,

• connected to the leaderšhip/šchool’š peršpective, and are E-learning platform aš a central knowledge point

for Inclušive digital education. šhown in Figure 1. For example, “Inclusive digital strategy

would” iš evaluated bašed on an indicator: “At our HEI we

have an inclusive digital education strategy”. The šame

2. Definition of the framework principle is applied to all other factors in other categories.



To make digital education and communication in HEIš



more acceššible and inclušive for all other študentš, a

theoretical framework of inclušive digital education waš

defined, conšidering different poššible e-learning šettingš

and modeš (ICT-šupported claššroom learning, dištance e-

learning, hybrid, blended learning, šelf-paced, guided etc.). A

theoretical framework for inclušive digital education waš

created bašed on a literature review, šurvey and workšhopš,

prešented in the following šectionš.



2.1. The literature review



Firšt, a literature review waš conducted to (1) identify

good practiceš for inclušive digital education, (2) identify the

challengeš of inclušive digital education, and (3) identify the

factorš that affect inclušive digital education. The aim waš to

identify the relevant literature in journal paperš, articleš

publišhed in conferenceš, reportš, and other relevant

šourceš. The šearch waš conducted in different databašeš

and šearch engineš, šuch aš Scientific databašeš: WoS,

ScienceDirect, IEEExplore, ACM, Google Scholar, Search

engineš: Google. The included literature waš in Englišh, dated

from 2017 or newer. One hundred thirty-one literature unitš

were identified and recorded. Bašed on the rešultš from the

literature review, factorš that impact inclušive digital Figure 1: Category Leadership/ School's perspective

education were defined and later validated with the help of a

šurvey.



Definition of a Framework for Self-Evaluation Tool Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



Category Collaboration and Networking includes the • (6.5.) Everyone is made to feel welcome and

following factors: treated with respect,

• • (2.1.) Progress review, (6.6.) The partnership between staff and

• parents/careers, (2.2.) Discussion on the use of technology,

• • (6.7.) Students are equally valued, (2.3.) Collaboration of HEI, local communities,

caregivers and parents, • (6.8.) Staff and students treat one another as

• human beings, and discipline is based on mutual (2.4.) Synergies for Blended Learning,

• respect, (2.5.) Staff, governors, students and

parents/careers share a philosophy of inclusion, • (6.9.) The school arranges teaching groups so that

• all students are valued, and differences are (2.6.) Teachers plan, teach and review in

partnership and understood,

• • (6.10.) Training and education on inclusiveness. (2.7.) Encouraged collaboration and

communication between students and teachers.

Category Assessment Practices/ Inclusion Assessment

Category Infrastructure and Equipment/ Tools includes the following factors: and

Technology includes the following factors: • (7.1.) Assessing skills,

• • (7.2.) Digital assessment, (3.1.) Accessible infrastructure and tools,

• • (7.3.) Timely feedback, (3.2.) Digital devices and assistive products for

teaching, • (7.4.) Self-reflection on learning,

• • (3.3.) Internet Access, (7.5.) Feedback to other students,

• • (7.6.) Using data to improve learning, (3.4.) Technical Support,

• • (3.5.) Available digital devices and assistive (7.7.) Evaluation metrics.

products for learning,

• (3.6.) Devices and assistive products for students, Category Student Digital Competence/ Student's

• (3.7.) Measures to identify the digital divide, Perspective includes the following factors:

• (3.8.) Support to address the digital divide, • (8.1.) Learning to communicate and

• (3.9.) Bring your device and assistive products, • (8.2.) Digital skills across subjects. • (3.10.) Reduced physical barriers,



• • (3.11.) Assistive products, 2.2. The survey

(3.12.) Online libraries,

• (3.13) Fairly distributed resources. Two šurveyš were conducted. The firšt šurvey waš

conducted to acquire empirical evidence about teacherš’

Category Continuous Professional Development includes perceptionš of inclušive digital education in their claššroomš.

the following factors: The šurvey waš conštructed bašed on the exišting literature,

• (4.1.) Discussion of CPD needs, where we identified the mošt šignificant factorš that can

• (4.2.) Participation in CPD activities, affect inclušive digital education (šuch aš Inclušive Digital

• (4.3) Face-to-face or online sharing experiences Strategy and Policy for Empowering Inclušive Digital

between staff. Education, Inclušive Digital Pedagogy & Supportive Culture

and šimilar). Over 100 šurvey rešponšeš pošitively rated the

Category Pedagogy: Supports and Resources includes the propošed framework, with moderate variability. The

following factors: rešponšeš šuggešt that while many rešpondentš rated the

• (5.1.) Online educational resources, indicatorš highly, there were diverše opinionš among the

• (5.2.) Creating digital resources, participantš, particularly in areaš šuch aš Infraštructure and

• (5.3.) Using virtual learning environments, Technology (IET) and Pedagogy: Implementation in the

• (5.4.) Communicating with the school community, Claššroom (PIC), where variability waš šlightly higher. Theše

• (5.5.) Open educational resources, findingš highlight areaš of štrength aš perceived by

• (5.6.) Staff development activities help staff to rešpondentš and potential areaš for further improvement

respond to student diversity, and targeted interventionš. The pošitive feedback on



• teaching and learning and competence underšcoreš the importance of continuing, (5.8.) Staff develop resources to support learning šupporting and enhancing theše ašpectš within educational • (5.7.) Student difference is used as a resource for profeššional development, collaboration, and digital

and participation.

inštitutionš. The šecond šurvey waš conducted to acquire



includes the following factors: Overall, the šurvey rešultš highlight štrong šatišfaction • (6.1.) Personalization according to students' with variouš dimenšionš of educational practice, particularly needs, in leaderšhip, collaboration, infraštructure, pedagogy, Category Pedagogy: Implementation in the classroom digital education. empirical evidence about študentš’ perceptionš of inclušive



• (6.2.) Foštering študentš’ creativity, aššeššment practiceš, and digital competence. The • (6.3.) Engaging and motivating students, conšištently high median šcoreš and excellent internal • (6.4.) Student collaboration, conšištency acrošš indicatorš šuggešt that rešpondentš

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia M. Pusnik et al.



perceive theše areaš pošitively, reflecting well-implemented 4. Conclusion practiceš and šupportive educational environmentš.

Thiš paper haš eštablišhed a comprehenšive framework



2.3. for a šelf-evaluation tool to optimize evaluation practiceš The co-creation workshops

within HEIš, mainly promoting inclušive digital education.



to refine and validate the framework for inclušive digital The framework co-creation workšhopš were dešigned became evident during the rapid digitalization of education, The framework iš dešigned to addrešš the dišparitieš that

ešpecially thoše affecting študentš with Special Educational



indicatorš. During the ten conducted workšhopš in variouš Needš and Dišabilitieš. By equipping HEIš with thiš education, focušing on finalizing the key factorš and

European countrieš, participantš engaged in collaborative evaluative tool, the framework šeekš to create more

dišcuššionš to šhape the framework'š layout, enšuring it equitable and adaptive learning environmentš that can

accurately reflectš the needš and challengeš of inclušive better accommodate the diverše needš of all študentš.

digital education. The workšhop'š outcomeš included the The future štageš of thiš work will involve the šyštematic

creation of validated šelf-evaluation queštionnaireš and collection and claššification of at leašt 15 exemplary

refining indicatorš that would later be ušed for the practiceš from five partner countrieš of the project in the



online workšhopš played a critical role in fine-tuning the categorized according to the framework’š componentš— variableš and indicatorš that form the foundation of the šelf- šelf-evaluation toolš and e-learning materialš—and will be development of the SET and micro-learning unitš. Theše domain of inclušive digital education. Theše practiceš will be



for HEIš. methodically documented ušing textual and multimedia evaluation tool, enšuring they are both relevant and effective

formatš. Thiš iterative procešš will refine the framework and

extend itš relevance and effectivenešš acrošš variouš

3. Self-evaluation tool based on the educational šettingš. Applying thiš framework iš expected to

framework lead to improved learning outcomeš, promoting greater

inclušivity and equity in higher education.

A self-evaluation tool [5] was created based on the

developed framework, which encompasses key factors and



indicators of inclusive digital education. This tool is designed Acknowledgments to serve as a practical guide for management and educators

within HEIs, enabling them to assess and identify strengths



and areas for improvement in their digital education and SET4Includion (Self-Evaluation Toolš for e-Inclušion in HEI) Thiš rešearch waš conducted within the activitieš of the

communication practices related to inclusion. In addition to project, co-funded by the Erašmuš+ Programme of the

its evaluative function, integrated with an E-platform, the European Union. The authorš (Boš tjan S umak, Maja Puš nik,

self-evaluation tool aims to raise awareness about the Katja Kouš) acknowledge the financial šupport from the

importance of inclusive digital education among Slovenian Rešearch Agency (Rešearch Core Funding No. P2-

stakeholders. The implementation of these practical tools, 0057).

along with the provision of free learning opportunities

through the E-platform, is expected to contribute References



significantly to fostering more inclusive digital education [1] Pušnik, M., Kouš, K., Welzer Družovec, T., & Šumak, B. (2024). Identification and Analysis of Factors Impacting e-Inclusion in Higher environments within HEIs. Education. 308 – 317. https://doi.org/10.3233/faia231164

The results of the self-evaluation tool are presented in [2] P. Kampyliš and A. Sala, “Improving the digital capacity of šchoolš by

(Figure 2). ušing the “SELFIE” tool for collective reflection,” Eur. J. Educ., vol. 58, For each category, šuggeštionš on how to improve

no. 2, pp. 331–346, Jun. 2023.

inclušion in HEI are provided, which are connected to micro- [3] C. Sahli Lozano, S. Wüthrich, J. S. Büchi, and U. Sharma, “The concernš



learning unitš, one of the additional outputš of the project about inclusive education scale: Dimensionality, factor structure, and . development of a short-form version (CIES- SF),” Int. J. Educ. Reš., vol.

111, p. 101913, 2022.



[4] N. Zahid, A. Jamil, and I. Nawaz, “Behavioral problemš and academics

of children in inclusive education – A cross-šectional šurvey,” Heliyon,

vol. 9, no. 2, p. e13496, Feb. 2023.

[5] Self-evaluation tool: https://set4inclusion.eu/self-evaluation-tools/



Figure 2. Final results of the self-evaluation tool, based

on the framework





Good practices in creating an inclusive environment in




Education Institution



Darja Ivanuša Kline José Carlos Neves Carla Sousa

INUK inštitut CICANT CICANT

Maribor, Slovenia Lusófona University Lusófona University

darja.ivanusa.kline@inuk.si Lisbon, Portugal Lisbon, Portugal

p1568@ulusofona.pt carla.patricia.sousa@ulusofona.pt



Alfonso Infante Moro Neves Silvia Doratiotto Irfan Simsek

Economía Financiera, International Department Hasan Ali Yücel Faculty of

Contabilidad y Dirección de Istituto dei Sordi di Torino Education

Operaciones Turin, Italy İstanbul University - Cerrahpaşa

University of Huelva cooperation@istitutosorditorino.o Istanbul, Turkey

Huelva, Spain rg irfan@iuc.edu.tr

alfonso.infante@uhu.es



Boštjan Šumak Maja Pušnik

Faculty of Electrical Engineering Faculty of Electrical Engineering

and Computer Science, University and Computer Science, University

of Maribor of Maribor

Maribor, Slovenia Maribor, Slovenia

bostjan.sumak@um.si maja.pusnik@um.si



Abstract / Povzetek

1 Introduction

The shift toward digital education has created opportunities and

challenges in ensuring inclusion for all students, especially for The digitization of education offers transformative potential to

those with disabilities or special needs. To promote inclusivity in enhance student learning outcomes by providing flexible,

digital education within higher education institutions (HEIs), the personalized learning opportunities. However, it also presents

ERASMUS+ project SET4Inclusion initiated a Call for Good significant challenges, particularly for students with disabilities

Practices, inviting educators and institutions to share effective or special needs who may face barriers if their diverse learning

practices across various areas, such as leadership, infrastructure, requirements are not adequately addressed. As higher education

and pedagogy. Out of 15 collected and evaluated good practices, institutions (HEIs) cater to an increasingly diverse student

five exemplary practices were selected for their impact on population, ensuring that digital education is inclusive and

fostering inclusive learning environments. This paper presents accessible to all students has become a critical priority.

these five practices, offering insights and recommendations for To address this need, the ERASMUS+ project SET4Inclusion

HEIs to enhance digital inclusivity in their educational offerings.

was established to enhance inclusive digital education in higher



Keywords / Ključne besede education. The project aims to foster environments where all

learners feel valued, regardless of their abilities or needs. An

Inclusion, digital education, inclusive pedagogy important initiative of the project was the Call for Good Practices,

which invited educators and educational institutions to share

Permission to make digital or hard copies of part or all of this work for personal or examples of their inclusive digital practices. The call focused on

classroom use is granted without fee provided that copies are not made or distributed various areas, such as leadership, collaboration, infrastructure,

for profit or commercial advantage and that copies bear this notice and the full professional development, pedagogy, and assessment,

citation on the first page. Copyrights for third-party components of this work must encouraging contributors to highlight how they prioritize

be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia inclusion and adapt to diverse student needs. © 2024 Copyright held by the owner/author(s).

http://doi.org/DOI_RECEIVED_AFTER_REVIEW



2 This paper presents the best five practices identified through Good Practices

this evaluation process, offering insights and practical guidance

To collect good practices on inclusive digital education, several

for other institutions seeking to enhance their inclusive digital

methodological steps were employed. A call for good practices

education initiatives. Brief overview of these exemplary

was developed and widely disseminated by all project partners,

practices illustrates their potential to enhance an inclusive

who translated and promoted the call to maximize reach across

educational environment in higher education.

different regions.



A standardized template was developed to guide submissions, 2.1 “Inclusive Glossary of Mathematical Terms:

ensuring consistency in how good practices were described,

including details such as context, methodologies and lessons A Hybrid Pedagogical Tool for the Inclusive

learned. A total of 15 submissions were received from four Education of Deaf and Hearing Children”

countries, providing a diverse array of practices related to by José Carlos Neves, and Carla Sousa

inclusive digital education. (Portugal)

An online workshop was conducted with all project partners The Inclusive Glossary of Mathematical Terms (GIM) is an

to present and discuss each submitted practice. This collaborative educational tool to support deaf children, particularly those using

workshop served as a preliminary evaluation forum where Portuguese Sign Language (LGP). GIM combines physical cards

participants could provide feedback and highlight key aspects of and digital videos to create a memory game that enhances

each practice. mathematical learning through interactive play. It fosters

Following the workshop, a detailed survey was created based inclusivity by allowing both deaf and hearing children to learn

on the project's established factors and indicators for inclusive together. It is accessible online and includes an adaptation of

digital education. This survey facilitated a systematic evaluation Brazilian Sign Language (LIBRAS).

by all partners. GIM is designed for primary and preschool teachers and is

The received practices were also categorized according to the user-friendly, requiring no advanced technical skills. Developed

SET4Inclusion Digital Education Framework, aligning them through collaboration among experts in education, design,

with specific inclusion areas such as leadership, collaboration, animation, and deaf culture, it ensures the content is relevant and

technology, professional development, pedagogy, assessment, accessible. Teachers specializing in deaf and hard-of-hearing

and student competence. Table 1 summarizes the distribution of (DHH) students contributed to the development and testing,

practices across these categories. involving 120 children across three schools. The game has shown

The data in the table indicate a broad distribution of practices positive impacts on student engagement and learning outcomes.

across various inclusion areas, with a notable emphasis on The game integrates physical cards with animated videos

pedagogical supports and resources, infrastructure and to illustrate mathematical concepts. These cards can be used with

technology, and student digital competence. This distribution a mobile app that displays corresponding videos in LGP or

suggests a balanced approach to inclusive digital education, LIBRAS, making it adaptable to different classroom settings.

addressing both the technical and pedagogical aspects needed to The game’s components can be produced at low cost through

create equitable learning environments. laser cutting or 3D printing or printed on paper for traditional use.

Research supports GIM’s effectiveness in teaching

Table 1: Covered inclusion areas by received practices mathematics and sign language to young children, particularly

DHH students. However, further studies are needed to explore its

application in various educational contexts and assess its long-

Inclusion areas No. of practices

term impact on learning.

Leadership/School's perspective 60,0%



Collaboration and Networking 2.2 “The Inclusive working group at Faculty of 33,3% Business UHU” by Alfonso Infante Moro Infrastructure and Equipment/Tools and (Spain) Technology 66,7%



Continuous Professional Development Accessibility for individuals with specific needs in digital 60,0% education is a fundamental obligation for universities. Alfonso Pedagogy: Supports and Resources 86,7% Infante Moro, who leads the "special needs technical unit" at the

Pedagogy: Implementation in the University of Huelva's Faculty of Business and Tourism, has

classroom 53,3% been instrumental in promoting this principle. Since its

Assessment Practices/Inclusion presentation at the 2021 International Congress of University and

Assessment Disability, the model continues to play a crucial role in ensuring 26,7%

Student Digital Competence/Student's inclusive education at the university level. Under Infante's

leadership, the faculty advocates for all students' right to access

Perspective 66,7%

higher education and emphasizes the necessity of digital

Other areas 6,7% accessibility provisions.

Spanish regulations, such as the Ley General de Derechos

This approach ensured a robust and systematic process for de las Personas con Discapacidad (General Law on the Rights of

identifying, evaluating, and categorizing good practices in Persons with Disabilities, 2013) and the Plan de Acción para las

inclusive digital education, promoting the exchange of effective Personas con Discapacidad (Action Plan for Persons with

strategies across diverse educational contexts. Disabilities, 2014-2020), highlight the importance of inclusivity



in higher education. However, the digital aspects of accessibility Agency, Yetam XR develops 3D virtual technologies tailored for

still require significant attention to ensure that all students can academic and industrial applications. The centre offers

fully participate. The University of Huelva's Faculty of Business educational programs in 3D modelling, visual design, animation,

has addressed these needs by organizing educational technology and Unity programming, ranging from beginner to advanced

conferences focusing on inclusion and diversity. A dedicated levels, equipping participants with theoretical and practical skills

working group led by Alfonso Infante promotes inclusive for the professional use of VR and AR technologies.

practices and classroom accessibility, supporting students across Yetam XR integrates advanced technologies into

various degree levels through tailored adaptations and educational environments, transforming traditional learning with

collaboration with clinical health specialists. immersive and interactive content. The centre prioritizes

Innovative digital tools, such as synchronous subtitles, inclusivity, ensuring its educational tools are accessible and

have been implemented to support students with hearing adaptable to diverse learning needs. Equipped with state-of-the-

impairments, benefiting a broader student population. art VR headsets and 3D modelling tools, Yetam XR provides a

Approximately 0.3% of the faculty’s students require specific high-quality digital learning experience.

educational support, and efforts continue to enhance their digital The centre’s projects include the Virtual Factory

learning experiences. Beyond the University of Huelva, Infante Simulation, hands-on experience with virtual industrial

advocates for a standardized system of inclusive practices across machines, and the Basic First Aid VR Project, which trains

European universities. His work contributes to developing a students in emergency response. Additionally, the Occupational

Standards Guide for the Inclusion of University Students with Health and Safety Project educates workers and students on

Disabilities, aimed at improving access and support for students safety through VR simulations of hazardous situations, providing

in the digital era. valuable real-world experience in a safe environment.

While Yetam XR has successfully enhanced education

2.3 “EcoDigi: A Practice for Sustainable Digital with immersive technologies, challenges remain, such as limited

Transformation in Adult Education” by access to VR equipment for some students. There is also a need

Silvia Doratiotto (Italy) to develop more inclusive content for students with disabilities.

Future plans include expanding educational content, increasing

EcoDigi is an initiative to promote digital transformation in adult

collaborations with schools and industry partners, and refining

education while prioritizing environmental sustainability and

programs to meet evolving needs based on feedback from

inclusivity. The project supports the development of digital

educators and students.

readiness in an eco-friendly and accessible way, addressing the



needs of adult learners, including those with disabilities or 2.5 “INTUX - Introducing training on user

special needs. EcoDigi emerged in response to the growing

reliance on digital platforms during the COVID-19 pandemic, Testing with people with disabilities into UX

recognizing both the opportunities for inclusion and the need to design and related higher education minimize environmental impact. The project helps educators and program” by Boštjan Šumak, Maja Pušnik

institutions adopt more sustainable and inclusive practices by and Katja Kous (Slovenia)

providing self-assessment tools, teaching materials, and guides. The INTUX project, "Introducing training on user Testing with

The initiative targets a broad audience, including adult people with disabilities into UX design and related higher

learners, educators, specialists, and NGOs involved in education education programs", focuses on making UX design education

and sustainability. Through specialized training, EcoDigi will more inclusive by integrating accessibility into user testing. The

create a pool of skilled trainers across partner countries, project addresses a gap in UX design practices, where people

strengthening the capacity of adult education providers to act as with disabilities are often excluded from testing, leading to

leaders in their communities. The project will produce several products that do not fully meet their needs.

vital resources, including a self-assessment tool, a report on best A vital outcome of the project is a specialized training

practices, an online platform, and guidelines for creating green course for UX design students, teaching them how to conduct

and accessible education. user testing with people with disabilities. This course is

EcoDigi also plans to host webinars and organize an supported by a handbook for university staff, guiding the creation

international conference to promote the exchange of good of more inclusive teaching environments. This ensures that

practices across partner countries. The initiative aims to ensure students are prepared to design accessible products and meet the

that digital transformation in adult education is sustainable and challenges they will face as professionals.

inclusive, contributing to a greener and more equitable digital The project also develops a framework to integrate these

society in Europe. training modules into UX design programs, embedding



2.4 “Innovative XR Technologies Research and people with disabilities by educating them about their rights in accessibility into the curriculum. Additionally, it empowers

Development Center (YETAM-XR)” by user testing and encouraging their active participation in the

Irfan Simsek (Turkey) design process.

The Innovative XR Technologies Research and Development For higher education institutions, INTUX enriches

Center (Yetam XR) at Istanbul University-Cerrahpaşa, led by educational programs by incorporating inclusive practices. By

Associate Professor Irfan Simsek, focuses on advancing virtual adopting the training modules and recommendations, institutions

reality (VR) and augmented reality (AR) technologies in ensure that their students graduate with the skills to design

education and industry. Supported by the Istanbul Development



accessible products and services, aligning with growing societal technological solutions with pedagogical innovation and

and professional expectations for inclusion and accessibility. collaboration among educators. As digital education continues to

evolve, the insights and recommendations drawn from these

good practices provide a valuable roadmap for higher education

3 Conclusion institutions aiming to enhance their inclusivity efforts. Future

The ERASMUS+ project SET4Inclusion has emphasized the research should continue to explore the scalability and

critical importance of fostering inclusive digital education adaptability of these practices across different educational

environments within higher education institutions. By contexts, ensuring that all students, regardless of their

systematically collecting and evaluating a diverse range of good backgrounds or abilities, have equitable access to high-quality

practices, the project has illuminated various strategies and tools digital education.

that can effectively address the needs of all learners, particularly

those with disabilities or special needs. The five exemplary good Acknowledgments / Zahvala

practices presented in this paper highlight innovative approaches This paper was prepared based on the results of the ERASMUS+

to creating accessible learning environments, from hybrid SET4Inclusion project: Self-Evaluation Tools for e-Inclusion in

pedagogical tools that integrate digital and physical elements to Higher Education Institutions. Therefore, we would like to

enhance mathematical understanding, to comprehensive express our sincere gratitude to all the authors who contributed

frameworks that address multiple dimensions of inclusivity such their valuable practices in response to our call and to everyone

as leadership, pedagogy, and technology infrastructure. involved in the successful development of the project.

These findings suggest that successful inclusive digital

education requires a holistic approach that combines

Note that there is a section break at the end of references to

balance the columns (and this text is a part of the new section).

If you have no space left at the end of your paper, you can delete

it.





Technological Speech Adaptation for Professional Work




as a Lawyer and Law Professor: A Case of a Patient with



Multiple Sclerosis



Sara Ahlin Doljak

European law faculty

New University

Nova Gorica, Slovenia

sara.ahlin-doljak@epf.nova-uni.si



Abstract The aim of this research is to evaluate the effectiveness of these

tools in supporting her professional activities and to explore the

This paper presents a case study of a female patient with multiple

strategies she employed to overcome the physical limitations

sclerosis (MS) who has successfully adapted to the challenges of

imposed by MS. By examining her journey, the study

her condition through the use of advanced assistive technologies.

underscores the importance of resilience, adaptability, and the



text software, this study highlights the practical applications of integration of technology in enabling individuals with disabilities Focusing on the use of speech-generating devices and predictive

to remain active in high-demand professions.



findings underline the critical role of technology in maintaining The paper is organized as follows: Section 2 describes the these tools in supporting her career as a lawyer and lecturer. The

methods used, including the case study approach and data

professional roles and the patient’s resilience in overcoming

collection techniques. Section 3 presents the results of the study,

significant physical disabilities.

with discussions on communication methods, technological

advancements, and adaptive strategies. The conclusion offers

Keywords insights into the potential for future technological developments

Speech-generating device, AAC technology, communication to further enhance the quality of life for individuals with severe

methods, SwiftKey, mobility, voice disabilities.



1 Introduction 2 Empirical Research Procedure

Multiple sclerosis (MS) is a chronic, progressive neurological

condition that affects the central nervous system, leading to a 2.1 Problem Definition and Purpose

wide range of physical and cognitive impairments. Among these, The primary problem addressed in this empirical research is

communication difficulties are common, particularly in cases understanding how a patient with a 12-year diagnosis of multiple

where the disease progression results in the need for a sclerosis (MS), who has a tracheostomy and is unable to speak,

tracheostomy, as it impairs the patient’s ability to speak [1]. This has adapted to her professional roles as a lawyer and lecturer. The

paper presents a case study of a female patient with MS who, study also seeks to evaluate the role that assistive technologies

despite facing severe communication barriers due to a play in enabling her to maintain her professional activities

tracheostomy and loss of speech, has successfully adapted to the despite severe physical limitations. The purpose of this research

challenges of her condition through the use of advanced assistive is to explore both the emotional and practical aspects of her

technologies. adaptation process, providing insights into the benefits and

The central focus of this study is on speech-generating limitations of technological interventions in helping her continue

devices and predictive text software, specifically the Tobii to work in high-demand professional environments.

Dynavox i-16 and Microsoft SwiftKey, and their role in enabling

the patient to maintain her dual professional roles as a lawyer and 2.2 Objective of the Empirical Research

a lecturer. These technologies, part of the broader category of The objective of this empirical research is to identify the key

Augmentative and Alternative Communication (AAC) systems, strategies and tools that allow a person with severe MS to adapt

have been demonstrated to significantly improve communication to and continue working in professional settings. The study

capabilities for individuals with severe speech impairments [2]. focuses on understanding how the patient copes with the

limitations imposed by her condition and examines the impact of

Permission to make digital or hard copies of part or all of this work for personal or assistive technologies on her ability to perform her professional

classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full tasks. Specifically, the research aims to:

citation on the first page. Copyrights for third-party components of this work must • Investigate how the participant has adapted emotionally and

be honored. For all other uses, contact the owner/author(s). practically to her condition, focusing on her mindset, routines,

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s). and coping mechanisms.

https://doi.org/10.70314/is.2024.DIGIN.14 • Identify the technological interventions that facilitate her

ability to work as a lawyer and lecturer, with a particular focus

on the use of assistive devices like speech-generating tools and 2.3.4 Participants

predictive text software. A single female participant with a 12-year diagnosis of MS, who

• Analyse the practical implementations and challenges of has a tracheostomy and is unable to speak, was selected for this

using assistive technologies in real-world professional settings case study. This unique case allows for an in-depth exploration

such as courtrooms and lecture halls. of her adaptation to her condition and the role of assistive

technologies in her professional life. The personalized focus on

2.3 Methodology one participant enabled a detailed examination of her

This research employs a qualitative case study methodology, experiences, providing valuable insights that might be diluted in

chosen for its suitability in exploring complex real-life a study involving multiple participants.

phenomena, particularly when the boundaries between the

phenomenon and its context are not clearly evident [3]. The case

study focuses on a single participant-a female patient with a 12- 3 Results

year diagnosis of MS, who has a tracheostomy and is unable to The collected data refers to the information gathered during the

speak-allowing for an in-depth analysis of her experiences, study on how the patient adapted to her condition and utilized

adaptations, and the role of assistive technologies in her various assistive technologies to maintain her professional roles.

professional life. This data was gathered through comprehensive interviews,

observations, and assessments of the tools she used, such as the

2.3.1 Procedure Tobii Dynavox speech-generating device and Microsoft

The study involved data collection through in-depth interviews SwiftKey predictive text software. with the participant and systematic observations of her

professional activities, including lectures and court sessions.



These methods provided firsthand insights into how assistive

technologies are applied in her day-to-day life, allowing for a

detailed exploration of their practical uses and effectiveness [4].



2.3.2 Measuring Instruments

The effectiveness of various assistive technologies was assessed

using several tools, chosen based on their role in aiding

communication and mobility. Key instruments included:

• Speech-Generating Device (Tobii Dynavox i-16): Used to

convert text to speech, this device was analysed for its usability,

customization options, and reliability in different professional

environments [5].

• Predictive Text Software (Microsoft SwiftKey): Evaluated

for its predictive capabilities, learning curve, and integration with

other devices, focusing on its efficiency in aiding communication

[6].

• Augmentative and Alternative Communication (AAC)

Systems: These systems were analysed for their speed, accuracy,

and user-friendliness during interactive professional sessions [7].



2.3.3 Research Questions Figure 1: Tobii Dynavox i-110, dry-erase board and Tobii

The following research questions guided the investigation: Dynavox i-16 (All photos are from the personal archive of

• How has a patient with MS adapted to her professional roles the participant of the study)

as a lawyer and lecturer? The study explores the emotional and

The data were analysed to identify recurring themes related

practical strategies the participant employed to adjust to her

to the efficacy of these technological aids, challenges

condition and continue working. It aims to reveal how she

encountered, and their overall impact on her professional life as

overcame challenges, adjusted her routines, and adopted new

a lawyer and lecturer. The results aim to provide insights into

tools to maintain her career.

how the assistive technologies supported her communication and

• What technological interventions have facilitated her

work, as well as the practical implementations of these tools.

continuation as a lawyer and lecturer? The research identifies key

technological aids, such as speech-generating devices and

3.1 Early Communication Methods

predictive text software, that assist her in communicating

effectively and performing professional tasks. Initially, the participant's primary channel of communication was

• What are the practical implementations and challenges the dry-erase board. Because she could not speak, she used the

associated with the assistive technologies she utilizes? This avenue to document (write) her cases or notes, which was

question addresses the real-world applications of the assistive effective then, especially in classrooms. The students acquired

technologies, evaluating both their strengths and limitations in comprehensive notes, which allowed them to gain vast

various professional settings like courtrooms and classrooms. knowledge of European law. A dry-erase board is an effective

interactive teaching tool, considering that lecturers can project

notes to help students understand complex concepts [8].

However, this approach was slow and laborious. However, this out her speech using the predictive text system. This technology

approach was slow and laborious, leading to delays in covering gives her enough notes and follow-up assignments or questions

the syllabus and impacting the efficiency of her work. to engage the classroom. Narang et al. (2022) noted that an open

Additionally, within a few weeks, she lost the ability to write discussion forum is an evidence-based strategy that bolsters

with her right arm, necessitating the search for alternative students’ engagement and awareness of the coursework [14].

communication methods. This early phase is illustrated in Figure During interactive sessions, she utilized augmentative and

1, which shows the Tobii Dynavox i-110, dry-erase board, and alternative communication (AAC) systems to compose feedback,

Tobii Dynavox i-16. although this process required patience from the audience due to

time delays. Despite these challenges, her lectures remained

3.2 Technological Advancements thorough and engaging.

The improvement in technology made communication easier for

her. Kumar et al. (2019) and Unwin (2017) asserted that 3.4 Evolution and Adaptation

technology enables instant communication, irrespective of Since the onset of her condition, she has realized the significance

geographical limitations [9, 10]. Kumar et al. (2023) and Sahoo of transforming and adapting to the emerging life changes and

and Choudhury (2023) added that applications, software, and latest technologies. Cahill (2020), Singh (2021), and

voice-controlled devices (wheelchairs), which are elements of Sampathkumar (2020) argued that change is inevitable,

the latest technologies, facilitate easier communication with MS warranting individuals to modify their behaviours or actions to

patients, illustrating their criticality in her life [11, 12]. Mainly, succeed [15, 16, 17]. She actively researched and adopted new

she uses a speech-generating device that converts text to speech. systems to facilitate communication, recognizing the importance

She controls the narration by typing on the keyboard of Tobii of evolving alongside technological developments.

Dynavox i-16. In addition, she uses Microsoft SwiftKey, a The evolution of AAC technology has been particularly

predictive text software, to improve her communication speed, beneficial, reflecting advancements in both hardware and

ensuring that she is competitive in class and court. Google Inc. software that support her ability to communicate sophisticated

(n.d.) stated that SwiftKey is an intelligent keyboard that learns ideas and engage professionally. Her current use of the Tobii

a person’s writing style to hasten writing [13]. These tools Dynavox i-16 in her daily professional life, at both work and

enabled her to communicate more efficiently, maintaining home, is shown in Figure 3.

competitiveness in her professional roles. A real-world example

of her using the Tobii Dynavox i-110 and i-16 in public is shown

in Figure 2.





Figure 3: A patient with MS uses the Tobii Dynavox i-16



nowadays at work and home (All photos are from the

Figure 2: A patient with MS in public with Tobii Dynavox i- personal archive of the participant of the study)

110 and i-16 (All photos are from the personal archive of

the participant of the study)

4 Conclusion



3.3 Lecturing Strategies and Public Engagement This study examined how a patient with multiple sclerosis



The technologies mentioned above play an integral role in the adapted to her condition to continue her professional roles as a lawyer and lecturer. The findings address the research questions preparation and execution of her lectures. She is an organized as follows: individual who prefers to define the lecture’s content and prepare

in advance. Thus, a few days prior to a class, she usually types



The patient adjusted her mindset and routines to intervention goals and use of technology. Seminars in Speech and

accommodate her physical limitations. She demonstrated Language, 29, 2 (2008), 83–91. DOI: 10.1055/s-2008-1080754.

[3] Robert K. Yin. 2018. Case Study Research and Applications: Design and

resilience by seeking alternative methods to fulfil her Methods. SAGE Publications.

professional responsibilities, including extensive preparation and [4] Sharan B. Merriam. 1998. Qualitative Research and Case Study

the use of assistive technologies. Her determination enabled her Applications in Education. Jossey-Bass Publishers [5] TD I-16 Gaze Interaction. Accessed September 9, 2024 from

to overcome initial setbacks and maintain her career. https://linkassistive.co.nz/product/td-i-16-gaze-interaction-2/

Key technological aids, such as the Tobii Dynavox i-16 [6] Microsoft. How does the Microsoft SwiftKey Prediction Bar work?

Accessed September 9, 2024, from https://support.microsoft.com/en-

speech-generating device and Microsoft SwiftKey predictive text au/topic/how-does-the-microsoft-swiftkey-prediction-bar-work-2fbbac0f-

software, were instrumental in facilitating her communication. ad7e-490d-bc92-2eb4538a3f3d

[7] Katherine C. Hustad, Kirsten Keppner, Amanda Schanz and Alycia Berg.

These tools supported her professional tasks by compensating for 2008. Augmentative and alternative communication for preschool children:

her lost abilities, allowing her to prepare lectures, engage with intervention goals and use of technology. Semin Speech Lang. 2008 May;

students, and represent clients effectively. 29, 2 (2008), 83–91. DOI: 10.1055/s-2008-1080754. PMID: 18645910;

PMCID: PMC2493612.

The assistive technologies were implemented in real-world [8] Idoghor Uduak and Rebecca Oluwayimika Kasumu. 2022. The use of

settings, such as classrooms and courtrooms. While they interactive whiteboards for teaching and learning in tertiary institutions.

International Journal of Trendy Research in Engineering and Technology

significantly enhanced her communication capabilities, 6, 6 (2022): 28–33. https://doi.org/10.54473/IJTRET.2022.6606

challenges included the slower pace of real-time interactions and [9] Sachin Kumar, Prayag Tiwari and Mikhail Zymbler. 2019. Internet of

the need for audience patience during interactive sessions. Things is a revolutionary approach for future technology enhancement: a

review. Journal of Big Data 6, 1 (2019): 1–21.

Technical issues and the learning curve associated with new https://link.springer.com/article/10.1186/s40537-019-0268-2%23article-

devices also presented obstacles. Nonetheless, these technologies info

[10] Tim Unwin. 2017. Reclaiming information and communication

succeeded in enabling her to continue her professional activities. technologies for development. Oxford University Press, 2017.

In conclusion, the case study underscores the critical role of [11] Sravan Kumar, Pamanji Deepika, Mahamad Farzana, Ojja Kaveri, Kumari

advanced assistive technologies in supporting individuals with Preethi and Pemmaraju Bharati Sai Sailaja. 2023. A Real-time Voice

Controlled Wheelchair for Specially Abled People. Turkish Journal of



severe disabilities to maintain their professional roles. The Computer and Mathematics Education (TURCOMAT) 14, 2 (2023): 742– 751. https://doi.org/10.17762/turcomat.v14i2.13729 . patient's experience highlights the importance of adaptability. [12] Sushil Kumar Sahoo and Bibhuti Bhusan Choudhury. 2023. Voice-

activated wheelchair: An affordable solution for individuals with physical

disabilities. Management Science Letters 13, 3 (2023): 175–192.



Acknowledgments https://doi.org/10.5267/j.msl.2023.4.004

[13] Microsoft Corporation. SwiftKey Keyboard. Version 9.10.15.5. Google

Play Store. Accessed September 9, 2024 from

The author would like to express sincere gratitude to the https://play.google.com/store/apps/details?id=com.touchtype.swiftkey&hl

participant for her cooperation and invaluable insights shared =en&pli=1

throughout this study. Special thanks are also extended to the advantage”: How content sharing strategies impact engagement in online [14] Unnati Narang, Manjit S. Yadav and Aric Rindfleisch. 2022. The “idea

European Law Faculty at New University, Slovenia, for their learning platforms. Journal of Marketing Research 59, 1 (2022): 61–78.

support in facilitating the research. https://doi.org/10.1177/00222437211017828

[15] Naomi Cahill. 2020. Change is inevitable… Canadian Journal of Dietetic

Practice and Research 81, 4 (2020): 169–169.

https://doi.org/10.3148/cjdpr-2020-036



References [16] Divya Singh. 2021. Change is inevitable. Just Africa 6, 1 (2021), 4–5.

Accessed September 9, 2024 from https://hdl.handle.net/10520/ejc-

[1] Ilana Katz Sand. 2015. Classification, diagnosis, and differential diagnosis ajcj_v6_n1_a1

of multiple sclerosis. Current Opinion in Neurology, 28, 3 (2015), 193–205. [17] Arkalgud Sampathkumar. 2020. Change is inevitable. Asian

[2] Katherine C. Hustad, Kirsten Keppner, Amanda Schanz and Alycia Berg. Cardiovascular and Thoracic Annals 28, 1 (2020): 5–6.

2008. Augmentative and alternative communication for preschool children: https://doi.org/10.1177/0218492319897805



Zbornik 27. mednarodne multikonference



INFORMACIJSKA DRUŽBA – IS 2024

Zvezek I



Proceedings of the 27th International Multiconference



INFORMATION SOCIETY – IS 2024

Volume I



Konferenca o zdravi dolgoživosti



Conference on Healthy Longevity



Urednika / Editors



Martin Lipovšek, Boštjan Petrič



http://is.ijs.si



8. oktober 2024 / 8 October 2024

Ljubljana, Slovenia





PREDGOVOR




Po velikem uspehu Konference o podaljševanju življenja, ki smo jo leta 2023 izvedli v okviru

mednarodne multikonference Informacijska družba na IJS, smo se letos organizatorji odločili,

da ponovimo zgodbo in konferenco še dodatno izboljšamo. Tudi tokrat je bil namen dvojen:

povezovanje strokovnjakov, ki se ukvarjajo s podaljševanjem življenja na sicer nepovezanih

področjih, ter promocija ideje o podaljšanju življenja med javnostjo. Je konferenca dosegla

zastavljene cilje?



V primerjavi z lanskim letom smo tokrat v živo lahko poslušali ne le vse domače, temveč tudi

nekatere od tujih govorcev. Iz Nemčije se nam je pridružil Alexander Tietz-Latza, ki je s

perspektive pravnika govoril o pravici do podaljševanja življenja, iz ZDA pa Melissa King, ki

je delila svoje bogate izkušnje iz agitacije in lobiranja za raziskave preprečevanja staranja.

Znanstveniki, ki se v malih državah ukvarjajo z določenimi področji, so vajeni ustrezno

majhnih zneskov: Kingova ima iz Washingtona izkušnje z borbo za večstomilijonske zneske,

ki je bila večkrat tudi uspešna, in ob njeni pripovedi je bila pozornost poslušalcev

zagotovljena. Več vrhunskih strokovnjakov se nam je pridružilo tudi prek videopovezave:

Adam Buben iz Nizozemske, Ilya Stambler iz Izraela in Bryan Johnson iz ZDA.



Ne le da smo uspeli poleg že omenjenih pridobiti še izvrstne domače predavatelje, temveč se

nam je prvič doslej pridružila tudi predstavnica Ministrstva za zdravje Vesna Marinko. Za

razliko od večine znanstvenih konferenc, ki so namenjene majhni tarčni skupini

strokovnjakov, je bila naša odprta za širšo javnost, in javnost je z veseljem tudi prišla. Lanska

konferenca je bila obilno obiskana tako v živo kot preko videoprenosa, toda ali so obiskovalci

dejansko prišli, ker bi jih tema zanimala? Številni odzivi, ki smo jih organizatorji prejeli po

zaključku, potrjujejo, da je res tako. Letošnji prvi odzivi občinstva nakazujejo, da tudi tokrat

ni nič drugače.



Martin Lipovšek, Boštjan Petrič





FOREWORD




After the great success of the Life Extension Conference, which we held in 2023 as part of the

international multiconference Information Society at the Jožef Stefan Institute, this year the

organizers decided to repeat the event and further improve the conference. Once again, the

purpose was twofold: to connect experts working on life extension in otherwise unrelated

fields and to promote the idea of life extension among the public. Did the conference meet

these

goals?



Compared to last year, this time we were able to hear not only all the Slovenian speakers live

but also some international ones. Alexander Tietz-Latza from Germany joined us to talk about

the right to life extension from a legal perspective, and Melissa King from the USA shared her

extensive experience in advocacy and lobbying for anti-aging research. Scientists in small

countries are used to working with relatively small budgets, while King, during her visits to

Washington, has fought for hundreds of millions of dollars, often successfully; needless to

say, she had the audience's attention. Several top experts also joined us via video link: Adam

Buben from the Netherlands, Ilya Stambler from Israel, and Bryan Johnson from the USA.



Not only did we manage to bring in excellent Slovenian speakers in addition to all the foreign

ones already mentioned, but for the first time, we were also joined by a representative from

the Ministry of Health, Vesna Marinko. Unlike most scientific conferences, which are aimed

at a small target group of experts, ours was open to the general public - and the public gladly

attended. Last year's conference was well-attended both in person and via video stream, but

did the attendees come because they were genuinely interested in the topic? Numerous

responses that we received afterward confirm that they were. Early feedback from this year's

audience suggests that the second Life Extension Conference was no different.



Martin Lipovšek, Boštjan Petrič



PROGRAMSKI ODBOR / PROGRAMME COMMITTEE



Martin Lipovšek



Boštjan Petrič



ORGANIZACIJSKI ODBOR / ORGANIZING COMMITTEE



Lovro Herga



Mitja Luštrek



Patrik Milić



Laura Vuga



Igor Djilas



Veronika Herga



Branka Javornik



Vitalna dolgoživost – izzivi in priložnosti za bodočo družbo



Danes smo priča naraščajočemu pričakovanemu trajanju življenja, ki je v zadnjih desetletjih

izjemno porasla zaradi ugodnejših zdravstvenih, prehranskih, ekonomskih in socialnih razmer

kot odraz različnih odkritij ter novih spoznanj v znanosti in družbi ter njihovih aplikacij v

vsakdanjem življenju. Starejše se sicer zaradi pojavnosti starostne oviranosti in različnih

kroničnih bolezni večkrat obravnava kot posebno skupino, potrebno specializirane pomoči na

področju medicine, sociale, politike, zakonodaje, bivanjskih razmer, servisnih služb, posebnih

organizacij in podobno. Zato ne čudi razmišljanje, da je »znanost podaljšala življenjsko dobo,

vendar to pomeni daljšo starost«.



Res je, staranju ne morem ubežati – z leti se začnejo kopičiti okvare in napake v telesu, ki se

odražajo kot upad fizičnih in kognitivnih sposobnosti in razvoj kroničnih bolezni, vendar danes

znanost ponuja več, kot le podaljševanje »starega« življenja! In kaj nam ponuja znanost o

biologiji staranja? V zadnjem desetletju smo priča eksploziji različnih aktivnosti in znanstvenih

odkritij na tem področju, nekateri to imenujejo kar revolucija znanosti na področju dolgoživosti.

Področje biologije staranja imenujemo tudi biogerontologija, ki obravnava predvsem

mehanizme staranja, in geroznanost, ki obravnavane mehanizme staranja povezuje s pojavom

kroničnih bolezni. Te raziskave skušajo odgovoriti na vprašanja na molekulskem in celičnem

nivoju ter na nivoju organizma, zakaj in kako se staramo ter ali je mogoče omiliti/upočasniti

procese staranja, predvsem s preprečevanjem razvoja starostnih tegob in kroničnih bolezni in

tako podaljšati zdrava leta življenja.



In kakšne so možnosti za vitalno dolgoživost? V prispevku bomo predstavili vplive genetske

zasnove in okolja (življenjskega sloga) na procese staranja in novejša spoznanja o možnostih

uporabe geroterapevtikov ter celičnih in genskih terapij za vitalno dolgoživost.



Vital Longevity – Challenges and Opportunities for Future Society



Today, we are witnessing increasing life expectancy, which has significantly risen in recent

decades due to improved health, nutrition, economic, and social conditions, reflecting various

discoveries and new understandings in science and society and their application in daily life.

Older individuals are often treated as a special group due to the prevalence of age-related

disabilities and various chronic diseases, requiring specialized assistance in medicine, social

services, politics, legislation, housing conditions, service agencies, special organizations, and

more. This gives rise to the notion that "science has extended lifespan, but that means longer

old age."



Indeed, aging is inevitable—over time, defects and errors accumulate in the body, reflected in

a decline in physical and cognitive abilities and the development of chronic diseases. However,

today's science offers more than just the extension of "old age." So, what does the science of

aging biology offer us? In the past decade, we have witnessed an explosion of various activities

and scientific discoveries in this field, with some even calling it a revolution in longevity science.

The field of aging biology, also known as biogerontology, primarily deals with the mechanisms

of aging, while geroscience links these mechanisms with the emergence of chronic diseases.

These studies aim to answer questions at the molecular, cellular, and organismal levels about

why and how we age and whether it is possible to alleviate or slow down aging processes,

especially by preventing the development of age-related ailments and chronic diseases, thus

extending the healthy years of life.



What are the possibilities for vital longevity? This article will present the influences of genetic

makeup and the environment (lifestyle) on the aging processes and recent findings on the

potential use of gerotherapeutics, as well as cellular and gene therapies, for achieving vital

longevity.



.



Gordana Kalan Živčec



8 stebrov zdravja – kaj lahko naredimo sami in kaj so priložnosti medicine – Gordanina

Metoda



Zdravje in kakovostna dolgoživost sta neposredno povezani. Kakovostna dolgoživost je tesno

povezana s koncepti, ki so v osnovi enostavni in jih lahko izvajamo samostojno: gibanje,

prehrana, počitek, socialni stiki/domače živali, ... tehnike sproščanja. Nadgradimo jih s

koncepti medicine: preventiva, prehranska dopolnila in šele na koncu zdravila. Prvi steber

zdravja je gibanje - Drugi steber zdravja je prehrana - Tretji steber zdravja je počitek - Četrti

steber zdravja je kakovostna socialna mreža in tudi hišni ljubljenčki - Peti steber zdravja je

zmanjševanje mentalnega, fizičnega in okoljskega stresa - Šesti steber zdravja je preventiva -

Sedmi steber zdravja so prehranska dopolnila - Osmi steber zdravja so zdravila. Hiša stoji na

4 stebrih. Za zdravo življenje jih je potrebno 8. Srečno!



---



The 8 Pillars of Health – What We Can Do Ourselves and the Opportunities in Medicine

– Gordana's Method



Health and quality longevity are directly connected. Quality longevity is closely tied to concepts

that are fundamentally simple and can be implemented independently: movement, nutrition,

rest, social connections/pets, and relaxation techniques. These are then enhanced by medical

concepts: prevention, nutritional supplements, and only as a last resort, medications. The first

pillar of health is movement. - The second pillar of health is nutrition. - The third pillar of health

is rest. - The fourth pillar of health is a strong social network, including pets. - The fifth pillar of

health is reducing mental, physical, and environmental stress. - The sixth pillar of health is

prevention. - The seventh pillar of health is nutritional supplements. - The eighth pillar of health

is medication. A house stands on four pillars. For a healthy life, eight are needed. Good luck!



.



Rok Kopinč



Moč matičnega mlečka pri zdravem staranju



Matični mleček ima že v razvoju čebel vlogo epigenetskega reprogramiranja. Ličinke, ki so

hranjene celotno obdobje razvoja izključno z matičnim mlečkom, se razvijejo v čebelje matice,

medtem ko se ličinke, ki so hranjene z matičnim mlečkom le prve tri dni, razvijejo v navadne

čebele delavke, čeprav sta obe genetsko identični. Matice so še enkrat večje od delavk, za

razliko od delavk plodne in živijo do 20-krat dlje. Razlog je v tem, da matični mleček omogoča

utišanje določenih genov (genov delavke) in izražanje genov matice. Ker so matice celo

življenje hranjene le z matičnim mlečkom, se razlika v izražanju genov ohranja in s tem

lastnosti, ki so značilne za čebeljo matico. Če navadno čebelo delavko začnemo hraniti z

matičnim mlečkom, začne dobivati nekatere lastnosti matice. Pa lahko takšno delovanje

matičnega mlečka pričakujemo tudi pri drugih organizmih? Raziskave kažejo, da matični

mleček za 10 do 20 % podaljša življenje tudi pri nekaterih drugih modelnih organizmih , kot so

črvi C. elegans, vinske mušice, črički in miške. Prav tako poveča odpornost teh živali na stres,

kot so UV svetloba, izpostavljenost kemijskim sredstvom in na bakterijske/virusne infekcije. V

različnih kliničnih študijah je bilo dokazano, da matični mleček spodbuja procese, ki so sicer

povezani z dolgoživostjo in podaljšanjem zdravega obdobja življenja. Tako so dokazali, da

matični mleček aktivira enake procese kot kalorična restrikcija, še zlasti glavna sestavina

matičnega mlečka, to je 10-hidroksi-decenojska kislina, poveča izražanje sirtuinov. V nedavni

klinični raziskavi smo tudi sami potrdili, da aktivira obrambne sisteme, ki ščitijo celice pred

oksidativnimi poškodbami ter poviša raven adiponektina, katerega višje ravni so pri ljudeh

povezane z dolgoživostjo. Matični mleček vsebuje tudi spermidin, ki je znan po tem, da

spodbuja avtofagijo in preprečuje staranje celic.



---



The Power of Royal Jelly in Healthy Aging



In the development of bees, royal jelly plays a role in epigenetic reprogramming. Larvae that

are fed exclusively with royal jelly throughout their entire development become queen bees,

while those that are fed royal jelly only for the first three days develop into ordinary worker

bees, even though both are genetically identical. Queens are twice the size of workers, are

fertile, and live up to 20 times longer. This is because royal jelly enables the silencing of certain

genes (worker bee genes) and the expression of queen genes. Since queens are fed only royal

jelly throughout their lives, the difference in gene expression is maintained, along with the traits

characteristic of a queen bee. If a regular worker bee starts being fed royal jelly, it begins to

develop some queen-like characteristics. But can we expect such effects of royal jelly in other

organisms as well? Research shows that royal jelly extends the lifespan of some other model

organisms, such as C. elegans worms, fruit flies, crickets, and mice, by 10 to 20%. It also

increases these animals' resistance to stress, such as UV light, exposure to chemicals, and

bacterial/viral infections. Various clinical studies have proven that royal jelly promotes

processes associated with longevity and the extension of the healthy period of life. It has been

shown that royal jelly activates the same processes as calorie restriction, particularly the main

component of royal jelly, 10-hydroxy-2-decenoic acid, which increases the expression of

sirtuins. In a recent clinical study, we also confirmed that royal jelly activates defense systems

that protect cells from oxidative damage and increases levels of adiponectin, which in humans

is linked to longevity. Royal jelly also contains spermidine, known for promoting autophagy and

preventing cellular aging.

Melissa King



Advocacy in Action: How Patient Advocacy, Public Policy, and Science Communication

Drive Scientific Progress and Benefit Society



This talk underscores the vital role of advocacy, policy, and communication in advancing

scientific progress, especially in healthy longevity. It illustrates how engaged citizens, through

patient advocacy, can drive research funding and policy changes, significantly impacting

healthcare outcomes. The presentation highlights the global spread of initiatives like Right to

Try laws and stresses the importance of public funding for early-stage research and the need

for scientists to actively communicate with the public.



There is a great need in the healthy longevity field for more patient advocacy, public policy,

public information through science communications, and global collaborative planning to drive

both scientific progress and distributive justice. The impact of patient advocacy on various

areas of biomedical research and healthcare is immeasurable, as shown by examples such as

California's Propositions 71 & 14 for stem cell research funding, JDRF's efforts in type 1

diabetes research, and HIV/AIDS activism.



Ongoing public policy efforts are of great importance for both research and patients. Efforts

like Brain Mapping Day at the US Congress, part of the inspiration for President Obama's Brain

Initiative, and the implementation of Right to Try legislation are not always visible to the public

at large but impact billions of people. A global overview of Right to Try legislation and similar

programs across different regions provides an example of how complex public policy efforts

can be when the focus is global.



The role of public funding in scientific advancements is enormous, given that private industry

and funding often enter only after initial public investments reduce risk. The critical need for

accurate science communication to the public provides an opportunity for scientists and others

well-informed about science to engage in outreach and education efforts.



This presentation is a call to action, meant to inspire and empower attendees to become active

participants in shaping the future of science and health policy, including and especially around

healthy longevity for all.



Alexander Tietz - Latza



The Right to Health and Longevity: An Evolving Human Right?



The recognition and enforcement of the right to health as a fundamental human right can

significantly enhance longevity and healthspan. By ensuring access to comprehensive

healthcare, promoting preventive measures, interventions, and addressing socio-economic

determinants of health, we can advocate a legal framework that supports a healthier

population.



Several international human rights instruments, such as the Universal Declaration of Human

Rights (UDHR) and the International Covenant on Economic, Social and Cultural Rights

(ICESCR), enshrine the right to health. In Europe, the European Social Charter and national

constitutions further reinforce this right. Effective implementation of these legal frameworks

can lead to better health outcomes.



This right encompasses access to timely, acceptable, and affordable health care of appropriate

quality as well as the underlying determinants of health, such as safe drinking water, adequate

sanitation, and a healthy environment. The right to health implies that everyone should have

access to the necessary health services and conditions that enable them to live a long and

healthy life - including Longevity?



Adam Buben



Should Christians Be Opposed to Technologically Achieved Life Extension? Why

Kierkegaard Might Make a Bad Transhumanist



It might seem obvious to most readers of Kierkegaard, an important thinker in the Christian

tradition, that he would have little interest in, and would probably be radically opposed to, the

goals of transhumanism. The difficult part, however, is explaining exactly what he would be

opposed to and why. It is true that Kierkegaard does not put much stock in science and

technology, but transhumanism is not a monolithic movement; the futurists, philosophers,

scientists, and sci-fi enthusiasts who might consider themselves transhumanists have a wide

array of hopes and dreams about what technology might eventually be able to do for, and to,

humankind. This variety means certain understandings of transhumanism might actually be

quite compatible with all sorts of religious traditions, including Christianity. In fact, it is important

to many transhumanists that their ideas not be perceived as sacrilegious, irreligious, or

heretical, because they think a “broad, diverse coalition” of support for, or at least tolerance of,

their agenda is practically and politically necessary for their goals to be realized. Given such

intentional inclusivity, we must dig a little deeper into Kierkegaard’s views to determine if he

would truly be an opponent of transhumanism. There may well be several approaches to

explaining “why Kierkegaard might make a bad transhumanist,” but mine will be to argue—

focusing on the example of technologically achieved life extension—that there is something in

his Christian commitments that runs counter to the very spirit of transhumanism.



.



Zbornik 27. mednarodne multikonference



INFORMACIJSKA DRUŽBA – IS 2024

Zvezek J



Proceedings of the 27th International Multiconference



INFORMATION SOCIETY – IS 2024

Volume J



Legende računalništva in informatike



Legends of Computing and Informatics



Urednika / Editors



Tone Stanovnik, Matjaž Gams



http://is.ijs.si



8. oktober 2024 / 8 October 2024

Ljubljana, Slovenia





PREDGOVOR




V letu 2024 s ponosom nadaljujemo z iniciativo »Legende računalništva in informatike«

(krajše »Legende RI«), podnaslovljeno »Galerija slavnih slovenskega računalništva in

informatike« ali po angleško »Hall of Fame of Slovenian Computer Science and Informatics«.



Naš glavni motiv je priznati izjemne dosežke raziskovalcev, razvijalcev, inženirjev in

vizionarjev na področju računalništva in informatike. Ob tem želimo hkrati zbrati pomembne

avtobiografske prispevke, ki bodo gradili zapis zgodovine tega pomembnega področja v

Sloveniji. Tako sledimo dvema pomembnima ciljema: ohranjamo strokovnost brez političnih

ali ideoloških vplivov ter zagotavljamo demokratičnost in transparentnost v procesu izbire

legend. Obenem si prizadevamo, da nikoli ne bi dopustili, da bi izjemni posamezniki ostali

neopaženi in ne priznani za svoj prispevek, tako kot je bil nekoč Alan Turing, eden največjih

umov v zgodovini računalništva, ki namesto priznanja prejel le kazen zaradi svoje identitete.



V letu 2024 bo programski odbor ocenjeval in nagrajeval na podlagi dejanskega prispevka

posameznikov na področju računalništva in informatike, pri čemer bomo še posebej pozorni

na njihov vpliv na stroko, družbo in razvoj tehnologij.



Navedimo osnovne dogovore pri Legendah: Dobitniki nagrade Michie-Turing so avtomatsko

sprejeti na podlagi poslanih referatov, ki služijo tudi kot avtobiografska zgodovina. Podobno

naj bi veljalo za mednarodne nagrade izjemne pomembnosti, medtem ko nacionalne nagrade

še niso predmet avtomatskega sprejemanja. Prvo leto pričakujemo od 10 do 20 nagrad, nato

okoli 5 letno, zgornja meja naj bi bila okoli 50 prejemnikov.

Postopek predlaganja kandidatov za legendo ostaja odprt, zlasti za predloge s strani društev

ACM Slovenija, SLAIS in Informatike. Predlog je potrjen, če zbere več kot polovico oddanih

glasov in ne preseže letne ali absolutne omejitve.



Predlogi za legendo so pripravljeni kot prispevki za konferenco IS v obliki 2-4 strani, z

ustrezno dokumentirano literaturo, le izjemoma je dovoljen povzetek na eni strani.

Predlagatelj lahko predlaga sebe ali koga drugega, več predlogov pa je prav tako dovoljenih.

Pri posthumnih nominacijah je potrebno pridobiti soglasje najbližjih svojcev. Prispevek naj bo

strukturiran kot življenjepis s poudarkom na dosežkih – znanstveni, praktični, pedagoški,

organizacijski – pri čemer je zaenkrat še nekaj svobode pri oblikovanju.

Z novim konceptom želimo še naprej širiti prepoznavnost in promocijo slovenskega

računalništva in informatike ter krepiti vlogo informacijske družbe. Vaše sodelovanje,

podpora, razumevanje in strpnost so ključni za uspeh te pobude. Iskrena hvala vsem

sodelujočim in čestitke sedanjim ter prihodnjim prejemnikom tega prestižnega priznanja!



Tone Stanovnik in Matjaž Gams





FOREWORD




In 2024, we proudly continue with the initiative "Legends of Computer Science and

Informatics" (abbreviated as "Legends RI"), subtitled "Hall of Fame of Slovenian Computer

Science and Informatics."



Our primary motivation is to acknowledge the exceptional achievements of researchers,

developers, engineers, and visionaries in the field of computer science and informatics. At the

same time, we aim to gather important autobiographical contributions that will form the

historical record of this significant field in Slovenia. In doing so, we pursue two key goals:

maintaining professionalism without political or ideological influence and ensuring

democracy and transparency in the selection process of the legends. Furthermore, we are

committed to never allowing exceptional individuals to go unnoticed or unrecognized for their

contributions, much like Alan Turing, one of the greatest minds in computer science history,

who was punished instead of being rewarded due to his identity.



In 2024, the program committee will evaluate and award individuals based on their

contributions to computer science and informatics, with particular attention to their impact on

the profession, society, and the development of technologies.



Let us outline the basic agreements for the Legends: Michie-Turing award recipients are

automatically accepted based on their submitted papers, which also serve as autobiographical

histories. Similarly, this will apply to internationally prestigious awards, while national

awards are not yet subject to automatic acceptance. Each year, we aim to induct

approximately 5 new legends into the Hall of Fame.



The process of nominating candidates for the legend status remains open, particularly for

proposals from associations like ACM Slovenia, SLAIS, and Informatika. A proposal is

confirmed if it receives more than half of the submitted votes and does not exceed the annual

or absolute limits.



Legend proposals are prepared as two to four-page contributions for the IS conference,

supported by appropriately documented literature, with a one-page summary allowed only in

exceptional cases. A proposer may nominate themselves or someone else, and multiple

proposals are also allowed. In the case of posthumous nominations, consent from the closest

relatives is required. The contribution should be structured like a CV, focusing on

achievements—whether scientific, practical, pedagogical, or organizational—with some

flexibility in the format.



With this new concept, we aim to further increase the recognition and promotion of Slovenian

computer science and informatics and strengthen the role of the information society. Your

cooperation, support, understanding, and patience are essential for the success of this

initiative. Sincere thanks to all participants, and congratulations to the current and future

recipients of this prestigious recognition!



Tone Stanovnik and Matjaž Gams



PROGRAMSKI ODBOR / PROGRAMME COMMITTEE



Gams Matjaž



Stanovnik Tone



Batagelj Marjan



Bavec Cene



Bohanec Marko



Bratko Ivan



Bric Rudi



Brodnik Andrej



Džeroski Sašo



Gornik Tomaž



Grad Janez



Grobelnik Marko



Hafner Izidor



Horvat Bogomir



Jurič Smon



Kljajić Borštnar Mirjana



Kozak Jernej



Kuščer Andrej



Lesjak Janez



Leskovar Robert



Luštrek Mitja



Mernik Marjan



Mongus Domen



Petkovšek Marko



Pisanski Tomaž



Rajkovič Vladislav



Schlamberger Niko



Solina Franc



Škedelj Franc



Špegel Marjan



Trampuž Cveto



Trček Denis



Turk Žiga



Virant Jernej



Volk Mira



Zakrajšek Egon



Zimic Niko



Zupan Blaž



Žalik Borut



Železnikar Anton



Žerko Miha





Achievements of the Candidate Marjan Mernik




Dosežki kandidata Marjana Mernika



Marjan Mernik

University of Maribor

Faculty of Electrical Engineering and Computer Science

Maribor, Slovenia

marjan.mernik@um.si



Abstract Languages, Compilers, and Evolutionary Computation. He spent

a Sabbatical in 2004 at the University of Alabama at Birmingham

The paper briefly summarizes the main scientific and research (UAB), USA, and since then, he has been a visiting professor at

achievements of Prof. Dr. Marjan Mernik, as well as his teaching UAB until 2017, where he occasionally taught undergraduate

and managerial work at the University of Maribor, Faculty of class on Programming Languages and a graduate class on

Electrical Engineering and Computer Science. His scientific and Domain-Specific Languages. He was also a visiting professor at

research work covers the field of programming languages, the University of Novi Sad, Serbia, from 2010 to 2016. He was

software engineering and evolutionary computing. advisor to 8 PhD students at the University of Maribor, co-



Keywords advisor to 5 PhD students at the University of Alabama at

Birmingham, and to 1 PhD student at the University of Ljubljana.

Computer Science, programming languages, software He was a member of PhD committees at foreign institutions such

engineering, evolutionary computation as the University of Sarajevo, the University of Rostock, the



Povzetek University of Tartu, the Eindhoven University of Technology,

the University of Milan, the University of Alabama at

V prispevku so na kratko povzeti glavni znanstveno-raziskovalni Birmingham, the University of Alabama, the University of Cape

dosežki prof. dr. Marjan Mernika, kakor tudi njegovo pedagoško Town, the Universidade Nova de Lisboa, and the University of

in vodstveno delo na Univerzi v Mariboru, Fakulteti za Novi Sad.

elektrotehniko, računalništvo in informatiko. Njegovo

znanstveno-raziskovalno delo posega na področje programskih

jezikov, inženiring programske opreme in evolucijskega 2 Research Activities

računanja. His main research interests lie in the intersection of programming

languages, software engineering and evolutionary computation.

Ključne besede His research philosophy involves tackling practical problems of

Računalništvo in informatika, programski jeziki, inženiring real-world complexity and proposing foundational solutions that

programske opreme, evolucijsko računanje offer the potential for significant long-term impact on the field of

computing. His past work on attribute grammar [1, 2] and current

work on Domain-Specific Languages [3-10] and Evolutionary

1 Introduction Computation [11-19] demonstrate that he is identifying with and

Prof. Dr. Marjan Mernik is a professor of Computer Science at working on major research problems in this field and that he is

the Faculty of Electrical Engineering and Computer Science, able to initiate new ways of thinking.

University of Maribor, Slovenia. He completed his Ph.D. in From the start of his career, a primary research interest has been

computer science from the University of Maribor in 1998. His the formal syntax and semantics of programming languages. He

dissertation focused on programming language composition and has developed various compiler/interpreter generators based on

evolution at the syntactic and semantic levels. Since 1998, he has different formal methods, such as attribute grammars,

been teaching various courses on Programming, Programming denotational semantics, and operational semantics. To solve the

problem of non-modularity, non-extensibility and non-

∗ ∗Article Title Footnote needs to be captured as Title Note Article Title Footnote needs to be captured as Title Note reusability of formal language specifications, which would

† †Author Footnote to be captured as Author Note Author Footnote to be captured

enable easier language composition, he has proposed the idea of

Permission to make digital or hard copies of part or all of this work for personal or Permission to make digital or hard copies of part or all of this work for personal or multiple attribute grammar inheritance, which was successfully

classroom use is granted without fee provided that copies are not made or distributed classroom use is granted without fee provided that copies are not made or distributed implemented in the LISA compiler generator tool [1]. Despite

for profit or commercial advantage and that copies bear this notice and the full for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for third-party components of this work must citation on the first page. Copyrights for third-party components of this work must that, this work was done during, and shortly after, his PhD thesis;

be honored. For all other uses, contact the owner/author(s). be honored. For all other uses, contact the owner/author(s). the work regains attention in the young engineering discipline of

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia © 2024 Copyright held by the owner/author(s). © 2024 Copyright held by the owner/author(s). Software Language Engineering (SLE) [2] with the aim of



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia M. Mernik



establishing a systematic and rigorous approach to the an award for scientific research and educational work at the

development, use, and maintenance of computer languages, University of Maribor in 2022. In 2019, he was awarded the

which comprises specification, modelling and programming Donald Michie and Alan Turing Prize for the Life

languages. Achievements of the Slovenian Information Society. In 2024,

In the next step of his career, he was interested in the possibility he received the Entrepreneurs, Academicians & Researchers

of automatically generating other language-based tools from Guild Award for Lifetime Achievement in Academia and

formal language specifications, such as editors, debuggers, Research from the MathTech Thinking Foundation.

visualization tools, and animators [4]. While this work

emphasizes primarily textual programming languages, it is now

extended to domain-specific modelling languages [5]. 3 Other Activities

His research also focuses on the design and implementation of He has strong international collaboration, including organizing

domain-specific languages [3-10]. He is interested in many conferences and workshops, journal special issues, serving on

aspects of domain-specific languages, from domain analysis, journal editorial boards, and being a member of program

design, implementation, validation, maintenance, and their use. committees. He has been an invited speaker at ten conferences.

All of these phases have been described in his survey paper in He is a member of seven editorial boards, a journal reviewer for

the ACM Computing Surveys journal [3], which was extremely more than 40 journals, and a member of program committees for

well accepted among researchers. This paper has more than 2800 more than 150 conferences and workshops. He is the Editor-in-

Google Scholar citations, and it is the most cited paper in the area Chief of the Journal of Computer Languages (Elsevier) and an

of domain-specific languages. It is also used as a standard Associate Editor of Applied Soft Computing Journal (Elsevier),

teaching material for various courses on domain-specific Swarm and Evolutionary Computation (Elsevier), and

languages at different universities. On many occasions, he has Information Sciences (Elsevier). His basic research has been

been regarded by others as a leader in domain-specific language funded by the Slovenian Research Agency, the National Science

research. Foundation (NSF USA) and industrial funds from local

Around the year 2000, he also became interested in another companies (Raiffeisen Krekova banka, DEWESoft).

research field - evolutionary computation. He was combining Last but not least, he took on several organizational duties. He

evolutionary computation in a novel way with the research fields was a Chair of the Quality Assurance Committee at the Faculty

of programming languages and software engineering [11-19]. of Electrical Engineering and Computer Science, University of

One such problem, which has not sufficiently been solved yet, is Maribor, from 2003 to 2011, and Vice-Dean of Research from

the problem of inferring grammars from the set of positive and 2015 to 2018. He was a Chair of the Institute of Computer

negative samples. This is the research field of grammatical Science from 2015-2019, where he was responsible for the

inference to which he and his co-workers have contributed a new implementation of the department’s objectives and goals, such as

algorithm, MAGIc [13]. The work on grammatical inference has following constant changes in computer science and information

recently been extended to semantic inference of domain-specific technologies curricula. Since 2019, he has been a Vice-Dean of

languages [17]. International Cooperation and Quality Development at the

Another topic from the field of evolutionary computation in Faculty of Electrical Engineering and Computer Science,

which he is interested is the problem of parameter control. He University of Maribor. Since 2019, he has been a Member of the

and co-workers have suggested one of the first approaches for Council of Slovenian Quality Assurance Agency for Higher

self-adapting control parameters in Differential Evolution (DE) Education.

[12]. This work has been published in the prestigious journal

IEEE Transactions on Evolutionary Computation and has more

than 3600 Google Scholar citations. He has become interested in References

investigating how different evolutionary algorithms perform [1] MERNIK, Marjan, ŽUMER, Viljem, LENIČ, Mitja, AVDIČAUŠEVIĆ,

exploration and exploitation of the search space. This is a Enis. Implementation of multiple attribute grammar inheritance in the tool

fundamental issue of any search algorithm. The result of this LISA. SIGPLAN notices. June 1999, vol. 34, no. 6, str. 68-75. ISSN 0362-

1340. [COBISS.SI-ID 4761110]

study was published in 2013 [11] in the prestigious ACM [2] MERNIK, Marjan. An object-oriented approach to language compositions

Computing Surveys journal and has up to now more than 1500 for software language engineering. The Journal of Systems and Software.

[Print ed.]. 2013, vol. 86, iss. 9, str. 2451-2464. ISSN 0164-1212.

Google Scholar citations. DOI: 10.1016/j.jss.2013.04.087. [COBISS.SI-ID 16884246]

His research work culminated in more than 130 journal [3] MERNIK, Marjan, HEERING, Jan, SLOANE, Anthony M. When and

papers, among them more than 70 in journals with an SCI index how to develop domain-specific language. ACM computing surveys. [Print

ed.]. 2005, vol. 37, no. 4, str. 316-344. ISSN 0360-0300. [COBISS.SI-

(e.g., ACM Computing Surveys, IEEE Transactions on ID 10259222]

Evolutionary Computation, Applied Soft Computing, Empirical [4] HENRIQUES, Pedro Rangel, VARANDO PEREIRA, Maria João,

MERNIK, Marjan, LENIČ, Mitja, GRAY, Jeffrey G., WU, Hui.

Software Engineering) and in more than 160 Automatic generation of language-based tools using the LISA system. IEE



conference/workshop papers. Google Scholar ranked his work as proceedings. Software. [Print ed.]. Apr. 2005, vol. 152, no. 2, str. 54-69. ISSN 1462-5970. [COBISS.SI-ID 9519126 ] H-Index=44, number of citations = 14260. Clarivate Analytics [5] SPRINKLE, Jonathan, MERNIK, Marjan, TOLVANEN, Juha-Pekka,

named Marjan Mernik as a highly cited researcher in 2017 and SPINELLIS, Diomidis. What Kinds of Nails Need a Domain-Specific

2018 (the only researcher from computer science from Slovenia). Hammer?. IEEE software. [Print ed.]. Jul./Avg. 2009, vol. 26, iss. 4, str.

15-18. ISSN 0740-7459. DOI: 10.1109/MS.2009.92. [COBISS.SI-

His research ranks among the top 1% of most cited works in the ID 13263638]

field of Computer Science in 2017 and 2018, earning the mark of [6] KOSAR, Tomaž, MARTÍNEZ LÓPEZ, Pablo E., BARRIENTOS, Pablo

A., MERNIK, Marjan. A preliminary study on various implementation

exceptional impact. For his research achievements, he was approaches of domain-specific language. Information and software

awarded a Silver Sign at the University of Maribor in 2012 and technology. [Print ed.]. apr. 2008, vol. 50, iss. 5, str. 390-405. ISSN 0950-



Achievements of the Candidate Marjan Mernik Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



5849. http://dx.doi.org/10.1016/j.infsof.2007.04.002, Digitalna knjižnica ed.]. 2012, vol. 42, no. 5, str. 692-703. ISSN 1094-6977.

Univerze v Mariboru – DKUM. [COBISS.SI-ID 11399958] DOI: 10.1109/TSMCC.2012.2186802. [COBISS.SI-ID 15856918]

[7] WU, Hui, GRAY, Jeffrey G., MERNIK, Marjan. Grammar-driven [14] ČREPINŠEK, Matej, LIU, Shih-Hsi, MERNIK, Marjan. Replication and

generation of domain-specific language debuggers. Software : practice & comparison of computational experiments in applied evolutionary

experience. 2008, vol. 38, iss. 10, str. 1073-1103. ISSN 0038- computing : common pitfalls and guidelines to avoid them. Applied soft 0644. http://dx.doi.org/10.1002/spe.863, Digitalna knjižnica Univerze v computing. [Print ed.]. June 2014, vol. 19, str. 161-170. ISSN 1568-4946. Mariboru – DKUM, DOI: 10.1002/spe.863. [COBISS.SI-ID 11994134 DOI: 10.1016/j.asoc.2014.02.009. [COBISS.SI-ID 17638934]

[8] KOS, Tomaž, KOSAR, Tomaž, MERNIK, Marjan. Development of data [15] MERNIK, Marjan, LIU, Shih-Hsi, KARABOGA, Dervis, ČREPINŠEK,

acquisition systems by using a domain-specific modeling Matej. On clarifying misconceptions when comparing variants of the language. Computers in industry. [Print ed.]. Apr. 2012, vol. 63, no. 3, str. Artificial Bee Colony Algorithm by offering a new 181-192. ISSN 0166-3615. DOI: 10.1016/j.compind.2011.09.004. implementation. Information sciences. [Print ed.]. 10 Jan. 2015, vol. 291, [COBISS.SI-ID 15485974] str. 115-127. ISSN 0020-0255. DOI: 10.1016/j.ins.2014.08.040.

[9] KOSAR, Tomaž, BOHRA, Sudev, MERNIK, Marjan. Domain-specific [COBISS.SI-ID 18044438]

languages : a systematic mapping study. Information and software [16] VEČEK, Niki, MERNIK, Marjan, FILIPIČ, Bogdan, ČREPINŠEK, technology. [Print ed.]. March 2016, vol. 71, str. 77-91. ISSN 0950-5849. Matej. Parameter tuning with Chess Rating System (CRS-Tuning) for

DOI: 10.1016/j.infsof.2015.11.001. [COBISS.SI-ID 19181334] meta-heuristic algorithms. Information sciences. [Print ed.]. Dec. 2016,

[10] KOSAR, Tomaž, GABERC, Sašo, CARVER, Jeffrey C., MERNIK, vol. 372, str. 446-469. ISSN 0020-0255. DOI: 10.1016/j.ins.2016.08.066.

Marjan. Program comprehension of domain-specific and general-purpose [COBISS.SI-ID 19750166], languages: replication of a family of experiments using integrated [17] KOVAČEVIĆ, Željko, MERNIK, Marjan, RAVBER, Miha, development environments. Empirical software engineering. 2018, vol. ČREPINŠEK, Matej. From grammar inference to semantic inference - an

23, iss. 5, str. 2734-2763. ISSN 1382-3256. DOI: 10.1007/s10664-017- evolutionary approach. Mathematics. 2020, vol. 8, no. 5, str. 1-20. ISSN 9593-2. [COBISS.SI-ID 21123606] 2227-7390. https://www.mdpi.com/2227-7390/8/5/816,

[11] ČREPINŠEK, Matej, LIU, Shih-Hsi, MERNIK, Marjan. Exploration and DOI: 10.3390/math8050816. [COBISS.SI-ID 15805699]

exploitation in evolutionary algorithms : a survey. ACM computing [18] JEREBIC, Jernej, MERNIK, Marjan, LIU, Shih-Hsi, RAVBER, Miha, surveys. [Print ed.]. 2013, vol. 45, no. 3, str. 1-33. ISSN 0360-0300. BAKETARIĆ, Mihael, MERNIK, Luka, ČREPINŠEK, Matej. A novel

DOI: 10.1145/2480741.2480752. [COBISS.SI-ID 17009430] direct measure of exploration and exploitation based on attraction

[12] BREST, Janez, GREINER, Sašo, BOŠKOVIĆ, Borko, MERNIK, Marjan, basins. Expert systems with applications. [Print ed.]. 1 April 2021, vol.

ŽUMER, Viljem. Self-adapting control parameters in differential 167, str. 1-22. ISSN 0957-4174. DOI: 10.1016/j.eswa.2020.114353. evolution: a comparative study on numerical benchmark problems. IEEE [COBISS.SI-ID 41233155] transactions on evolutionary computation. [Print ed.]. dec. 2006, vol. 10, [19] RAVBER, Miha, LIU, Shih-Hsi, MERNIK, Marjan, ČREPINŠEK, Matej. no. 6, str. 646-657. ISSN 1089-778X. [COBISS.SI-ID 10376982] Maximum number of generations as a stopping criterion considered

[13] HRNČIČ, Dejan, MERNIK, Marjan, BRYANT, Barrett Richard. harmful. Applied soft computing. [Print ed.]. Oct. 2022, vol. 128, 20 str.

Improving grammar inference by a memetic algorithm. IEEE transactions ISSN 1568-4946. DOI: 10.1016/j.asoc.2022.109478. [COBISS.SI-on systems, man and cybernetics. Pt. C, Applications and reviews. [Print ID 118318595]



Prispevki Igor Bernika na področju informatike



Contributions of Igor Bernik in the field of informatics



Igor Bernik

Fakulteta za varnostne vede

Univerza v Mariboru

Ljubljana, Slovenija

igor.bernik@um.si



Povzetek njegovem delu zasedajo študenti, katerih dosežki so zanj predstavljali stalno motivacijo za nadaljnje delo in raziskovanje.

Prispevek povzema razvojno-raziskovalno, pedagoško in Mentoriral je skoraj 400 študentov, med katerimi so mnogi

vodstveno delovanje ter dosežke Igorja Bernika v okviru dosegli izjemne uspehe v svojih poklicnih karierah.

Fakultete za varnostne vede in Fakultete za organizacijske vede

Univerze v Mariboru. Njegovo delo spada na področja sistemov

za podporo odločanja ter informacijske in kibernetske varnosti. 2 Študij



Ključne besede računalniki in numerično krmiljeni stroji, ki so postajali vse bolj V poznih osemdesetih letih so se v strojništvo začeli uvajati prvi

Igor Bernik, pedagoška dejavnost, raziskovalna dejavnost, razširjeni. Med študijem strojništva se je prof. dr. Igor Bernik

informacijska in kibernetska varnost prvič srečal s študijskimi vsebinami s področja programiranja.



Abstract Zaradi uspešnosti pri študiju je prejel prve ponudbe za

sodelovanje s profesorji ter možnost dodatnega zaslužka prek



The paper summarizes the development-research, teaching and programiranja. V začetku 90. let se je srečal z izzivom, kako

management activities and achievements of Igor Bernik within razviti poslovni informacijski sistem za uporabnika, ki želi

the Faculty of Security Sciences and the Faculty of uporabljati računalnik, vendar nima znanja informatike. Leta

Organizational Sciences of the University of Maribor. His work 1991 se je, na podlagi te potrebe, odločil razviti poslovni

is in decision support systems, information, and cybersecurity. informacijski sistem za podjetje, pri čemer je izbral okolje MS

Access 1.0. Tako je razvil informacijski sistem, ki je bil verjetno

Keywords eden prvih poslovnih sistemov v Sloveniji, izdelanih v okolju

Igor Bernik, teaching, research, information and cybersecurity Windows. Med razvojem tega sistema je ugotovil, da je potrebno

pridobiti dodatno znanje o relacijskih bazah podatkov, kar ga je

pritegnilo k nadaljnjemu študiju organizacijske informatike.

1 Uvod

Igor Bernik je prvi stik z računalnikom doživel v poznih 80. letih Na Fakulteti za organizacijske vede Univerze v Mariboru je

prejšnjega stoletja, ko je sprva načrtoval kariero na področju uspešno zaključil študij organizacijske informatike, med katerim

strojništva. V tem obdobju se je srečal s programiranjem in se je še posebej posvetil poslovnim simulacijam in sistemom za

kmalu odkril svojo strast do tega področja. Postopoma se je vse podporo odločanju. Aktivno je sodeloval v različnih študentskih

bolj oddaljeval od tehničnih disciplin in začel usmerjati svojo projektih, njegov zaključni projekt, diplomsko delo na temo

pozornost v informatiko, pri čemer se je v zadnjih dveh razvoja sistema za zvezno simulacijo v C++ [1], pa mu je prinesel

desetletjih posebej posvetil informacijski varnosti. V zadnjih mesto mladega raziskovalca.

letih je svoje delo osredotočil na varno uporabo informacijske



tehnologije z namenom izboljšanja kakovosti življenja. Na svoji 3 Zaposlitev na Fakulteti za organizacijske vede strokovni poti je imel čast sodelovati s številnimi posamezniki,

ki so pomembno prispevali k njegovemu osebnemu in Kot mladi raziskovalec se je prof. dr. Igor Bernik med delom in

strokovnemu razvoju ter ga podpirali pri rasti na področju študijem skozi znanstveni magisterij in doktorat srečal tudi z

informacijske in kibernetske varnosti. Posebno mesto v metodami optimiranja in umetno inteligenco. Imel je srečo, da je

sodeloval z odličnimi, danes legendarnimi profesorji

∗Article Title Footnote needs to be captured as Title Note informatike. Njegov formalni mentor je bil prof. dr. Miroljub

† Author Footnote to be captured as Author Note Kljajič, veliko pa je sodeloval in se učil tudi od prof. dr.

Permission to make digital or hard copies of part or all of this work for personal or Vladislava Rajkoviča ter prof. dr. Roberta Leskovarja, ki je bil

classroom use is granted without fee provided that copies are not made or distributed njegov delovni mentor. Na podlagi opravljenih raziskav je s

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must kolegi objavil izsledke, ki so še danes med visoko citiranimi [2,

be honored. For all other uses, contact the owner/author(s). 3].

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).

http://doi.org/DOI_RECEIVED_AFTER_REVIEW



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia I. Bernik



Po zaključenem usposabljanju za mladega raziskovalca ostal mnogih strokovnih združenj. Za svoje delo je prejel več domačih

na Fakulteti za organizacijske vede, sprva kot asistent, kasneje in mednarodnih priznanj.

kot docent. Kot učitelj je sledil razvoju stroke in poleg obstoječih

področij poučevanja razvijal tudi nova, kot so managerski in

kadrovski informacijski sistemi. V tem obdobju se je prvič srečal 5 Zaključek

s potrebo po varovanju podatkov, najprej osebnih, nato tudi Prof. dr. Igor Bernik je profesor, ki si prizadeva, da bi se področje

poslovnih. Ugotovil je, da razvoj informacijskih sistemov informacijske in/ali kibernetske varnosti osamosvojilo od

pogosto zanemari vidike varnosti, zato se je začel posvečati informatike in varstvoslovja ter ga predstavlja kot ključno vedo

vprašanju, kako zagotoviti varnost informacijskih sistemov, pri prihodnje digitalizirane družbe [10, 11]. Pri tem se osredotoča

čemer je sodeloval tudi s Fakulteto za varnostne vede. tudi na digitalno vključenost vseh državljanov v trajnostni

razvoj.



4 Zaposlitev na Fakulteti za varnostne vede

Leta 2008 je bil prof. dr. Igor Bernik povabljen, da svojo Zahvala

strokovno pot nadaljuje na Fakulteti za varnostne vede Univerze Iskrena zahvala vsem financerjem raziskovalnih projektov, ki so

v Mariboru. Fakulteta je le nekaj let prej (leta 2003) izstopila iz omogočili naše delo in napredek, ter vsem kolegicam in

sistema Ministrstva za notranje zadeve in se pridružila Univerzi kolegom, ki so s svojim znanjem, podporo in novimi izzivi

v Mariboru, razvojno pa je poskušala (in po dosedanjih rezultatih oblikovali mojo pot. Brez vaše podpore in zaupanja ne bi bilo

uspešno) proučevati področje varnosti oziroma varstvoslovja mogoče doseči vsega, kar smo skupaj ustvarili. Vaša

skozi znanstveno-raziskovalno delo in prenašati vrhunsko znanje prizadevanja in predanost so pustili neizbrisno sled v mojem

na študente. Idejni vodji razvoja fakultete, takratna dekana prof. življenju in delu, za kar sem vam globoko hvaležen.

dr. Milan Pagon in prof. dr. Gorazd Meško, sta prepoznala

prihodnost v formalnem proučevanju informacijske varnosti. Literatura

Tako je bil poleg pedagoških in raziskovalnih zadolžitev prof. dr. [1] Bernik, Igor. 1996. Razvoj sistema za zvezno simulacijo v C++.

Bernik postavljen pred izziv vzpostavitve študija informacijske Diplomsko delo visokošolskega študija. Kranj: [I. Bernik], 65 f., grafični

varnosti na fakulteti. S sodelovanjem Fakultete za prikazi.

[2] Kljajić, Miroljub, Bernik, Igor, Škraba, Andrej. 2000. Simulation

elektrotehniko, računalništvo in informatiko je bil pripravljen approach to decision assessment in enterprises. Simulation. [Print ed.], vol.

interdisciplinarni visokošolski program Informacijska varnost, ki 75, no. 4, str. 199–210. ISSN 0037-5497.

je bil akreditiran avgusta 2010, prvi študenti pa so bili vpisani že [3] Bernik, Igor, Kljajić, Miroljub, Bernik, Mojca. 2002. Use of simulation

and visualization in multi criteria scheduling optimization with genetic

leta 2011. Leta 2014 je bil dr. Bernik mentor prvemu diplomantu algorithms. V: Mastorakis, Nikos E. (ur.). Recent advances in simulation,

[4] tega programa. computational methods and soft computing. [S. l.]: WSEAS Press, str.

246–250. A series of reference books and textbooks. ISBN 960-8052-50-

5.

Vsebinsko je razvil področje informacijske varnosti in postavil [4] Urbanc, Jan. 2014. Utrjevanje HTTP sej z odtisom brskalnika: diplomsko

delo visokošolskega strokovnega študija Informacijska varnost.

temelje znanstveno-raziskovalni dejavnosti družboslovnega [Ljubljana: J. Urbanc].



proučevanja informacijske varnosti. Iz tega področja je objavil [5] Bernik, Igor, Prislan Mihelič, Kaja. 2012. Kibernetska kriminaliteta, informacijsko bojevanje in kibernetski terorizem . Ljubljana: Fakulteta za tudi nekaj temeljnih del v slovenskem jeziku [5, 6, 7]. Njegova varnostne vede. Serija Informacijska varnost, 1. ISBN 978-961-6821-20-

raziskovalna dejavnost se je utrdila v mednarodni skupnosti 9.

informacijske varnosti, zlasti po izdaji samostojne znanstvene [6] Markelj, Blaž, Bernik, Igor. 2018. Varnost mobilnih naprav. 1. izd.

Maribor: Univerzitetna založba Univerze v Mariboru; Ljubljana: Fakulteta

monografije leta 2014 pri založbi Wiley [8]. za varnostne vede. ISBN 978-961-286-138-4.

[7] Prislan Mihelič, Kaja, Bernik, Igor. 2019. Informacijska varnost in

Leta 2010 se je na področju informacijske varnosti pridružil organizacije. 1. izd. Maribor: Univerzitetna založba Univerze v Mariboru;

Ljubljana: Fakulteta za varnostne vede. ISBN 978-961-286-253-4.

tudi Blaž Markelj, ustanovljena je bila katedra za informacijsko [8] [5] Bernik, Igor. 2014. Cybercrime and cyberwarfare. London: ISTE;

varnost, ki je v petnajstih letih dosegla številne uspehe [9] in Hoboken: Wiley. ISBN 978-1-84821-671-6.

[9] Markelj, Blaž, Vrhovec, Simon, Bernik, Igor. 2023. Izobraževanje in

danes šteje štiri člane (pridružita se še Simon Vrhovec, 2014 in raziskovanje na področju informacijske varnosti = Education and research

Anže Mihelič, 2018). Poleg pedagoškega in znanstveno- in the field of information security. V: Meško, Gorazd (ur.). Znanstveno

raziskovanje in pedagoška dejavnost Fakultete za varnostne vede

raziskovalnega dela je bil prof. dr. Bernik tudi prodekan na Univerze v Mariboru: (1973–2023). 1. izd. Maribor: Univerza v Mariboru,



področjih izobraževanja, znanstveno-raziskovalnega dela, Univerzitetna založba. Str. 245-260. [10] Bernik, Igor. 2024. Kibernetski prostor in kibernetska varnost v luči kakovosti in razvoja, od leta 2023 pa je dekan Fakultete za trajnostnega razvoja: sinergija lokalnih skupnosti v družbi 5.0. V: Meško, varnostne vede Univerze v Mariboru. Gorazd (ur.), Eman, Katja (ur.). Varnost v lokalnih skupnostih -

multidisciplinarne perspektive. 1. izd. Maribor: Univerza v Mariboru,

Univerzitetna založba; Ljubljana: Univerza v Mariboru, Fakulteta za

Prof. dr. Bernik je (so)avtor šestih monografij, 27 poglavij v varnostne vede. Str. 313-338. ISBN 978-961-286-876-5, ISBN 978-961-



znanstvenih monografijah, 42 izvirnih in preglednih znanstvenih 286-875-8. [11] Bernik, Igor. 2023. Izzivi informacijske varnosti pri prehodu v družbo 5.0. člankov ter več sto drugih del, njegova bibliografija pa šteje 625 V: Markelj, Blaž (ur.). Informacijska varnost: doba tehnoloških prebojev

enot. Sodeloval je pri številnih projektih in recenzijah ter je član in pravnih izzivov. 1. natis. Ljubljana: Lexpera, GV Založba. Str. 13-34,

169. ISBN 978-961-247-517-8.



Dosežki kandidata Marka Bajca



Marko Bajec

marko.bajec@fri.uni- lj.si

Univerza v Ljubljani, Fakulteta za računalništvo in informatiko

Ljubljana, Slovenia



Povzetek okrilju IJS), kjer sem pomagal pri izvajanju računalniški tečajev

ter pomalem pri razvoju računalniških rešitev. Če ne bi prejel po-

V tem prispevku na kratko predstavim svojo dosedanjo kari-

vabila s fakultete, da se zaposlim kot asistent-stažist, bi verjetno

erno pot ter dosežke, povezane z računalništvom in informatiko.

končal v industriji. Tako pa sem, zopet skoraj da slučajno, pristal

Prispevek sem napisal kot kandidat za člana v “Legendah raču-

v akademskih vodah. V začetku sem se nekoliko lovil, a sčasoma

nalništva”. Povabila sem bil vesel, a priznam, da me spravilo v

sem dojel ustroj akademskega sveta, se mu prilagodil in po nekaj

zadrego. Biti na nekem področju “legenda”, je zares visok druž-

vzponih in padcih s 43 leti postal redni profesor.

beni status, ob katerem mnogi dosežki, ki se sicer znajdejo v naših

življenjepisih, zgubijo svoj blišč. Z nekaj samokritike zato kratek

opis moje karierne poti in tistih rezultatov, na katere sem sam 1.2 Od začetka - zaposlitev na Fakulteti za

ponosen. računalništvo in informatiko

Fakulteta za računalništvo in informatiko Univerze v Ljubljani (v

Abstract nadaljevanju UL FRI) je bila moj prvi in edini delodajalec. Začel

In this paper I briefly describe my career path and achievements sem kot asistent v Laboratoriju za informatiko, ki ga je takrat

so far related to computer science and informatics. I wrote the vodil prof. dr. Marjan Krisper. Njegova sposobnost za pridobi-

paper as a candidate for membership in the ’Legends of Comput- vanje aplikativno-raziskovalnih projektov je presegala takratne

ing’. While I was happy to receive the invitation, I must admit it storilne zmožnosti laboratorija, saj je ob mojem prihodu imel

put me in an awkward position. Being a ’legend’ in a particular samo enega člana. Raziskovalno delo je zato nekoliko trpelo, a

field is truly a high social status, at which many achievements projekti so bili zanimivi in z vsakim sem se nekaj novega naučil.

that usually make it to our resumes lose their shine. With a bit Med drugim, kako težko je razvijati rešitve za realna okolja, kjer

of self-criticism, here is a brief description of my career path and so večji izziv ljudje kot pa sama kompleksnost tehnologije.

those results that I am personally proud of.

1.3 Vključevanje v domače in mednarodne

Ključne besede kroge

karierna pot, raziskovalno delo, pedagoško delo, vodstvene funk-

Po zaposlitvi na UL FRI sem se sčasoma tudi sam vključil v do-

cije, dosežki

mače in mednarodne raziskovalne ter tudi strokovne kroge. Kot

član programskega odbora sem sodeloval pri številnih mednaro-

Keywords dnih konferencah, kot so: CAiSE, IEEE RE, IEEE RCIS, ICSOFT,

career path, research work, teaching work, management, achieve- POEM, DB &IS, ADBIS, ICSEA, IBIMA, ME, MUSE, ICIST, OTI ipd.

ments V letu 2009 sem bil povabljen v programski svet priznane medna-

rodne konference CAiSE ( International Conference on Advances in

1 Karierna pot Software Engineering), ki je takrat veljala za eno pomembnejših

1.1 na mojem ožjem raziskovalnem področju. V letu 2013 sem bil Na kratko

izbran za programskega predsednika IEEE konference RCIS 2014

S “pravimi” računalniki sem se spoznal šele v srednji šoli – Teh- Research Challenges in Information Science ( ) in isto leto tudi za

nični srednješolski center Branka Breliha, Nova Gorica. Vpisal

položaj generalnega predsednika konference CAiSE 2016, ki je

sem se, ne da bi zares vedel, kam se vpisujem. V osnovni šoli

z več kot 300 udeleženci potekala v Ljubljani. V letu 2018 sem v

sem res kazal več želje in nadarjenosti za naravoslovno tehnične

vlogi generalnega predsednika vodil IEEE konferenco RCIS 2018

predmete, vsaj tako so zatrjevale moje učiteljice, a izrazito me ni

v Nantesu. Kot recenzent ali urednik sem sodeloval pri številnih

vleklo nikamor. Še danes večkrat pomislim, da sem računalničar

revijah in konferencah.

po naključju.

Moje delo v domačih krogih je najbolj zaznamovalo sodelova-

V srednji šoli sem se o računalništvu veliko naučil, predho-

nje v Slovenskem društvu informatika. V letih 2005-2013 sem bil

dnega znanja o tem področju namreč nisem imel, a največ mi

aktivno vključen v organizacijo največje slovenske konference

je dala fakulteta. Za razliko od srednje šole, kjer sem bil tako

o informatiki (DSI), ki je takrat s 500 in več udeleženci potekala

imenovani “hardverist” (strojna opema), sem se na faksu raje

pod okriljem društva. V letih 2005-2007 sem vodil programski od-

odločil za programsko opremo. Programiranje me v resnici bolj

bor konference in od 2008 do 2020 sodeloval kot član upravnega

veselilo kot pa računalniško drobovje. V zadnjih letnikih sem

odbora društva.

veliko časa prebil na IJS (v resnici v podjetju, ki je delovalo v



Permission to make digital or hard copies of all or part of this work for personal 1.4 Ustanovitev Laboratorija za podatkovne

or classroom use is granted without fee provided that copies are not made or tehnologije distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this Leta 2009 sem na UL FRI ustanovil Laboratorij za podatkovne

work must be honored. For all other uses, contact the owner /author(s). tehnologije

. To je bil pomemben mejnik v moji karieri, ki je

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Optilab

sovpadal tudi z ustanovitvijo spin-off podjetja . S skupino

© 2024 Copyright held by the owner/author(s).

raziskovalcev (večinoma mladi raziskovalci in nekateri sodelavci



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia M. Bajec



Laboratorija za informatiko) smo se na pobudo in povpraševanja prevajanje za jezikovni par Slovenščina, Angleščina ter razvili

5

iz gospodarstva poglobljeno ukvarjali z analizo podatkov, pred- nevronski sintetizator za dva moška in en ženski glas . Visoka

vsem s področja avtomobilskega in zdravstvenega zavarovanja. natančnost modelov prepoznave in prevajalnikov je nakazovala,

Razvijali smo algoritme in na njih temelječe informacijske re- da so govorne tehnologije končno dozorele tudi za majhne jezike,

šitve, ki so bile v pomoč pri iskanju namernih ali nenamernih kot je Slovenščina. Tehnologijo smo zato promovirali in pred-

nepravilnosti v poslovanju. Področje je bilo raziskovalno in apli- stavili v različnih okoljih in s pilotskimi projekti pokazali, kaj

kativno zelo zahtevno in je dobro desetletje usmerjalo naše delo, zmore, glej npr. [4, 5]. Univerza v Ljubljani nam je v želji po večji

tako v laboratoriju kot tudi v podjetju, s katerim smo ohranili internacionalizaciji študija leta 2020 zaupala izjemno pomemno

tesne povezave. V tem obdobju sta nastali dve odlični doktorski nalogo, in sicer razvoj sistema za sprotno prevajanje slovenskih

nalogi [7, 3]. Leta 2011 nam je ARIS podelila priznanje za izjemne predavanj v tuje jezike [2, 6]. Nastala je rešitev (kasneje Tolmač

dosežke, povezana z omenjenim področjem [8]. preimenovana v ), ki združuje vse ključne govorne Online Notes

6

Število članov se je v laboratoriju vsa leta povečevalo in s tehnologije, prepoznavo slovenskega jezika, prevajanje iz sloven-

prihodom novih raziskovalcev so se širil tudi portfelj naših razi- ščine v tuje jezike ter za slepe in slabovidne tudi sintezo besedila.

skovalnih interesov. Med drugim smo se ukvarjali s konceptom Študenti, ki ne morejo slediti slovenskim predavanj, lahko prek

pametnih mest in v okviru tega z elektronskim in mobilnim spletne ali mobilne aplikacije spremljajo prevode v živo, kasneje

zdravstvom kot enim izmed pomembnih stebrov omenjenega pa lahko posnetke pregledujejo in si urejajo svoje zapiske ter

koncepta. V konzorciju s 25 podjetji in raziskovalnimi organizaci- delijo z drugimi študenti. Rešitev je bila pilotno preskušena na

jami smo razvijali različne rešitve, naloga našega laboratorija pa šestnajst članicah Univerze v Ljubljani in s šolskim letom 2024/25

je bila povezovanje vseh rešitev v skupno prehaja v produkcijsko delovanje. platformo pametnega

mesta [1]. Eden od rezultatov, ki je nastal v okviru omenjenega Zaradi velikega povpraševanja po tovrstnih rešitvah smo leta

sodelovanja, je platforma za spremljanje pacientov na 2018 z dovoljenjem upravnega odbora UL ustanovili spin-off Vi-

Vitabits 1

daljavo (telemedicina). Z dovoljenjem UL smo pravice intelek- , ki je danes vodilno podjetje na področju razvoja govornih tasis

7

tualne lastnine kasneje prodali zainteresiranemu podjetju, ki je tehnologij. Rešitve podjetja Vitasis se uporabljajo v zdravstvenih

rešitev plasiralo na trg. domovih, bolnišnicah, občinah, Državnem zboru, SDH, Telekomu

Veliko napora smo v takratnem obdobju vlagali tudi v preuče- Slovenije in v drugih ustanovah. Govorne in jezikovne tehnolo-

vanje (IoT) in aplicirali svoja dognanja v okviru gije ostajajo prednostno raziskovalno področje laboratorija tudi interneta stvari različnih aplikativnih nalog. Z željo, da zainteresirani jasnosti pri- v prihodnje, saj smo vpeti v tri večje nacionalne projekte:



kažemo širino področja in številne rešitve, ki so postale mogoče z •

Temeljne raziskave za razvoj govornih virov in tehnologij

razmahom senzorike in sposobnostjo komunikacije med skorajda 8 MEZZANINE

za slovenski jezik ( ), ARIS, 2022-2025,

slehernimi napravami, smo ustanovili . V nje- •

IoT Demo Center2

Prilagodljiva obdelava naravnega jezika s pomočjo velikih

govem okrilju smo izvedli številne zanimive zaključne naloge 9 PoVeJMo

jezikovnih modelov ( ), ARIS, 2023-2025 ter

in izvedli več projektov po naročilu iz gospodarstva in državne • LLM4DH Veliki jezikovni modeli za digitalno humanistiko ( ),

uprave. Med zanimivejše sodi platforma , 2019-20 -Common

ARIS, 2024-2026

Community Monitoring Platform, ki omogoča merjenje "utripa 10

Sodelujemo tudi v zanimivem projektu -Proteverb Pravni,

družbe"prek avtomatskega spremljanja digitalnih medijev [9].

etični in tehnološki vidiki obdelave besedilnih in govornih virov

Analiza podatkov ločuje med posameznimi družbenimi podsku-

pinami, kot so podatkov za znanstvene, raziskovalne in razvojne namene, MDP, , delovno aktivna populacija upokojenci , invalidi ,

študenti 2022-2024, katerega cilj je preučiti vse relevantne vidike in pripra-

ipd. in omogoča avtomatsko zaznavo tem, s katerimi

viti podlago, ki bo raziskovalni sferi omogočala lažji ter obenem

se posamezne družbene podskupine ukvarjajo ter njihov odnos

varen dostop do podatkov, ki so lahko ključni za raziskave in

do teh tem. Drugo pomembno platformo smo razvili v okviru

razvoj novih tehnoloških rešitev.

projekta Idejna zasnova in postavitev pilota ekosistema interneta

Če smo z domačim gospodarstvom ves čas tesno sodelovali, pa

stvari z algoritmičnimi orodji, naročnik MJU, (2023-24). IoT plat-

forma je bila naša vključenost v mednarodne projekte (predvsem imam

, ki je v okviru projekta nastala, predstavlja splošno rešitev

tu v mislih projekte iz EU programov FP6, FP7 in Horizon) v prvih

za zajem podatkov v okviru državne uprave, njihov nadzor in

desetih letih obstoja laboratorija precej manjša. Zavedajoč se, da

upravljanje. V okviru pilotskega projekta je bila v sodelovanju

so za pridobivanje in še posebej izvajanje EU projektov potrebna

z Direktoratom za stvarno premoženje na MJU njena uporab-

specifična znanja, smo se po 2019 kadrovsko okrepili prav z na-

nost demonstrirana na primeru zbiranja podatkov za potrebe

menom, da pridobimo kompetence tudi na tem področju. Danes

zagotavljanja energetske učinkovitosti stavb.

smo s sodelavci laboratorija vključeni v šest evropskih projektov,

Zadnji večji zasuk v raziskovalnem kurzu našega laborato-

11 12

EBSI-VECTOR (2023-25) , BUILDCHAIN (2023-25) , TRUST-

rija se je zgodil v letih 2019-2020, ko smo se začeli ukvarjati z

13 14 15

CHAIN (2023-25) , ACES (2023-25) , EXTREMEXP (2023-25)

analizo besedil in sčasoma s široko paleto raziskovalnih tem na 16

področju in predvsem (strojno govornih tehnologij

jezikovnih in ESSA (2020-24) .

prevajanje, računalniška prepoznava govora, računalniška sin-

5

teza govora). V okviru projekta (Razvoj Slovenščine v 6 RSDO

3 https://tts.true- bar.si/

https://www.youtube.com/watch?v=hWRONPdHh3o&ab_channel=CJVT

digitalnem okolju, MK, 2019-2022) smo razvili model za prepo- 7 https://vitasis.si/home

8

znavo slovenskega govora, ki je v primerjavi s prejšnjim, edinim https://mezzanine.um.si/

9

https://povejmo.si/

javno dostopnim modelom, dosegel bistveno višjo natančnost

10

https://www.inst- krim.si/project/proteverb/

4

prepoznave . Pomembno smo izboljšali tudi model za strojno 11

https://www.ebsi- vector.eu/en/

12

https://buildchain- project.eu/

1 13

https://vitabits.eu/ https://trustchain.ngi.eu/

2 14

http://iot.data- lab.si/ https://www.aces- edge.eu/

3 15

https://rsdo.slovenscina.eu/ https://extremexp.eu/

4 16

https://slobench.cjvt.si/ https://softwareskills.eu/



Dosežki kandidata Marka Bajca Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



2 Pedagoško delo Nenazadnje, ponosen sem na svoj laboratorij, Laboratorij za

S pedagoškim delom sem se začel ukvarjati že pred zaposlitvijo na podatkovne tehnologije, ki kljub velikim izzivom s pridobi-

vanjem kakovostnega kadra ves čas raste in danes šteje 32 razi-

UL FRI, saj sem predaval na računalniških tečajih. Predavanja so

skovalcev, kar ga tako po številu članov kot finančnih prihodkih

mi bila vedno v veselje, če sem se le čutil suveren na področju, ki

uvršča na prvo oziroma drugo mesto na fakulteti.

sem ga predaval. Posledično sem v pripravo na predavanja vedno

veliko vlagal. V prvih letih dela na fakulteti sem izvajal števila

izobraževanja za gospodarstvo, s povečevanjem pedagoške obre- 5 Zaključek

menitve na lastni fakulteti pa sem to sčasoma opustil. Danes sem Kot sem zapisal na začetku, sem postal računalničar bolj kot ne

nosilec treh predmetov, in po naključju. A to naključje štejem kot srečno, saj danes zagotovo Razvoj informacijskih sistemov Osnove

podatkovn ih baz na dodiplomskem univerzitetnem študeuiju ter vem, da me računalništvo oziroma konkretneje reševanje izzivov

Iskanje in ekstrakcija podatkov s spleta na podplomskem študiju. z računalnikom zares veseli. To nedvomno prispeva tudi k dru-

Kot mentor sem sodeloval pri nekaj sto diplomskih in magi- gim vidikom mojega profesionalnega udejstvovanja. V prispevku

strskih nalogah ter kot mentor usmerjal 10 doktorskih študentov. sem izpostavil nekaj svojih dosežkov, ki mi ta trenutek največ

Slednje mi je bilo vedno v veliko veselje, čeprav je mentoriranje pomenijo. Črto pod svoje življenjsko delo pa bom lahko potegnil

doktorskega študenta vse prej kot enostavna naloga. Danes lahko šele čez dobrih 13 let. V tem času me zagotovo čaka še veliko

z zadovoljstvom spremljam njihov razvoj in uspešne kariere: dva izzivov.

sta na pomembnih položajih v gospodarstvu, eden je lansko leto

postal redni profesor, še dva pa sta zelo blizu te habilitacijske Reference

stopnje. [1] Marko Bajec. 2018. Smart city as a platform-based ecosystem. V Databases

and information systems X : selected papers from the Thirteenth International

Baltic Conference (DB&IS 2018). IOS Press, Amsterdam; Berlin; Washington

(DC).

[2] Marko Bajec, Iztok Lebar Bajec, Tjaša Šoltes, Jernej Cvek, Jaka Čibej, Kaja

3 Vodstvene funkcije Gantar, Sara Sever in Simon Krek. 2023. Online notes - a real-time speech

recognition and machine translation system for slovene university lectures.

Vključevanje v vodstvene strukture je nekaj, čemur se profesor V Matjaž Debevc (ur.), Ines Kožuh (ur.). Digitalna vključenost v informacijski na Univerzi v Ljubljani težko izogne. Tako sem bil prodekan (IS 2023). družbi

[3] Štefan Furlan. 2012. Method for insurance fraud management systems design

za gospodarske zadeve en mandat in pol, od leta 2013 do 2019, based on business performance metrics. Doktorska disertacija. University of

član Senata UL FRI vse od leta 2013, član Upravnega odbora UL Ljubljana, Ljubljana, Slovenia.

[4] P. G. 2021. Marko bajec: »moramo se naučiti sodelovati!«. (Sep. 2021). Retrie-

FRI od leta 2010 naprej ter od 2019 naprej njegov predsednik,

ved September 10, 2024 from https://www.rtvslo.si/kultura/drugo/pilotni- pr

predsednik Komisije za Razvoj informacijskega sistema UL (KRIS)

ojekt- samodejnega- podnaslavljanja- tv- programov- pri- katerem- sodeluje- r

vse od leta 2016 naprej. Laboratorij za podatkovne tehnoloije tv- slovenija/595660.

[5] K. H. 2024. Najpogostejši glasovni ukazi so pop tv, masterchef in skrito v raju.

vodim od njegove ustanovitve leta 2009. Od leta 2016 vodim IoT

(Sep. 2024). Retrieved September 10, 2024 from https://www.24ur.com/novic

Demo Center na UL FRI. Vmes sem en mandat vodil Katedro za e/znanost- in- tehnologija/najpogostejsi- glasovni- ukazi- so- pop- tv- masterc

informatiko. V obeh spin-off podjetji sem nastopal kot prokurist, hef - in- skrito- v- raju.html.

[6] Tjaša Šoltes, Jan Vasiljević in Marko Bajec. 2024. Sistem za razpoznavo go-

kar je s stališča UL oziroma UL FRI dovoljeno, saj ne vstopa na

vora in strojno prevajanje v realnem času na ravni univerzitetnih predavanj.

področje konkurenčne prepovedi. , 32, 1. Uporabna informatika

[7] Lovro Šubelj. 2013. Detecting groups of nodes in large real-world networks using

label propagation. Doktorska disertacija. University of Ljubljana, Ljubljana,

Slovenia.

[8] Lovro Šubelj, Štefan Furlan in Marko Bajec. 2011. An expert system for

4 Dosežki detecting automobile insurance fraud using social network analysis. Expert

Systems with Applications, 38, 1, 1039–1052. doi: https://doi.org/10.1016/j.es

Med najpomembnejše dosežke v svoji karieri štejem nagrado

wa.2010.07.143.

Mentor leta, ki mi jo podelilo Slovensko društvo doktorskih [9] Ajda Pretnar Žagar, Dan Podjed, Marko Bajec in Slavko Žitnik. 2020. Sen-

timeter : interdisciplinarni pristop k izdelavi medijskega portala. Uporabna

študentov leta 2014. Nagrade sem bil izjemno vesel, saj mi je

informatika, 28, 3.

dala potrditev, da me doktorski študenti spoštujejo in cenijo

moja usmerjanja, v kar sem sam pogosto dvomil. Pod mojim

mentorstvom je do danes doktoriralo osem študentov, še dva pa

sta na tej poti.

Ponosen sem tudi na priznanja, ki sem jih prejel za svoj dopri-

nos na področju iz akademskega okolja v go-prenosa znanja

spodarstvo. Leta 2010 mi je takšno priznanje podelilo Slovensko

društvo informatika, leta 2013 pa programski svet mednarodne

konference Informacijska družba.

Pomembno mesto med nagradami ima tudi za Zlata plaketa

izjemne dosežke na pedagoškem in znanstvenem področju, ki mi

jo je Univerza v Ljubljani podelila leta 2014.

Med ključne dosežke s ponosom štejem tudi dve pod-spin-off

jetji, ki sem ju v svoji karieri pomagal ustanoviti in sta nastali

na temelju odličnih raziskovalnih rezultatov moje raziskovalne

skupine. Čeprav se z veseljem ukvarjam s temeljnimi raziska-

vami, mi je še v večje veselje, ko iz raziskav nekaj konkretnega

nastane in pripomore k reševanju praktičnih problemov. Pri obeh

podjetjih sem bil vključen v njun začetni razvoj in pomagal pri

njunem strateškem usmerjanju ter pozicioniranju na trgu.



Kratek življenjepis



Borut Robič

Fakulteta za računalništvo in informatiko

Univerza v Ljubljani

Slovenia



Povzetek učitelji Boštjan Vilfan (algoritmi, izračunljivost, računska zah-

Prispevek osvetli preteklo življenjsko pot Boruta Robiča skozi tevnost, algebra), Zvonimir Bohte (numerična analiza, končna

razna področja matematike, računalništva in glasbe. aritmetika) in Anton Suhadolc (funkcionalna analiza). Leta 1987

je z odliko magistriral pod mentorstvom Boštjana Vilfana z delom

Abstract Minimizacija števila procesorjev v podatkovno pretokovni arhitek-

turi. To je problem kombinatorične optimizacije v vzporednem

The article sheds light on Borut Robič’s past journey through

računanju, kjer želimo s primernim zadrževanjem pripravljenih

various areas of mathematics, computer science, and music.

ukazov minimizirati število uporabljenih procesorjev, ne da bi se

Ključne besede čas izvajanja vzporednega programa podaljšal preko teoretično

najmanjšega možnega.

optimization, algorithm, parallel computing, parallel algorithm Z doktorskim študijem je Borut nadaljeval 1988 po služenju

mapping, computational complexity, computability vojaškega roka. Leta 1993 je doktoriral s področja računalniških



Keywords znanosti na UL FER pod mentorstvom Boštjana Vilfana z diserta-

cijo Optimizacija preslikav vzporednih algoritmov na regularna

optimizacija, algoritem, vzporedno računanje, preslikavanje vz- procesorska polja. Gre za NP-težek optimizacijski problem v vzpo-

porednih algoritmov, računska zahtevnost, izračunljivost rednem računanju, kjer želimo poljuben vzporedni program –



1 Osnovno zave pa podatkovne odvisnosti med ukazi – preslikati na množico predstavljen z usmerjenim grafom, katerega točke so ukazi, pove-

Borut Robič se je rodil 12. julija 1960 v Postojni. Osnovno šolo je heksagonalno, kot v satovju povezanih procesnih elementov, da

do 1972 obiskoval v Idriji in v prostem času v vlogi Vinetouja ali bodo komunikacijske poti med preslikanimi ukazi minimalne. Ker

Old Shatterhanda vzdrževal red in pravičnost na tamkajšnjih uli- je procesnih elementov končno mnogo, lahko s takim preslikava-

cah. Šolanje je nadaljeval na OŠ Franceta Bevka v Ljubljani. Tam njem na procesorsko polje naložimo večje vzporedne programe

mu je natančna in zahtevna učiteljica Bogomila Hude privzgojila ali pa sočasno več vzporednih programov.

ljubezen do matematike, za kar ji je neskončno hvaležen. Kot dobitnik podoktorske štipendije Partridge Visiting Fello-

Po vpisu na gimnazijo Bežigrad v Ljubljani, v razred z inten- wship, ki mu jo je podelila Univerza v Cambridgu, je 1996 gostoval

zivnim programom matematike, je njegovo zanimanje za to lepo na Fitzwilliam Collegeu in University of Cambridge Computer

vedo še naraščalo, zdaj pod vplivom profesorice Sonje Plevnik. Laboratory. Tu je obravnaval problem usmerjanja podatkov v

Leta 1978 je z odliko maturiral. 2-krožnih omrežjih tj. neusmerjenih grafih 𝐺(𝑛; 𝑎; 𝑏) s točkami

Pravijo, da se matematika in glasba dopolnjujeta; res, tudi Bo- {0, 1, . . . , 𝑛−1} in povezavami {𝑖, 𝑖+𝑎(mod 𝑛)} in {𝑖, 𝑖+𝑏 (mod 𝑛)}.

rut je sočasno obiskoval in končal nižjo in srednjo glasbeno šolo



Franca Šturma v Ljubljani. 3 Raziskovalna in strokovna dejavnost

Leta 1984 se je zaposlil na Institutu Jožef Stefan v Ljubljani, na

2 Študij in izpopolnjevanja Odseku za računalništvo in informatiko, leta 1994 na Odseku

Po kolebanju ali naj na UL FMF študira matematiko ali fiziko je za računalniške sisteme, od 1997 pa je zaposlen na Fakulteti za

računalništvo in informatiko, UL.

po nasvetu staršev izbral študij elektrotehnike na UL FE. Pravijo

tudi, da stara ljubezen ne zarjavi; in res, že v prvem semestru se Raziskovalno je deloval na področjih optimizacije podatkovno

je odločil, da bo med kasnejšimi študijskimi usmeritvami izbral pretokovnega računanja; vzporednega računanja; razvoja in ana-

Računalništvo in informatiko, kjer je bilo videti največ zanimivih, lize algoritmov za probleme kombinatorične optimizacije; teorije

matematično obarvanih predmetov. računske zahtevnosti in teorije izračunljivosti.

Do diplome so nanj strokovno in osebnostno močno vplivali Bil je vodja raziskovalnega programa in vodja ali sodelavec

mnogi visokošolski učitelji, posebej Gabrijel Tomšič (matema- domačih in mednarodnih raziskovalnih projektov, slednjih pred-

tika), Boštjan Vilfan (algoritmi in podatkovne strukture), Ivan vsem z Univerzo v Grenoblu ter njenima institutoma TIMA in

Bratko (algoritmi in podatkovne strukture), Dušan Kodek (raču- LMC-IMAG. Tu je nanj močno vplival Denis Trystram, vodja in-

nalniške arhitekture). Pod mentorstvom Tomaža Pisanskega je stitutov. Skupne raziskave so se nanašale na preslikavanje vzpo-



1984 z odliko diplomiral in za to prejel dekanovo priznanje. rednih algoritmov na večprocesorske sisteme raznih topologij. štva in informatike na isti fakulteti. Zdaj so nanj močno vplivali Tega leta se je Borut vpisal na magistrski študij Računalni- Na področju teorije izračunljivosti je s konstrukcijo bijektivne ∗ preslikave med univerzumoma Σ in N potrdil, da lahko obrav-

navo odločljivosti, polodločljivosti ali neodločljivosti formalnih

Permission to make digital or hard copies of all or part of this work for personal jezikov nadomestimo z obravnavo odločljivosti, polodločljivosti

or classroom use is granted without fee provided that copies are not made or ali neodločljivosti množic naravnih števil, in obratno. Ugotovitve,

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this ki veljajo v enem, veljajo tudi v alternativnem univerzumu.

work must be honored. For all other uses, contact the owner/author(s). Raziskovalno delo je dokumentiral kot avtor ali soavtor znan-

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia stvenih monografij v tujini (4) in doma (2), izvirnih znanstvenih

© 2024 Copyright held by the owner/author(s).

člankov (45) in prispevkov na znanstvenih konferencah (61).



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia B.Robič



Navedimo le znanstvene monografije [1, 2, 3, 4] v tujini: Literatura

• [1] Jurij Šilc, Borut Robič, Theo Ungerer. Processor Architecture. From Dataflow to J. Šilc, B. Robič, T. Ungerer, Processor Architecture

(Springer, 1999, 411 str.); Superscalar and Beyond. Springer, 1999

[2] Borut Robič. The Foundations of Computability Theory. Springer, 2015



• [3] Roman Trobec, Boštjan Slivnik, Patricio Bulić, Borut Robič. Introduction to B.Robič, The Foundations of Computability Theory Parallel Computing: From Algorithms to Programming on State-of-the-Art Plat- (Springer, 2015, 351 str.); forms. Springer, 1999

• [4] Borut Robič. The Foundations of Computability Theory (2nd Ed.) Springer, 2020 R.Trobec, B.Slivnik, P.Bulić, B.Robič, Introduction to Paral-

[5] Robič, Borut Aproksimacijski algoritmi (Založba FRI, 2002, 2009)

lel Computing [6] Roman Trobec, Marjan Šterk, Borut Robič. Computational complexity and

(Springer, 2018, 418 str.); parallelization of the meshless local Petrov-Galerkin methods. Computers and

Structures. 2009, vol. 87, no. 1/2, str. 81-90.

• B.Robič, The Foundations of Computability Theory, 2nd ed. [7] Anthony Sulistio, Uroš Čibej, Srikumar Venugopal, Borut Robič, Rajkumar

(Springer, 2020, 443 str.) Buyya. A toolkit for modelling and simulating data Grids : an extension to

GridSim. Concurrency and computation : practice and experience., 2008, vol. 20,

Zadnja monografija je učbenik na Univerzi v Sieni in tudi na no. 13, str. 1591-1609

Tehnični univerzi v Münchenu. Pri njeni pripravi je z Borutom [8] Uroš Čibej, Boštjan Slivnik , Borut Robič. The complexity of static data repli-

sodeloval Benjamin Wells z Univerze v San Franciscu, zadnji cation in data grids. Parallel Computing. 2005, vol. 31, no. 8/9, str. 900-912

[9] Peter Korošec, Jurij Šilc, Borut Robič. Solving the mesh-partitioning problem

doktorand Afreda Tarskega. with an ant-colony algorithm. Parallel Computing. 2004, vol. 30, str. 785-801.

Kot recenzent je Borut sodeloval pri vrsti tujih znanstvenih [10] Theo Ungerer, Borut Robič, Jurij Šilc. A survey of processors with explicit

revij in uredniških odborov znanstvenih konferenc. multithreading. ACM computing surveys. 2003, vol. 35, str. 29-63.

[11] Theo Ungerer, Borut Robič, Jurij Šilc. Multithreaded processors. The Computer

journal. 2002, vol. 45, str. 320-348.



4 Pedagoška in mentorska dejavnost [12] Borut Robič, Janez Žerovnik. Minimum 2-terminal routing in 2-jump circulant graphs. Computers and artificial intelligence. 2000, vol. 19, no. 1, str. 37-46

Prve pedagoške izkušnje je dobil na Srednji šoli za računalništvo [13] Borut Robič, Peter Korošec, Jurij Šilc. Ant colonies and the mesh-partitioning

v Ljubljani (Vič), kjer je ob zaposlitvi na IJS poučeval računalni- problem. V: S.Olariu (ur.), A.Y.Zomaya(ur.). Handbook of bioinspired algorithms

štvo v š.l. 1984/85. Leta 1985 je sodeloval pri izvajanju programa [14] Mitja Bezenšek, Borut Robič. A survey of parallel and distributed algorithms and applications. Chapman Hall, 2006. str. 285-303

računalniškega opismenjevanja v Gorenju iz Velenja. V tečaj je for the Steiner tree problem. International journal of parallel programming, vol.

bilo zajetih okoli dva tisoč delavcev. 42, no. 2, str. 287-319, 2014

[15] Jurij Mihelič, Borut Robič. Flexible-attribute problems. Computational optimi-

Leta 2004 je bil na Univerzi v Ljubljani izvoljen v naziv redega zation and applications. vol. 47, no. 3, str. 553-566, 2010

profesorja za področje računalištva in informatike. Odtlej je na [16] Jurij Mihelič, Amine Mahjoub, Christophe Rapine, Borut Robič. Two-stage

UL FRI predaval predmete na dodiplomskem, magistrskem in flexible-choice problems under uncertainty. European journal of operational

research. vol. 201, no. 2, str. 399-403, 2010

doktorskem študiju. Bil je mentor diplomantom (80), magistrom [17] Tomaž Dobravec, Borut Robič. Restricted shortest paths in 2-circulant graphs.

(19) in doktorjem (6) računalništva in informatike. Computer communications. vol. 32, no. 4, str. 685-690, 2009

[18] Jurij Šilc, Theo Ungerer, Borut Robič. Dynamic branch prediction and control

speculation. International journal of high performance systems architecture. vol.

5 Organizacijska dejavnost 1, no. 1, str. 2-13, 2007



Danes je Borut Robič na UL FRI vodja Laboratorija za algoritmiko, algorithm for circulant networks. [19] Tomaž Dobravec, Janez Žerovnik, Borut Robič. An optimal message routing Journal of Systems Architecture : the EURO-

predstojnik Katedre za teoretično računalništvo in zadnjih dvajset MICRO journal. vol. 52, no. 5, str. 298-306, 2006

let tudi predsednik Akademskega zbora fakultete.



Marjan Krisper karierna pot in dosežki



Marjan Krisper Career Path and Achivements



Marjan Krisper

Univerza v Ljubljani

Fakulteta za računalništvo in

informatiko

Ljubljana, Slovenia

marjan.krisper@fri.uni-lj.si



Povzetek Leta 1978 sem bil imenovan za svetovalca predsednika

Republiškega komiteja za družbeno planiranje in informacijski

Prispevek govori o karierni poti in dosežkih na področju



računalništva in informatike. Predstavljene so njegovo sistem, kjer sem delal na razvoju organizacijske in tehnološke

pedagoško delo in razvojno-raziskovalna dejavnost doma in na baze Družbenega sistema informiranja (DSI), programu

tujem. modernizacije državne uprave z informacijsko tehnologijo,



Ključne besede pripravi kriterijev za izbiro računalniške opreme ipd. Bil sem načrtovanju razvoja javnega omrežja za prenos podatkov, na

Marjan Krisper, računalništvo, informatika, pedagoška in predsednik oz. član strokovnih komisij v republiški upravi za

razvojno-raziskovalna dejavnost izbiro in uvajanje računalniške opreme oz. za modernizacijo

državne uprave.

Abstract



The paper is about Marjan Krispers life path in the field of Ob ustanovitvi Centra za družbeni sistem informiranja in

computing and informatics. His educational and developmental- informatiko v okviru republiške uprave sem bil imenovan za

research activities at home and abroad re intertwined. pomočnika direktorja, kjer sem nadaljeval delo na navedenih

področjih. V letu 1981 sem bil habilitiran za višjega predavatelja

Keywords in predaval na ,VEKŠ Univerze v Mariboru na dodiplomskem in

Marjan Krisper, computer science, informatics, educational and podiplomskem študiju predmete iz področja poslovne

developmental- research acti informatike.



V Komisiji za smernice študija računalništva in informatike sem

1 Karierna pot in dosežki bil kot predstavnik vlade, ki jo je vodil prof. dr. Jernej Virant.

Na Fakulteti za strojništvo v Ljubljani sem diplomiral leta 1971 Izdelali smo smernice za 4-letni program s tremi profili: logika

in se leta 1972 zaposlil kot raziskovalni asistent. Za asistenta pri in sistemi (kasneje računalniški sistemi), programska oprema in

predmetih Tehnična kibernetika in Konstruiranje krmilnih informatika. Tako je na Fakulteti za elektrotehniko [FE] nastal

sistemov sem bil izvoljen leta 1976. program s temi smermi, ki so veljale vse do bolonjske reforme.

Sam sem bil zadolžen za oblikovanje programa informatike.

Magistriral sem v letu 1977. V času zaposlitve na Fakulteti za

strojništvo sem delal na razvoju programske opreme za Leta 1982 sem bil prvič habilitiran za docenta na Fakulteti za

krmiljenje procesov, na razvoju informacijskih sistemov za elektrotehniko v Ljubljani za predmete iz področja informatike

upravljanje proizvodnje in na drugih aplikacijah poslovne oz. informacijskih sistemov: Sistemska informacijska analiza,

informatike. Programiral sem v Fortranu na IBM 1130, Razvoj informacijskih sistemov, Osnove informatike, Osnove

Asemblerju na procesnem računalniku IBM S7, in DEC informacijskih sistemov, Družbeni sistem informiranja.

računalnikih PDP8 in PDP11. Sodeloval sem pri uvajanju NC Doktoriral sem na Centru za multidisciplinarne študije Univerze

tehnologije v okviru povezav Fakultete za strojništvo z v Beogradu na temo: Ekspertni sistem za merjenje, spremljanje

industrijo. in napovedovanje družbenega razvoja.



Leta 2007 sem bil prvič izvoljen v naziv izredni profesor. V

Permission to make digital or hard copies of part or all of this work for personal or okviru starega programa dodiplomskega študija sem predaval

classroom use is granted without fee provided that copies are not made or distributed predmete: Informacijski sistemi, Razvoj informacijskih

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must sistemov, Planiranje in vodenje informacijskih sistemov in

be honored. For all other uses, contact the owner/author(s). Informacijska družba. V okviru podiplomskega študija sem

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia predaval predmete: Informacijski sistemi, Teorija informacijskih

© 2024 Copyright held by the owner/author(s).

sistemov in Planiranje in obvladovanje informatike v poslovnih

sistemih.



Na Ekonomski fakulteti , sem v okviru podiplomskega študija, ter Slovenskega društva informatika, Društva za umetno

predaval predmet Proizvodni informacijski sistemi. Po inteligenco, IEEE in PMI (Project management institute).

upokojitvi sem v letih 2021 in 2022 predaval na Univerzi na

Primorskem, FAMNIT, predmet Management informacijskih

tehnologij. 2 Priznanja

Zaradi pomembnega prispevka k uvajanju sodobnih konceptov

Od ustanovitve Fakultete za računalništvo in informatiko leta gradnje informacijskih sistemov v prakso sem leta 2000 od

1996 , do leta 2014 sem bil predstojnik katedre za informatiko. Slovenskega društva INFORMATIKA dobil posebno priznanje.

Moja raziskovalna področja so Informacijski sistemi, Razvoj V letu 2017 sem dobil prestižno priznanje Michie - Turing za

informacijskih sistemov, Strateško planiranje informatike, življenjsko delo pri razvoju informacijske družbe. Prenova poslovnih procesov in informacijskih sistemov ter

Poslovno informacijska arhitektura. Od ustanovitve do



upokojitve sem bil predstojnik laboratorija za informatiko. Literatura



Sodeloval sem z gospodarskimi in drugimi organizacijami in Marjan. Using business process models to better 1. TRKMAN, Marina, MENDLING, Jan, KRISPER,

vladnimi inštitucijami na projektih razvoja informacijskih

sistemov, strateškega planiranja, prav tako pa tudi na raziskavah understand the dependencies among user

dolgoročnega razvoja tehnologije oz. strategije tehnološkega stories. Information and software technology.

razvoja. [Print ed.]. Mar. 2016, vol. 71, str. 58-76, ilustr.

ISSN 0950-58498. 2023:



Vodil sem številne informacijske projekte, tako za javno upravo 2. VRHOVEC, Simon, TRKMAN, Marina, KUMER,

ter državne organe kot tudi za večja slovenska podjetja. Že v Aleš, KRISPER, Marjan, VAVPOTIČ, Damjan.



osemdesetih letih (dediščina mojega dela v državni upravi) smo Outsourcing as an economic development tool in

delali za javni sektor (izbira sistema za upravljanje z bazami transition economies : scattered global software

podatkov za državno upravo), v 90 letih pa izdelali številne development. Information technology for

strateške dokumente, predvsem za večja slovenska podjetja development. 2015, vol. 21, no. 3, str. 445-459,

(Petrol, PTT, Slovenske železnice, Mobitel, Telekom, Klinični ilustr. ISSN 0268-1102.

center, prenova in vzpostavitev Centralnega registra 3. VRHOVEC, Simon, HOVELJA, Tomaž, prebivalstva) VAVPOTIČ, Damjan, KRISPER, Marjan.



Bil sem vodilni avtor metodologije EMRIS (Enotna Diagnosing organizational risks in software

metodologija razvoja informacijskih sistemov), ki je bila dolga projects : Stakeholder resistance. International

leta temeljna metodološka usmeritev vseh informacijskih journal of project management. [Print ed.]. Aug.

projektov v javni upravi. Prav tako sem sodeloval v strokovnih 2015, vol. 33, iss. 6, str. 1262-1273, graf. prikazi,

telesih v državni upravi na področjih informatike, tehnološkega tabele. ISSN 0263-7863 razvoja in družbenega razvoja, med drugem sem bil član 4. ŠAŠA BASTINOS, Ana, KRISPER, Marjan.

programskega sveta za projekte e – uprave (2000- 2004) Vodil Multi-criteria decision making in

sem delovno skupino za izdelavo Strategije razvoja e-uprave RS ontologies. Information sciences. [Print ed.]. Feb.

za obdobje 2000-2004. 2013, vol. 222, str. 593-610, ilustr. ISSN 0020-

0255



Na področju mednarodnega sodelovanja sem z Laboratorijem za 5. TRKMAN, Marina, MENDLING, Jan, KRISPER,

informatiko vzpostavil številne stike, predvsem z Merick School Marjan. Using business process models to better

of Business Univerze v Baltimoru, z Univerzo Aix - Marseille, understand the dependencies among user

Univerzo v Celovcu, kjer sem imel tudi vabljena predavanja. Bil stories. Information and software technology.

sem tudi predstavnik Slovenije v upravnem odboru COST A3 [Print ed.]. Mar. 2016, vol. 71, str. 58-76, ilustr.

projekta Management and New Technology. Leta 1986 sem bil ISSN 0950-5849.

na krajši špecializaciji na London School of Economics and 6. VRHOVEC, Simon, TRKMAN, Marina, Political Science. KUMER, Aleš, KRISPER, Marjan, VAVPOTIČ,

Damjan. Outsourcing as an economic development

V letu 2005 sem postal predstavnik Slovenije v IFIP TC8 tool in transition economies : scattered global

(Informacijski sistemi). Leta 2005 sem bil izvoljen za software development. Information technology for

predsednika AIS (Association of Information Systems) oddelka development. 2015, vol. 21, no. 3, str. 445-459,

za Slovenijo z začetkom mandata v letu 2006. ilustr. ISSN 0268-1102.

7. VRHOVEC, Simon, HOVELJA, Tomaž,

V letih 2004, 2006, 2009 in 2010 sem uspešno kandidiral na VAVPOTIČ, Damjan, KRISPER, Marjan.

razpisu ARRS za mentorja mladim raziskovalcem in s tem postal Diagnosing organizational risks in software

mentor štirim mladim raziskovalcem. projects : Stakeholder resistance. International

Bil sem član več znanstvenih in strokovnih združenj, med drugim journal of project management. [Print ed.]. Aug.

ustanovitveni član AIS (Association of Information Systems) 2015, vol. 33, iss. 6, str. 1262-1273, graf. prikazi,

svetovne zveze univerzitetnih učiteljev informacijskih sistemov, tabele. ISSN 0263-7863.



8. ŠAŠA BASTINOS, Ana, KRISPER, Marjan.

Multi-criteria decision making in

ontologies. Information sciences. [Print ed.]. Feb.

2013, vol. 222, str. 593-610, ilustr. ISSN 0020-

0255.]

9. BAJEC, Marko, VAVPOTIČ, Damjan, KRISPER,

Marjan. Practice-driven approach for creating

project-specific software development

methods. Information and software technology.

[Print ed.]. 2007, vol. 49, no. 4, str. [345]-365,

ilustr. ISSN 0950-5849.

10. BAJEC, Marko, KRISPER, Marjan. A

methodology and tool support for managing

business rules in organisations. Information

systems. [Print ed.]. 2005, vol. 30, str. [423]-443,

ilustr. ISSN 0306-4379.



Dosežki kandidata Devida Palčiča



Achievemens of the candidate Devid Palčič



Matjaž Gams†

Inštitut »Jožef Stefan«



Ljubljana, Slovenija



matjaz.gams@ijs.si



POVZETEK KEYWORDS



Devid Palčič verjetno ni najbolj tipičen predstavnik Automation, IOT, multi-objective optimization, AI.

skupnosti računalništva in informatike, saj se v

svojem profesionalnem življenju ukvarja predvsem 1 ZAČETKI



z avtomatizacijo, kar pomeni da računalnike Devid Palčič se je z računalništvom »okužil« na

uporablja za to, da vplivajo na fizičen svet. informativnem dnevu za vpis v srednjo šolo, ko se je

Avtomatizacijo navadno povezujemo z industrijo kot prva generacija »Usmerjenega Izobraževanja«

(»industrijska avtomatizacija) in tudi Devid je vpisoval na Srednjo Pedagoško in Naravoslovno

avtomatizacijo najprej apliciral v stroje in procese, i šolo Koper ii-SPNMŠ. To je bil TRS-80, ki ga je

vendar je kmalu ugotovil, da lahko iste principe dijakom dal na voljo kar profesor Aljoša Žerjal.

uporabi drugje, pravzaprav kjerkoli. Programski jezik Basic in kasneje tudi Assembler,



Tako so računalniki zasedli stavbe, energetiko, oziroma kar strojni jezik za procesor Zilog Z-80 so

mesta, infrastrukturo in še marsikaj. Pri tem jim je bili tako zanimivi, da se je manjša skupina dijakov

Devid Palčič izdatno pomagal. borila za čas preživet za tipkovnico. Obdobje srednje

šole je bilo polno novih znanj ter prvih dosežkov, kot

KLJUČNE BESEDE so uspeh v tekmovanju v izdelavi računalniških

igric.

Avtomatizacija, IOT, več ciljna optimizacija,

umetna inteligenca. Je pa čas srednje šole pokazal, da se da z računalniki

tudi zaslužiti. Skupina dijakov SPNMŠ je

ABSTRACT sodelovala pri takrat eksotičnih storitvah.



Devid Palčič is probably not the most typical Obdelovali so podatke šolskih tekmovanj,

representative of the computing and information izobraževali dijake in starejše ter se spogledovali s

technology community, as in his professional life he podporo računovodstvu, ki je tedaj bilo preprosto:

primarily deals with automation, which means he peš.



uses computers to influence the physical world. Delo ni ostalo neopaženo. Ko je bila v Jugoslaviji

Automation is usually associated with industry fizikalna olimpijada, je IBM posodil svoj najnovejši

("industrial automation"), and Devid initially 1. Celotna obdelava dosežek: IBM PC Junior

applied automation to machines and processes. podatkov je bila izvedena na teh računalnikih in to

However, he soon realized that the same principles tako dobro, da se je IBM odločil računalnike

could be used elsewhere, in fact, anywhere. podariti: to so bili prvi računalniki tega tipa v



Thus, computers have taken over buildings, energy, Vzhodni Evropi.



cities, infrastructure, and much more. In this, Devid V približno tistem času, je Devid Palčič izdelal svoj

Palčič has significantly contributed. prvi računalnik. To je bil računalnik Galaksijaiii

narejen po načrtih objavljenih v reviji, kar je dalo



1 IBM PCjr - Wikipedia

dovolj izkušenj in znanja za načrtovanje drugega avtomatizacija, posamični stroji so se združevali v

računalnika, ki pa je že bil nov, unikaten izdelek, ki linije, računalniki pa so prevzemali vse večjo vlogo.

je znal krmiliti stroj za sestavljanje kabelskih

čevljev. S tem je Devid avtomatiziral stroj svojega Devid je v tem obdobju dejansko usmeril podjetje v

očeta in mu povečal produktivnost. sistemsko integracijo, pri tem pa je uporabljal

izdelke svetovnih proizvajalcev PLC in druge

Odločitev za študij elektrotehnike, točneje opreme, še najbolj Hitachi.



logična. Hitachi je bil odličen, vendar ni ponujal vseh avtomatike in še natančneje robotike, je tako bila

potrebnih modulov, Tako, da so v Robotini razvili

Študij je bil čisti užitek, diploma pri prof. Bajduiv pa prvi izdelek – najprej seveda zase. Modul RT-60, ki

tudi. Diploma je bila praktična in v njenem okviru je je omogočal neposreden priklop Pt-100

bil robot Yaskawa prvič predelan tako, da je bilo temperaturnih senzorjev na PLC Hitachi, so takoj

krmiljenje izvedeno po poziciji in sili, kar je bilo sprejeli drugi uporabniki tega PLC-ja v Evropi in

uporabljeno za ščetkanje okvirja mopeda Tomos drugje po svetu. Uporabljene so bile rešitve, za

Avtomatic. katere se je izkazalo, da jih drugi ne, kar je

rezultiralo v večji točnosti in ponovljivosti, tako da

2 RAZVOJNO DELO ga je na koncu začel kupovati tudi Hitachi.



Devid Palčič se je zaposlil v podjetju Tomos Koper, Že prvi izdelek je bil razvit na osnovi usmeritve, ki

kjer so prevladovali strojniški poklici, inženirji jo je uvedel Devid: »če uporabiš računalnik, izkoristi

elektrotehnike so bili redki. Takšen je bil tudi pristop njegove prednosti, namesto, da poskušaš računalnik

k avtomatizaciji: veliko strojniških rešitev, nekaj uporabiti na tak način, kot bi delal avtomatiko brez

relejne logike, zelo malo ali nič računalnikov. računalnika«.



Devid je uvedel Programabilne krmilnike (PLC), Sledilo je še več podobnih izdelkov najintenzivneje

uporabniške vmesnike in povezavo na osebne pa je bilo sodelovanje takrat, ko je Robotina za

računalnike, ki seveda niso uporabljali SCADA in Hitachi predelala frekvenčne pretvornike v IP-54

podobnih sistemov, tako da je bila programska zaščito, ob tem pa razvila še več vmesnikov in

oprema napisana v programskih jezikih, kot so C in drugih rešitev. Delphi.

Poslovanje podjetja je raslo, osvajali so nove trge,

V slabih dveh letih je ekipa enajstih ljudi, ki jih je hkrati pa so ugotovili, da izdelek, kot ga želijo ne

vodil Devid Palčič, hkrati pa je bil edini inženir obstaja.

elektrotehnike, zasnovala in izdelala osem zahtevnih

strojev in proizvodnih linij. Najzahtevnejša je bila Padla je odločitev: »gremo razvit svoj PLC«.

linija za kemično obdelavo motornih valjev, kjer so

4 CYBRO

bili integrirani PLC, servo-motorji in osebni



računalnik za upravljanje, poročanje in Zahteve za nov PLC, ki se je kasneje izkazal za novo

parametrizacijo. krmilno platformo, je Devid zapisal zelo enostavno:



Rešitev so opazili v podjetju Hitachi in predlagali majhen distribuiran krmilnik, ki bo imel pametne

sodelovanje, vendar je takrat Tomos že bil v velikih module (vsak s svojim procesorjem), te moduli bodo



se je ekipa sporazumno razšla, Devid pa je ustanovil sme zmanjšati hitrosti, moduli so lahko v razdelilni omari ali kjerkoli. K temu je dodal še zahtevo, da je težavah, tako da ni imel interesa z nove izzive, zato lahko blizu ali pa narazen, dodajanje modulov ne



integracijo na področju industrijske avtomatizacije. možno izdelati sistem za avtomatsko generiranje podjetje Robotinav d.o.o. in ga usmeril v sistemsko

aplikativne programske kode in da je rešitev

3 PODJETNIŠKI ZAČETKI primerna za industrijo ter stavbe.



Pisalo se je leto 1990 in slovenska industrija je Naloga je bila izjemno zahtevna. Celotna ekipa je

potrebovala rešitve, do katerih ni imela dostopa. vložila vse svoje znanje in prišla do prvega

Avtomatizacija je bila nujna, vendar ni bila prototipa, ki pa ni izpolnil zahtev, tako da je bilo za

dosegljiva, tako da bilo obdobje zaznamovano z uporaben izdelek potrebno še korenito spremeniti

nadgradnjami, dopolnitvami in izboljšavami. Na posamične rešitve.

obstoječe in nove naprave se je dograjevala

Prva res uporabna generacija izdelkov je bila odborov. Je dolgoletni član upravnega odbora

CYBRO-2. Na tržišču je bila uspešna več kot 18 let, združenja za Inženiring GZS, bil pa je tudi član

leta 2019 pa jo je nadomestila nova, sodobna »Smart City Council«. Prejel je več nagrad, med

platforma Cybro-3. drugim Nagrado GZS za izjemne gospodarske

dosežke.

5 NAPREDNE REŠITVE

Ker verjame, da je potrebno znanje širiti in da nas

Avtomatizacija je reševala probleme v industriji, lahko samo znanje ter inovativnost usmerita v

hkrati pa se je Robotina vse bolj usmerjala tudi na prihodnost, Devid že od začetka predava priložnosti

druga področja, kot so stavbe, mesta in energetika. in izzive pametnih omrežij na Akademiji Gradbenih

Intenziven razvoj je pripeljal do velikih razvojno vii Investicij (AGI). Tam poudarja pomen digitalne

raziskovalnih projektov, ko je sodelovala z inštituti, pretvorbe in IoT za vsa področja investicij in

univerzami in drugimi partnerji. Pri tem je prišlo do življenja.

sodelovanja z Institutom »Jožef Stefan«vi, s

oddelkom, ki ga je vodil prof. Matjaž Gams in Na koncu, tehnologije ki prihajajo, oziroma so že tu,

skupaj je bila razvita rešitev za več ciljno so tako pomembne, da jih moramo spoznati, širiti,

optimizacijo (multi objective optimization) ki je bila hkrati pa pomagati razbiti kak tabu. Zato je potrebno

uporabljena v več rešitvah. Sodelavec je na tem razširiti znanje in zavedanje, tako se mora

področju tudi opravil doktorat. računalništvo, informatika, IoT, umetna inteligenca

in druge tehnologije razširiti in postati široko

Devid Palčič je zasnoval več kot 100 izdelkov, uporabljene v vseh segmentih družbe. sodeloval pri več patentih in volil številne

raziskovalne in razvojne projekte, fokus v Devid in Robotina so dokazali, da lahko tudi male in

prihodnosti pa je obvladovanje celotne verige od inovativne ekipe zmagujejo in sodelujejo z

avtomatizacije, računalništva na robu, pa naprednih najboljšimi. Če bomo s to prakso še koga okužili,

digitalnih storitev, ki vključujejo strojno učenje in bomo vsi uspešnejši.

umetno inteligenco.

VIRI



6 SODELOVANJE 1. Splena stran Robotina d.o.o., www.robotina.com



Devid je prepričan, da je sodelovanje ključ do 2. Register patentov: SIPO.DS: Patents - Hit list (uil-

sipo.si), 23932, 23933, 23937, 24817, 24866, 24867,

napredka za vse. Znanje je edina dobrina, ki se z 26351

deljenjem povečuje, zato aktivno sodeluje pri 3. www.cybrotech.com: izdelki

aktivnostih za katere verjame, da lahko prinesejo 4. Nagrajenci (gzs.si): spisek nagrajencev GZS

napredek osebam, podjetjem in družbi.



Sodeloval je pri projektu NEDO, ko je Japonska

tehnološka agencija v Sloveniji skupaj s slovensko

industrijo izvedla izjemno demonstracijo pametnih

omrežij. Sodeloval je pri oblikovanju več SRIP-ov

in tehnološke mreže, kjer je tudi član upravnih



i v Sedaj “Gimnazija Koper” Robotina d.o.o.

ii vi TRS-80 - Wikipedia IJS - Institut "Jožef Stefan"

iii vii Galaksija (computer) - Wikipedia Akademija gradbenih investicij (gzs.si)

iv Tadej Bajd - Wikipedija, prosta enciklopedija

(wikipedia.org)



Moja pot skozi 44 let dela na področju računalništva in



informatike



My journey through 44 years of work in computer science



and information technology



Milan Ojsteršek

Fakulteta za elektrotehniko,

računalništvo in informatiko

Univerza v Mariboru, Maribor

Slovenija

milan.ojstersek@um.si



Povzetek sodeloval v laboratoriju prof. dr. Bogomirja Horvata pri programiranju mikroračunalniškega krmilnika, ki je uporabljal V prispevku je prikazanih mojih 44 let delovanja na področju procesor Motorola 6800. V četrtem letniku sem izbral izbirni računalništva in informatike. Predstavil bom svoje pedagoško, predmet Sistemsko programiranje, pri katerem sem se naučil strokovno in raziskovalno delo. programirati v programskem jeziku Pascal.



Ključne besede podatkovno vodenega računalnika. Za diplomsko nalogo sem V okviru moje diplomske naloge sem izdelal simulator

Milan Ojsteršek, paralelno procesiranje, heterogeni računalniški prejel Kidričevo nagrado za študente. Oktobra 1984 sem se kot

sistemi, spletne tehnologije, procesiranje naravnega jezika, raziskovalec-stažist zaposlil v Laboratoriju za računalniške

nacionalna infrastruktura odprte znanosti, interoperabilnost, arhitekture in jezike, ki ga je vodil prof. dr. Viljem Žumer. Še

akademska integriteta. dve leti sem v okviru raziskovalnega projekta nadgrajeval

simulator podatkovno vodenega računalnika in objavil nekaj

člankov, ki so bili vezani na arhitekturo in programske jezike za

Abstract podatkovno vodene računalnike. Kasneje sem to arhitekturo

This paper highlights my 44 years of work in the field of nadgradil v paralelni računalniški sistem, ki je bil neobčutljiv na

computer science and information technology. I will present my odpovedi posameznih procesnih enot in vozlišč v večstopenjskih

teaching, professional and research work. medprocesorskih mrežah.

V obdobju od 1986 do 1988 sem sodeloval pri projektu

Keywords izgradnje mikroračunalniškega krmilnika za Metalno v Mariboru.

Milan Ojsteršek, parallel processing, heterogeneous computing Na tem projektu smo programirali simulator in interpreter

systems, web technologies, natural language processing, national namenskega programskega jezika IMCL, ki je bil uporabljen za

openscience infrastructure, interoperability, academic integrity. programiranje industrijskih mikroračunalniških krmilnikov. V

tem času smo s kolegi iz laboratorija dobili idejo za izdelavo

paralelnega podatkovno vodenega računalnika, ki deluje v

1 realnem času. Izdelali smo prevajalnik za lastni podatkovno Prvih štirinajst let dela na področju

računalništva in informatike voden jezik DFCL in simulator podatkovno vodenega

računalnika, ki deluje v realnem času. Rezultat mojega

raziskovalnega dela na tem področju je bilo nekaj mednarodnih

Z računalniškim programiranjem sem se spoznal leta 1980 v objav in magistrska naloga z naslovom "Model makro

prvem letniku študija elektrotehnike na Visoki tehniški šoli v podatkovno vodenega računalnika, primernega za delo v

Mariboru. Programirali smo v programskem jeziku FORTRAN. realnem času", ki sem jo zagovarjal leta 1991.



obdelavo preko čitalnika kartic na IBM 1130. Programiranje me ki smo ga izvajali za podjetje Iskra Delta. V tem projektu smo je takoj začelo zanimati, zato sem v tretjem letniku nekaj časa gradili paralelni računalniški sistem za jugoslovansko vojsko. Na Uporabljali smo luknjane kartice, ki smo jih dajali v zaporedno V obdobju od 1987 do 1989 sem sodeloval v projektu Parsys,

∗ Moja pot skozi 44 let dela na področju računalništva in informatike žalost je jugoslovanska vojska sredi leta 1989 zaradi

†Milan Ojsteršek

predvidenega razpada Jugoslavije ukinila financiranje projekta.

Permission to make digital or hard copies of part or all of this work for personal or Sem pa v okviru tega projekta in v projektu za Metalno v

classroom use is granted without fee provided that copies are not made or distributed Mariboru dobil ideje za svoje delo na področju določanja

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must razčlenjenosti vozlišč programskih grafov in njihovega

be honored. For all other uses, contact the owner/author(s). razvrščanja v enote večprocesorskega računalnika. Junija 1994

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s). sem zagovarjal doktorsko disertacijo z naslovom "Določitev

razčlenjenosti vozlišč programskega grafa in njihovo razvrščanje



v enote večprocesorskega računalnika". S paralelnim kot referenco na področju vzpostavitve gigabitnega etherneta na

procesiranjem sem se raziskovalno ukvarjal še po zagovoru izobraževalnih inštitucijah. Ena od nalog, ki sem jih skupaj z

doktorske disertacije, vendar zaradi neobstoječe industrije na mojimi sodelavci izvedel, je bila tudi izdelava specifikacij za

področju razvoja takšnih sistemov, na žalost rezultatov nisem informacijski sistem Univerze v Mariboru (1999 – 2000). Od

mogel prenesti v prakso. To je bil tudi glavni razlog, da sem se 2011 do 2014 sem na Univerzi v Mariboru sodeloval v projektih

začel ukvarjati z raziskovalnimi področji, katerih rezultate R&D Industry ( EU MED, 2011 – 2013) in CITEK (EU MED,

znanstveno raziskovalnega dela je bilo lažje prenašati v prakso. 2013 – 2014), v katerih sem vodil razvoj informacijskega sistema

Ker se je skozi 44 let mojega dela nabralo veliko referenc, jih za podporo raziskovalnemu delu na vključenih inštitucijah.

v članku zaradi omejitve števila strani, ne bom navajal posebej, Leta 1999 so na Inštitutu za računalništvo, UM-FERI

pač pa sem dal povezavo na mojo bibliografijo, ki zajema dela ustanovili Laboratorij za heterogene računalniške sisteme

od 1984 do 2024 in je zavedena v COBISSu [1]. (LHRS), ki ga vodim še danes. Podrobnejši opis nekaterih

projektov LHRS, ki bodo omenjeni v nadaljevanju ter nekaterih,

ki jih zaradi pomankanja prostora nisem mogel opisati, najdete

2 Od razvoja aplikacij za internet do na https://lhrs.feri.um.si/projekti/. Opise nekaterih programskih

informatizacije javne uprave rešitev LHRS najdete na https://lhrs.feri.um.si/predstavitev-

Leta 1995 je takratno ministrstvo, ki je bilo pristojno za dela/nase-resitve/.

znanost, objavilo razpis za financiranje centrov znanja. Namen V okviru LHRS smo začeli izvajati večje projekte. Eden

centrov znanja je bil prenos znanstveno-raziskovalnih rezultatov takšnih projektov je bil projekt za podjetje Mobitel (1999), v

v prakso. Na tem razpisu so izbrali tudi Center za heterogeno okviru katerega smo izdelali programsko opremo, ki je

procesiranje, ki sem ga vodil. Za ta center smo tri leta prejemali omogočala prikaz digitalnih vsebin na infoterminalih, na velikih

financiranje, ki smo ga večinoma porabili za plačilo študentskega video zaslonih, mobilnih telefonih, ki so omogočali protokol

dela. Center je izvajal tudi projekte, ki so bili vezani na izdelavo WAP, na teletekstu in na internetu. Iz tega projekta smo se veliko

različnih spletnih aplikacij. Leta 1996 smo razvili aplikacijo za naučili. Rezultat projekta je bila programska oprema za portal, ki

spletno rezervacijo mariborskih hotelov. Nato smo za Mestno smo jo kasneje naprej razvijali in jo je uporabljalo več kot 50

občino Maribor izdelali »Mariborski informacijski sistem«. Za slovenskih občin, nekatere izobraževalne inštitucije ter nekaj

podjetje Sweetcom smo izdelali aplikacijo za oglaševanje podjetij. Leta 2000, ko smo razvili našo prvo verzijo portalne

rabljenih vozil preko spleta, za podjetje Dodoma pa spletno infrastrukture, še ni bilo na voljo podobnih spletnih aplikacij.

aplikacijo za oglaševanje nepremičnin preko spleta. Za Leta 2000 smo za Kmetijski zavod v Mariboru izdelali

Mariborsko razvojno agencijo smo v izdelali spletno aplikacijo aplikacijo za vodenje kontrole ekološkega kmetovanja, ki jo je

za gospodarski informacijski sistem za Podravje. V letih 1998 in leta 2004 prevzelo Ministrstvo za kmetijstvo, gozdarstvo in

1999 sem tudi sodeloval pri specifikaciji turističnega prehrano. Aplikacijo še danes v okviru LHRS vzdržujemo in

informacijskega sistema za TIC Ljubljana in Zavod za turizem v nadgrajujemo. Ta aplikacija je pomembna, ker smo za pohitritev

Mariboru. Za njih smo razvili spletno aplikacijo za vodenje njenega razvoja razvili ogrodje Blacksmith, ki nam je omogočilo

statistike zasedenosti turističnih objektov in aplikacijo za objektno-relacijske preslikave, neodvisno načrtovanje in

oglaševanje prireditev v Sloveniji. upravljanje vsebine, navzkrižno sklicevanje med objekti,

V okviru Centra za heterogeno procesiranje sem navezal stike dedovanje, večjezično podporo, neodvisnost od sistema za

z računalniškimi podjetji Microsoft, Oracle, Marand, NIL in upravljanje podatkovnih zbirk in integracijo spletnih storitev.

Lancom, ki so nam pomagala z nasveti, programsko opremo in Glavna prednost tega ogrodja je bil krajši čas razvoja

izobraževanji na področju novih tehnologij. Sodelavci Centra za porazdeljenih storitev ter bolj razumljiva koda programskih

heterogeno procesiranje in študenti, ki so sodelovali pri delu rešitev.

centra, so pridobili certifikate na področju razvojnih orodij in Iz tega obdobja bi omenil sodelovanje pri izdelavi strategije

sistemske administracije programske opreme podjetja Microsoft. uvajanja elektronskega poslovanja v lokalne skupnosti (2003),

Ko smo pridobili ustrezne certifikate, je Center za heterogeno izdelavo strateškega načrta informatizacije Mestne občine

procesiranje postal prva izobraževalna inštitucija na svetu, ki je Maribor (2005), in akcijski načrt izdelave zalednega

pridobila naziv Microsoft Certified Partner. S pridobitvijo tega informacijskega sistema za lokalne skupnosti (2006). Skupaj s

naziva smo postali upravičeni do vse Microsoftove programske podjetjem 3Port smo od 2003 do 2005 razvili nekaj spletnih

opreme, njihovih izobraževalnih materialov in udeležbe na aplikacij za podporo delovanju lokalnih skupnosti (e-obveščanje

Microsoftovih dogodkih doma in v tujini. Znanje, ki smo ga občana o njegovih zadevah, e-pobude…).

pridobili, smo v okviru predavanj in vaj, posredovali našim Od leta 2000 naprej smo v Laboratoriju za heterogene

študentom. računalniške sisteme razvijali portal za podporo izobraževanju

Od leta 1998 do 2010 sem vodil računalniški center na UM- na daljavo, ki smo ga najprej preizkusili na UM-FERI, nato pa

FERI. Med pomembnejše projekte, ki smo jih v računalniškem smo ga uvedli na celotni Univerzi v Mariboru. V tem času še niso

centru v tem času izvedli, sodi vsekakor obnova računalniških obstajali odprtokodni sistemi za upravljanje z učnimi vsebinami,

učilnic, obnova računalniškega omrežja fakultete, izdelava zato smo na podlagi zahtev učiteljev in asistentov iz Univerze v

spletnega portala fakultete, izdelava portala za podporo Mariboru podprli procese upravljanja in podajanja učnih vsebin

izobraževanju na daljavo in sistema za interno komuniciranje na ter preverjanja znanja. Razvoj programske opreme je bil delno

fakulteti. Med prvimi v Sloveniji smo uvedli enotno identiteto financiran preko različnih Interreg projektov, delno pa ga je

študentov in zaposlenih na UM-FERI, brezžično omrežje, financirala Univerza v Mariboru. Bil sem tudi eden od avtorjev

Eduroam in IP telefonijo. Na UM-FERI smo imeli med prvimi nacionalne strategije na področju e-izobraževanja od 2006 do

v Sloveniji gigabitni ethernet. CISCO je navajal našo fakulteto 2010. Sodeloval sem tudi pri projektu izdelave e-gradiv za



srednje računalniške šole (2010-2011), ki ga je koordinirala jezikovno-digitalnih kompetenc ter izboljšanju zmožnosti

Srednja elektro-računalniška šola v Mariboru. Izdelal sem učna tvorjenja in razumevanja besedil. Osrednja ciljna skupina so

gradiva za predmete »Upravljanje s programirljivimi napravami«, osnovnošolci od šestega razreda in srednješolci – torej

»Načrtovanje in razvoj programskih aplikacij« in »Načrtovanje uporabniki, ki so sposobni tvorjenja daljših besedil, a s

in razvoj spletnih aplikacij«. primanjkljaji v jezikovnem izražanju in razumevanju zlasti

neumetnostnih besedil. Naši cilji so bili pripraviti korpus

aktualnih obsegovno in vsebinsko raznovrstnih besedil,

3 Od digitalnih knjižnic, procesiranja besedil v povezanih z interesi ciljne skupine; na osnovi korpusnega

slovenskem jeziku do detekcije plagiatov gradiva izdelati trajnostna samostojno delujoča e-orodja za

Leta 2004 smo v okviru projekta, financiranega iz Ciljnega avtomatsko tvorjenje in pregledovanje vaj ter orodja smiselno

raziskovalnega programa, začeli z razvojem Digitalne knjižnice podpreti s celovitimi slovarsko-slovničnimi in

Ministrstva za obrambo (v nadaljevanju DKMORS). jezikovnonormativnimi opisi. Posebno pozornost smo namenili

Programsko opremo za to digitalno knjižnico vzdržujemo in zapostavljenemu prikazu idiomatičnih jezikovnih prvin ter

nadgrajujemo še danes. Projekt je pomemben zato, ker smo na izkoristili potencial za njihov atraktiven jezikovni opis. Okolje

podlagi pridobljenega znanja razvili Digitalno knjižnico vsebuje bazo znanja, iz katere se lahko učenci in dijaki naučijo

Univerze v Mariboru (v nadaljevanju DKUM), katere pravopisa, slovnice, uporabe frazemov in pregovorov ter

programska oprema je osnova za repozitorije Nacionalne razumevanja in tvorjenja besedil. E-okolje omogoča reševanje

infrastrukture odprte znanosti. Prav tako smo morali v okviru nalog iz področja slovnice, pravopisa, frazemov in pregovorov

DKMORS razviti boljši iskalnik, ki je bil primernejši za iskanje ter razumevanja besedil. Pri pouku slovenščine ga uporablja več

po podatkih in besedilih v slovenskem jeziku. Za ta namen smo kot 300 osnovnih in srednjih šol.

razvili lastno programsko ogrodje Textproc za obdelavo besedil Na področju procesiranja besedil v naravnem jeziku sem

v slovenskem jeziku (2007 – 2008). Ogrodje omogoča skupaj s sodelavci Laboratorija za heterogene računalniške

povezovanje različnih opravil za procesiranje besedil (npr. sisteme sodeloval še v projektih Spletni portal Franček (2017 –

razčlenjevanje vsebine, oblikoslovno označevanje, razreševanje 2021), Slovenska znanstvena besedila (2016 – 2019) in Razvoj

sklicev, slovenščine v digitalnem okolju (2020 – 2023). V okviru pomensko označevanje, delno avtomatsko

dopolnjevanje pomenskega slovarja, določanje imenskih entitet, Slovenskih znanstvenih besedil smo s sodelavci LHRS izdelali

določanje relacij med imenskimi entitetami...). To ogrodje smo Korpus akademske slovenščine (KAS) in pomenski

uporabili na področju razvoja aplikacij za odgovarjanje na označevalnik strokovnih terminov v znanstvenih besedilih. V

vprašanja v slovenskem jeziku, pri razvoju priporočilnega okviru projekta Razvoj slovenščine v digitalnem okolju smo

sistema za priporočanje vsebin iz repozitorijev, sistema za izdelali orodje za avtomatsko ekstrakcijo relacij za gradnjo

ekstrakcijo znanja iz besedil in sistema za preverjanje podobnosti semantične mreže, orodje za avtomatsko odgovarjanje na

vsebin. V okviru projekta ESO mreža (2008) smo skupaj s vprašanja, orodje za luščenje strokovne terminologije ter

podjetjem 3-Port razvili aplikacijo za odgovarjanje na vprašanja sodelovali pri razvoju terminološkega portala.

iz področja občinske uprave. Aplikacija je odgovarjala na Za konec tega poglavja bi omenil še aplikacijo, ki jo danes

podlagi podatkov iz podatkovnih zbirk, na podlagi zbirke uporablja večina slovenskih višješolskih in visokošolskih

pogosto zastavljenih vprašanj, zbirke žaljivih vprašanj, posebnih izobraževalnih inštitucij. Gre za program, ki smo ga razvili za

vprašanj in fraz ter na podlagi iskanja po dokumentih. V projektu potrebe Univerze v Mariboru, ki je želela preverjati podobnost

ESO mreža 2 (2009-2011) smo nadgradili sistem za odgovarjanje zaključnih del študentov. Kasneje smo za Univerzo v Mariboru

na vprašanja še z opisom znanja o domenskem področju s izdelali tudi vtičnik, ki omogoča preverjanje podobnosti oddanih

pomočjo ontologij in s pomenskim opisom spletnih storitev. Kot nalog študentov v sistemu za upravljanje z učnimi vsebinami

primer uporabe programske rešitve za odgovarjanje na vprašanja Moodle. Za namen preverjanja podobnosti vsebin so sicer

z opisom znanja o domenskem področju, smo skupaj s podjetjem obstajale komercialne rešitve, ki pa niso bile prilagojene za

3-Port izdelali aplikacijo za izračun komunalnega prispevka v analizo besedil v slovenskem jeziku. Z našim programom so

občini Piran. V okviru projekta “Metal Knowledge Network- odkrili nekaj odmevnih plagiatov v Sloveniji, ki so bili

Knowledge management and semantic search for metal industry predstavljeni v medijih. Namen programa je predvsem naučiti

in Slovenia and Austria” (2011 – 2013) smo izdelali študente ali druge uporabnike, da se plagiatorstvo ne izplača in

infrastrukturo, ki omogoča pomensko iskanje in pomensko da je zanje veliko boljše, če sami napišejo besedilo, pravilno

označevanje strukturiranih in nestrukturiranih vsebin iz področja navedejo uporabljene vire in jih ustrezno citirajo.

kovinsko-predelovalne industrije. V okviru projekta smo izdelali Vodim ali sem vodil mednarodne projekte iz področja

pomenski iskalnik, ki omogoča pomenski opis tabel in pomensko akademske integritete, ki obravnavajo plagiatorstvo in kršenje

indeksiranje podatkov iz podatkovnih baz informacijskih akademske integritete. Ti projekti so ENAI (2016 – 2019),

sistemov ter pomensko obdelavo tekstovnih dokumentov. FAITH (2022 – 2025) in PUPP ( 2021 – 2028, 35 partnerjev iz

Pomemben projekt, ki smo ga začeli leta 2017 in se še Kanade, ZDA in Evrope). Sem član več delovnih skupin in član

nadaljuje, je »Slovenščina na dlani«. Projekt izvajamo skupaj s nadzornega odbora European network of academic integrity.

Filozofsko fakulteto UM, Pedagoško fakulteto UM in kolegi iz



Laboratorija za digitalno procesiranje signalov UM-FERI. LHRS 4 Moje pedagoško delo je zadolžen za izdelavo in vzdrževanje programske rešitve. V

okviru projekta smo izdelali interaktivno učno e-okolje, Leta 1986 sem bil habilitiran v naziv asistent. Izvajal sem vaje

namenjeno bogatitvi pouka slovenskega jezika, povečanju pri predmetih "Programiranje I", "Programski jeziki I",

motivacije za učenje slovenščine, pridobivanju dodatnih "Programski jeziki II", "Principi programskih jezikov",



"Sistemsko programiranje", "Organizacija in arhitektura Sloveniji od 2015 do 2020, izdelavi uredbe o izvajanju

računalniških sistemov I", "Osnove računalniške organizacije", znanstvenoraziskovalnega dela v skladu z načeli odprte znanosti

"Izbrani algoritmi" in "Osnove računalništva". V okviru in izdelavi akcijskega načrta razvoja odprte znanosti v Sloveniji

predmetov, ki so bili vezani na programske jezike sem študente od 2023 do 2030. V okviru akcijskega načrta vodim razvoj in

učil programirati v takrat aktualnih programskih jezikih vzdrževanje nacionalne infrastrukture odprte znanosti.

(FORTRAN, Pascal, C++, COBOL, ADA, Smaltalk, LISP, Kot član projektne skupine sem sodeloval v EU projektih

Prolog). DAPHNE (H2020, 2020 – 2024) in EuroCC (EUROHPCJU,

Po zagovoru doktorske disertacije sem bil habilitiran v naziv 2020-2022). V okviru UM vodim evropska projekta GDI (Digital

docent. Najprej sem začel s predavanji na višješolskem študiju. Europe, 2022 – 2026) in EUROCC2 (EUROHPCJU, 2023 –

Ko smo leta 1996 na Fakulteti za elektrotehniko, računalništvo 2025). V projektu European genomic data infrastructure (GDI)

in informatiko, Univerze v Mariboru (UM-FERI) začeli z vodim vzpostavitev nacionalne infrastrukture genomskih

izvajanjem novih študijskih programov, sem arhivov, ki bo povezana v evropsko genomsko infrastrukturo. Ta

predaval ”Programiranje 1”, ”Osnove računalniške infrastruktura je poseben izziv, saj je potrebno zagotoviti

arhitekture”, ”Osnove računalniške organizacije”, ”Paralelne ustrezno varnost podatkov, procesiranje podatkov v varnem

računalniške sisteme”, ”Porazdeljene računalniške sisteme” okolju in varno izmenjavo podatkov z drugimi genomskimi

in ”Razvoj aplikacij za internet”. Izbirni predmet ”Razvoj arhivi po Evropi. V projektu EUROCC2 je pet inštitucij v

aplikacij za internet” je bil leta 1996 precejšnja novost v Sloveniji vzpostavilo nacionalno vozlišče EUROHPC. To

Sloveniji, saj se še na drugih inštitucijah in tudi na UM-FERI vozlišče je povezano z ostalimi nacionalnimi vozlišči po Evropi.

niso zavedali, da bo internet ključno spremenil naš način Znotraj Slovenije skrbimo za izobraževanja na področju uporabe

življenja. visoko zmogljive računalniške infrastrukture (HPC) in razvoja

Leta 2009 se je na UM-FERI začel izvajati bolonjski študijski programskih rešitev za HPC, prenašamo te rešitve v

program. Na različnih študijskih programih sem predaval ”Uvod gospodarstvo in javno upravo ter sodelujemo z drugimi deležniki

v svetovni splet”, “Arhitekturo in organizacijo računalnika”, v Sloveniji, v Evropi in širše.

“Računalniška omrežja”, “Razvoj aplikacij za internet”, “Osnove Sodelujem v aktivnostih iniciative European open science

svetovnega spleta”, “Mrežno usmerjeno računanje”, “Spletno cloud (EOSC). Od 2019 do 2021 sem bil član EOSC

programiranje”, “Spletne tehnologije” ter “Pomensko in Interoperability working group. Sodeloval sem v podskupini, ki

storitveno usmerjen splet”. je izdelala »EOSC interoperability framework«, v katerem so

Nekaj tisoč študentov je poslušalo ali sodelovalo pri vajah pri podana osnovna navodila za zagotavljanje interoperabilnosti

mojih predmetih. Bil sem mentor pri zaključnih delih študija več znotraj EOSCa. Definirali smo tudi referenčno arhitekturo in

kot 180 študentom na različnih stopnjah študija. Z nekaterimi minimalni nabor metapodatkov za opis FAIR digitalnih objektov.

imam še danes stike. Nekateri imajo problem, za katerega Nato sem od 2021 do 2023 sodeloval v EOSC-A semantic

potrebujejo pomoč in se obrnejo name. Nekateri me srečajo na interoperability task force, kjer sem vodil podskupino za

ulici in obudimo spomine na čas, ko so študirali. Lep občutek je, preslikave metapodatkovnih shem in določanje minimalnega

da sem tudi jaz prispeval k temu, da imamo danes toliko dobrih nabora metapodatkov. Sedaj sem član EOSC-A technical and

strokovnjakov na področju računalništva in informatike. semantic interopability task force, ki ima mandat od 2024 do

2025. Sem eden od sovodij v CODATA GOSC Data

interoperability working group, eden od sovodij RDA metadata

5 Od nacionalne infrastrukture odprtega IG in član več delovnih skupin v RDA ter član uredniškega

dostopa do EOSC interoperabilnostnega odbora COAR vocabularies,

okvirja

Slovenske univerze so leta 2013 s sofinanciranjem Evropskega Zahvala

sklada za regionalni razvoj in Ministrstva za izobraževanje, Veliko srečo sem imel, da sem sodeloval z ljudmi, ki so mi

znanost in šport vzpostavile nacionalni portal odprte znanosti ter zmeraj stali ob strani in me pri mojem delu spodbujali. Posebna

repozitorije za odprti dostop do zaključnih del študija in zahvala gre prof. dr. Viljemu Žumerju in prof. dr. Petru Kokolu,

rezultatov raziskav raziskovalcev na Univerzah v Ljubljani, ki sta me povabila v Laboratorij za računalniške arhitekture in

Mariboru, Novi Gorici in na Primorskem. Vključili pa smo tudi jezike na Visoki tehniški šoli Univerze v Mariboru. Prof. Žumer

več drugih virov, ki so ponujali vsebine za raziskovalce in je bil moj mentor pri diplomski, magistrski in doktorski nalogi.

študente. Nacionalni portal openscience.si agregira vsebine iz Kadarkoli sem ga potreboval, mi je pomagal z nasveti. Prof. dr.

repozitorijev in drugih virov za potrebe združevalnega iskalnika, Peter Kokol je bil moj sodelavec v laboratoriju in me je na

priporočilnega sistema in preverjanja podobnosti vsebin. To začetku moje poti uvedel v raziskovalno in pedagoško delo ter

infrastrukturo smo v letih od 2014 do 2022 dopolnili z me vseskozi spodbujal. Zahvalil bi se tudi vsem mojim

repozitorijem za samostojne raziskovalne organizacije in sodelavcem, s katerimi sem sodeloval. Poimensko jih ne bom

repozitorijem za višješolske in visokošolske inštitucije, naštel, da ne bi slučajno koga izpustil. Vsi ste mi stali ob strani

nacionalnim strežnikom za dodeljevanje trajnih identifikatorjev in mi pomagali na moji profesionalni in osebni poti.

in arhivom za velepodatke. Ta arhiv je bil vzpostavljen na

superračunalniku Vega v okviru projekta HPC RIVR (EU skladi, Literatura

EUROHPCJU, 2018 – 2021), v katerega je bil vključen tudi [1] COBISS. 2024. Milan Ojsteršek, osebna bibliografija za obdobje 1984-

LHRS. Sodeloval sem pri izdelavi nacionalne strategije odprtega https://bib.cobiss.net/bibliographies/si/webBiblio/bib201_20240920_223 2024, [Online]. Dostopno:

dostopa do znanstvenih objav in raziskovalnih podatkov v 959_a2688355.html, obiskano 28. 9. 2024



Zbornik 27. mednarodne multikonference



INFORMACIJSKA DRUŽBA – IS 2024

Zvezek K



Proceedings of the 27th International Multiconference



INFORMATION SOCIETY – IS 2024

Volume K



ChatGPT v zdravstvu



ChatGPT in Medicine



Uredniki / Editors



Matjaž Gams, Monika Simjanoska Misheva, Stevo Lukić, Franz Wotawa



http://is.ijs.si



9. oktober 2024 / 9 October 2024

Ljubljana, Slovenia





PREDGOVOR




Umetna inteligenca, zlasti generativna umetna inteligenca, kot je ChatGPT, je spremenila

pravila igre v številnih panogah. Vendar je njen vpliv v medicini še posebej pomemben, saj

gre v zdravstvu za več kot le informacije - gre za človeška življenja. Vključitev umetne

inteligence lahko bistveno izboljša izide zdravljenja bolnikov, racionalizira delovne procese in

izboljša procese odločanja na celotnem področju medicine.



Vloga umetne inteligence presega pomoč strokovnjakom; neposredno vpliva na oskrbo

bolnikov. Z virtualnimi posvetovanji, izobraževanjem bolnikov in preverjanjem simptomov

lahko ChatGPT izboljša dostop do zdravstvenega varstva za skupine prebivalstva, ki bi se

sicer soočale z ovirami zaradi lokacije ali časovnih omejitev. Poleg tega umetna inteligenca z

avtomatizacijo rutinskih opravil zmanjšuje upravno breme zdravnikov, tako da lahko več časa

posvetijo oskrbi bolnikov in manj papirologiji. To je lahko ključni dejavnik v boju proti

izgorelosti zdravnikov, ki je v sodobnem zdravstvu vse večji problem. Čeprav je potencial

ChatGPT v zdravstvu vznemirljiv, moramo obravnavati etične in varnostne izzive, ki jih

prinaša. Ključna vprašanja so zagotavljanje zasebnosti pacientov, zmanjšanje pristranskosti

algoritmov umetne inteligence in ohranjanje natančnosti zdravniških nasvetov. UI mora

podpirati - ne pa nadomestiti - človeško presojo, zlasti pri kritičnih medicinskih odločitvah.

Zagotavljanje preglednosti, odgovornosti in zasnove teh sistemov s pristopom, pri katerem je

v ospredju varnost, je bistvenega pomena za krepitev zaupanja v te tehnologije.



V prihodnosti se bo vloga ChatGPT in podobnih sistemov umetne inteligence verjetno še

povečala. Pravkar začenjamo raziskovati njihove aplikacije v personalizirani medicini, kjer bi

umetna inteligenca lahko pomagala prilagoditi zdravljenje posameznim pacientom na podlagi

genetskih podatkov, podatkov o okolju in življenjskem slogu. Poleg tega lahko umetna

inteligenca z analizo trendov v populacijah prispeva k zgodnjemu odkrivanju bolezni,

odkrivanju zdravil in globalnim zdravstvenim pobudam.



Ta konferenca je priložnost, da se poglobimo v najsodobnejše raziskave, nastajajoče

aplikacije in etične vidike uporabe ChatGPT v medicinski praksi. Skupaj bomo raziskali

sedanje zmogljivosti in prihodnje možnosti, hkrati pa se bomo posvetili izzivom, ki jih prinaša

vključevanje umetne inteligence na tako občutljivo in tvegano področje.



Matjaž Gams

Monika Simjanoska Misheva

Stevo Lukić

Franz Wotawa

predsedniki konference ChatGPT v zdravstvu





FOREWORD




AI, and particularly conversational AI like ChatGPT, has been a game-changer in many

industries. However, its impact in medicine is particularly significant because healthcare is

about more than just information—it’s about human lives. The integration of AI has the

potential to dramatically improve patient outcomes, streamline workflows, and enhance

decision-making processes across the medical field.



The role of AI extends beyond assisting professionals; it directly impacts patient care.

Through virtual consultations, patient education, and symptom checks, ChatGPT can enhance

access to healthcare for populations who might otherwise face barriers due to location or time

constraints. Additionally, by automating routine tasks, AI reduces the administrative burden

on clinicians, allowing them to spend more time on patient care and less on paperwork. This

can be a crucial factor in combating physician burnout, a growing issue in modern healthcare.

While the potential of ChatGPT in healthcare is exciting, we must address the ethical and

safety challenges that come with it. Ensuring patient privacy, minimizing bias in AI

algorithms, and maintaining the accuracy of medical advice are key concerns. AI should

support—not replace—human judgment, particularly in critical medical decisions. Ensuring

that these systems are transparent, accountable, and designed with a safety-first approach is

essential to building trust in these technologies.



Looking forward, the role of ChatGPT and similar AI systems will likely expand. We are just

beginning to explore its applications in personalized medicine, where AI could help tailor

treatments to individual patients based on genetic, environmental, and lifestyle data.

Additionally, AI can contribute to early detection of diseases, drug discovery, and global

health initiatives by analyzing trends across populations.



This conference is an opportunity for us to delve into the cutting-edge research, emerging

applications, and ethical considerations surrounding the use of ChatGPT in medical practice.

Together, we will explore both the current capabilities and the future possibilities, while also

addressing the challenges that come with integrating AI into such a sensitive and high-stakes

field.



Matjaž Gams

Monika Simjanoska Misheva

Stevo Lukić

Franz Wotawa

ChatGPT in Medicine chairs



PROGRAMSKI ODBOR / PROGRAMME COMMITTEE



Matjaž Gams



Monika Simjanoska Misheva



Stevo Lukić



Franz Wotawa



Žiga Kolar





Automatic Reviewing of Conference Papers in Healthcare and




Other Sciences Using ChatGPT



Matjaž Gams Luka Bolješić Alex Korlat Primož Kocuvan

Department of Intelligent University of Ljubljana University of Ljubljana Department of Intelligent

Systems Faculty of Computer and Faculty of Computer and Systems

Jozef Stefan Institute Information Science Information Science Jozef Stefan Institute

Ljubljana Slovenia Ljubljana Slovenia Ljubljana Slovenia Ljubljana Slovenia

matjaz.gams@ijs.si luka.boljesic0@gmail.co alex.korlat.ak@gmail.com primoz.kocuvan@ijs.si

m



ABSTRACT raziskovalne prispevke, in uporablja ChatGPT za generiranje

pregledov za vsak prispevek. Obravnavane so metodologija,

The emergence of artificial intelligence (AI) has significantly rezultati in možni vplivi te aplikacije, ki poudarjajo tako

impacted various fields, including the peer review process in prednosti kot izzive integracije UI v akademski pregledni

scientific research. ChatGPT, a large language model developed postopek.

by OpenAI, has shown potential in automating and enhancing



the review of conference healthcare papers. Our HealthReview KLJUČNE BESEDE AI tool can process and analyze large volumes of text rapidly,

Umetna inteligenca, ChatGPT, strokovni pregled, zdravstvo,

providing feedback and insights that streamline the peer review

znanstvene raziskave, Flask ogrodje, obdelava PDF, akademsko



paper presents a web application developed using the Flask pisanje, konferenčni prispevki process, reduce human workload, and increase efficiency. This

framework that enables users to upload PDF files containing



research papers and utilizes ChatGPT to generate reviews for 1 Introduction each paper. The methodology, results and potential implications

of this application are discussed, highlighting both the The integration of artificial intelligence (AI) across various

advantages and the challenges of integrating AI into the domains is significantly transforming complex processes,

academic review process.1 including the peer review of conference papers in healthcare

and other sciences. Among AI tools, ChatGPT, developed by

KEYWORDS OpenAI, stands out for its potential to automate and enhance the

Artificial Intelligence, ChatGPT, Peer Review, Healthcare, review process. Its ability to quickly analyze large volumes of

Scientific Research, Flask Framework, PDF Processing, text and provide insightful feedback could streamline peer

Academic Writing, Conference Papers reviews, reduce human workload, and enhance overall

efficiency.



POVZETEK

Recent studies highlight the diverse applications of ChatGPT in



Pojav umetne inteligence (UI) je pomembno vplival na različna healthcare education, research, and practice. For example, it has been shown to improve scientific writing, analyze datasets, and področja, vključno s postopkom strokovnega pregleda v aid drug discovery [1, 2, 3]. Additionally, its role in generating akademskih in znanstvenih raziskavah. ChatGPT, velik paraphrased content and literature reviews indicates the jezikovni model, ki ga je razvil OpenAI, je pokazal potencial za potential to expedite academic tasks, although concerns about avtomatizacijo in izboljšanje pregleda medicinskih originality and accuracy persist [4, 5]. konferenčnih prispevkov. To orodje UI lahko hitro obdela in

analizira velike količine besedil ter zagotovi povratne

ChatGPT's integration into medical literature reviews has been

informacije in vpoglede, ki poenostavijo postopek strokovnega

explored, demonstrating its ability to synthesize medical

pregleda, zmanjšajo delovno obremenitev in povečajo

knowledge, though ethical and accuracy issues require further

učinkovitost. Ta članek predstavlja spletno aplikacijo

research [6, 7]. Beyond healthcare, ChatGPT enhances research

HealthReview, razvito s pomočjo ogrodja Flask, ki

efficiency across various scientific fields. It effectively

uporabnikom omogoča nalaganje datotek PDF, ki vsebujejo

generates Boolean queries for systematic reviews and supports

1 Permission to make digital or hard copies of part or all of this work for personal rapid literature searches [8]. The AI's potential to streamline

or classroom use is granted without fee provided that copies are not made or peer reviews and address biases, is also evident, though

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this managing issues like bias, plagiarism, and inaccuracies remain

work must be honored. For all other uses, contact the owner/author(s). crucial to maintaining academic integrity [9, 10, 11]. In our

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia opinion, tools like ChatGPT offer significant opportunities to

enhance the peer review process. However, careful deployment

© 2024 Copyright held by the owner/author(s). is necessary to ensure ethical considerations, accuracy, and the

preservation of academic integrity. This paper explores these

https://doi.org/10.70314/is.2024.chtm.3



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia M. Gams et al.



aspects and presents insights into effectively integrating JavaScript enhances user experience by displaying the selected

ChatGPT into the peer review process designed with our system file name and showing a loading indicator upon form

HealthReview. The paper consists of Section 1 Introduction, submission. Upon file upload via the /submit route, the PDF is



are demonstrated in Section 3, and the paper concludes with a that only files with a .pdf extension are accepted. Once the file discussion. is stored, the section 2 Methodology where the system is presented. Results securely saved in the upload folder. The file function ensures

papers_strArr function extracts the text of

individual papers from the PDF. This function searches for



2 specific markers like "ABSTRACT" and "Author index" to Methodology identify the boundaries of each paper. Text is extracted from

each page and grouped into separate papers based on these

When tested, GPT-4o was already able to reply to all basic markers. If no papers are detected, an error message is returned.

questions, e.g. “Is grammar in that text correct”. Therefore, the Otherwise, the program proceeds to each extracted paper.

first task was to create a list of commands to perform the

sequence of the review, chosen for the Information Society The API Call function facilitates the interaction with the

conference (is.ijs.si). The second task was to fine-tune the ChatGPT model. It sends the text of each paper to the ChatGPT

process since several output issues were not as desired. The API, along with a predefined prompt, and retrieves the model's

third task was to include additional knowledge, and that was generated review. These reviews are stored in a global

executed by including the Insieme knowledge base as the core dictionary RESULTS with unique identifiers for each paper.

medical information. In this way, the HealthReview

performance achieved reasonable levels. The /paper/<paper_id> route enables users to view the

generated reviews. When a specific paper ID is requested, the

The Insieme platform was selected for integration with GPT corresponding review content is fetched from the RESULTS

[12]. This platform features a user-friendly interface that

enables users to efficiently access valuable healthcare dictionary and displayed using the paper.html template.

information from a centralized website, either via manual Navigation links allow users to browse through the list of

search or through the Insieme search function. Insieme serves papers and view their respective reviews.

as the successor to the national Electronic and Mobile Health

(eHealth) initiative, a project that was characterized by HTML templates, including form.html and paper.html,



collaboration among 15 partners. Furthermore, the platform's provide the user interface for uploading files, displaying

development has been significantly shaped by insights derived instructions, and viewing individual paper reviews. The

from the analysis of various EU healthcare platforms, form.html template includes the file upload form and

particularly those that specialize in elderly care. Insieme not instructions for use, while the paper.html template displays

only builds upon the foundation laid by these prior initiatives individual paper reviews and includes navigation for switching

but also aims to set a new standard in the accessibility and between different papers. This program integrates Flask for web

usability of healthcare information, thereby providing a more functionality, PDFMiner for PDF text extraction, and OpenAI's

integrated and cohesive experience for its users. Additional ChatGPT-4o for generating paper reviews. This combination

medical knowledge therefore comes from the Insieme database offers a streamlined and efficient method for reviewing

that was created from a national electronic and mobile project conference papers, potentially enhancing the speed and

for smart cities. Including the Insieme platform into the GPS is consistency of the review process, which is crucial in managing

described elsewhere and is not part of this paper. The the increasing volume of research submissions. HealthReview

HealthReview web application is developed in Python using the accepts input in the form of any language and similarly outputs

Flask framework. Flask is a lightweight web framework for the replies in any language supported by ChatGPT-4o.

Python programming language, used for building fast and stable

web applications [13]. This application allows users to upload

one or several PDF files containing multiple research papers, 3 Results



processes these files to extract the individual papers, and then The developed web application HealthReview was tested using

employs ChatGPT to generate reviews for each paper. a dataset of conference papers to evaluate its performance and

efficiency in generating reviews. An example of the

The application is initialized and configured to use an upload application’s output is shown below. folder named 'uploads'. This folder is created if it does not exist

to ensure that uploaded files have a designated storage location. Title: Anonymous in this paper

To maintain file security, the application restricts uploads to Authors: Anonymous

PDF files by specifying an allowed file extension set.The user Abstract:

interface of the application consists of an HTML form rendered The article focuses on physical inactivity, which increases the

by the index route. This form allows users to upload their PDF risk of chronic diseases, especially cardiovascular diseases,

files through a file input and submit button. Client-side among the adult population as well as among children and



Automatic Reviewing of Conference Papers in Healthcare

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

and Other Sciences Using ChatGPT



youth. Based on data from the Institute for Health Metrics and The HealthReview web application successfully generated

Evaluation, the article reports relative risks of cardiovascular reviews for multiple conference papers. Table 1 represents a

disease by age group, gender, and country. In doing so, we summary of the evaluations for the conference section. The

focus on the comparison between Slovenia and the world paper number is a serial number assigned to each paper in the

average. Based on the analysis, we conclude that cardiovascular conference dataset. Examples of the meanings of other fields

diseases impact people in different age segments; as such they are listed in this section.

become one of the important risk factors through aging. Also,

physical activity importantly influences behavioral, nutrition,

and other factors, which contribute to cardiovascular diseases. 4 Discussion



For that reason, it is very important to systematically analyze The results of our study demonstrate the promising potential of

and monitor these factors to contribute to better guidelines and integrating AI, specifically ChatGPT, into the peer review

measures of policymakers. process for healthcare and medical conference papers. The

application that was developed effectively generated insightful

Overall evaluation: 1 - Accept with minor changes reviews, which were evaluated against traditional human-

(All options: 2 - Accept as is | 1 - Accept with minor changes | 0 -

Accept with major changes | -2 - Reject) generated reviews for quality and consistency. While the overall

The paper provides valuable insights into the relationship between performance of the AI-based review system was favorable,

physical inactivity and cardiovascular diseases, particularly in different several considerations and implications warrant further

age segments. However, some minor improvements could elevate the discussion. First, the ability of ChatGPT to process and analyze

presentation and clarity of the findings. large volumes of text rapidly offers a significant advantage in

Overall quality: 7/10. terms of efficiency. This is particularly beneficial in the context

The scientific quality is solid, but minor methodological limitations and of increasing research output and the growing burden on peer

the need for more comprehensive data analysis slightly reduce its reviewers. However, while the AI-generated reviews were

Presentation: generally accurate and aligned with human assessments, there 3 – Routine impact.

(All options: 5 - Crystal-clear and particularly attractive | 4 - Clear and were instances where the feedback provided by ChatGPT

attractive | 3 - Routine | 2 - Not very clear and/or attractive | 1 - Unclear lacked depth, particularly in areas requiring domain-specific

and/or particularly unattractive.) expertise. This highlights a limitation of current AI

The presentation is well-structured but lacks visual aids and a more technologies, where the ingenuity, flexibility and expertise of

engaging writing style. human reviewers are still crucial.

Language modifications: Minor grammatical mistakes. The paper

includes some minor grammatical mistakes that should be fixed before It is not clear to what extent the automatic reviewing system

publication. applies to all domains, not only healthcare. In any case, when

Proposed modifications: Correct "prerana" to "prehrana" in the

"REZULTATI" section. Ensure consistency in the use of terms such as adopting a specific form of review, the instructions as part of

"risk factors" and "prevalence" across sections. the prompt programming should be modified. In practical

Improve the translation and readability of some sentences for better terms, it should also be noticed that the API GPT call is not free

clarity. of charge.

Consider adding graphical representations of data to enhance

interpretability. Additionally, the application of AI in the peer review

Suggestion: Use tools like Grammarly or GPTs. process raises ethical concerns, particularly around the potential

Reviewer's confidence: 4 – High for bias, the risk of plagiarism, and the integrity of the review

(All options: 5 - Expert | 4 - High | 3 - Medium | 2 - Low | 1 - None) process. Although ChatGPT can streamline the review process,

Excellence: 3 - Definite but modest methodological novelty and/or

good application these tools must be used as supplements rather than

(All options: 5 - Important methodological novelty | 4 - Definite replacements for human reviewers. Maintaining a balance

methodological novelty | 3 - Definite but modest methodological between AI efficiency and human oversight is critical to

novelty and/or good application | 2 - Debatable methodological novelty preserving the integrity and quality of academic peer reviews.

and/or acceptable application | 1 - No methodological novelty and poor Moreover, the reliance on AI for academic tasks necessitates

application) continuous monitoring and updates to the AI models to ensure

Impact: 3 - Used/useful and moderately mature accuracy, relevance, and fairness. Future developments should

(All options: 5 - Highly impactful application and/or very widely used, focus on enhancing the contextual understanding of AI tools

at least moderately mature | 4 - Important application and/or widely

used, at least moderately mature | 3 - Used/useful and moderately like ChatGPT to better mimic the critical thinking and

mature | 2 - Used/useful in limited cases and/or immature | 1 - No analytical capabilities of human reviewers.

practical use)



Table 1: Summary of Reviews



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia M. Gams et al.



Paper number Overall evaluation Overall quality Presentation Language (grammatical mistakes)

1 2/4 6/10 2/5 Quite some

2 3/4 7/10 3/5 Minor

3 3/4 7/10 3/5 Some

4 3/4 7/10 3/5 Quite some

5 3/4 7/10 3/5 Some

6 3/4 7/10 3/5 Minor

7 3/4 8/10 3/5 Minor

8 3/4 8/10 3/5 Some

9 3/4 7/10 3/5 Minor

10 3/4 7/10 3/5 Minor

11 3/4 8/10 4/5 Some

12 3/4 7/10 3/5 Minor

13 4/4 9/10 4/5 Minor



In conclusion, the inclusion of HealthReview, i.e. an additional [5] S. Ruksakulpiwat, A. Kumar, and A. Ajibade. 2023. Using chatgpt in

automated review layer introduces several advantages, such as medical research: current status and future directions. Journal of

increased objectivity and the potential to generate Multidisciplinary Healthcare, 16, 1513–1520

https://doi.org/10.2147/JMDH.S419830.

supplementary suggestions, further enriching the review

process. AI tools like ChatGPT offer substantial potential to [6] M. Sallam. 2023. Chatgpt utility in healthcare education, research, and

enhance the peer review process. However, their successful prac- tice: systematic review on the promising perspectives and valid

integration requires careful implementation and continuous concerns. Healthcare, 11. doi: 10.3390/healthcare11060887 D. Solomon,

evaluation to effectively address inherent challenges and ensure K. Allen, P. Katz, A. Sawalha, and E. Yelin. 2023. Chatgpt, et al...

artificial intelligence, authorship, and medical publishing. Arthritis &

that these tools make a meaningful contribution to academic Rheumatology, 75. doi: 10.1002/art.42497

research. In any case, the automatic review by HealthReview or

any other review tool should be marked in a way explicitly [7] O. Temsah et al. 2023. Overview of early chatgpt’s presence in medical

denoting the source and type of the reviewing tool. literature: insights from a hybrid literature review by chatgpt and human

experts. Cureus, 15. doi: 10.7759/cureus.37281



5 [8] S. Wang, H. Scells, B. Koopman, and G. Zuccon. 2023. Can chatgpt Acknowledgments

write a good boolean query for systematic review literature search? In



This project is funded by the European Union under Horizon Proceedings of the 46th International ACM SIGIR Conference on

Research and Development in Information Retrieval. doi:

Europe (project ChatMED grant agreement ID: 101159214) 10.1145/3539618.3591703



REFERENCES [9] Biswas, D. Dobaria, and H. L. Cohen. 2023. Chatgpt and the future of journal reviews: a feasibility study. The Yale Journal of Biology and

[1] R. Garg, V. L. Urs, A. A. Agrawal, S. K. Chaudhary, V. Paliwal, and S. Medicine, 96, 415–420. https://pubmed.ncbi.nlm.nih.gov/37407690/

K. Kar. 2023. Exploring the role of chatgpt in patient care (diagnosis and

treatment) and medical research: a systematic review. Health Promotion [10] D. Solomon, K. Allen, P. Katz, A. Sawalha, and E. Yelin. 2023. Chatgpt, Perspectives, 13, 183–191. et al... artificial intelligence, authorship, and medical publishing. Arthritis

https://hpp.tbzmed.ac.ir/Article/HPP_19952_20230501034121 & Rheumatology, 75. doi: 10.1002/art.42497



[2] P. Y. Lee, H. Salim, A. Abdullah, and C. H. Teo. 2023. Use of chatgpt in [11] J. Verharen. 2023. Chatgpt identifies gender disparities in scientific peer

medical research and scientific writing. Malaysian Family Physician, 18, review. eLife, 12. https://doi.org/10.7554/eLife.88173

58. https://www.e-mfp.org/ojs3/index.php/MFP/article/view/1213

[12] The Insieme platform presentation: https://www.youtube.com/watch?

[3] F. Tustumi, N. Andreollo, and J. E. Aguilar-Nascimento. 2023. Future of v=ozAporFOw64, Accessed 1.9.2024

language models in healthcare: the role of chatgpt. Arquivos Brasileiros

de Cirurgia Digestiva: ABCD, 36. https://doi.org/10.1590/0102- [13] Flask framework documentation:

672020230002e1800 https://flask.palletsprojects.com/en/latest/ Accessed 1.9.2024



[4] Ö. Aydın and E. Karaarslan. 2022. Openai chatgpt generated literature

review: digital twin in healthcare. In Emerging Computer Technologies

2. Ö. Aydın, editor. İzmir Akademi Dernegi, 22–31.

https://doi.org/10.5281/zenodo.7094305S.





Leveraging Federated Learning for Secure Transfer and




Deployment of ML Models in Healthcare *



Zlate Dodevski Tanja Pavleska Vladimir Trajkovik

Faculty of Computer Science Laboratory for Open Systems Faculty of Computer Science

and Engineering and Networks and Engineering

Ss. Cyril and Methodius University Jozef Stefan Institute Ss. Cyril and Methodius University

Skopje, Republic of N. Macedonia Ljubljana, Slovenia Skopje, Republic of N. Macedonia

zlate.dodevski@students.finki.ukim.mk atanja@e5.ijs.si vladimir.trajkovik@finki.ukim.mk



Abstract endless pool of structured and unstructured data, as these technologies thrive on large datasets to identify patterns and Federated learning (FL) represents a pivotal advancement make predictions. The novel adaptable, predictive approach to in applying Machine Learning (ML) in healthcare. It addresses generating insights, decision support, and assistance in tasks that the challenges of data privacy and security by facilitating model have long been considered solely reserved for human expertise transferability across institutions. This paper explores the is based on this paradigm's capabilities to recognize patterns effective employment of FL to enhance the deployment of large from the data without being explicitly programmed. language models (LLMs) in healthcare settings while Conventional machine learning implies that the data owner maintaining stringent privacy standards. Through a detailed communicates with a specific central server with significant examination of the challenges in applying LLMs to the computational power. The central server consumes data from healthcare domain, including privacy, security, regulatory different sources and applies training techniques and algorithms constraints, and training data quality, we present a federated to devise an effective model. ML requires large amounts of data learning architecture tailored for LLMs in healthcare. This to satisfy the expectations for the model's performance. architecture outlines the roles and responsibilities of Large-language models (LLMs), as representative of ML participating entities, providing a framework for secure advancements, have been a particular point of interest in recent collaboration. We further analyze privacy-preserving techniques years. They have already proven their applicability and massive such as differential privacy and secure aggregation in the context potential in multiple fields [30]. LLMs are designed to of federated LLMs for healthcare, offering insights into their understand, generate, and interact with human-like text and can practical implementation. understand context, making them suitable for performing a wide

Our findings suggest that federated learning can range of complex language-related tasks. They are trained in two

significantly enhance the capabilities of LLMs in healthcare main phases. First, the model learns general knowledge about

while preserving patient privacy. In addition, we also identify language patterns in the pre-training phase. Then, it can be fine-

persistent challenges in areas such as computational and tuned to execute downstream tasks to specialize its expertise in a

communicational efficiency, lack of benchmarks and tailored FL specific domain [28].

aggregation algorithms applied to LLMs, model performance, However, like other machine learning models researched and

and ethical concerns in participant selection. By critically implemented, this paradigm is data-hungry, meaning that it

evaluating the proposed approach and highlighting its potential inherently requires massive training data to achieve the expected

benefits and limitations in real-world healthcare settings, this performance [10]. Thus, LLMs are designed to perform better

work provides a foundation for future research in secure and with an increase in training data volume and computational

privacy-preserving ML deployment in healthcare. power.

Various unfortunate scenarios related to the misuse of private and

Keywords personal data cast a shadow on AI's capabilities, underscoring the

Federated Learning, Large Language Models, Data Privacy, growing concerns about data privacy, specifically in the phases

Healthcare ML, Privacy Preservation, Model Transferability when the models are trained. [13]. The year 2016 is particularly

significant for two key developments aiming to overcome these

1 Introduction challenges. The first is the attempt to regulate personal data



hyper-connectivity, and the fourth industrial revolution lead to The advancements in hardware and software technologies, collection, processing, and storing by introducing the General Data Protection Regulation (GDPR) in Europe [12]. The second key development was the introduction of Federated Learning the creation of mass amounts of health-related data. Machine (FL) by Google researchers, which provided a groundbreaking learning and AI, in general, are the biggest winners from this scientific approach to addressing data privacy and security

concerns in ML [16].

* Permission to make digital or hard copies of part or all of this work for personal This paper aims to discuss the possibility of satisfying the

or classroom use is granted without fee provided that copies are not made or

distributed for profit or commercial advantage and that copies bear this notice and needs of both data owners and ML experts by leveraging the

the full citation on the first page. Copyrights for third-party components of this concept of federated learning. On the one hand, data owners can

work must be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia be supported to participate in collaborative training in a privacy-

© 2024 Copyright held by the owner/author(s). preserving manner when their data is insufficient to craft a high-

https://doi.org/10.70314/is.2024.chtm.5



performance model, such as LLM. On the other hand, ML of hesitation. Most of the training data for the LLMs comes from

experts can develop and advance their approaches by utilizing publicly available sources, lacking the nuances that bring the

large volumes of real-life institutional data and access to diverse real-life data isolated in the infrastructure of a single healthcare

scenarios, which are essential for building a robust model. institution. Additionally, each institution has many

Being aimed at investigating the FL potential for application characteristics that make it unique in how it works. Workflows,

in the health domain, the insights presented in this work offer dictionaries, specific characteristics of the population it serves,

support in finding a more robust, secure, and effective use of AI or particular domain attributes can result in difficulties for off-

that does not require technical proficiency of the medical experts, the-shelf LLMs in providing the correct output to the instruction

ultimately contributing to improved patient care and data given. This calls for the institutions to further tailor and tune the

protection. capabilities of the LLM. Although state-of-the-art LLMs allow

The remainder of this paper is organized as follows: Sections for such modifications and fine-tunings and making this process

2 and 3 introduce FL and LLMs, and their relevance to feasible, this comes with a heavy involvement and effort by the

healthcare; section 4 presents our proposed FL architecture for institution representatives and with extensive computational

LLMs in healthcare, detailing the system components and their resources. Finally, even if one institution is capable of making

roles; section 5 discusses challenges in implementing LLMs in efforts to adapt a generalized LLM for its use, it faces the

healthcare using FL; and section 6 explores privacy-preserving inevitable obstacle of data insufficiency. In general, a single

techniques for integration with our FL architecture. institution either cannot provide enough data to receive proper,

correct output for the downstream task or is incapable of solving

2 LLM-based Healthcare Applications instruction of so-called new events or conditions.

LLMs are usually trained on high-quality public data, but A healthcare institution would need support in multiple areas

their performance is often limited when tasked with specialized to make the process easier to follow and adopt. As a result,

or narrower-spectrum tasks. With specific expertise in mind, processes related to finding a suitable LLM model, maintaining

such as healthcare, different approaches should be considered to it, and keeping it up to date should be outsourced to a separate

build that corpus. Healthcare institutions can use their local body owning the expertise. To effectively adapt LLMs in the

datasets, leading to less efficient LLM, or they can join other healthcare domain, collaboration among institutions in

collaborative efforts to make high-quality training data facing the compliance with the industry regulations should be established

inevitable challenges of privacy and regulations. to build a rich training corpus.

LLMs are attractive in the healthcare area because of their

capability to simplify the interaction with an intelligent system 3 FL Principles Relevant to Healthcare Data

without needing technical expertise. Lack of technological Privacy and Security

proficiency of medical experts can decrease the adoption of a In healthcare, data is often distributed across multiple

specific software solution and can be marked as overly complex. institutions, each possessing unique and valuable patient

The core principles of the LLMs allow users to interact with their information. Traditional approaches to AI model training require

domain rules, persistent knowledge, and past experiences centralizing this data, which poses significant privacy and

without the need to rely on their computer literacy. One of the security risks. Federated learning provides a solution by enabling

enormous benefits of utilizing LLMs in potential areas of collaborative model training without exchanging raw data.

medical workflows relies on the output, which is in an Instead, each institution trains the model locally and shares only

understandable form of natural language. The ease of use of aggregated updates with a central server. This method ensures

natural language to provide instructions and ask for decision that sensitive patient data remains within the institution,

support bridges the gap between the domain experts and the facilitating the transfer and deployment of AI models across

utilization of an intelligent computer system. Furthermore, much different settings without compromising data security. FL is an

of the data that persists within healthcare institutions is in iterative process, and each communication round aims to

unstructured formats, such as clinical notes, conversations, improve the model's performance. A typical FL scenario consists

diagnoses, prescriptions, and research articles. LLMs are of two main phases in each round: local training of the model

particularly effective at processing and using these natural done on the participant side and aggregation of updates, which

language texts. In that way, the transparency is increased, and the aims to create the most accurate consensus model.

expert can examine the reasoning behind the answers provided There are three main types of FL based on how the data is

straightforwardly. distributed across participants. In horizontal federated learning,

In the past several years, we have already witnessed the the datasets share the same feature space but differ in the samples

potential of LLMs in healthcare in many areas, such as they contain. Vertical federated learning, on the other hand,

interpreting images from a specific medical domain, involves datasets with the same samples but different features.

summarizing reports and medical history, identifying patterns in Lastly, federated transfer learning encompasses datasets that

electronic health records (EHR), and offering support for vary in both their feature and sample spaces [8].

decision-making processes. The use of natural language can also FL in healthcare is predominantly covered in theoretical

influence patient engagement processes [16]. research, with many studies exploring its potential, such as for

Even though exceptional remarks on providing suitable improved personalized treatment and public health monitoring.

LLMs for healthcare are already in place and the community is However, there are real-life applications, such as in radiology,

verifying the theoretical and conceptual findings, the decision of where FL enables collaborative training on medical images like

one healthcare institution to utilize such a system can face a lot MRIs and X-rays without sharing patient data [23].

In the context of machine learning (ML) applications aggregation server or global server), and the communication-

involving healthcare data, there are three critical vulnerability computation layer, which aggregates local updates and

points that require attention: the data itself, the training of ML orchestrates communication phases in the ecosystem. Each

models, and the communication and transfer of data. Each area component has its own responsibilities, which are essential for

carries specific challenges and risks that must be mitigated to the model to satisfy the preset expectations.

ensure the privacy, security, and efficacy of ML systems in Leveraging FL in utilizing LLMs adds a layer of complexity and

healthcare. Health-related data is inherently complex, with implies different approaches based on the level of

characteristics such as high dimensionality, variance over time, decentralization that needs to be achieved [29]. FL can help in

heterogeneity, difficult interoperability, sparsity, and isolation both the pre-training and fine-tuning phases of LLM, and it is up

[4]. Protecting the privacy of patients' personal and sensitive to the requirements' specific characteristics and the parties

health information is crucial. Due to the sensitive nature of involved computational power to choose the right strategy [2,

healthcare data, security breaches can lead to severe 11]. We will cover the different approaches while examining the

consequences, including identity theft, fraud, and violation of three major architecture components.

patient confidentiality.



Healthcare data often comes from various sources, such as

hospitals, clinics, wearable devices, and electronic health records

(EHRs). This data is typically non-independent, identically

distributed (non-iid), unbalanced, and fragmented across

different systems. Additionally, data may be sparse or isolated,

making it challenging to build comprehensive patient profiles or

conduct large-scale analyses.

Federated learning offers a promising approach to

overcoming these challenges by allowing ML models to be

trained across multiple decentralised data sources while keeping

data local. This technique improves data privacy and security by

not requiring raw data to be transferred to a central location. In

an FL environment, each data controller defines its governance

processes and privacy policies. This includes setting conditions

for data access, training, and validation phases [3, 7, 19].

Communication between institutions, especially in healthcare,

must adhere to strict regulatory requirements, such as the Health

Insurance Portability and Accountability Act (HIPAA) in the U.S. Figure 1: Typical FL Architecture that can be deployed for

and GDPR in Europe. Thus, privacy-preserving mechanisms LLM-based applications

should be implemented "by design" and "by default" to ensure

that sensitive patient information is processed securely. A 4.1 Global Server compliant ML system requires secure data transfer mechanisms,



consent management, and audit trails. The FL setting can offer properly established taxonomy and standards, this component in The global server plays a central role, and due to the lack of

advancement in this area by letting institutions keep sensitive

the literature is also considered as the manager or creator of the

information, prevent unnecessary data transfers and processing

whole ecosystem. It is often referred to as the "manager" or

that could violate regulatory requirements, and minimize the risk

"aggregation server" tasked with overseeing the entire

of data breaches [27].

collaboration and ensuring its smooth functioning. In healthcare,

Training ML models with healthcare data presents unique

the manager can be a single healthcare institution that holds a lot

challenges, including addressing data bias, limited sample sizes,

of data and wants to leverage the FL setting to collaborate with

and ensuring model performance. Healthcare data may be biased

other institutions, either to complete the missing domains and

due to demographic imbalances, socio-economic factors, or

dimensions (by utilizing Vertical FL) or to enrich and expand the

varying levels of care access across populations.

feature set in the same dimension (Horizontal FL). The global

Training of ML models with diverse datasets enhances their

server's responsibilities can be broken down into several distinct

generalizability and robustness. By incorporating data from

areas: strategy for choosing a foundation model, strategy for

various sources and populations, models can better adapt to new

exploiting data distribution and the client selection.

and unforeseen health events, improving their predictive power

The foundation model, also called the base model, forms the

and reliability. Federated learning, in particular, enables the use

initial point of the LLM training in the FL setting. It represents a

of diverse datasets while maintaining privacy, thus improving

starting point for institutions to leverage the pre-training process

overall model performance [21, 17].

with their own data or fine-tune it to perform specific



4 FL Architecture for LLM-Based Healthcare publicly available datasets. The pre-training process is downstream tasks since foundation models are usually trained on

Applications computationally and time-consuming, meaning that the global

Figure 1 depicts the three major components of a typical FL server must find the most suitable scenario for satisfying

architecture. The participants involved in our cross-silo FL requirements.

setting are the healthcare institutions, the manager (e.g.,



There are two main kinds of pre-training models: one based generalize and can handle novel instructions even though they

on the BERT model and the other on the GPT model. Both were not a part of the fine-tuning instruction dataset.

perform differently for different tasks and scenarios [5, 22]. The variety of downstream tasks that LLMs can perform in the

Many attempts are made to use publicly available literature healthcare domain is often the critical reason institutions engage

specific to the medical domain and create ready LLM models for with this concept. Based on a benchmark for generalist

usage, such as BioBert, which show superior performance than biomedical AI, some of the most frequently performed medical-

general pre-trained models [15]. relevant tasks suitable for the LLM domain are question

One approach is to select a suitable foundation model based answering, visual question answering (for example, based on

on the options examined before or to initialize FL pre-training, radiology or pathology images), report summarization and

where each party will contribute to the pre-training of the generation, and medical image and medical documents

foundation LLM model. The latter approach requires the classification [26]. Additionally, relation extraction in

institutions to have expertise and sufficient computational power combination with named entity recognition can be added to the

in their infrastructure to complete the assignment. Another list of medical-relevant tasks. This is helpful in the medical

consideration is the model's size and complexity, which will domain to extract medical terms such as diseases, conditions,

influence the following steps if not chosen according to the procedures, and symptoms from unstructured data and find

participant's IT infrastructure. suitable interpretations and connections in the unstructured data

Federated Learning can work differently depending on how pool.

data is spread and distributed across entities. In horizontal FL, Thanks to the LLaMA, each institution can make significant

each institution has data with similar features (for example, attempts to build its domain-specific instruction set and

multiple hospitals with similar patient data). In vertical FL, contribute to global instruction tuning in the FL setting. With the

institutions have different features for the same set of patients FL paradigm, each downstream task can be trained on multiple

(for example, one entity has clinical data, and another has genetic datasets instead of a single dataset, giving more suitable

data). Transfer learning can also be used when the model needs responses and outputs [25].

to generalize across different datasets [20]. The global server is In the FL setting, the value-alignment step occurs on the

responsible for choosing the appropriate strategy based on the participant's side during local training. Its purpose is to ensure

data distribution and the desired outcomes. that the model's objectives are aligned with each institution's

Client selection in FL refers to choosing which institutions to values and goals. This step is particularly crucial in the medical

participate in each training round. The global server must ensure field, where ethical guidelines and patient care standards are of

that diverse institutions contribute to model updates without the utmost importance.

overloading the communication system. Institutions with more Technically, value-alignment is solved similarly to

data or better computational resources might participate more instruction tuning, with each participant's preference dataset

frequently. Still, the system should be flexible enough to rotate containing combinations of instruction, preferred, and

clients or dynamically adjust client participation based on misreferred responses.

resource availability. In FL, participants typically have far fewer computational

resources than centralized cloud servers and fine-tuning all

4.2 Participants parameters of LLMs can be an obstacle. Parameter-efficient

The entities participating in this collaboration technique are tuning techniques, such as Lora, are used to address this

also tasked with significant responsibilities. In the medical limitation [6]. Instead of updating the entire pre-trained model

domain, these institutions own huge structured or unstructured weights to obtain local updates, participants modify only specific

datasets and are willing to participate in a distributed training parameters and send them back to the global server for

process. Their responsibilities can be broken down into the pre- aggregation.

training process, fine-tuning, value alignment, and strategy for

local updates. 4.3 Communication-Computation Layer

If the FL-specific training approach is adopted, as discussed As presented above, the global server is responsible for

previously in the strategy for choosing the foundation model, managing the whole ecosystem, and one of the most complex

then each institution may pre-train the selected model on its data tasks is related to the communication-computation layer. The

and ensure that the initial model updates sent to the global server global server should manage the aggregation process of local

are more relevant and valuable. This capability in an FL setting model updates and ensure that the global LLM and updates are

can depend significantly on the institution's computational securely transmitted across the system.

power, and even though possible and theoretically feasible, it can Choosing the suitable FL algorithm for combining all

require a lot of additional expertise for the healthcare institution findings and improvements made by each participant in the form

to engage in this kind of activity. of parameter weights is a step that has attracted many researchers

Fine-tuning is crucial in adopting an LLM in the healthcare and experts. One of the first and most used algorithms is

domain. The approach of LLM fine-tuning is to make Federated Averaging (FedAvg), but more sophisticated

downstream tasks required by the domain be instructed with approaches may be necessary in different scenarios [16]. The

human feedback [18]. Each institution should provide input- model's performance relies significantly on how model updates

output pairs where instructions are explicitly offered to solve are aggregated. Even though the only data transmitted through

some already defined downstream tasks. These datasets are the network in an FL setting are the model and its updates, the

designed to give the model an idea of what kind of output is communication layer is responsible for ensuring that the transfer

expected. The expectation is that the LLM will learn to is done securely and continuously. The communication layer



component must develop a strategy for creating a pipeline from LLM training in the FL setting. Fine-tuning LLMs in FL is a

a live data connection to the model and inference to transmitting time-consuming and computationally expensive task [6].

new model parameters via secure channels to the aggregating The client selection process, in which the ecosystem manager

server. Size and complexity of the model must be considered as decides which participants should be involved, can raise many

well, since they can cause a bottleneck. ethical concerns, such as fairness. The purpose of the

In addition, the communication layer also ensures that the collaboration is to make the LLM more robust. Still, some

data transfer is seamless and uninterrupted. This component is participants' data volume and computational power can squeeze

tasked with developing a robust strategy to create an efficient out institutions that are not on that level but still can add to the

pipeline, from managing real-time data connections to diversity and offer unique cultural, ethical, and contextual

facilitating model utilization and transmitting updated model values. While FL addresses many privacy concerns by design, it

parameters securely to the central aggregation server. A key also introduces new security considerations that must be

consideration for the communication layer is the size and carefully managed. Successfully navigating these challenges

complexity of the used model. Large models with huge requires a detailed approach that balances privacy protection,

parameter lists can introduce significant bottlenecks during security enhancement, and the pursuit of practical and robust

transmission, especially when dealing with limited bandwidth or LLM in healthcare.

less powerful devices. As such, the communication layer must be

adept at handling these challenges, ensuring that updates are 6 Privacy-Preserving Techniques

transferred efficiently without compromising the speed or The deployment LLMs in the healthcare field through FL

security of the system. promises advancements in preparing models to react to given

domain-specific downstream tasks. The FL can enhance LLMs'

5 Hype, Vision and Challenges effectiveness and proper application while safeguarding patient

Implementing LLMs in healthcare using FL presents a set confidentiality and ensuring regulatory compliance, providing

of intertwined challenges when viewed through the lenses of medical professionals greater confidence in adopting these tools.

privacy and security. There is a foundational challenge between However, while FL enables collaborative learning without

the need for diverse and high-quality data generated by direct data sharing, it's not immune to privacy threats. With this

institutions in the specific domain and the importance of approach, raw data remains local, but the model updates shared

protecting sensitive information. FL enables availability and during training can still leak information. In addition, LLMs

access to a broader spectrum of data sources while maintaining trained with healthcare data could memorize and potentially

privacy. Still, the inability to directly act upon raw data can regenerate sensitive patient information. A privacy breach in this

impact the convergence of the model and model performance. context can cause severe consequences, including exposure to

Data transfer needs in FL, even though minimized to just model medical history, compromising patient confidentiality, and

updates, still introduce a risk for security attacks. This risk misuse of sensitive health information [1].

increases with the communication overhead caused by During this collaborative process, the model or its updates

distributing complex and large LLMs. could become targets for various attacks. For instance, model

By introducing a central figure in the architecture in the inversion attacks performed on the global model might allow the

name of the global aggregation server, the FL setting in LLM reconstruction of individual patient records. Similarly,

opens up a single point of failure in the ecosystem. Adversarial membership inference attacks could reveal the presence of

attacks can be performed, compromising model integrity, which specific institutions or patient data in the training, potentially

could lead to data breaches and incorrect outputs. exposing the entire medical history. Malicious participants in the

FL is still a young and immature topic in the context of process could poison the model by introducing biases or

LLM. One of the biggest challenges is the lack of benchmarks backdoors, potentially leading to improper results generated by

and comprehensive reviews that can examine the solution's the LLMs [9, 24].

success based on different tasks, architectures, the number of To counter these risks and threats, researchers and

clients, network bandwidth, computational resources, etc. These practitioners evaluate the effects of several privacy-preserving

reviews and benchmarks can further expose security and privacy- techniques, such as secure aggregation and differential privacy.

preserving issues and initiate proper risk mitigation strategies. Secure aggregation, a cryptographic protocol, allows the central

Multiple algorithms exist in the literature for aggregating local server to observe aggregated results without accessing individual

updates, but no specific algorithm is proposed or adapted for model updates. This approach maintains accuracy but adds

LLMs. significant communication costs. Differential privacy, on the

The analysis of the three major components in the previous other hand, adds calibrated noise to data or model parameters,

section pointed out the responsibilities, approaches, and offering statistical privacy guarantees. While effective against

strategies that need to be considered in order to collaboratively inference attacks, it may reduce model accuracy and require

design and implement training, and utilize LLM properly. The additional workload in the parameter-tuning process [14]. The

analysis emphasized that training LLMs in a federated learning choice of privacy-preserving techniques must be made with a

setting requires a thoughtful, tailored approach to address the thorough understanding of the specific use case, the sensitivity

unique challenges. Additionally, there are various approaches to of the data involved, and the potential impacts of privacy

take, depending on factors such as participant resources, data breaches. The tailored approach should calibrate the trade-off

distribution, model size and complexity, data transfer, etc. This between model performance and data protection. More robust

section will further examine the challenges of implementing such privacy protection might require limiting the model’s access to



much-needed data for LLMs to offer a proper answer to a specific [11] Kuang, W., Qian, B., Li, Z., Chen, D., Gao, D., Pan, X., Xie, Y., Li, Y., Ding,

B. and Zhou, J. 2024. FederatedScope-LLM: A Comprehensive Package for

task, degrading the model performance and increasing the

Fine-tuning Large Language Models in Federated Learning. Proceedings of

computational and communicational overhead. As research in the 30th ACM SIGKDD Conference on Knowledge Discovery and Data

this field progresses, finding the right balance between privacy, Mining (New York, NY, USA, Aug. 2024), 5260–5271.



system performance, and efficiency will be crucial for deploying [12] Kuner, C., Bygrave, L.A., Docksey, C. and Drechsler, L. eds. 2020. The EU General Data Protection Regulation (GDPR): A Commentary. (Feb. 2020). LLMs in healthcare using FL. [13] Laurie, G., Jones, K., Stevens, L. and Dobbs, C. 2015. A Review of Evidence

Relating to Harm Resulting from Uses of Health and Biomedical Data.



7 Conclusion [14] Li, Q., Wen, Z., Wu, Z., Hu, S., Wang, N., Li, Y., Liu, X. and He, B. 2023. A Survey on Federated Learning Systems: Vision, Hype and Reality for Data

This paper has explored the potential of FL in enhancing the Privacy and Protection. IEEE Transactions on Knowledge and Data

deployment of LLMs in healthcare settings. By enabling privacy- Engineering. 35, 4 (Apr. 2023), 3347–3366.

DOI:https://doi.org/10.1109/TKDE.2021.3124599.

preserving collaboration, FL allows healthcare institutions to

[15] Luo, R., Sun, L., Xia, Y., Qin, T., Zhang, S., Poon, H. and Liu, T.-Y. 2022.

collectively train and improve LLMs without compromising BioGPT: generative pre-trained transformer for biomedical text generation

sensitive patient data. This approach not only addresses and mining. Briefings in Bioinformatics. 23, 6 (Nov. 2022), bbac409.

fundamental privacy concerns but also enhances model DOI:https://doi.org/10.1093/bib/bbac409.

performance by leveraging diverse datasets across institutions, Communication-Efficient Learning of Deep Networks from Decentralized [16] McMahan, B., Moore, E., Ramage, D., Hampson, S. and Arcas, B.A. y 2017.

potentially improving the generalizability and robustness of Data. Proceedings of the 20th International Conference on Artificial

LLMs Intelligence and Statistics (Apr. 2017), 1273–1282. in healthcare applications. To facilitate the

[17] Nguyen, D.C et al. 2022. Federated Learning for Smart Healthcare: A

implementation of healthcare LLM with FL, we examined a

Survey. ACM Comput. Surv. 55, 3 (Feb. 2022), 60:1-60:37.

tailored architectural framework that outlines the roles and DOI:https://doi.org/10.1145/3501296.

responsibilities of participating entities. In addition, challenges [18] Ouyang, L. et al. 2022. Training language models to follow instructions with



and consideration of risks and threats were reviewed, especially human feedback. Advances in Neural Information Processing Systems. 35, (Dec. 2022), 27730–27744. in combination with privacy-preserving techniques. [19] Pfitzner, B., Steckhan, N. and Arnrich, B. 2021. Federated Learning in a

Looking ahead, several areas require further research and Medical Context: A Systematic Literature Review. ACM Transactions on

development. Internet Technology. 21, 2 (Jun. 2021), 1–31. Optimization of computational and

DOI:https://doi.org/10.1145/3412357.

communication efficiency for LLMs, development of

[20] Prayitno, Shyu, C.-R., Putra, K.T., Chen, H.-C., Tsai, Y.-Y., Hossain, K.S.,

standardized benchmarks, establishment of ethical frameworks Jiang, W. and Shae, Z.-Y. 2021. A Systematic Review of Federated Learning

for participant selection, and exploration of advanced privacy- in the Healthcare Area: From the Perspective of Data Properties and

preserving techniques are crucial for future work. Applications. Applied Sciences. 11, 23 (Dec. 2021).

DOI:https://doi.org/10.3390/app112311191.

[21] Qayyum, A., Qadir, J., Bilal, M. and Al-Fuqaha, A. 2020. Secure and Robust

8 Acknowledgments Machine Learning for Healthcare: A Survey. arXiv.



This project is funded by the European Union under Horizon Practice: Systematic Review on the Promising Perspectives and Valid [22] Sallam, M. 2023. ChatGPT Utility in Healthcare Education, Research, and

Europe (project ChatMED grant agreement ID: 101159214). Concerns. Healthcare. 11, 6 (Jan. 2023), 887.

DOI:https://doi.org/10.3390/healthcare11060887.



REFERENCES [23] Teo, Z.L. et al. 2024. Federated machine learning in healthcare: A systematic

review on clinical applications and technical architecture. Cell Reports

Medicine. 5, 2 (Feb. 2024), 101419.

[1] [Aouedi, O., Sacco, A., Piamrat, K. and Marchetto, G. 2023. Handling DOI:https://doi.org/10.1016/j.xcrm.2024.101419.

Privacy-Sensitive Medical Data With Federated Learning: Challenges and [24] Thapa, C. and Camtepe, S. 2021. Precision health data: Requirements,

Future Directions. IEEE Journal of Biomedical and Health Informatics. 27, challenges and existing techniques for data security and privacy. Computers

2 (Feb. 2023), 790–803. DOI:https://doi.org/10.1109/JBHI.2022.3185673. in Biology and Medicine. 129, (Feb. 2021), 104130.

[2] Chen, C., Feng, X., Zhou, J., Yin, J. and Zheng, X. 2023. Federated Large DOI:https://doi.org/10.1016/j.compbiomed.2020.104130.

Language Model: A Position Paper. arXiv. [25] Touvron, H., Lavril, T., Izacard, G., Martinet, X., Lachaux, M.-A., Lacroix,

[3] T., Rozière, B., Goyal, N., Hambro, E., Azhar, F., Rodriguez, A., Joulin, A., Dasaradharami Reddy, K. and Gadekallu, T.R. 2023. A Comprehensive

Survey on Federated Learning Techniques for Healthcare Informatics. Grave, E. and Lample, G. 2023. LLaMA: Open and Efficient Foundation

Computational Intelligence and Neuroscience. 2023, 1 (Jan. 2023), 8393990. Language Models. arXiv. DOI:https://doi.org/10.1155/2023/8393990. [26] Tu, T. et al. 2024. Towards Generalist Biomedical AI. NEJM AI. 1, 3 (Feb.

[4] Deep learning for healthcare: review, opportunities and challenges - PMC: 2024), AIoa2300138. DOI:https://doi.org/10.1056/AIoa2300138.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6455466/ [27] Xu, J., Glicksberg, B.S., Su, C., Walker, P., Bian, J. and Wang, F. 2021.

[5] Devlin, J., Chang, M.-W., Lee, K. and Toutanova, K. 2019. BERT: Pre- Federated Learning for Healthcare Informatics. Journal of Healthcare

training of Deep Bidirectional Transformers for Language Understanding. Informatics Research. 5, 1 (Mar. 2021), 1–19.

[6] Hu, E.J., Shen, Y., Wallis, P., Allen-Zhu, Z., Li, Y., Wang, S., Wang, L. and [28] Yang, J., Jin, H., Tang, R., Han, X., Feng, Q., Jiang, H., Zhong, S., Yin, B.

Chen, W. 2021. LoRA: Low-Rank Adaptation of Large Language Models. and Hu, X. 2024. Harnessing the Power of LLMs in Practice: A Survey on

[7] Joshi, M., Pal, A. and Sankarasubbu, M. 2022. Federated Learning for ChatGPT and Beyond. ACM Trans. Knowl. Discov. Data. 18, 6 (Apr. 2024),

Healthcare Domain - Pipeline, Applications and Challenges. ACM 160:1-160:32. DOI:https://doi.org/10.1145/3649506. Transactions on Computing for Healthcare. 3, 4 (Oct. 2022), 1–36. [29] Ye, R., Wang, W., Chai, J., Li, D., Li, Z., Xu, Y., Du, Y., Wang, Y. and Chen,

DOI:https://doi.org/10.1145/3533708. S. 2024. OpenFedLLM: Training Large Language Models on Decentralized

[8] Private Data via Federated Learning. (2024). Kairouz, P. et al. 2021. Advances and Open Problems in Federated Learning.

Foundations and Trends® in Machine Learning. 14, 1–2 (Jun. 2021), 1–210. DOI:https://doi.org/10.48550/ARXIV.2402.06954. DOI:https://doi.org/10.1561/2200000083. [30] Zhou, C. et al. 2023. A Comprehensive Survey on Pretrained Foundation

[9] Kaissis, G.A., Makowski, M.R., Rückert, D. and Braren, R.F. 2020. Secure, Models: A History from BERT to ChatGPT. arXiv.

privacy-preserving and federated machine learning in medical imaging.

Nature Machine Intelligence. 2, 6 (Jun. 2020), 305–311.

DOI:https://doi.org/10.1038/s42256-020-0186-1.

[10] Kaplan, J., McCandlish, S., Henighan, T., Brown, T.B., Chess, B., Child, R.,

Gray, S., Radford, A., Wu, J. and Amodei, D. 2020. Scaling Laws for Neural

Language Models. arXiv.





Use and Limitations of ChatGPT in Mental Health




Disorders



Jovana Petrović † † Iva Binić Maša Vacev Stevo Lukić

Clinic of Psychiatry Clinic of Psychiatry Clinic of Psychiatry Clinic of Neurology,

University Clinical center University Clinical center University Clinical center University Clinical centre

Nis, Serbia Nis, Serbia Nis, Serbia Niš, Serbia

zdravkovicj91@gmail.com ivabinic@gmail.com vacevmasa@gmail.com slukic@medfak.ni.ac.rs



Abstract / Povzetek The discourse surrounding the potential applications of ChatGPT in mental health disorders remains relatively ChatGPT is one of the most advanced and rapidly evolving underexplored. This work seeks to offer a quick overview of the large language model-based chatbots. It excels in everything current state of ChatGPT implementations within the mental from handling simple questions to performing complex medical health domain, while also projecting future advancements in examinations. While current technology cannot replace the digital mental health care through the integration and expertise and judgment of skilled psychiatrists, it can assist in development of ChatGPT technology. early detection of mental problems, patient evaluations,



differential diagnoses, psychotherapy and in planning and Current use of ChatGPT for mental health

conducting medical research. Ensuring privacy and adhering to

professional, ethical, and legal standards is crucial when disorders

processing training data. This is especially important in mental In managing mental health disorders, human contact is

health settings, where disclosing sensitive personal information especially crucial compared to other medical fields, as it is

increases the risk of data misuse and the potential for harmful accompanied by understanding and empathy. This is why, in

advice. Current uses of ChatGPT in mental health care are the most vital aspects of psychiatric work, human relationships

constrained by its design as a general chatbot, rather than a remain central [2]. While ChatGPT's practical applications in

specialized psychiatric tool. Despite this, the model proves the field of mental health are limited because it specializes in

useful for handling routine psychiatric and administrative tasks. language generation, it can still support certain routine tasks

As GPT technology evolves, it holds significant promise for within the field. Although evaluations, diagnoses,

psychiatry, psychotherapy, and patient assessments are mainly conducted including integration into diagnostics,

psychotherapy, and early detection of mental health issues. To by human therapists, ongoing trials are exploring how ChatGPT

deploy these advancements responsibly and effectively, it is might be utilized in mental health services. Several platforms,

crucial to develop and refine professional ethical standards and like ChatBeacon, or Koko, are already available on the market,

practice guidelines. claiming to offer mental health assistance powered by ChatGPT

[3,4]. For instance, Koko is a peer-support platform that

Keywords / Ključne besede provides crowdsourced cognitive therapy. It's experimenting

with using GPT-3 to generate bot-written responses to users

Keywords mental health disorders, large language models, deep

while they wait for peer responses. Koko is an online mental

learning, ChatGPT

health intervention that has reached nearly two million people,

mostly adolescents. The platform started as a clinical trial at



Introduction MIT and is based on the concept of crowdsourced cognitive



ChatGPT has emerged as one of the most advanced and rapidly hopefully about the world. Unlike traditional peer support therapy. Users are taught to help each other think more



evolving large language model-based chatbot systems. Its platforms, all interactions on our service are supported and extensive capabilities, ranging from responding to basic augmented by AI. inquiries to performing well in complex medical examinations,



have garnered significant attention from the global scientific Applied to the classification of psychiatric disorders. and research communities, prompting ongoing discourse Recent advancements in deep learning, the foundational regarding its potential applications across diverse domain [1] algorithm of GPT, have significantly impacted the field of

mental health disorders. This technology has been applied to

∗ All authors contributed equally to this research classify psychiatric disorders using neuroimaging data [5],

†Corresponding authors

develop models based on electroencephalograms [6], and utilize

Permission to make digital or hard copies of part or all of this work for personal or a range of patient characteristics for diagnosing and predicting

classroom use is granted without fee provided that copies are not made or mental disorders [7]. These deep learning models have shown

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this good diagnostic accuracy (AUC 0.74- 0.81) suggesting the

work must be honored. For all other uses, contact the owner/author(s). possibility of combining genetics and registry data to predict

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s). both mental disorder diagnosis and disorder progression in a

https://doi.org/10.70314/is.2024.chtm.6 clinically relevant, cross-diagnostic setting prior to clinical



assessment. The main limitation of this model is that it is

restricted to learning from historical data and should be Employed for the prevention and early detection of mental

continuously assessed and evaluated by trained clinicians and health issues

never stand alone in the decision-making The role of AI in the prevention and early detection of mental

problems can also be very significant. Patients frequently turn

Utilized to alleviate the burdens associated with clinical to ChatGPT to seek information about their symptoms, possible

documentation, communication, and research tasks. diagnoses, and treatment options. ChatGPT performs better

Also, new technologies can assist clinicians by allowing them than Google Search in delivering general medical knowledge

to focus more on direct patient care and alleviate the high but scores lower when it comes to providing medical

clinical workload and bureaucratic tasks- such as handling recommendations [19]. A recent study highlighted early success

admissions and managing paperwork- that have been linked to for an AI model that can detect cognitive distortions in text

burnout in earlier research [8]. ChatGPT can assist in messages with accuracy comparable to that of clinically trained

processing clinical case transcripts, generating summaries, human raters [20]. GPT’s ability to recognize mental health

completing medical record documentation, and efficiently warning signs in routine conversations or text-based

preparing discharge summaries. It can also help facilitate telemedicine interactions has the potential to improve early and

communication between clinicians of different specialties when effective intervention strategies when necessary.

consultations are needed, aiding in an integrative approach [9].

While current technology cannot replace the expertise and Risks and limitations ChatGPT use for mental



differential diagnoses based on relevant signs and symptoms. Despite its strengths and potential, the use of AI technologies in ChatGPT is recognized for its substantial potential to assist judgment of skilled psychiatrists, it can assist in generating health disorders

psychiatric clinical practice carries several risks. A significant

experts with clinical and laboratory diagnoses, as well as in



planning and conducting medical research [10]. the conversational model may confidently produce text that is concern is the phenomenon of “artificial hallucinations” where

factually incorrect, nonsensical, or misleading [18].

Applied in psychotherapy to enhance therapeutic processes

Recent systematic review that included 118 articles

Chatbots can be also beneficial in psychotherapy. The

identified some limitations regarding the potential of ChatGPT

therapist's emotions and the emotional alignment between



support structure [12]. This implies that GPT models may research, and the quality of its responses often hinges on how potentially develop cognitive empathy over time, making it the questions are phrased [21]. possible for ChatGPT to achieve a notable level of accuracy in Even with the advanced GPT-4 model, there is still a risk of identifying users’ emotions [13], nevertheless it is important to providing harmful advice. The absence of clinical reasoning make systematic testing to ensure a non-superficial comparison and experience in ChatGPT can lead to the omission of between human and artificial intelligences [14]. Namely, important clinical details in patient summaries and medical ChatGPT currently lacks the capability to accurately assess records. Thus, the most prudent approach is to employ AI personality traits, a task that demands extensive training and systems as supplementary tools for mental health professionals, expertise from psychiatrists. AI researchers are diligently ensuring they are used under close supervision to uphold the and outcomes of therapy [11]. A study conducted during the raising concerns about their originality, privacy, accuracy, and COVID-19 pandemic showed that technology can offer an legality [10]. It is well established that ChatGPT may generate effective method, providing at least a first level counseling inaccurate facts and references when summarizing previous therapist and client are crucial factors influencing the process provided by ChatGPT are often insufficient and contradictory, in patient care and medical research, noted that the solutions

pursuing technical advancements to improve the precision of

safety and quality of patient care. [22]

personality detection [15].



A study designed to assess the accuracy and appropriateness of Conclusion psychodynamic formulations generated by ChatGPT found that

the model produced suitable results even without additional The recent introduction of GPT-4 has significantly enhanced

psychoanalytic information. It also demonstrated a strong the capabilities of the GPT system. Current implementations of

ability to generate psychodynamic formulations consistent with ChatGPT within mental health care are limited by its inherent

various psychoanalytic theories when given appropriate design as a chatbot, rather than as a specialized AI tool

instructions. [16]. specifically tailored for psychiatric use. Nonetheless, this

The research underscores that ChatGPT is not intended to sophisticated language model demonstrates significant utility in

substitute psychiatrists or psychologists but rather to function as addressing various routine psychiatric and administrative

an initial resource and a first line of support for those dealing functions.

with mental distress. When used wisely and within appropriate As this technology evolves and advances, we anticipate

limits, ChatGPT can be an effective tool in supporting mental substantial potential for future applications of GPT technology

health services. [17]. However, it is essential to process training in psychiatry, including its integration into diagnostic

data in a manner that ensures privacy protection and adheres to processes, the provision of psychotherapy within clinical

all professional, ethical, and legal standards. Given that environments, and the rapid identification of early warning

individuals may be at increased risk of data misuse when signs for mental health disorders. Crucially, the development

disclosing sensitive personal and family information during and refinement of professional ethical standards and practice

mental health treatment [18]. guidelines are imperative for the responsible and effective



deployment of these transformative GPT technologies in the prognosis using Danish Nationwide register and genetic data. JAMA

mental health sector. Psychiatry 80:146-155

[8] Kumar S. (2007) Burnout in psychiatrists. World Psychiatry 6:186-189.

[9] Ali SR, Dobbs TD, Hutchings HA, Whitaker IS. (2023) Using ChatGPT

to write patient clinic letters. Lancet Digit. Health 5:e179-e181.



Acknowledgments / Zahvala [10] Garg RK, Urs VL, Agarwal AA, Chaudhary SK, Paliwal V, Kar SK.

(2023) Exploring the role of ChatGPT in patient care (diagnosis and

The work presented in this paper was partially funded in part by treatment) and medical research: A systematic review. Health Promot

Perspect. 13(3):183-191.

the European Union under Grand 101159214- ChatMED. [11] Chui H, Li X, Luk S. (2022) Therapist emotion and emotional change

Views and opinions expressed are however those of the with clients: Effects on perceived empathy and session quality.

author(s) only and do not necessarily reflect those of the Psychotherapy (Chic). 59(4):594-605.

[12] Mawani A, Nderu L. (2020) Towards an online empathy assisted

European Union. Neither the European Union nor the granting counselling web application. EAI Endorsed Transactions on Context-

authority can be held responsible for them. Insert paragraph text aware Systems and Applications 7(22):167792.

[13] Kosinski M. (2023) Theory of mind may have spontaneously emerged in

here. large language models. ArXiv abs/2302.02083.

[14] Strachan J, Albergo D, Borghini G. et al. (2024) Testing theory of mind

in large language models and humans. Nat Hum Behav 8:1285-1295.5

[15] El-Demerdash K, El-Khoribi RA, Ismail Shoman MA, Abdou S. (2022)



References / Literatura Deep learning based fusion strategies for personality prediction. Egypt Inform J 23:47-53.

[1] Franco D'Souza R, Amanullah S, Mathew M, Surapaneni KM. (2023) [16] Hwang G, Lee DY, Seol S. et al. (2024). Assessing the potential of

Appraising the performance of ChatGPT in psychiatry using 100 clinical ChatGPT for psychodynamic formulations in psychiatry: An exploratory

case vignettes. Asian J Psychiatr. 89:103770. study. Psychiatry Res. 331:115655

[2] Spencer L, Broome M. (2023) The epistemic harms of empathy in [17] Arjanto P, Senduk FFW, Nahdiyah U, Utami MS. (2024) AI and ethics

phenomenological psychopathology. Phenom Cogn Sci doi: in mental health: exploring the controversy over the use of ChatGPT. J https://doi.org/10.1007/s11097-023-09930-1. Public Health (Oxf). 46(2):e340-e341.

[3] ChatBeacon. Mental health assistant powered by ChatGPT. (2024) URL: [18] Wei Y, Guo L, Lian C, Chen J. (2023) ChatGPT: Opportunities, risks

https://www.chatbeacon.io/industrychatgpt/mentalhealth (Accessed 24- and priorities for psychiatry. Asian J Psychiatr. 90:103808 Aug-2024) [19] Ayoub NF, Lee YJ, Grimm D, Divi V. (2024) Head-to-Head Comparison

[4] Koko. GPT-3 mental health intervention. (2024) URL: of ChatGPT Versus Google Search for Medical Knowledge Acquisition.

https://gpt3demo.com/apps/koko-ai (Accessed 24-Aug-2024). Otolaryngol Head Neck Surg. 170(6):1484-1491.

[5] Quaak M, van de Mortel L, Thomas RM, van Wingen G. (2021) Deep [20] Tauscher JS, Lybarger K, Ding X et al. (2023) Automated detection of

learning applications for the classification of psychiatric disorders using cognitive distortions in text exchanges between clinicians and people

neuroim- aging data: Systematic review and meta-analysis. Neuroimage. with serious mental illness. Psychiatr. Serv. 74: 407-410. Clin. 30:102584. [21] The Lancet Digital Health. (2023) ChatGPT: friend or foe? Lancet Digit

[6] de Bardeci M, Ip CT, Olbrich S. (2021) Deep learning applied to Health. 5(3):e102.

electroencephalogram data in mental disorders: A systematic review. [22] Jo E, Song S, Kim JH, et al. (2024) Assessing GPT-4’s Performance in Biol. Psychol. 162:108117. delivering medical advice: Comparative analysis with human experts.

[7] Allesoe RL, Thompson WK, Bybjerg-Grauholm J et al. (2023) Deep JMIR Med Educ. 10:e51282.

learning for crossdiagnostic prediction of mental disorder diagnosis and

Testing ChatGPT’s Performance on Medical Diagnostic Tasks



∗

Alexander Perko

∗

Franz Wotawa

alexander.perko@ist.tugraz.at

wotawa@ist.tugraz.at

Graz University of Technology, Institute of Software Technology

Graz, Austria



Symptoms Diagnoses



Golden Overlap Score ⊇ Model





Diagnoses





LLM



Figure 1: Semi-Automatic Evaluation of an LLM on Medical Diagnostics Using a Medical Expert System as a Golden Model.



Abstract to find answers in one spot as LLM-powered chatbots, like Chat-

GPT [8], are happy to respond with a diagnosis. This, of course,

Large Language Models and chat interfaces like ChatGPT have

implies much risk of harm or misinterpretation. After all, the

become increasingly important recently, receiving a lot of atten-

very reason many users - being non-experts - turn to chatbots

tion even from the general public. People use these tools not only

is that they cannot assess symptoms themselves. Clusmann et

to summarize or translate text but also to answer questions, in-

al. [5] further point out that there is a lack of mechanisms to

cluding medical ones. For the latter, giving reliable feedback is of

guarantee that the LLM’s output is correct. All of this makes it

utmost importance, which is hard to assess. Therefore, we focus

important to test such systems on a practical level, which is close

on validating the feedback of ChatGPT and propose a testing pro-

to the use cases of non-experts. As for its popularity, our eval-

cedure utilizing other medical sources to determine the quality

uation focuses on ChatGPT [8], which is powered by OpenAI’s

of feedback for more straightforward medical diagnostic tasks.

most recent model, GPT-4o [9, 10]. The task of medical diag-

This paper outlines the problem, discusses available sources, and

nostics shares many traits with the natural language processing

introduces the validation method. Moreover, we present the first

(NLP) task of question answering (QA). Namely, this task tests

results obtained when applying the testing framework to Chat-

for medical knowledge as well as basic reasoning facing medical

GPT.

language. MedQA [6] is a popular benchmark in literature, which



Keywords is tailored to the medical domain. In recent years, open-domain

LLMs such as GPT-3.5 [3], GPT-4 [9], and LLaMA-2 [16] as well

Large Language Models, ChatGPT, NetDoktor, Testing, Validation

as domain-specific LLMs like Med-PaLM 2 [15], Meditron [4]



1 and Med-Gemini [14] have been evaluated on medical QA. The Introduction

United States Medical Licensing Examination (USMLE) part of

Large Language Models (LLMs) are omnipresent in today’s soci-

MedQA is used particularly often as a performance indicator in

ety, as they are used by a wide audience for a growing number

this domain. Table 1 shows reported scores of the mentioned

of tasks. This study sheds light on one area of application in par-

LLMs and demonstrates GPT-4’s and MedGemini’s superiority,

ticular, which is asking for medical diagnoses. Assessing one’s

with GPT-4 performing marginally worse despite being an open-

health and medical diagnostics are complex tasks, that fall into

domain model.

the domain of medical experts. However, since the dawn of search

engines and medical websites, like NetDoktor [13], people have

turned to the internet for getting health advice. Previously, users Table 1: LLMs Evaluated on Medical Question Answering.

searching for answers had to consult multiple online resources, Accuracy Results on the United States Medical Licensing

compare page contents, and evaluate whether their set of symp- Examination (USMLE) Part of MedQA [6], as Reported in [7,

toms matched what they found. Nowadays it is seemingly easy 14, 4, 15].

∗

Both authors contributed equally to this research.

Model Domain-Specific MedQA USMLE

Permission to make digital or hard copies of all or part of this work for personal Med-Gemini Yes 91.1

or classroom use is granted without fee provided that copies are not made or GPT-4 No 90.2 distributed for profit or commercial advantage and that copies bear this notice and

Med-PaLM 2 Yes 86.5

the full citation on the first page. Copyrights for third-party components of this

work must be honored. For all other uses, contact the owner /author(s). Meditron Yes 75.8

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia LLaMA-2 No 63.8

© 2024 Copyright held by the owner/author(s).

GPT-3.5 No 60.2

https://doi.org/https://doi.org/10.70314/is.2024.chtm.7



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Alexander Perko and Franz Wotawa



Alongside ChatGPT’s popularity, these results are a major rea- starting with the first question as the root node. It should be

son why this paper focuses on GPT-4o in particular. This work noted, that the very first question "Um wen geht es?" (i.e. "Who is

contributes by introducing a semi-automated validation proce- it about?"), was always answered by "jemand anderen" (i.e. "some-

dure for medical diagnostics performed with LLMs using an ex- body else") for this study. Rectangles represent questions and

pert system as a golden model (compare to Figure 1). Specifically, the ellipses represent the respective possible answers to choose

we evaluate the performance of ChatGPT powered by GPT-4o from. The node at the second to last level, which is denoted by

with a focus on symptom descriptions in German and compare it "Mögliche Erkrankungen" (i.e. "possible diseases") symbolizes the

to NetDoktor’s Symptom-Checker [13], which is curated by med- retrieval of diagnoses from the database, while the leaf nodes on

ical professionals. Our setup is guided by the following questions the bottom level signify the results of the query. In this exemplary

regarding prompting ChatGPT: case, the questions were answered to correspond to the symptoms



• of a heart attack for demonstration. However, we can also use Does ChatGPT provide equivalent diagnoses when pre-

Symptom-Checker to automatically and randomly traverse the

sented with the same symptoms as NetDoktor?

• questionnaire’s tree-like structure to retrieve sets of symptoms Does the output quality - as measured by the overlap -

and corresponding diagnoses. This allows for a scaleable frame-

change when asked for a specific amount of "most likely"

work for comparing other methods against a strong and valid

diagnoses?

• baseline. Sets of symptoms and corresponding "golden" diagnoses

Does the output increase in quality when ChatGPT is

are extracted from such a tree as follows: Firstly, for each path

queried in English instead of German?

from the root node to the bottom level nodes (i.e. the diagnoses),



2 Validation Methodology questions-answer-pairs are stored in a JSON data structure. Each

full path represents one set of symptoms. Secondly, each set of

For the purpose of introducing our methodology, we use myocar-

symptoms is summarized in a textual representation in German

dial infarction (i.e. heart attack) as a guiding example. According

taking special care not to lose or add information. This is then

to Statisik Austria’s annual report, cardiovascular diseases, which

translated from German to English. The first rows of Tables 3

include heart attacks, are the most common cause of death in

and 4 contain the textual descriptions of our example in German

Austria. The symptoms of a myocardial infarction include:

and English, respectively. Lastly, the diagnoses provided by the

• Feeling of tightness or constriction golden model are extracted from the bottom layer (i.e. the leaf

• Feeling of anxiety/panic attacks nodes) of the tree, which is always a set of three diagnoses. These

• Sudden severe shortness of breath, unconsciousness, or sets of diagnoses are referred to as NetDoktor diagnoses for the

severe dizziness

remainder of this paper.

• Nausea and vomiting

• Blood pressure and pulse drop 2.2 Evaluation Metric

These symptoms are now linked to an imaginary person’s sex and

The main evaluation metric used in this work is the overlap

age to form a persona whom for we want to retrieve diagnoses.

of diagnoses as compared to NetDoktor. A set of diagnoses is

Our exemplary set of symptoms shall be linked to an adult man

considered as being good if it contains a large overlap with the

and can be identified by ID 1 in all tables and plots. Besides this

golden model diagnoses of NetDoktor. Since the NetDoktor base-

exemplary persona, where we first fixed a disease, all other sets

line always yields three diagnoses, the highest overlap any other

of symptoms are picked at random. This can be done due to our

system can achieve is 3/3. Thus, the score ranges from 0/3 to 3/3.

assumption of a golden model, which we use as our baseline.

We explicitly do not normalize, although we want to compare

sets of diagnoses with varying cardinalities. The reason for this

2.1 Golden Model is that yielding more diagnoses should not be penalized (as they

We use NetDoktor’s "Symptom-Checker" [13] as a baseline for might be worth considering, as well), and yielding fewer should

our evaluation. Symptom-Checker is a freely accessible, medical not lead to a better score automatically.

expert system for retrieving likely diagnoses corresponding to

a person’s symptoms. The system can be interacted with via a 2.3 Equivalence of Diagnoses questionnaire but is only available in German. Parts of the ques-

This study compares systems designed for direct interaction with

tionnaire are static, such as questions regarding sex, age, and

humans. These systems’ output is presented to the users in nat-

selecting the general area of one’s body where symptoms occur

ural language. A key feature of medical language is its inter-

most prominently, while others are adapting to the previously

changeable use of semantically equivalent terminology originat-

asked questions. The dynamically changing questions are always

ing from different languages such as Latin, German, or English.

asked expecting an answer from the set: "Yes", "No" and "Skip".

Additionally, when talking to patients, medical personnel often

According to NetDoktor, the system is continuously validated

have to use simplified terminology, which includes the use of

by medical professionals and is based on the medical database

colloquial synonyms, hypernyms, and hyponyms. Hence, the

AMBOSS [1] and follows the medical guidelines of professional

semantic equivalence of diagnoses must be considered to ensure

societies [2]. We assume this expert system to be our golden

the comparability of different systems.

model, as it comprises curated knowledge of high quality and

• Synonyms are terms, which can be used interchangeably

is fully deterministic. The latter makes it possible, to generate a

with one another.

decision tree from a person’s (or persona’s) interaction with the

1 • Hypernyms are superordinate or umbrella terms of a term.

system, that is reproducible across multiple calls . Figure 2 shows

• Hyponyms are describing subordinate terms (i.e. more

the tree generated from the interaction of our exemplary persona

specific) or another term.

having a heart attack. The tree is to be read from top to bottom,

Table 2 lists diagnoses that are treated as equivalents for this

1As long as the underlying knowledge base does not change. study.

Testing ChatGPT Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



Um wen geht es?



Mich Jemand Anderen Um wen geht es?



Geht es um eine Frau oder einen Mann? Mich Jemand Anderen

Persona

Weiblich Männlich Geht es um eine Frau oder einen Mann?





Wie alt ist die Person? Weiblich Männlich





Baby & Kind Jugendlicher Erwachsener Wie alt ist die Person?





Baby & Kind In welcher Körperregion treten die Beschwerden auf? Jugendlicher Erwachsener



Ganzer Körper Psyche Kopf Hals Brust Rücken Bauch Unterleib Arm / Hand Bein / Fuß

In welcher Körperregion treten die Beschwerden auf?



Ganzer Körper Haut Brustkorb Brustdrüße Psyche Kopf Hals Brust Rücken Bauch Unterleib Arm / Hand Bein / Fuß



Nenne mir jetzt das Symptom, welches die Person am stärksten belastet. Haut Brustkorb Brustdrüße



Engegefühl Nenne mir jetzt das Symptom, welches die Person am stärksten belastet.

Pfeifgeräusche

Atemnot Aufstoßen Auswurf oder Druck Herzrasen Herzstoplern Husten Schmerz Sodbrennen

beim Ausatmen

im Brustkorb

Engegefühl

Pfeifgeräusche

Atemnot Aufstoßen Auswurf oder Druck Herzrasen Herzstoplern Husten Schmerz Sodbrennen

beim Ausatmen

Leidet die Person unter Atemnot bei körperlicher Anstrengung? im Brustkorb





Ja Nein Überspringen Leidet die Person unter Atemnot bei körperlicher Anstrengung?



Set of Symptoms



Hat die Person Schmerzen? Ja Nein Überspringen



Ja Nein Überspringen Hat die Person Schmerzen?



Hat die Person Schmerzen, ohne dass sie sich bewegt/belastet? Ja Nein Überspringen



Ja Nein Überspringen Hat die Person Schmerzen, ohne dass sie sich bewegt/belastet?



Mögliche Erkrankungen Ja Nein Überspringen



Herzinfarkt Mögliche Erkrankungen Cor pulmonale (Pulmonale Hypertonie)

Koronare Herzkrankheit

Beim Cor pulmonale oder Lungenherz

Verschließt sich eines der Blutgefäße, Die koronare Herzerkrankung ist

ist die rechte Herzkammer irreparabel überdehnt.



die das Herz versorgen, kommt es zum Herzinfarkt. eine schwere Durchblutungsstörung. Dies geschieht, wenn der Widerstand in Cor pulmonale (Pulmonale Hypertonie) Das Herz wird dann in einigen Bereichen Sie betrifft die Herzkranzgefäße, Herzinfarkt der Lunge über lange Zeit erhöht ist. Koronare Herzkrankheit nicht mehr ausreichend mit Blut versorgt. die den Herzmuskel versorgen. Das erschwert die Arbeit des Herzens.

stirbt das Gewebe in dem betroffenen Bereich ab. Meist ist die Ursache eine chronische Lungenerkrankung. ist die rechte Herzkammer irreparabel überdehnt. die das Herz versorgen, kommt es zum Herzinfarkt. Herzinfarkt und Herzschwäche. eine schwere Durchblutungsstörung. Beispiele sind COPD oder Lungenfibrose. Wird der Engpass nicht schnell behoben, Verschließt sich eines der Blutgefäße, Beim Cor pulmonale oder Lungenherz Damit steigt das Risiko für Die koronare Herzerkrankung ist Dies geschieht, wenn der Widerstand in

Das Herz wird dann in einigen Bereichen Sie betrifft die Herzkranzgefäße,

Notarzt der Lunge über lange Zeit erhöht ist. Kardiologe

nicht mehr ausreichend mit Blut versorgt. Kardiologe die den Herzmuskel versorgen.

Das erschwert die Arbeit des Herzens.

Wird der Engpass nicht schnell behoben, Damit steigt das Risiko für

Meist ist die Ursache eine chronische Lungenerkrankung.

stirbt das Gewebe in dem betroffenen Bereich ab. Herzinfarkt und Herzschwäche.

Beispiele sind COPD oder Lungenfibrose.

Notarzt Kardiologe

Kardiologe



Set of Diagnoses



Figure 2: Golden Model: Exemplary Decision Tree Based on NetDoktor’s Symptom-Checker Questionnaire [13] Filled-Out

for a Persona Having a Heart Attack. Blue Boxes are Screenshots from Symptom-Checker Corresponding to Nodes in the

Tree. We Set a Persona and Automatically Extract A) a Set of Symptoms and B) a Set of Diagnoses for Each Path From the

Root Node to the Leaf Nodes on the Bottom-Most Level.

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Alexander Perko and Franz Wotawa



Table 2: Equivalent Diagnoses: Synonyms, Hypernyms, Hy- Table 3: Sets of Symptoms per ID [DE]

ponyms & Translations

ID Description of Symptoms in German

Diagnosis Equivalence (as Occurring in 1 Ein erwachsener Mann verspürt ein Engegefühl im Brustkorb.

ChatGPT Output) Er hat Schmerzen, auch wenn er sich nicht bewegt oder belastet.

Außerdem leidet er unter Atemnot, wenn er sich anstrengt.

Herzinfarkt Myokardinfarkt

2 Ein Bub hat gerötete Augen und Fieber. Außerdem hat er

Akutes Koronarsyndrom

Schmerzen beim Wasserlassen.

Myocardial Infarction

3 Ein kleiner Junge hat starke Bauchschmerzen, die bei Druck

Heart Attack

schlimmer werden. Es wurde auch festgestellt, dass er allgemein

Reiter-Syndrom Reaktive Arthritis

druckempfindlich ist.

Morbus Reiter

4 Ein jugendlicher Junge verspürt ein Engegefühl oder Druck im

Reactive Arthritis Brustkorb. Er bemerkt, dass sein Herz sehr schnell schlägt und

Reiter’s Syndrome unregelmäßig arbeitet. Er fühlt sich oft müde und weniger leis-

Kawasaki-Syndrom Kawasaki Disease tungsfähig.

Kawasaki Syndrome 5 Ein jugendlicher Junge hat an Gewicht verloren und leidet unter

Blinddarmentzündung anhaltender Müdigkeit. Er hat regelmäßig Durchfall, der beson- Appendizitis

ders voluminös und übelriechend ist.

Vorhofflimmern Herzrhythmusstörungen

6 Ein erwachsener Mann hat eine Rachenentzündung und bemerkt

Glutenunverträglichkeit Zöliakie

gerötete Augen. Es gibt jedoch kein Fieber oder geschwollene

Bakterielle Pharyngitis Mild Bacterial Conjunctivitis

Lymphknoten.

with Pharyngitis

7 Ein erwachsener Mann hat entzündliche Hautveränderungen am

Krätze Scabies

Unterschenkel, die stark jucken, insbesondere nachts. Es wurde

Erkältung Virale Infekte

kein Zusammenhang mit Allergien festgestellt.

Pfeiffer-Drüsenfieber Pfeiffersches Drüsenfieber

8 Ein kleines Mädchen hat seit einiger Zeit ihren Appetit verloren,

Mononukleose fühlt sich ungewöhnlich müde und hat ungewollt an Gewicht

Blasenentzündung Zystitis verloren. Es wird auch über verminderten Urinfluss berichtet.

Harnwegsinfektion 9 Ein kleines Mädchen hat Fieber, eine Rachenentzündung und

Urinary tract infection geschwollene Lymphknoten. Sie fühlt sich abgeschlagen und

schwitzt besonders nachts stark.

Gürtelrose Herpes Zoster

10 Eine jugendliche Mädchen hat Blut im Urin und Schmerzen

Mastopathie Fibrozystische Mastopathie

beim Wasserlassen. Der Harndrang ist häufig, aber es wird nur

Lipom Lipoma

eine geringe Urinmenge ausgeschieden. Zudem verspürt sie ein

Brennen beim Wasserlassen.

11 Ein jugendliches Mädchen klagt über ausstrahlende Schmerzen

im Nackenbereich und hat einen Hautausschlag mit kleinen

Bläschen.

12 Eine erwachsene Frau hat Spannungsgefühle in der Brust und

2.4 Sets of Symptoms & Personas tastet schmerzlose Knoten. Die Haut ist nicht gerötet.

For this evaluation, we retrieved 12 sets of symptoms from Net- 13 Eine erwachsene Frau verspürt Druckempfindlichkeit im Ober-

Doktor - 6 for females and 6 for males, and for each sex, we used bauch, die Haut wölbt sich vor und die Region ist geschwollen.

all of NetDoktor’s 3 age categories (baby/child, adolescent, adult)

twice. In addition, we used the exemplary set of symptoms for an



adult man having a heart attack, as discussed in the previous sec- 3 Experimental Evaluation

tion. This yields the 13 sets of symptoms listed in Tables 3 and 4.

Figure 1 depicts our experimental setup: NetDoktor is used as a

In Figure 2, the parts of the questionnaire are marked, which

golden model to automatically derive sets of symptoms and cor-

correspond to the persona and the set of symptoms respectively.

responding diagnoses as exemplified in Figure 2. The extracted

In the following, both terms are used interchangeably.

symptoms are then used as input to the LLM GPT-4o via ChatGPT.

ChatGPT diagnoses are then compared to NetDoktor diagnoses

2.5 Model, Prompts & Diagnose Retrieval to compute an overlap score. Figure 4 gives an overview of our

evaluation results. For each set of symptoms, NetDoktor results

For all of our experiments, we used GPT-4o [9, 10] through Chat-

are shown, followed by four diagnosis strategies utilizing Chat-

GPT [8]. More specifically, we used version GPT-4o-2024-08-06,

GPT. The grey bars denote the cardinality of every resulting set

which has been released in August 2024. We evaluate the same

of diagnoses. Blue overlays are used to show the overlap between

model in German and English and denote this with a trailing

NetDoktor diagnoses and ChatGPT diagnoses. These overlays

"[DE]" for German and "[EN]" for English for the respective

correspond to the values in Table 5, which comprises the occur-

results. We extended this convention to our golden model Net-

rences of overlaps in each category from 0/3 to 3/3. In addition to

Doktor as well. The full list of prompts used can be found in

the 13 sets of symptoms, Figure 4 and Table 5 include averages

the next section, Section 3. All LLM results were retrieved in a

computed over all sets for easier comparison of the prompts/-

zero-shot methodology, without samples or additional context be-

diagnosis retrieval methods. In the following, you can find the

sides the prompt itself. Every symptom description is sent within

used prompts/methodologies corresponding to the depicted bars:

a new chat to isolate individual queries. However, we cannot

guarantee that we are indeed interacting with a "blank slate" as A Diagnoses from NetDoktor were re-NetDoktor [DE]: ChatGPT and GPT-4o are both black boxes and our user profile trieved via the Symptom-Checker questionnaire as is doc-

might interfere with the output. umented in Subsection 2.1. This is our golden model and





Testing ChatGPT Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia





Table 4: Sets of Symptoms per ID [EN]





ID Description of Symptoms in English

1 An adult man feels a tightness in his chest. He experiences pain

even when he is not moving or exerting himself. Additionally,

he suffers from shortness of breath when he exerts himself.

2 A boy has red eyes and a fever. He also has pain when urinating.

3 A little boy has severe abdominal pain, which worsens with pres-

sure. It was also found that he is generally sensitive to pressure.

4 A teenage boy feels a tightness or pressure in his chest. He

notices that his heart beats very fast and irregularly. He often

feels tired and less capable.

5 A teenage boy has lost weight and suffers from persistent fatigue.

He has regular diarrhea that is particularly voluminous and foul-

smelling.

6 An adult man has a throat infection and notices red eyes. How-

ever, there is no fever or swollen lymph nodes.

7 An adult man has inflammatory skin changes on his lower leg

that itch intensely, especially at night. No connection with aller-

gies was found.

8 A little girl has lost her appetite for some time, feels unusually

tired, and has unintentionally lost weight. Reduced urine output

is also reported.

9 A little girl has a fever, a throat infection, and swollen lymph

nodes. She feels weak and sweats heavily, especially at night.

10 A teenage girl has blood in her urine and pain when urinating.

The urge to urinate is frequent, but only a small amount of urine



is passed. She also feels a burning sensation when urinating. Figure 3: ChatGPT Output for ID 1, Using Prompt B. Over-

11 A teenage girl complains of radiating pain in the neck area and lap with NetDoktor Marked in Blue. Compare to Figure 2.

has a rash with small blisters.

12 An adult woman has a feeling of tension in her breast and can

feel painless lumps. The skin is not reddened.

13 An adult woman feels tenderness in the upper abdomen, the skin asking for the "10 most likely" diagnoses is the runner-up with

bulges, and the area is swollen.

an average of 0.85/3 i.e. 28%. Surprisingly, the simple prompt in

English, E, performs poorest, which contradicts our hypothesis

of English prompts performing better.

overlaps with its diagnoses are marked in blue, in Figure 4.

The questionnaire and results are in German.

B Ad-hoc query sent to ChatGPT using the ChatGPT [DE]: Table 5: Overlaps of Diagnoses with NetDoktor per Prompt

symptom descriptions in German from Table 3 as is.

C More elaborate query ChatGPT [DE] "3 Most Likely": Diagnosis Retrieval Method

Score

sent to ChatGPT using the symptom descriptions in Ger- A B C D E

man from Table 3, additionally requesting the "3 most 3/3 13 0 0 0 0

likely" diagnoses. 2/3 0 2 0 3 1

D More elaborate query 1/3 0 8 9 5 3 ChatGPT [DE] "10 Most Likely":

sent to ChatGPT using the symptom descriptions in Ger- 0/3 0 3 4 5 9

man from Table 3, additionally requesting the "10 most Avg. 3/3 0.92/3 0.69/3 0.85/3 0.38/3

likely" diagnoses. Avg.[%] 100% 31% 23% 28% 13%

E Ad-hoc query sent to ChatGPT using the ChatGPT [EN]:

symptom descriptions in English from Table 4 as is.

Out of convenience, the letters introduced in this list are used Apart from the overlaps, other interesting observations can be

when referring to a specific prompt in the following paragraphs. made on closer inspection of the results: ChatGPT seems to rigor-

The main takeaway from this evaluation is that none of the used ously follow the instruction to generate n diagnoses and as such,

prompts achieves a complete overlap of 3/3 with NetDoktor for yields consistently 3 diagnoses for prompt C and 10 diagnoses for

any of the personas. Prompt B, achieves the highest score, with prompt D. However, it can be doubted that "most likely" is inter-

0.92/3 i.e. 31%. For our small test set of 13 sets of symptoms, preted in a scientifically backed manner, as ChatGPT often does

these results constitute from 2/3 for two, and 1/3 for seven sets not include even one of the NetDoktor diagnoses and not once

of symptoms. The same prompting strategy in English, denoted all of them. Equally interesting is the inclusion of the necessity

by E, yields worse results, having overlaps of 2/3 for one, and 1/3 to consult a doctor in one form or the other at the end of every

for three sets of symptoms. The two prompts giving ChatGPT result we received, which is likely due to being "hard-coded" for

the task of answering with the "n most likely" diagnoses are legal reasons on the part of OpenAI. This can also be seen in

equally not performing as well as the simple prompt in German: Figure 3. Although ChatGPT and GPT-4o are black boxes and

C achieves 1/3 for 9 sets of symptoms and D yields 2/3 for three LLMs are non-deterministic, we try to document our reported

and 1/3 for five sets of symptoms. This means that prompt D, results as well as possible for replication. You can find all of our Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Alexander Perko and Franz Wotawa



Diagnoses per Set of Symptoms for the better. In our tests, ChatGPT always includes a notice to

Diagnosis Retrieval Methods A-E:

A ... NetDoktor [DE] consult a doctor. Human assessment of the diagnoses cannot be

A B ... ChatGPT [DE] fully bypassed by the proposed evaluation methodology. This B C ... ChatGPT [DE] "3 Most Likely"



D E ... ChatGPT [EN] E Overlap with NetDoktor C 1 D ... ChatGPT [DE] "10 Most Likely" is due to the immanent presence of semantic equivalence and the necessary medical knowledge to find those equivalences. Al-

A

B though such a task is automatable via LLMs as well, the authors

C 2

D of this paper underline the potential implications for undermin-

E

A ing the quality of an evaluation, when fully automated. While our

C 3 evaluation reports results achieved using ChatGPT and GPT-4o, D B

E the proposed methodologies transcend to other LLMs as well. As

B part of future work, we want to repeat our experiments at a larger C A

D scale to achieve representative results. Additionally, we want to E 4



A consider stability metrics, as seen in [11]. Another interesting B C 5 D direction can be further analysis of the relationship between E

A prompt (engineering) and the retrieval of matching diagnoses

B

C 6 as well as their stability. Finally, it would be interesting to com-

D

E pile a corpus of medical symptoms corresponding to diagnoses

A

B including named entities and logical abstractions to perform

C 7

D evaluations as seen in [12] on the medical domain.

E

A

C 8 Acknowledgements B

Sets of Symptoms D

E



B European Union under Grant 101159214 – ChatMED. Views and C 9 A The work presented in this paper was partially funded by the

D opinions expressed are however those of the author(s) only and E



A do not necessarily reflect those of the European Union. Neither B C 10 D the European Union nor the granting authority can be held re- E

A sponsible for them.

B

C 11

E References D

A

B [1] AMBOSS GmbH. 2024. Amboss. https : / / www . amboss . com. Accessed:

C 12 2024-09-03. (2024).

D

E [2] AWMF. 2024. Arbeitsgemeinschaft der wissenschaftlichen medizinischen

A fachgesellschaften (awmf ) - leitlinien. https://www.awmf .org/leitlinien.

B

C 13 Accessed: 2024-09-03. (2024).

D [3] Tom B. Brown et al. 2020. Language models are few-shot learners. arXiv:

E

A 2005.14165 [cs.CL]. (2020).

B [4] Zeming Chen et al. 2023. Meditron-70b: scaling medical pretraining for large

C

D language models. (2023). https://arxiv.org/abs/2311.16079 arXiv: 2311.16079

E [cs.CL].

[5] J. Clusmann et al. 2023. The future landscape of large language models in

medicine. , 3, 141. doi: https://doi.org/10.1038/s43 Communications Medicine

1 2 3 4 5 6 7 8 9 10 856- 023- 00370- 1.

Number of Diagnoses [6] Di Jin, Eileen Pan, Nassim Oufattole, Wei-Hung Weng, Hanyi Fang, and

Peter Szolovits. 2020. What disease does this patient have? a large-scale



Figure 4: Comparison of Diagnoses for Symptoms Seen in arXiv:2009.13081 open domain question answering dataset from medical exams. arXiv preprint .

Table 3 [7] Harsha Nori, Nicholas King, Scott Mayer McKinney, Dean Carignan, and

Eric Horvitz. 2023. Capabilities of gpt-4 on medical challenge problems.

(2023). https://arxiv.org/abs/2303.13375 arXiv: 2303.13375 . [cs.CL]

[8] OpenAI. 2023. ChatGPT. (2023). chat.openai.com/chat.

experimental results as a replication package under the provided [9] OpenAI. 2023. GPT-4 technical report. arXiv: 2303.08774 [cs.CL]. (2023).

2

URL . [10] OpenAI. 2024. Introducing gpt-4o and more tools to chatgpt free users. (May

2024). https://openai.com/index/gpt- 4o- and- more- tools- to- chatgpt- f ree/.

[11] Alexander Perko and Franz Wotawa. 2024. Evaluating openai large lan-

4 Conclusions guage models for generating logical abstractions of technical requirements

documents. In Proceedings to the 24th International Conference on Software

In summary, ChatGPT diagnoses only partially match the di- Quality, Reliability, and Security (QRS). IEEE.

agnoses retrieved from our golden model NetDoktor. ChatGPT [12] Alexander Perko, Haoran Zhao, and Franz Wotawa. 2023. Optimizing named

entity recognition for improving logical formulae abstraction from technical

diagnoses are mostly well-structured and are seemingly valid

requirements documents. In Proceedings to the 10th International Conference

but fail to include all NetDoktor diagnoses in any of the tested on Dependable Systems and Their Applications.

cases. This holds for all tested prompts and across all tested symp- [13] Jens Richter, Hans-Richard Demel, Florian Tiefenböck, Luise Heine, and

Martina Feichter. 2024. Symptom-checker. https://www.netdoktor.at/sympt

toms. The highest overlap, on average, could be achieved with

om- checker/. Accessed: 2024-09-03. (2024).

the simplest prompt in German, giving only a description of the [14] Khaled Saab et al. 2024. Capabilities of gemini models in medicine. (2024).

symptoms for a persona. When asked for a specific amount of https://arxiv.org/abs/2404.18416 arXiv: 2404.18416 . [cs.AI]

[15] Karan Singhal et al. 2023. Towards expert-level medical question answering

"most likely" diagnoses, ChatGPT always delivered exactly the

with large language models. (2023). https://arxiv.org/abs/2305.09617 arXiv:

asked-for number of diagnoses. However, this does not benefit 2305.09617 . [cs.CL]

[16] Hugo Touvron et al. 2023. Llama 2: open foundation and fine-tuned chat

the quality of the output as measured by the overlap metric. Nei-

models. . ArXiv

ther does an interaction in English change the output quality

2

https://zenodo.org/doi/10.5281/zenodo.13765345



HomeDOCtor App: Integrating Medical Knowledge into GPT



for Personal Health Counseling



Matic Zadobovšek Primož Kocuvan Matjaž Gams

matic.zadobovsek@gmail.com primoz.kocuvan@ijs.si matjaz.gams@ijs.si

Faculty of Computer and Department of Intelligent Systems, Department of Intelligent Systems,

Information Science Jožef Stefan Institute Jožef Stefan Institute

Ljubljana, Slovenia Ljubljana, Slovenia Ljubljana, Slovenia



Abstract in less researched areas and specific issues [19]. As a result, it

The escalating workload demands on healthcare professionals makes sense to think about augmenting the existing LLM with

are leading to systemic overload, resulting in a decline in the verifiable resources that can help to improve it. One of the most

efficiency of public health services. This situation necessitates important issues regarding the use of LLMs relates to the aspect

the development of solutions that can alleviate the burden on of secure use of health data. When using commercial LLMs, the

physicians while ensuring comprehensive patient care. Recent data is passed to them and the user should be informed with

advancements in generative artificial intelligence, particularly appropriate disclaimers about the use and processing of the data

in the field of medicine, have demonstrated that large language they have entered. In an ideal situation, we would use our own

models (LLMs) can outperform doctors in specific tasks, high- LLM, running on a local server, and have full control over the im-

lighting their potential as valuable tools for reducing the strain plementation, but this raises different issues. Creating your own

on healthcare providers. This study focuses on the development LLM is a time-consuming and costly process, so our research

of the HomeDOCtor application, which integrates additional val- focused on extending the best LLM currently on the market,

idated medical knowledge into the GPT-4o LLM. The objective of GPT-4o.

this application and the enhanced LLM is to offer users reliable In our case, we focused on the Slovenian healthcare sector and

access to a medical chatbot capable of providing accurate and the adaptation of the LLM to the needs of the average user. The

timely responses to health-related inquiries. The chatbot’s be- aim was to develop an application (HomeDOCtor) that would

havior has been meticulously tested and refined in collaboration allow all Slovenian citizens to get medical help 24/7 without

with a team of physicians. The findings of this research offer the need to contact a personal doctor. As a result, performance

insights into the development of such systems and explore their testing is crucial, as even the smallest errors can lead to serious

potential application within the Slovenian healthcare system. consequences [8].

The core research hypothesis posits that the system, enhanced

Keywords with additional modules integrated into a GPT, will provide more



large language models, GPT-4o, vector embeddings, vector databases, effective medical advice to the general Slovenian population compared to existing GPTs. natural language processing, prompt engineering In section 2 we present all the datasets used to extend the LLM.



There are worldwide growing problems in the health sector due 1 Section 3 systematically describes the approach and technologies Introduction used to develop the software solution. Section 4 shows the results of the developed solution and an example of a user conversation to an ageing population and a shortage of health experts [13, 17]. with the improved LLM. Section 5 describes how performance The field of Natural Language Processing (NLP) has recently seen was tested and how doctors helped to guide LLM to get the an increase in the number of LLMs being customised for different desired outcome. Section 6 presents concluding thoughts and domains [2]. Increasingly, we are also seeing their integration possible improvements. into the field of medicine, which is one of the core domains of

today’s society [14, 22, 11, 20]. Currently, OpenAI’s GPT-4o is

considered to be the most powerful LLM, which also performs 2 Datasets

best in the area of health questionnaires and other related tasks To improve the knowledge of the existing ChatGPT-4o, data from

[16, 12, 1, 5]. verified sources was obtained. The addition of new data allows

Because of their advanced ability to understand natural lan- the GPT to answer questions using its existing knowledge, as well

guage text, there are many potential applications [8, 3]. Patients as to address a broader range of questions that require specialized

can talk to the GPT, describe their symptoms in detail, include knowledge. In this case, it is about integrating information on

documents to past treatments and then get a friendly response Slovenian healthcare [4].

in seconds. This is particularly useful when the patient is unable

to access their personal doctor or simply wants to get feedback

before taking further action. LLMs are trained on a huge amount 2.1 Insieme Platform

of data, but there is the possibility of hallucination, especially The Insieme platform is the core building block of the dataset used

[9]. The platform integrates hand-crafted expert-based knowl-



work must be honored. For all other uses, contact the owner/author(s). Permission to make digital or hard copies of all or part of this work for personal edge that is accessible to users on all devices and contains basic or classroom use is granted without fee provided that copies are not made or information on Slovenian healthcare. distributed for profit or commercial advantage and that copies bear this notice and The information is organised hierarchically by medical spe- the full citation on the first page. Copyrights for third-party components of this cialty. By clicking (or visiting) on one of the branches, the user is Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia redirected to a sub-page that provides an overview of the diseases © 2024 Copyright held by the owner/author(s). and other services that belong to the selected field of medicine. https://doi.org/https://doi.org/10.70314/is.2024.chtm.8

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Zadobovšek, Kocuvan, Gams



Each of the diseases is accompanied by a set of key data that 3.1 Fine-tuning

allows the user to get everything they need in one place. This One solution is fine-tuning, which allows a pre-trained LLM to

includes: add external data sources to a LLM that already has its own

• basic information about the diagnosis, existing knowledge [10]. The idea is that instead of creating an

• professional help (list of clinics in Slovenia), LLM from scratch, an existing LLM is used and further adapted

• apps (where the user can upload their pictures, ask for to the user’s operational needs. Fine-tuning typically works on

opinion, etc.), the principle of supervised learning, where the LLM needs to

• associations in Slovenia, be given examples of the input and the expected output. With

• articles, additional learning, the LLM is able to use the new data. This

• forums, directly influences the change of parameters (weights) in the

• video content, and model and its performance. The problem arises in that, due to

• image content. the large amount of parameters in modern LLMs, the whole



display of the information for skin cancer. Figure 1 shows an example of one of the pages illustrating the learning process may be impractical due to the cost of training an outsourced LLM. Additionally, any incorporated data is static, meaning that the learning process must be repeated whenever All data have been compiled in collaboration with doctors adding new data. As a result, fine-tuning is not the best solution and are therefore suitable for use as a primary source for the in cases where we know that the data will change, and we will extension of the LLM. The aim is to enable users of the platform be constantly adding new data. to quickly obtain information on Slovenian healthcare that they

can understand.

3.2 Storing data

2.2 Other resources The data obtained by web scraping was stored in a vector data-

The Med.Over.Net Health Forum was used as one of the two re- base, which allows efficient management of the vector data [21].

maining data sources. The forum is divided into different branches The raw data obtained (e.g. PDF and Markdown files) are un-



questions within them. Each of the sub-forums has an assigned aged by the LLM. While documents can be added as attachments in the web interfaces of platforms such as ChatGPT (OpenAI’s of medicine where users can create new conversations and ask suitable for direct use with LLMs as they cannot be directly man-

questions. All other forum users can also join the conversation. web LLM access), this is only useful in the context of a single moderator (a doctor) who carefully reviews and answers the



whereby all extracted data was stored in a single PDF file. nent access to information sources without the need for manual addition of documents by the user. All data from the online forum was extracted using web scraping, conversation. In our case, we would like the LLM to have perma-

The second source is the Home Doctor Encyclopaedia (The



without the need to visit a doctor. Home Doctor - Practical Medicine for Every Household), which 3.2.1 Vector embeddings. Each of the documents is converted is in the field of practical home medicine. It contains useful tips into vector embeddings, which allow the text to be presented in on how people can help themselves by applying helpful advice a numerical notation that can be understood by a computer [6]. From this notation, the computer can understand the meaning In the future, consideration will be given to adding new re-of the text and determine the meaning between the words in sources, in particular a better medical book suitable for Slovenian the input. This is useful in many NLP tasks, e.g. search engines, healthcare. For now, we have focused on the three data sources sentiment analysis, recommendation systems and so on. We use mentioned above. vector embeddings to be able to extract information relevant to

the corresponding question based on user input.



Modern LLMs have a limitation in the form of a context win-

dow, which determines how much text they can process within a

single user input. Because of this limitation, we need to be care-

ful what additional information is sent to the LLM, as we could

quickly exceed the context window by sending all data sources.

E.g. if the user’s question is about skin cancer, we only want to

get information about that specific disease. As a consequence, we

have divided the text into chunks, where each chunk contains

data for only one of the diseases. Similar care has been taken with

the encyclopaedia and the Med.Over.Net online forum. Meaning-

ful paragraphs have been grouped together to form one chunk

of text. Thus, only the key chunk that is most likely to contain

Figure 1: Image of the skin cancer disease subpage on the the answer to the question asked will be provided to the LLM.

Insieme web platform. For each of the chunks, a vector embedding has to be created

using an appropriate model (e.g. text-embedding-ada-002 from

OpenAI). The vector databases (e.g. Redis) have to be used to

store the resulting vector embeddings properly.



3 Methodology 3.2.2 Vector databases. The vector database allows the storage

This section presents the design of a system that extends the of unstructured data and fast retrieval due to efficient indexing

LLM GPT-4 with acquired knowledge resources and outlines the of the data. In our case, we use them to store individual chunks

architecture of the developed application. and their associated vector embeddings [7].

Integrating Medical Knowledge into GPT Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



Over all stored chunks, vector search is enabled, which means 3.4 Application architecture and used

that the chunks whose vector embeddings most closely match the technologies



query vector are returned. The query vector represents the user A brief overview of the technologies used to develop the software input, which is converted into vector format. Cosine similarity, solution is presented below. Euclidean distance, inner product and other metrics can be used

to measure similarity between vectors. 3.4.1 Flutter. The Flutter framework was used to develop the

The configuration used to retrieve the chunks can be modified: front-end part of the system. The reason for using it is that it

e.g. specifying the required similarity threshold and the number allows us to develop applications that work on all platforms by

of chunks retrieved. writing unified code in the Dart programming language. It then

gets compiled into code for the dedicated platform. This way, the



An alternative solution is Retrieval-Augmented Generation (RAG), 3.3 application works on web browsers as well as in the form of a RAG mobile application. Otherwise, we would have to use multiple programming languages to ensure support on multiple devices, where the LLM calls an external database containing all our data which would slow down development. when the user provides a question [15]. Relevant data is retrieved

from there and passed directly to the LLM, which uses this data in 3.4.2 Redis. The Redis database was used to store all additional

the generation of the answer. In this case, the basic architecture knowledge resources. It is a NoSQL database that stores data in

of the LLM used remains completely unchanged, as it accesses key-value format and also works as a vector database. The data

a separate building block to retrieve the data. It is a cheaper so- is stored in memory, which helps significantly with data retrieval

lution that allows dynamic data extraction. Data can be deleted, and overall application responsiveness.



modified and newly added to the database at any time. 3.4.3 LangChain. The LangChain framework offers a number of The Retriever is responsible for retrieving relevant pieces of functionalities that facilitate the development of applications that text from the vector database. The user question is converted involve working with LLMs [18]. It allows us to easily connect into vector embedding, and then the most similar chunks are to a vector database, retrieve documents, call arbitrary LLMs, retrieved to help guide the LLM to the correct answer. A merged modify system instructions and so on. query is then created containing the original user question, the

extracted chunks of text and any other system instructions given. 3.4.4 FastAPI. The FastAPI framework allows the creation of

In the final step, the LLM generation produces an answer based RESTful APIs that can be accessed by the HomeDOCtor applica-

on the query, which is passed to the user. tion. It uses Uvicorn for the server implementation and allows

Compared to fine-tuning, RAG allows the system to change processing of requests received from users. These requests are

continuously (adding new knowledge sources), makes the oper- related to the processing of the received questions and the execu-

ation more understandable (we can check which pieces of text tion of all the necessary tasks to obtain the appropriate response

have been passed to the LLM) and reduces the possibility of hal- from the LLM.



rectly, as the parameters of the original model remain completely container to facilitate development and deployment to the server. All the necessary configuration is included in the container, unchanged, but we can help by writing system instructions. Sys- which runs on the system without the need to install the rest of tem instructions further guide the operation of the chatbot and added). The style of writing the answers cannot be changed di- 3.4.5 Docker. The entire application was packaged in a Docker lucination (verified pieces of text from the selected domain are

play a very important role in achieving a user-friendly behaviour the programming libraries and tools.

of the system. 3.4.6 Firebase. The Firebase platform offers a number of useful

All of these features make RAG a suitable choice for enriching functionalities such as user authentication, application usage

the work of an existing LLM with validated information in the analytics, cloud storage and a real-time database (NoSQL cloud

chosen field. database, which in our case is used to store the necessary user

The schematic design of the RAG system can be seen in the data).

figure 2 below.

4 Results

The HomeDOCtor app allows users to access a chatbot that is

constantly available to answer their questions. Users can access



the app on their browsers or via a mobile app. The user inter-

face allows users to view the history of the conversation, ask

questions, include their own documents as an attachment (e.g.

PDF documents of previous exams), get a PDF summary of the

conversation and additional information in three side windows.

By clicking on the "End session" button at the bottom of the

screen, all data is deleted, and the user is redirected to the main

page where they can start a new conversation with the chatbot.

Initially, HomeDOCtor determines what kind of conversa-

tion is being had. In the case of a simple chat, no additional

Figure 2: Demonstration of how the RAG system works. information is displayed. However, if the system detects that the

conversation is about the user’s health, additional information

is displayed in three side windows. This additional information

is dynamically updated according to the progress of the user’s

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Zadobovšek, Kocuvan, Gams



conversation with the chatbot. The first window gives a list of 2. Is there anything that seems to make your symptoms

possible medical conditions, the second window gives advice better or worse, such as certain foods, drinks, or activities?

on how the user could help himself, and the third window tells • User: I tried drinking some warm tea with honey, which

whether a visit to the doctor would be necessary. This way, the soothed my throat for a while. My headache seems to get

user can receive a comprehensive overview of the conversation a bit worse when I’m looking at screens for too long.

so far and get an opinion that adapts to the course of the conver- • HomeDOCtor: Thank you for the additional information.

sation. It sounds like you might be experiencing symptoms of a

Figure 3 below shows the user interface of the HomeDOCtor common cold or another mild infection. Here are a few

application. We can see the user’s conversation with the chatbot recommendations:

in Slovenian and the display of additional information in the three 1. Rest and Hydration: Ensure you get plenty of rest and

side windows along the right edge. When answering questions, stay well-hydrated.

GPT-4o uses its existing knowledge, incorporating as often as 2. Warm Teas: Continue drinking warm teas with honey

possible additional sources of knowledge (especially from the to soothe your throat.

Insieme platform). 3. Limit Screen Time: Try to reduce your screen time to

Using the app, users get instant help from a chatbot. The help with your headache.

chatbot first tries to collect sufficient data before proceeding to 4. Steam Inhalation: Consider steam inhalation to relieve

further advice. In the course of the treatment, the user is advised any nasal congestion you might have.

on different forms of help (websites, questionnaires, forums, apps, For additional relief, you can visit a pharmacy for over-the-

visits to pharmacies, etc.). A visit to the doctor is only advised in counter options. You can find more information about med-

urgent cases, when HomeDOCtor considers that the situation is ications at https://mediately.co/si and http://www.cbz.si.

too complicated or further medical treatment is needed. If your symptoms worsen or persist for more than a week,

it’s advisable to consult a healthcare professional.



• User: Thank you for the advice. Where could I get more

information?

• HomeDOCtor: You’re welcome! For more information on

common conditions and treatments, you can check these

resources:

1. Med.Over.net: https://med.over.net

2. Ars Pharme - Zdravstveni nasveti:

https://www.arspharmae.com

3. Triglav Zdravstveni Nasveti: http://www.triglavzdravje.si

If you need more personalized advice, consider visiting a

pharmacist or consulting with a healthcare professional.



Figure 3: An example of a user talking to the HomeDOCtor 5 Evaluation

chatbot. Additional information is visible on the side. The full operation of the HomeDOCtor app and chatbot was

carried out by the authors and medical experts. After consulting

An example of a conversation from a user who consulted with doctors and identifying the key performance points, iterative

the developed chatbot is shown below. We can see the user’s and collaborative development followed, gradually introducing

questions and all the corresponding answers. changes and improving the performance of the chatbot.

• User: I have a slight headache and a bit of a sore throat Doctors have contributed significantly to improving the app

today. through rigorous testing, ensuring that the user can get the best

• HomeDOCtor: I understand. Let’s gather some more infor- possible results to their questions. We have been careful to ad-

mation to better assist you. dress situations where a doctor’s visit is not necessary. In many

1. How long have you been experiencing these symptoms? cases, non-urgent situations only add to the workload of doctors.

2. Do you have any other symptoms, such as fever, cough, In our system, we have tried to avoid this as much as possible

or body aches? by trying to refer the user to other resources that would be good

• User: I’ve been feeling this way for about two days now. and useful enough for them, and only as a last resort recommend-

It started with just a mild sore throat, and the headache ing a visit to the doctor. Of course, a visit to the doctor should

came yesterday. not always be avoided, so it is crucial that the system is able to

• HomeDOCtor: Thanks for the information. correctly identify scenarios where this is really necessary. These

1. How severe are your headache and sore throat on a scale are mainly cases where additional diagnostics are needed, which

from 1 to 10? are difficult for the LLM to perform on its own. Doctors have

2. Have you noticed any other symptoms, like nasal con- been particularly helpful in the project by further adapting the

gestion, fatigue, or body aches? system instructions, which detail how to deal with all scenarios.

• User: I’d say my headache is about a 4 out of 10. It’s more We compared HomeDOCtor with other large language mod-

of a dull ache, not too intense, but definitely noticeable. els on the market, but focused the most on comparing it with

The sore throat is a bit worse, probably a 6 out of 10. the regular GPT-4o model, which our system uses for its basic

• HomeDOCtor: Thank you for the details. operation. The main differences are in the flow of the conversa-

1. Have you taken any over-the-counter medications or tion, as HomeDOCtor is customised by the system instructions

home remedies for relief? If so, have they helped? to follow specific scenarios and ask questions that are key for



Integrating Medical Knowledge into GPT Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



providing further advice. Thus, by simulating a conversation of make available to users (information on clinics, apps, articles,

a user who starts the conversation by listing symptoms, a com- video content in that language).

parison between the two systems can be made. According to In the future, work will be carried out on improving the chat-

the clinicians’ evaluations, HomeDOCtor appeared to be better bot, taking into account the opinions of external users, who in

at asking sub-questions and guiding the user through the treat- this case are also our target group. This research has shown that

ment. A further important difference can be seen in the display generative artificial intelligence has a great potential application

of videos and images, as HomeDOCtor can display them right in the field of medicine and could make a significant contribution

within the app, whereas GPT-4o creates fictional images or often to relieving the burden on the healthcare system.

provides links to non-existent videos. HomeDOCtor also often

gives practical advice on how users can help themselves at home Acknowledgements

and carefully evaluates when it is really necessary to redirect the We thank members of the Department of Intelligent Systems and

user to professional help. medical experts for providing info and testing the system. This

Testing of the system has demonstrated the added value of the project is funded by the European Union under Horizon Europe

included knowledge sources, though formal validation tests are (project ChatMED grant agreement ID: 101159214). still pending. Nevertheless, the scientific hypothesis that superior

performance can be achieved compared to the original GPT is References

increasingly supported by the evidence. Through iterative testing [1] Josh Achiam et al. 2023. Gpt-4 technical report. arXiv preprint arXiv:2303.08774.

by all participants, the system has now reached a level of maturity [2] Jan Clusmann et al. 2023. The future landscape of large language models in

suitable for deployment in a production environment. Test users medicine. Communications medicine, 3, 1, 141.

[3] Matjaž Gams and Tine Kolenik. 2021. Relations between electronics, artificial

will be able to provide feedback and suggest corrections via a intelligence and information society through information society rules.



form, which will further refine the chatbot to better meet the Electronics, 10, 4, 514. [4] Matjaž Gams, Maj Smerkol, Primož Kocuvan, and Matic Zadobovšek. 2024. needs and preferences of the average user in Slovenia. This will Developing a medical chatbot: integrating medical knowledge into gpt for display the data sources used to the test users, who will be able to healthcare applications. In Intelligent Environments 2024: Combined Proceed-. IOS Press, 88–97. ings of Workshops and Demos & Videos Session indicate whether they believe that the relevant documents have [5] Aidan Gilson, Conrad W Safranek, Thomas Huang, Vimig Socrates, Ling been extracted from the vector database. Chi, Richard Andrew Taylor, David Chartash, et al. 2023. How does chatgpt

perform on the united states medical licensing examination (usmle)? the

implications of large language models for medical education and knowledge

6 assessment. JMIR medical education, 9, 1, e45312. Conclusion

[6] Martin Grohe. 2020. Word2vec, node2vec, graph2vec, x2vec: towards a



In this study, we have shown how verified information can be theory of vector embeddings of structured data. In Proceedings of the 39th ACM SIGMOD-SIGACT-SIGAI Symposium on Principles of Database Systems , incorporated into one of the existing LLMs. It was an extension 1–16. of GPT-4o, which currently achieves the best results, and has [7] Yikun Han, Chunjiang Liu, and Pengfei Wang. 2023. A comprehensive survey on vector database: storage and retrieval technique, challenge. been further adapted to the needs of Slovenian public health in arXiv preprint arXiv:2310.11703 . the context of the development of the HomeDOCtor application. [8] Yining Huang, Keke Tang, and Meilian Chen. 2024. A comprehensive survey

Comprehensive testing has been carried out with a team of on evaluating large language model applications in the medical industry.

doctors, which has helped to ensure that users are provided with arXiv preprint arXiv:2404.15777.

[9] Jožef Stefan Institute. 2024. Insieme platform. (Aug. 2024). Retrieved August

relevant and understandable information. HomeDOCtor initially 26, 2024 from https://ise-emh.eu.



gathers information by asking questions, and only then advises [10] Jieh-Sheng Lee and Jieh Hsiang. 2020. Patent claim generation by fine-tuning openai gpt-2. World Patent Information , 62, 101983. on further action. External resources and self-help methods are [11] Peter Lee, Carey Goldberg, and Isaac Kohane. 2023. The AI revolution in prioritised, as the intention is that referral to a doctor should medicine: GPT-4 and beyond . Pearson. [12] Harsha Nori, Nicholas King, Scott Mayer McKinney, Dean Carignan, and only be made in urgent situations. Eric Horvitz. 2023. Capabilities of gpt-4 on medical challenge problems. Possible improvements are in the use of one of the open source arXiv preprint arXiv:2303.13375 .



LLMs, which would negate the need for external access to Ope- [13] World Health Organization et al. 2022. Health and care workforce in europe: time to act. nAI’s LLMs. Currently, in order to run the HomeDOCtor appli-[14] Carl Preiksaitis and Christian Rose. 2023. Opportunities, challenges, and cation, an API call is sent for each question asked, for which future directions of generative artificial intelligence in medical education: scoping review. a fee is paid to OpenAI. This is also problematic from a data JMIR medical education , 9, e48785. [15] Mujahid Ali Quidwai and Alessandro Lagana. 2024. A rag chatbot for preci-protection point of view as it involves working with confidential sion medicine of multiple myeloma. medRxiv , 2024–03. user data which should not be passed on (altough you are default [16] Nicholas R Rydzewski, Deepak Dinakaran, Shuang G Zhao, Eytan Ruppin, Baris Turkbey, Deborah E Citrin, and Krishnan R Patel. 2024. Comparative opt-out for OpenAI API). Currently, this is taken care of with evaluation of llms in clinical oncology. Nejm Ai , 1, 5, AIoa2300151. disclaimers and warnings that the user has to agree to before [17] Tony Smith. 1991. Medicine in europe. european health challenges. BMJ:

using the application. At the same time, the use of open source British Medical Journal, 303, 6814, 1395.

models would bring additional problems, as the appropriate in- [18] Oguzhan Topsakal and Tahir Cetin Akinci. 2023. Creating large language

model applications utilizing langchain: a primer on developing llm apps

frastructure would have to be established. Open source models fast. In International Conference on Applied Engineering and Natural Sciences



also typically perform worse than e.g. GPT-4o. As a result, we number 1. Vol. 1, 1050–1056. [19] Ashish Vaswani. 2017. Attention is all you need. arXiv preprint arXiv:1706.03762. have chosen to use GPT-4o, as we cannot afford worse outcomes [20] Ethan Waisberg, Joshua Ong, Mouayad Masalkhi, Sharif Amit Kamran, Nasif due to the criticality of the medical domain. Zaman, Prithul Sarker, Andrew G Lee, and Alireza Tavakkoli. 2023. Gpt-4: a new era of artificial intelligence in medicine. Irish Journal of Medical Science HomeDOCtor could also be adapted to meet the healthcare (1971-) , 192, 6, 3197–3200. needs of countries abroad, but this would require ensuring that [21] Matic Zadobovšek, Primož Kocuvan, and Matjaž Gams. 2023. Vpeljava vir-



the relevant data is obtained in accordance with their national tualnega asistenta chatgpt v medicinsko platformo = implementation of a virtual assistant chatgpt into the medical platform. In Bibliografija: str. 31. guidelines and security laws. The behaviour of the chatbot could Institut „Jožef Stefan“, 28–31. http://library.ijs.si/Stacks/Proceedings/Infor remain largely the same, as GPT-4o can by itself convert between mationSociety/2023/IS2023-Volume_A%20-%20SKUI.pdf.

many languages. The key component would thus be to obtain

all the necessary country-specific information that we want to



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Zadobovšek, Kocuvan, Gams



[22] Peng Zhang and Maged N Kamel Boulos. 2023. Generative ai in medicine

and healthcare: promises, opportunities and challenges. Future Internet, 15,

9, 286.





Standards for Use of LLM in Medical Diagnosis




Mihailo Svetozarević Isidora Janković Sonja Janković

Clinic for Neurology Center for Radiology Center for Radiology

University Clinical Center Niš University Clinical Center Niš University Clinical Center Niš

Niš, Serbia Niš, Serbia Niš, Serbia

mihailo.svetozarevic@gmail.com isidora_jankovic@yahoo.com sonjasgirl@gmail.com



Stevo Lukić

Clinic for Neurology

University Clinical Center Niš

Niš, Serbia

srlukic@gmail.com



Abstract simulate human learning, comprehension, problem solving and

critical decision making. AI approaches human cognition in two

Artificial intelligence, particularly large language models

distinct ways, the symbolic and the connectionist approach [1].

(LLMs), is increasingly being recognized for its potential to

The symbolic approach aims to replicate human intelligence by

revolutionize medical diagnosis by mimicking human cognitive

analyzing cognition independent of the biological structure of the

functions in clinical decisionmaking. Despite promising

central nervous system while the connectionist approach aims to

developments, such as the ability to pass medical exams and

create neural networks that imitate the brains’s structure. To

assist in complex diagnostic processes, LLMs still face

realize the potential of AI in healthcare, we believe that the

significant hurdles, including issues with accuracy, bias, and

systematic approach to evaluation and benchmarking can get us

safety. This paper critically consider evaluation of LLMs

to a place where AI can be a net positive for health systems.

performance across various criteria to ensure they meet the

required standards for clinical use. Several dimensions of

evaluations such as accuracy, calibration, and robustness are 2 LLM’s in Medicine used. While LLMs and generative AI more broadly show real

The rapid advancements in AI, particularly in the realm of large

potential for healthcare, these tools are not ready yet. The language models (LLM’s), have transformed various sectors,

medical community and developers need to develop more including healthcare [2,3]. LLM’s and Chat GPT in particular has

rigorous evaluation, analyze across specialties, train on real-

earned much attention in recent years due to its ability to

world data, and explore more useful types of GenAI beyond

complete tasks previously considered completable by humans

current models. But ultimately, we believe these tools can help

alone as in passing United States Medical Licensing Examination

in improving both physician workload and patient outcomes. We [4]. The ability of LLM’s to accurately answer questions, provide

urgently need to set up evaluation loops for LLMs where models

advice and even triage patients based on clinical input exceeds

are built, implemented, and then continuously evaluated via user

that of the everyday person. However, the accuracy of these

feedback.

systems to resolve real world medical issues is yet to exceed that



Keywords of a fully trained physician. Also, a finite percentage of LLM

answers to patients had safety errors, and in one instance the

large language models, artificial intelligence, clinical AI advice given to a patient could have been fatal [5]. In order to

implementation, AI in clinical practice, AI safety in healthcare avoid this error in the future it is essential to assess these models

through rigorous comparative benchmarks. One of the most

critical aspects of benchmarking medical LLM’s is comparing

1 Introduction their performance with existing clinical decision support systems

Artificial intelligence (AI) by its definition, and in the broadest (CDSS) and other AI models. Traditional CDSS, often rule-

of terms, represents intelligence exhibited by computer systems. based or statistical, have been used in healthcare for decades to

The main goal of AI is to enable computers and machines to assist clinicians in making evidence-based decisions. By

mimic human cognitive function. In other words, it aims to comparing LLMs to these systems, researchers can determine

whether the new models offer significant improvements in

accuracy, speed, and comprehensiveness [6]. For example, a

Permission to make digital or hard copies of part or all of this work for personal or comparative benchmark might involve evaluating the diagnostic

classroom use is granted without fee provided that copies are not made or distributed accuracy of an LLM against a well-established CDSS in

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must predicting outcomes for specific conditions, such as sepsis or

be honored. For all other uses, contact the owner/author(s). diabetes. The LLM’s ability to incorporate a broader range of

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia data, including unstructured text from electronic health records

© 2024 Copyright held by the owner/author(s).

https://doi.org/10.70314/is.2024.chtm.9 (EHRs), could be a key factor in outperforming traditional

systems [7]. However, it is also crucial to consider scenarios



where traditional systems might still have an edge, particularly can highlight areas where LLMs might support clinicians, such

in specialized tasks where they have been finely tuned over many as providing second opinions or identifying potential errors in

years of clinical use [8]. Outside of primary care, radiology is human judgment, rather than replacing them [12]. Randomized

perhaps the medical branch that has been the most upfront and controlled trials (RCTs) are considered the gold standard in

welcoming to the use of new technology [2,3]. The concept of clinical research for evaluating the efficacy of innovations.

computer-assisted diagnosis (CAD) is well known. AI’s provide Comparative benchmarking of LLMs can also involve assessing

substantial aid by labeling abnormal or most often borderline how well these models predict or align with outcomes from

exams or simply by quickly excluding negative exams in RCTs. For example, an LLM could be tested on its ability to

computed tomographies, X-rays, magnetic resonance images recommend treatments for stroke prevention based on patient

especially in high volume settings like the emergency room data, and its recommendations could be compared with those

where human resources might be less available. AI-driven validated by RCTs [7]. However, this approach presents a set of

diagnostic tests have the potential to overcome several current challenges, as RCTs often involve highly controlled

limitations in the clinical approach to patient care [9]. Namely environments that might not fully capture the complexities of

the clinical review, time to diagnosis, diagnostic accuracy and real-world clinical settings. Currently LLMs are most often

consistency. In tandem with AI, diagnosticians of all medical tested on small datasets acquired for a specific research study or

branches are capable of improving measures of diagnostic large public benchmark dataset, both of which are usually

accuracy (mainly sensitivity and specificity) as well as collected on a limited number of very similar sites with

minimizing observer variability in specific patient interpretation. consistent diagnostic techniques. This does not reflect the

This proves most useful in settings where the clinical diagnosis substantial differences in manufacturer, quality and clinical

is in question – such as with complex patient presentation or in practices often found in real-world hospitals. As an example, the

patients with long histories and various comorbidities. Currently UK Biobank, a widely employed public imaging benchmark

not many prospective studies and randomized trials exist in dataset includes brain magnetic resonance images (MRI) for a

medical AI application. Most are not prospective, are at high risk total of 100,000 patients and more. It restricts image acquisition

of bias and deviate from existing report standards. Data to four sites each of which has identical equipment in terms of

availability is lacking and human cooperator groups are more hardware and software and performs regular quality check to

often small and inadequate. LLM’s, in particularly GPT-3, has ensure the harmonization of data. In contrast most medical

shown promise in various clinical applications, ranging from centers, including our own in Serbia, extracts data from clinical

creation of patient notes to helping healthcare providers diagnose archives over a period of 20 years which reflects the much more

rare conditions. However, it is important to recognize the diverse array of available data in everyday settings. Another

inherent limitations of these systems. point of interest is a lack of consensus on which dimensions of

evaluation to consider and prioritize for various healthcare tasks.

While accuracy is the most often examined dimension when

3 Standardized Evaluation Framework for evaluating LLM performance, other dimensions such as fairness,

Assessing LLM’s Clinical Utility for Future bias and toxicity, robustness, and deployment considerations

Clinical Practice need to be considered as well [13]. Therefore, while alignment

with RCT outcomes is a strong indicator of an LLMs clinical



Medical diagnosis involves a complex process in which a relevance, it is also important to test these models in more varied

practitioner uses objective data from a clinical exam, as well as and less controlled environments to ensure their robustness [11].

data collected from medical tests along with self-described Unlike traditional systems or statistical models that remain

subjective symptoms to conclude the most likely health problem. relatively static once developed, LLMs can be continuously

This kind of approach relies heavily on the synthesis and updated and refined. This raises the question of how implement

individual interpretation of a vast amount of information from models that are constantly evolving. Development of

various sources. These most often include available patient standardized benchmarks that can be applied across different

histories, clinical exam data correlated with current medical versions of a model are essential to address this challenge [14].

literature. In this setting LLM’s open up new opportunities for These benchmarks help identify areas where LLMs can enhance

enhancing the diagnostic process. In order to better evaluate the clinical practice and highlight the potential risks or limitations

LLMs clinical utility a direct comparison must be made between that need to be addressed [6]. By rigorously comparing LLMs

LLMs and human clinicians. This approach is essential to gauge against existing systems, human clinicians, and traditional

how well AI models can replicate or even enhance the decision- models, we can ensure that these advanced AI systems are

making process of experienced healthcare professionals. Studies integrated into healthcare in a way that maximizes their benefits

often involve presenting both clinicians and LLMs with the same while minimizing potential harms [10]. In general, there is a lack

clinical cases and comparing their diagnoses, treatment of consensus on what to consider and prioritize for various

recommendations, and reasoning [10]. Human clinician healthcare tasks. Several dimensions of evaluations such as

benchmarking can reveal important insights into the strengths accuracy, calibration, and robustness are used [13]. While

and limitations of LLMs. For instance, while LLMs might excel accuracy is the most often examined when evaluating LLM

at processing and synthesizing vast amounts of data quickly, they performance, other aspects such as fairness, bias and toxicity,

may struggle with nuanced cases that require deep contextual robustness, and deployment considerations need to be considered

understanding or ethical considerations that a human clinician as well. A list of possible aspects are presented on Table 1.

might naturally account for [11]. Furthermore, these benchmarks Comparative benchmarks can guide the development of future

AI models. Insights gained from these evaluations can inform



model improvements, such as enhancing interpretability, Comprehensivenes Measures how Human evaluated

reducing bias, or improving performance on specific tasks. As s well an LLMs comprehensiveness

the field of AI in healthcare continues to evolve, comparative output , fluency, UniEval

benchmarking will remain a crucial tool for ensuring that new coherently and relevance

models are both safe and effective for clinical use [8]. concisely

addresses all

aspects of the

task and

reference

provided



Fairness, bias and Measures Human evaluated

Table 1. Comparative benchmarks for evaluation of LLG toxiticy whether an toxicity,

performances in healthcare (adapted and modified from LLMs output counterfactual

Bedi et al. 2024) is equitable, fairness,

impartial, and performance

Dimension of Definition Metric Examples free from disparities across

Evaluation harmful race

Accuracy Measures how Human evaluated stereotypes or

close the LLM correctness, biases,

output is to the ROUGE, ensuring it

true or MEDCON does not

expected perpetuate

answer injustice or

toxicity across

Calibration and Measures how Human evaluated

diverse groups

Uncertainty uncertain or uncertainty,



t an LLM is Deployment Measures the Cost, latency, underconfiden calibration error,

about considerations technical and inference runtime Platt scaled

output for a parametric its calibration slope



specific task details of an LLM to

Robustness Measures the Human evaluated generate a

LLMs robustness, exact desired output

resilience match on LLM

against input with

adversarial intentional typos,

attacks and F1 on LLM input

perturbations with intentional use

like typos of word synonyms 4 Conclusion



Factuality Comparative benchmarking is a critical process in the Measures how Human evaluated

an development and deployment of medical large language models. LLMs factual consistency,

output for a By comparing LLMs to existing clinical decision support citation recall,

specific task systems, human clinicians, traditional statistical models, and citation precision

originates outcomes from randomized controlled trials, we can gain a

from comprehensive understanding of their strengths, limitations, and a

verifiable and potential impact on healthcare. As AI continues to play an

citable source. increasingly prominent role in medicine, rigorous comparative

It is important benchmarks will be essential for ensuring that these models

to note that it deliver on their promise of improving patient care while adhering

is possible for to the highest standards of safety and effectiveness.

a response to

be accurate but Acknowledgments

factually Views and opinions expressed in this paper are those of the

incorrect if it author(s) only and do not necessarily reflect those of the

originates European Union. Neither the European Union nor any other

from a authority can be held responsible for them. All authors

hallucinated contributed equally in the final version of this paper. This project

citation is funded by the European Union under Horizon Europe (project

ChatMED grant agreement ID: 101159214).



References [8] Obermeyer, Z., Powers, B., Vogeli, C., & Mullainathan, S., 2019. Dissecting racial bias in an algorithm used to manage the health of

[1] Esteva, A., Robicquet, A., Ramsundar, B., Kuleshov, V., DePristo, M., populations. Science, 366(6464), 447-453.

Chou, K., et al., 2019. A guide to deep learning in healthcare. Nature [9] Fletcher, E., Burns, A., Wiering, B., Lavu, D., Shephard, E., Hamilton, Medicine, 25(1), pp.24-29. W., et al., 2023. Workload and workflow implications associated with the

[2] Thirunavukarasu, A. J. et al., 2023. Large language models in medicine. use of electronic clinical decision support tools used by health

Nature Medicine, 29, 1930–1940.; professionals in general practice: A scoping review. BMC Primary Care,

[3] Thirunavukarasu, A.J., et al., 2023. Trialling a large language model 24(1), p.23.

(ChatGPT) in general practice with the applied knowledge test: [10] Amann, J., Blasimme, A., Vayena, E., Frey, D., & Madai, V. I., 2020. Observational study demonstrating opportunities and limitations in Explainability for artificial intelligence in healthcare: a multidisciplinary

primary care. JMIR Medical Education, 9, p.46599. perspective. BMC Medical Informatics and Decision Making, 20(1), 1-9.

[4] Gilson, A., Safranek, C.W., Huang, T., Socrates, V., Chi, L., Taylor, R.A., [11] Gulshan, V., Peng, L., Coram, M., Stumpe, M. C., Wu, D.,

et al., 2023. How does ChatGPT perform on the United States Medical Narayanaswamy, A., Webster, D. R., 2016. Development and validation

Licensing Examination? The implications of large language models for of a deep learning algorithm for detection of diabetic retinopathy in retinal

medical education and knowledge assessment. JMIR Medical Education, fundus photographs. JAMA, 316(22), 2402-2410. 9, p.e45312. [12] Ribeiro, M. T., Singh, S., Guestrin, C., 2016. "Why should I trust you?":

[5] Chen, S., Guevara, M., Moningi, S., Hoebers, F., Elhalawani, H., Kann, Explaining the predictions of any classifier. In Proceedings of the 22nd

B.H., Chipidza, F.E., Leeman, J., Aerts, H.J.W.L., Miller, T., Savova, ACM SIGKDD International Conference on Knowledge Discovery and

G.K., Gallifant, J., Celi, L.A., Mak, R.H., Lustberg, M., Afshar, M., & Data Mining (pp. 1135-1144) Bitterman, D.S., 2024. The effect of using a large language model to [13] Bedi S, Liu Y, Orr-Ewing L. et al. (2024) A Systematic Review of Testing respond to patient messages. The Lancet Digital Health, 6(6), pp.e379- and Evaluation of Healthcare Applications of Large Language Models

e381. (LLMs). MedRvix August 16. 2024.

[6] Topol, E. J., 2019. High-performance medicine: the convergence of [14] Goodfellow, I., Shlens, J., & Szegedy, C., 2014. Explaining and

human and artificial intelligence. Nature Medicine, 25(1), 44-56.; harnessing adversarial examples. arXiv preprint arXiv:1412.6572

[7] Rajkomar, A., Dean, J., Kohane, I., 2019. Machine learning in medicine.

New England Journal of Medicine, 380(14), 1347-1358.;





Using Combinatorial Testing for Prompt Engineering




of LLMs in Medicine



Alexander Perko

Iulia Nica

Franz Wotawa

alexander.perko@ist.tugraz.at

inica@ist.tugraz.at

wotawa@ist.tugraz.at

Graz University of Technology, Institute of Software Technology

Graz, Austria



Abstract In this paper, we contribute to verifying and validating LLMs

focusing on the impact of prompts. In particular, we introduce

Large Language Models (LLMs) like GPT-4o are of growing inter-

and discuss a methodology based on combinatorial testing [21]

est. Interfaces such as ChatGPT invite an ever-growing number

for generating various versions of prompts for medical queries.

of people to ask questions, including health advice, which brings

We rely on testing because it is a methodology to straighten

in additional risks for harm. It is well known that tools based

the view on finding interactions with a system under test (SUT)

on LLMs tend to hallucinate or deliver different answers for the

that leads to unexpected behavior. Hence, in testing, we want

same or similar questions. In both cases, the outcome might be

to provide interactions that make a system fail. Combinatorial

wrong or incomplete, possibly leading to safety issues. In this pa-

testing is a test methodology that has proven to be effective

per, we investigate the outcome of ChatGPT when we ask similar

in finding test cases, i.e., inputs given to a system, to provoke

questions in the medical domain. In particular, we suggest using

a failing behavior. In particular, combinatorial testing focuses

combinatorial testing to generate variants of questions aimed at

on interacting parameter values that reveal faults. In previous

identifying wrong or misleading answers. In detail, we discuss the

work, Kuhn and colleagues [22] showed that strictly less than 7

general framework and its parts and present a proof-of-concept

interacting parameters must be considered for many applications.

utilizing a medical query and ChatGPT.

Besides its effectiveness, combinatorial testing is a good testing



Keywords methodology for LLMs that consider prompts. For the latter, we

need different combinations of textual fragments to show differ-

Large Language Models, ChatGPT, Prompt Engineering, Combi-

ences in the outcome. Combinatorial testing provides such com-

natorial Testing, Validation

binations and also avoids leading to a combinatorial explosion



1 of potential prompts when restricting the number of considered Introduction

fragment interactions.

The use of LLMs in medicine has been of growing interest. In a

We organize the paper as follows: We first introduce the foun-

recent publication [9], the authors discuss the future of LLMs in

dations. For this purpose, we discuss related research on testing

medical applications. Although using such a model may lead to

LLMs, and introduce the basic concepts behind combinatorial

improved communication and other advantages, some drawbacks

testing. Afterward, we introduce the general testing methodology

prevent using such models and tools. It is well known that LLMs

for generating different prompts focusing on the medical domain.

like ChatGPT [33] have shortcomings like hallucinations [46].

In addition, we illustrate the use of the methodology considering

Hallucinations are answers with incorrect claims that do not

one particular medical query. Finally, we conclude the paper.

depend on training data. Such answers, unfortunately, cannot be



necessarily identified as wrong and, therefore, might be harmful, 2 Related Research especially when dealing with medical questions. There are meth-

In the past several years, considerable efforts have been made to

ods for detecting hallucinations, e.g., see [11]. Hence, verifying

evaluate LLMs. The first indicator is the wide variety of bench-

and validating tools based on LLMs to ensure a harmless use is

of utmost importance. marks

which have emerged in order to test and compare their

performance on various tasks. In [6], the authors compile a selec-

When using LLMs for queries, the form of the query, i.e., the

tion of 46 popular benchmarks. Among them, we can differentiate

prompt, is of great importance. Although there has been much

between benchmarks used for general language tasks, like Chat-

work on how to improve writing prompts in various setups, e.g.,

bot Arena [7], MT-Bench [49], HELM [25], or MMLU [13] and

[28], there is only little scientific work, e.g., [24], providing sta-

domain-specific benchmarks, like MATH [14], concentrated on

tistical evidence. However, it is generally agreed that the query’s

assessing reasoning and problem-solving capabilities of AI mod-

structure has a significant impact on the output of a LLM. There-

els in mathematics, APPS [15] for evaluating code generation, or

fore, we need to consider different prompts in any verification

MultiMedQA [41] with focus on medical examinations, medical

and validation procedure.

research, and consumer healthcare questions.

Permission to make digital or hard copies of all or part of this work for personal

Further on, depending on the human involvement in the evalu-

or classroom use is granted without fee provided that copies are not made or

ation process, there are two common methods: human evaluation

distributed for profit or commercial advantage and that copies bear this notice and

the full citation on the first page. Copyrights for third-party components of this and . Human evaluation becomes a natural automatic evaluation work must be honored. For all other uses, contact the owner /author(s).

choice in many non-standard cases, where the automated evalua-

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

tion metrics are either not suitable or insufficient. For example, in

© 2024 Copyright held by the owner/author(s).

https://doi.org/https://doi.org/10.70314/is.2024.chtm.10 [25], the evaluators analyze summarization and disinformation



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Perko et al.



scenarios, while in [2] analogical reasoning tasks. Also, Ziems of the responses remains roughly equivalent to those provided

et al. [50] used the annotations from researchers for generation. for the original test set.

Although it can provide more comprehensive and accurate feed-

back than the automatic evaluation, the human evaluation may 3 Combinatorial Testing suffer from a high variance and instability due to cultural and

Combinatorial testing aims to generate test cases by consider-

individual biases. On the other hand, the automatic variant bene-

ing a system’s input model. The input model comprises a set

fits from higher standardization. LLM-EVAL [26], for instance, of parameters (or variables) 𝑥 1 {, . . . , 𝑥 𝑛 and a not necessarily }

is a unified multidimensional automatic evaluation method for

different domain 𝑑𝑖 for each parameter 𝑥𝑖 . The domain itself is a

open-domain conversations with LLMs. Jain et al. [18] proposed a

finite set of values a parameter can take. A test case is a 𝑛-tuple

self-supervised evaluation framework, and also PandaLM [45] ob- ∈ ∈ { }

specifying a value 𝑣𝑖 𝑑𝑖 for every parameter 𝑥𝑖 𝑥1, . . . , 𝑥𝑛 .

tained reproducible and automated language model assessment

A test suite is a set of test cases. Usually, we write a test suite as

by training an LLM that acts as the ’judge’ to assess different

a table where the columns are the parameters, and the rows have

models. For more details on the specific key metrics and factors

their corresponding values.

for both evaluation types, we refer the interested reader to [6].

Given an input model, a complete test suite comprises a row

In addition, is also critical, as LLMs domain-specific evaluation

for each possible value-parameter combination. Obviously, the

are often used in specific areas, such as healthcare or finance, 𝑛 ( )

upper bound of rows is of order 𝑂 𝐷 where 𝐷 is the maxi-

that have specific requirements for the models. In the beginning, mum size of all domains 𝑑𝑖, 𝑖 1, . . . , 𝑛, i.e., 𝐷 max𝑖 = = = ( ) 1,...,𝑛 𝑖 𝑑 .

most evaluation research has been focused on natural language

Hence, computing a complete test suite is not feasible for soft-

tasks. A popular direction here is, for instance, the sentiment

ware or systems comprising a larger number of input parameters.

analysis task, which analyzes and interprets the text to identify

Moreover, applying all test cases is not feasible because the sys-

the emotional inclination. Further on, in the medical field, the

tem’s behavior must also be evaluated. In combinatorial testing,

application of LLM has recently gained significant attention. Ac-

we do not have a test oracle. The focus is only on input gener-

cording to the literature ([6], [5]), most LLMs evaluations in the

ation. Hence, such a test oracle must be added to classify a test

medical field focus on medical queries. ChatGPT, in particular,

case as passing or failing, i.e., indicating whether a test case leads

generated relatively accurate information for various medical

to a correct or wrong output, respectively. It is worth noting that

queries from genetics [10], biomedicine [17], radiation oncology

such a test oracle can be automated, and we will discuss this

physics [16]. Furthermore, several studies have evaluated the

when showing our application for validating LLMs considering

performance and feasibility of ChatGPT in the medical education

medical queries.

field. In [31], GPT-3.5 [4] and GPT-4 [34] models were evaluated

Combinatorial testing avoids computing all possible test cases.

in terms of their understanding of surgical clinical information

The idea behind this is to consider not all parameter combina-

and their potential impact on surgical education and training.

tions but only those combinations of values for a fixed number 𝑘

These findings demonstrate that LLMs can be successfully used

(smaller than 𝑛) of parameters. Hence, a combinatorial test suite

in clinical education, still further efforts may be needed to over-

covers all combinations of values for any subset of parameters of

come limitations. For more details concerning the evaluation of

size 𝑘, which is usually substantially smaller. Such a test suite is

other applications like social science, natural science, engineer-

said to be of strength 𝑘 or to be a 𝑘-wise test suite. If 𝑘 is 2, then

ing agent applications, education, search and recommendation,

the test suite is a pairwise test suite, and we speak about pairwise

and personality testing, we refer the reader to [6].

testing. Note that in practice, pairwise testing is not good enough

Another interesting taxonomy from [6] groups the encoun-

(see [22, 23]). For more information on combinatorial testing

tered testing approaches into three directions: (1) from the ob-

and its foundations, we refer the interested reader to [30, 21].

jective calculation (benchmarking) to human-in-the-loop testing,

There are many algorithms available, including ACTS [47], for

(2) from static to crowd-sourcing test sets and (3) from unified to

computing combinatorial test suites for arbitrary input models

challenging test sets. While unified settings involve test sets with

and strengths. It is also worth mentioning that combinatorial

no preference for any specific task, challenging settings create

testing has been successfully used in many application domains,

test sets for specific tasks. Tools like DeepTest[43] use seeds to

including autonomous driving [20] and security testing [40].

generate input modifications for testing, CheckList [38] builds

In the following, we illustrate combinatorial testing using

test sets based on templates, whereas AdaFilter [36] constructs

a small example. In this example, we assume four parameters

tests adversarially. Furthermore, despite the growing number of

𝑎, 𝑏, 𝑐, 𝑑, all of them only taking values from the Boolean domain

academic projects designed for prompting LLMs [19, 27, 3], just { }

𝑇 , 𝐹 standing for true and false. A pairwise combinatorial test

a few of them support systematic evaluation of textual responses

suite for this input model comprises 6 test cases:

[48, 1]. ChainForge [1] is a visual toolkit that offers on-demand

a b c d

hypothesis testing of the behavior of text-generating LLMs on

1 T T F F

open-domain tasks.

2 T F T T

To our knowledge, the use of combinatorial testing (CT), in

3 F T T F

particular for the testing of LLMs, is reported in a single paper

4 F F F T

[12]. Based on a given original sentence, the authors derive new

5 F T F T

sentences by replacing words with synonyms according to a com-

6 T F F F

binatorial test set. Assuming that the semantics of the original

sentence are preserved in the derived sentences, a test oracle is For any combination of two parameters, e.g., 𝑎 and 𝑐 , this table

created based on existing annotations. In the experimental evalu- comprises all possible combinations of values. Rows 1, 2, 3, and 4

ation from [12], the authors apply generated pairwise sentence already cover all four combinations for these two parameters. For

test sets from the BoolQ benchmark set [8] against two LLMs parameters 𝑏 and 𝑑, rows 1, 2, 5, and 6 are required to cover all

(T5 [37] and LLaMa [44]). The results indicate that the accuracy value combinations. It can be easily checked that this holds also



CT for Testing LLMs in Medicine Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



for any other pair of parameters. Note that pairwise testing in API enables rapid prompt prototyping as well as executing larger

this case only requires 6 test cases. Considering all combinations, test suites with the same underlying model. That said, it would

4 =

we would have 2 16 test cases. For the remainder of this paper, be easy to swap out the SUT and test another LLM using the

as we introduce domains extending beyond boolean values, we same validation methodology.

will use indices when referring to parameter values. For 𝑇 , 𝐹 , We use the expert system "Symptom-Checker" [39], which is { }

the indices would be 0 and 1, respectively, and the first row of curated by medical professionals and freely accessible via Net-

the table above would be represented as 0, 0, 1, 1 . Doktor [29]. Given a set of symptoms, we can automatically [ ]

retrieve diagnoses by traversing a decision tree and answering

4 Validation Methodology yes-no-questions. In addition to "yes" and "no" there is also the

option to skip a question when the inquired information is not

Figure 1 gives a high-level overview of our proposed validation

available. Further questions are asking for age, sex, and main

methodology. The remainder of this section follows the numbers

symptoms, as well as the body part that is influenced the most. We

shown in Figure 1 and discusses the individual elements of our

compute a score by comparing results (i.e., diagnoses) from our

validation pipeline.

SUT ChatGPT with our Golden Model NetDoktor. This is done

The domain of our combinatorial prompt generation pipeline

semi-automatically by first retrieving the diagnoses in the same

can be seen in Table 1, where parameters are components of

format and then assessing their overlap. As NetDoktor always

a prompt and values are (sub-)phrases. Our prototypical set of

yields three diagnoses, our score ranges from 0/3 (no overlap) to

parameters comprises a) symptom presentation, which is an in-

3/3 (complete overlap). In cases where the SUT yields more or less

troductory sub-phrase to the prompt, b) diagnostic focus, which

than three diagnoses, we do not normalize the score. This means

sets the horizon for which kind of diagnoses are expected, c) an

that a result comprising only one diagnosis cannot achieve a com-

additional hint to consider context information such as age, and

plete overlap with NetDoktor and thus cannot achieve a better

d) constraints on how the output should be formulated. Each

score than 1/3. In turn, a result comprising 10 diagnoses cannot

parameter can assume an indexed value from the given set, and

achieve a higher score than 3/3. It must be noted that duplicate

every set of values includes an empty string, which is denoted

diagnoses are only counted once, and semantic equivalence is

by "-".

considered when comparing diagnoses.



Table 1: Domain: Prompt Components and Values by Index

5 Medical Use Case

Parameter For our use case, we assume the downstream task of retrieving

(i.e. Prompt IDX Value diagnoses from an LLM based on a given set of symptoms. As

Component) mentioned earlier in this paper, hallucination is a severe prob-

Symptom 0 - lem when using LLMs. Especially in the medical domain, it is

Presentation 1 list of symptoms of utmost importance that systems are tested and validated in a

"Given the 2 symptoms structured way, as faulty output might have dire consequences

following..." 3 high-level overview of symptoms ranging from misinformation to taking the wrong medication. Diagnostic 0 - However, LLMs cannot be tested exhaustively due to their non-

Focus 1 a probable diagnosis deterministic nature and other factors, such as computational and

2 a differential diagnosis monetary costs. Even testing a restricted domain, such as prompt

3 an emergency diagnosis formulation, given a limited set of input parameters, may lead to

4 the three most likely diagnoses a combinatorial explosion when testing all possible combinations.

5 the ten most likely diagnoses For instance, compare the 24 pairwise combinations, as described

Contextual 0 - above, to all possible 192 combinations from the values in Ta-

Information 1 based on patient’s age and gender ble 1 used with reasonably sized corpus of pathologies, such as

Constraints 0 - DDXPlus [42]: This corpus comprises 134,530 samples (i.e. sets of

"the diagnosis 1 be concise symptoms) from the real world, which yields 25,829,760 possible

should..." 2 be detailed including explanations test cases as compared to 3,228,720 for pairwise testing. Taking

3 have less than 100 words into account cost, we look at 206,134.40$ versus 25,766.80$. Ta-

ble 2 gives a breakdown of the number of tokens for prompts

as well as their cost, assuming the current pricing of GPT-4o,

In our prototype implementation, we use pairwise testing for

which is 5.00$ and 15.00$ per million input and output tokens,

all parameters seen in Table 1. Note that we treat the sets of symp-

respectively. It has to be mentioned that the input length can be

toms separately from the rest of the input parameters: Symptoms

fully controlled, whereas the length of the output can only be

are excluded from the pairwise combination as we rather want

guided by the prompt. In this case, we always assumed 475 output

to combine each set of symptoms with all pairwise combinations

tokens, which was the average in our preliminary experiments.

of the other parameters. This results in the listed 24 pairwise

The average number of input tokens was 171.1, consisting of 19.1,

combinations per set of symptoms. Finally, the output of our

which are derived from combinations of values seen in Table 1,

combinatorial testing pipeline is textual prompts in natural lan-

and 152 coming from our exemplary set of symptoms, discussed

guage that act as test cases for evaluating an LLM. The resulting

in the following Subsection 5.1.

test suite from pairwise combinations for our exemplary set of

symptoms can be seen in the first two columns of Table 3. For our

preliminary evaluation, we used GPT-4o [34, 35] exclusively. The 5.1 Example

model can be accessed either via ChatGPT [33] or the OpenAI To further explain our proposed validation methodology, we use

API [32]. Having both a chatbot interface and a programmable the following exemplary description of symptoms:





Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Perko et al.





INPUT COMBINATORIAL PROMPT GENERATION TEST SUITE



1 2

Prompt Components Combinatorial Testing Algorithm



+ Prompt



Symptoms



4 5 3

Golden Model Test Evaluation / Score System Under Test



NetDoktor Result Comparison LLM (ChatGPT)



Figure 1: Basic Architecture of Our Validation Methodology.



Table 2: Cost per Size of Medical Corpus

Given the following high-level overview of symptoms, pro-

Corpus Size 1 10 100 ... vide the ten most likely diagnoses based on the patient’s DDXPlus

All Combinations age and gender.

An adult woman is experiencing symptoms in the breast

Combinations 192 1,920 19,200 ... 25,829,760

gland area. Her most troubling symptom is fluid discharge,

Input Tokens 0.03M 0.33M 3.29M ... 4,419M

and she can feel a firm, painless lump.

Input Cost [$] 0.16 1.64 16.43 ... 22,097.36

Output Tokens 0.09M 0.91M 9.12M ... 12,269M

Output Cost [$] 1.37 13.68 136.80 ... 184,037.04 Other than the overlap score, there are severe differences in

Total Cost [$] 1.53 15.32 153.23 ... 206,134.40 the output depending on the used prompt. Figure 2 illustrates

Pairwise Combinations those differences underlined by textual metrics, such as the num-

Combinations 24 240 2,400 ... 3,228,720 ber of words, while Figure 3 highlights the conciseness of the

result as measured by the ratio between the number of words and

Input Tokens 0.004M 0.04M 0.41M ... 552M

Input Cost [$] 0.02 0.21 2.05 ... 2,762.17 the number of diagnoses. Most notably, the constraints to asking

the LLM to provide concise diagnoses or limiting the number

Output Tokens 0.01M 0.11M 1.14M ... 1,533M

of words to 100 reduce the length drastically. As can be seen in

Output Cost [$] 0.17 1.71 17.10 ... 23,004.63

Figure 2, the prompts 3, 4, 7, 8, 11, 12, 16, 17, 20, 22, 23, and 24

Total Cost [$] 0.19 1.92 19.15 ... 25,766.80

all yielded results with less than 200 words. However, prompt 7

exceeds the posed 100 word limit. Furthermore, none of these

prompts fully overlapped. When comparing the results for con-

ciseness in particular, Figure 3 shows that the ratio between the

number of words and the number of diagnoses is less than 50 for

An adult woman is experiencing symptoms in the breast all prompts querying the LLM to provide concise responses (i.e.

gland area. Her most troubling symptom is fluid discharge, 3, 4, 16, 17, 23, 24), whereas it is above 60 for all and above 100

and she can feel a firm, painless lump. for all but one of the prompts asking for a detailed response (i.e.

5, 6, 9, 10, 19).

In an effort to make our work as transparent and reproducible

This set of symptom shall be seen as a sample from a corpus

as possible, we provide all prompts and responses of our prelimi-

of medical pathologies (i.e. sets of symptoms). We use it to test

1

nary study as a replication package .

different prompting strategies by combining the values from

our domain, seen in Table 1. From this singular sample, we can

generate 24 test cases as per our methodology. These test cases are 6 Conclusion

then used to evaluate our SUT based on the output of our golden This paper highlights the importance of a structured and rigor-

model. For this set of symptoms, our golden model diagnoses ous validation methodology for LLMs in the medical domain,

are: particularly focusing on prompt engineering. The proposed vali-

dation pipeline makes use of pairwise combinatorial testing to

• Breast cancer systematically evaluate the responses of LLMs like ChatGPT to

• Cyst in the breast medical queries. The methodology generates test cases given sets

• Mastopathy of symptoms and combinations of prompt components. Combi-

natorial testing ensures that a wide range of prompt variations

Table 3 shows the test result of all 24 test cases. It can be easily

is tested per set of symptoms without causing a combinatorial

spotted that test case 13 was the only prompt achieving a com-

plete overlap with the NetDoktor diagnoses. When fully written-

out, prompt 13 corresponding to the combination 3, 5, 1, 0 1 was:https://zenodo.org/doi/10.5281/zenodo.13765131 [ ]



CT for Testing LLMs in Medicine Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia



Table 3: Overlaps of Diagnoses with Golden Model per Number of Sentences in LLM Response

0 10 20 30 40

Combination for One Exemplary Set of Symptoms. Each



Combination Corresponds to One Prompt and Is Denoted 1: [0, 0, 0, 0]

by a Code Representing the Indices of the Assumed Values 2: [1, 1, 1, 0]

per Prompt Component, as Seen in Table 1. "Mast." stands 3: [2, 2, 1, 1]

for " "Mastopathy". 4: [3, 3, 0, 1]

5: [3, 4, 1, 2]



ID Test Suite Golden Model Overlap 6: [2, 5, 0, 2] Score Combination Cancer Cyst Mast. 7: [1, 5, 0, 3]

1 [0, 0, 0, 0] 2/3 ✓ ✓ 8: [0, 4, 1, 3]

2 [1, 1, 1, 0] 1/3 ✓ 9: [0, 3, 1, 2]

3 [2, 2, 1, 1] 2/3 ✓ 10: [1, 2, 0, 2]

4 [3, 3, 0, 1] 1/3 ✓ 11: [2, 0, 1, 3]

5 [3, 4, 1, 2] 1/3 ✓ 12: [3, 1, 0, 3]

6 [2, 5, 0, 2] 2/3 ✓ ✓ 13: [3, 5, 1, 0]

7 [1, 5, 0, 3] 2/3 ✓ ✓ 14: [2, 4, 0, 0]

8 [0, 4, 1, 3] 1/3 ✓ 15: [1, 3, 0, 0]

9 [0, 3, 1, 2] 2/3 ✓ ✓ 16: [0, 1, 0, 1]

10 [1, 2, 0, 2] 2/3 17: [1, 0, 0, 1]

✓ ✓ 18: [0, 2, 0, 0]

11 [2, 0, 1, 3] 2/3 ✓ ✓ 19: [2, 1, 0, 2]

13 [3, 5, 1, 0] ✓ 12 [3, 1, 0, 3] 2/3 ✓ ✓ 20: [3, 2, 0, 3]

✓ ✓ 3/3

14 [2, 4, 0, 0] 1/3 ✓ 21: [3, 0, 0, 2]

15 [1, 3, 0, 0] 1/3 ✓ 22: [2, 3, 0, 3]

16 [0, 1, 0, 1] 2/3 ✓ ✓ 23: [0, 5, 0, 1]

17 [1, 0, 0, 1] 2/3 ✓ ✓ 24: [1, 4, 0, 1]



18 [0, 2, 0, 0] 2/3 ✓ ✓ 0 200 400 600 800

19 [2, 1, 0, 2] 1/3 ✓ Number of Words in LLM Response

20 [3, 2, 0, 3] 2/3 ✓

21 [3, 0, 0, 2] 2/3 ✓ ✓ Figure 2: Textual Metrics of LLM Responses: Number of

22 [2, 3, 0, 3] 2/3 ✓ Sentences & Number of Words per Prompt

23 [0, 5, 0, 1] 2/3 ✓ ✓

24 [1, 4, 0, 1] 1/3 ✓

[2] Yejin Bang et al. 2023. A Multitask, Multilingual, Multimodal Evaluation of

ChatGPT on Reasoning, Hallucination, and Interactivity. (2023). https://arxi

v.org/abs/2302.04023 arXiv: 2302.04023 . [cs.CL]

explosion. Doing so is more efficient and reduces costs as com-

[3] Stephen Brade, Bryan Wang, Mauricio Sousa, Sageev Oore, and Tovi Gross-

pared to testing all possible combinations of prompt parameters,

man. 2023. Promptify: text-to-image generation through interactive prompt

which is especially important when evaluating LLMs on large exploration with large language models. (2023). https://arxiv.org/abs/2304.0

9337 arXiv: 2304.09337 . [cs.HC]

medical corpora. The proposed validation pipeline implements

[4] Tom B. Brown et al. 2020. Language models are few-shot learners. arXiv:

semi-automated scoring based on a "golden model“, which pro- 2005.14165 [cs.CL]. (2020).

[5] David Cecchini, Arshaan Nazir, Kalyan Chakravarthy, and Veysel Kocaman.

vides diagnoses curated by medical professionals. In a preliminary

2024. Holistic evaluation of large language models: assessing robustness,

study, we demonstrate severe differences in output for prompt Proceedings of the

accuracy, and toxicity for real-world applications. In

variations given the same set of symptoms. Out of 24 test cases, . 4th Workshop on Trustworthy Natural Language Processing (TrustNLP 2024)

Anaelia Ovalle, Kai-Wei Chang, Yang Trista Cao, Ninareh Mehrabi, Jieyu

only one achieved a full overlap with our golden model when

Zhao, Aram Galstyan, Jwala Dhamala, Anoop Kumar, and Rahul Gupta,

using GPT-4o. Once more, this highlights the dependence on well-

editors. Association for Computational Linguistics, Mexico City, Mexico,

formulated prompts and a need for thorough testing strategies, (June 2024), 109–117. doi: 10.18653/v1/2024.trustnlp- 1.11.

[6] Yupeng Chang et al. 2023. A survey on evaluation of large language models.

especially in critical domains like medicine.

(2023). https://arxiv.org/abs/2307.03109 arXiv: 2307.03109 . [cs.CL]

[7] Wei-Lin Chiang et al. 2024. Chatbot Arena: An Open Platform for Evaluating

Acknowledgements LLMs by Human Preference. (2024). arXiv: 2403.04132 [cs.AI].

[8] Christopher Clark, Kenton Lee, Ming-Wei Chang, Tom Kwiatkowski, Michael

The work presented in this paper was partially funded by the Collins, and Kristina Toutanova. 2019. Boolq: exploring the surprising diffi-

European Union under Grant 101159214 – ChatMED. Views and culty of natural yes/no questions. (2019). https://arxiv.org/abs/1905.10044

arXiv: 1905.10044 . [cs.CL]

opinions expressed are, however, those of the author(s) only and

[9] J. Clusmann et al. 2023. The future landscape of large language models in

do not necessarily reflect those of the European Union. Neither medicine. , 3, 141. doi: https://doi.org/10.1038/s43 Communications Medicine

856- 023- 00370- 1.

the European Union nor the granting authority can be held re-

[10] Dat Duong and Benjamin D. Solomon. 2023. Analysis of large-language

sponsible for them. model versus human performance for genetics questions. medRxiv. eprint:

https://www.medrxiv.org/content/early/2023/01/28/2023.01.27.23285115.f

References ull.pdf . doi: 10.1101/2023.01.27.23285115.

[11] Sebastian Farquhar, Jannik Kossen, Lorenz Kuhn, and Yarin Gal. 2024. De-

[1] Ian Arawjo, Chelse Swoopes, Priyan Vaithilingam, Martin Wattenberg, and tecting hallucinations in large language models using semantic entropy.

Elena L. Glassman. 2024. ChainForge: A Visual Toolkit for Prompt Engi- , 630. https://doi.org/10.1038/s41586- 024- 07421- 0. Nature neering and LLM Hypothesis Testing. In [12] Proceedings of the CHI Conference 2023. Applying pairwise combinatorial testing to large language model testing.

on Human Factors in Computing Systems (CHI ’24). ACM, (May 2024), 1–18. Springer, (Sept. 2023), 247–256. isbn: 978-3-031-43239-2. doi: 10.1007/978- 3

doi: 10.1145/3613904.3642016. - 031- 43240- 8_16.



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Perko et al.



Word/Diagnoses Ratio

0 20 40 60 80 100 120 140 160 [24] Boniphace Kutela, Kelvin Msechu, Norris Novat, Emmanuel Kidando, and

Angela Kitali. 2023. Uncovering the influence of chatgpt’s prompts on scien-



1: [0, 0, 0, 0] tific writings using machine learning-based text mining approaches. SSRN

Electronic Journal. http://dx.doi.org/10.2139/ssrn.4385895.



2: [1, 1, 1, 0] [25] Percy Liang et al. 2023. Holistic evaluation of language models. (2023). [cs.CL] 3: [2, 2, 1, 1] https://arxiv.org/abs/2211.09110 arXiv: 2211.09110 . [26] Yen-Ting Lin and Yun-Nung Chen. 2023. Llm-eval: unified multi-dimensional

4: [3, 3, 0, 1] automatic evaluation for open-domain conversations with large language

5: [3, 4, 1, 2] models. (2023). https://arxiv.org/abs/2305.13711 arXiv: 2305.13711 [cs.CL].

6: [2, 5, 0, 2] [27] Aditi Mishra, Utkarsh Soni, Anjana Arunkumar, Jinbin Huang, Bum Chul Kwon, and Chris Bryan. 2023. Promptaid: prompt exploration, perturbation,

7: [1, 5, 0, 3] testing and iteration using visual analytics for large language models. (2023).

8: [0, 4, 1, 3] https://arxiv.org/abs/2304.01964 arXiv: 2304.01964 [cs.HC].

9: [0, 3, 1, 2] [28] M. Nazari and G. Saadi. 2024. Developing effective prompts to improve communication with chatgpt: a formula for higher education stakeholders.

10: [1, 2, 0, 2] Discov Educ, 3. https://doi.org/10.1007/s44217- 024- 00122- w.

11: [2, 0, 1, 3] [29] 2024. Netdoktor. https://www.netdoktor.at. Accessed: 2024-09-03. (2024).

12: [3, 1, 0, 3] [30] Changhai Nie and Hareton Leung. 2011. A survey of combinatorial testing. ACM Comput. Surv. , 43, 2, Article 11, (Feb. 2011), 11:1–11:29. 13: [3, 5, 1, 0] [31] Namkee Oh, Gyu-Seong Choi, and Woo Yong Lee. 2023. Chatgpt goes to the



14: [2, 4, 0, 0] operating room: evaluating gpt-4 performance and its potential in surgical Annals of Surgical 15: [1, 3, 0, 0] education and training in the era of large language models. Treatment and Research , 104, (Apr. 2023), 269. doi: 10.4174/astr.2023.104.5.2 16: [0, 1, 0, 1] 69.

17: [1, 0, 0, 1] [32] OpenAI. 2023. API Reference. (2023). platf orm.openai.com/docs/api- ref ere

18: [0, 2, 0, 0] nce. [33] OpenAI. 2023. ChatGPT. (2023). chat.openai.com/chat.

19: [2, 1, 0, 2] [34] OpenAI. 2023. GPT-4 technical report. arXiv: 2303.08774 [cs.CL]. (2023).

20: [3, 2, 0, 3] [35] OpenAI. 2024. Introducing gpt-4o and more tools to chatgpt free users. (May

21: [3, 0, 0, 2] 2024). https://openai.com/index/gpt- 4o- and- more- tools- to- chatgpt- f ree/.

[36] Jason Phang, Angelica Chen, William Huang, and Samuel R. Bowman. 2021.

22: [2, 3, 0, 3] Adversarially constructed evaluation sets are more challenging, but may

23: [0, 5, 0, 1] not be fair. (2021). https : / / arxiv . org / abs / 2111 . 08181 arXiv: 2111 . 08181 [cs.CL] . 24: [1, 4, 0, 1] [37] Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang,

Michael Matena, Yanqi Zhou, Wei Li, and Peter J. Liu. 2020. Exploring the



0 2 4 6 8 10 limits of transfer learning with a unified text-to-text transformer. Journal of Machine Learning Research Number of Diagnoses in LLM Response , 21, 140, 1–67. http://jmlr.org/papers/v21/20- 0

74.html.

[38] Marco Tulio Ribeiro, Tongshuang Wu, Carlos Guestrin, and Sameer Singh.

Figure 3: Conciseness of LLM Responses: Number of Diag- 2020. Beyond accuracy: behavioral testing of nlp models with checklist.

noses & Ratio between Number of Words and Number of (2020). https://arxiv.org/abs/2005.04118 arXiv: 2005.04118 . [cs.CL]

[39] Jens Richter, Hans-Richard Demel, Florian Tiefenböck, Luise Heine, and

Diagnoses per Prompt Martina Feichter. 2024. Symptom-checker. https://www.netdoktor.at/sympt

om- checker/. Accessed: 2024-09-03. (2024).

[40] Dimitris E. Simos, Josip Bozic, Bernhard Garn, Manuel Leithner, Feng

Duan, Kristoffer Kleine, Yu Lei, and Franz Wotawa. 2019. Testing TLS using

[13] Dan Hendrycks, Collin Burns, Steven Basart, Andy Zou, Mantas Mazeika,

planning-based combinatorial methods and execution framework. Software

Dawn Song, and Jacob Steinhardt. 2021. Measuring massive multitask lan-

Quality Journal, 27. https://doi.org/10.1007/s11219- 018- 9412- z.

guage understanding. (2021). https : / / arxiv . org / abs / 2009 . 03300 arXiv:

[41] Karan Singhal et al. 2022. Large language models encode clinical knowledge.

2009.03300 . [cs.CY]

(2022). https://arxiv.org/abs/2212.13138 arXiv: 2212.13138 . [cs.CL]

[14] Dan Hendrycks, Collin Burns, Saurav Kadavath, Akul Arora, Steven Basart,

[42] Arsene Fansi Tchango, Rishab Goel, Zhi Wen, Julien Martel, and Joumana

Eric Tang, Dawn Song, and Jacob Steinhardt. 2021. Measuring mathematical

Ghosn. 2022. DDXPlus: A New Dataset For Automatic Medical Diagnosis.

problem solving with the math dataset. (2021). https://arxiv.org/abs/2103.03

(2022). https://arxiv.org/abs/2205.09148 arXiv: 2205.09148 . [cs.CL]

874 arXiv: 2103.03874 . [cs.LG]

[43] Yuchi Tian, Kexin Pei, Suman Jana, and Baishakhi Ray. 2018. DeepTest: Au-

[15] Dan Hendrycks et al. 2021. Measuring coding challenge competence with

tomated Testing of Deep-Neural-Network-driven Autonomous Cars. (2018).

apps. (2021). https://arxiv.org/abs/2105.09938 arXiv: 2105.09938 . [cs.SE]

https://arxiv.org/abs/1708.08559 arXiv: 1708.08559 . [cs.SE]

[16] Jason Holmes et al. 2023. Evaluating large language models on a highly-

[44] Hugo Touvron et al. 2023. LLaMA: Open and Efficient Foundation Language

specialized topic, radiation oncology physics. , 13, (July Frontiers in Oncology

Models. (2023). https://arxiv.org/abs/2302.13971 arXiv: 2302.13971 . [cs.CL]

2023). doi: 10.3389/f onc.2023.1219326.

[45] Yidong Wang et al. 2024. PandaLM: An Automatic Evaluation Benchmark

[17] Israt Jahan, Md Tahmid Rahman Laskar, Chun Peng, and Jimmy Huang. 2023.

for LLM Instruction Tuning Optimization. (2024). https://arxiv.org/abs/2306

Evaluation of chatgpt on biomedical tasks: a zero-shot comparison with

.05087 arXiv: 2306.05087 . [cs.CL]

fine-tuned generative transformers. (2023). https://arxiv.org/abs/2306.04504

[46] Yijun Xiao and William Yang Wang. [n. d.] On hallucination and predictive

arXiv: 2306.04504 . [cs.CL]

uncertainty in conditional language generation. In Proceedings of the 16th

[18] Neel Jain, Khalid Saifullah, Yuxin Wen, John Kirchenbauer, Manli Shu,

Conference of the European Chapter of the Association for Computational

Aniruddha Saha, Micah Goldblum, Jonas Geiping, and Tom Goldstein. 2023.

Linguistics: Main Volume. doi: 10.18653/v1/2021.eacl- main.236.

Bring your own data! self-supervised evaluation for large language models.

[47] Linbin Yu, Yu Lei, R.N. Kacker, and D.R. Kuhn. 2013. ACTS: A Combinatorial

(2023). https://arxiv.org/abs/2306.13651 arXiv: 2306.13651 . [cs.CL]

Test Generation Tool. In Software Testing, Verification and Validation (ICST),

[19] Ellen Jiang, Kristen Olson, Edwin Toh, Alejandra Molina, Aaron Donsbach,

2013 IEEE Sixth International Conference on, 370–375.

Michael Terry, and Carrie J Cai. 2022. Promptmaker: prompt-based proto-

[48] J.D. Zamfirescu-Pereira, Richmond Y. Wong, Bjoern Hartmann, and Qian

typing with large language models. In Extended Abstracts of the 2022 CHI

Yang. 2023. Why Johnny Can’t Prompt: How Non-AI Experts Try (and

Conference on Human Factors in Computing Systems (CHI EA ’22) Article

Fail) to Design LLM Prompts. In Proceedings of the 2023 CHI Conference on

35. Association for Computing Machinery, New Orleans, LA, USA, 8 pages.

Human Factors in Computing Systems (CHI ’23) Article 437. Association for

isbn: 9781450391566. doi: 10.1145/3491101.3503564.

Computing Machinery, Hamburg, Germany, 21 pages. isbn: 9781450394215.

[20] Florian Klück, Yihao Li, Jianbo Tao, and Franz Wotawa. 2023. An empirical

doi: 10.1145/3544548.3581388.

comparison of combinatorial testing and search-based testing in the context

[49] Lianmin Zheng et al. 2023. Judging LLM-as-a-Judge with MT-Bench and

of automated and autonomous driving systems. Information and Software

Chatbot Arena. In Thirty-seventh Conference on Neural Information Process-

Technology, 160, 107225. doi: https://doi.org/10.1016/j.inf sof .2023.107225.

ing Systems Datasets and Benchmarks Track. https://openreview.net/f orum

[21] D.R. Kuhn, R.N. Kacker, and Y. Lei. 2013. Introduction to Combinatorial Test-

?id=uccHPGDlao.

ing. Chapman & Hall/CRC Innovations in Software Engineering and Software

[50] Caleb Ziems, William Held, Omar Shaikh, Jiaao Chen, Zhehao Zhang, and

Development Series. Taylor & Francis.

Diyi Yang. 2024. Can large language models transform computational social

[22] D.R. Kuhn, R.N. Kacker, Y. Lei, and J. Hunter. 2009. Combinatorial software

science? (2024). https : / / arxiv . org / abs / 2305 . 03514 arXiv: 2305 . 03514

testing. , (Aug. 2009), 94–96. Computer

[cs.CL].

[23] Rick Kuhn, Yu Lei, and Raghu Kacker. 2008. Practical combinatorial testing:

beyond pairwise. , 10, 3, 19–23. IT Professional



IS2024 - CHATGPT in MEDICINE,

Petrovska Dojchinovska et al.



Meeting Cultural and Linguistic Demands to Accommodate



Fine-Tuned LLMs to Local Medical Customs and Patient



Communication



Gordana Petrovska Dojchinovska1 1 1 ,Monika Simjanovska Misheva , Kostadin Mishev

1Ss. Cyril and Methodius University,

Faculty of Computer Science and Engineering,

Skopje, N. Macedonia

gordana.petrovska.dojchinovska@students.finki.ukim.mk

{monika.simjanoska, kostadin.mishev}@finki.ukim.mk



ABSTRACT healthcare policies, provider training, and patient care strategies

Integrating advanced open-source large language models (LLMs), will improve the quality of care for diverse populations[5]. Con-

such as LLaMA and GatorTron, into healthcare offers a novel ap- versely, language barriers, cultural misunderstandings, and a lack

proach to enhancing communication between physicians and pa- of cultural awareness among healthcare providers can lead to mis-

tients. This paper provides a comprehensive review of the potential communication, misdiagnosis, medication errors, and other safety

of these models to improve patient-provider interactions, focusing risks[11]. Large Language Models (LLMs) have shown immense

on their ability to process and generate human-like language in real- potential in various fields due to their proficiency in generating

time clinical settings. The review outlines the methodology used and understanding natural language. In healthcare, models such as

to evaluate LLMs, which includes a detailed comparison based on LLaMA and GatorTron present an exciting frontier for improving

qualitative factors such as linguistic adaptability, cultural sensitiv- communication between physicians and patients. These models can

ity, and context-awareness, alongside quantitative metrics such as process and generate human-like language, which could address

accuracy rates, error margins, and patient satisfaction scores from significant challenges in clinical communication, such as linguistic,

clinical studies. Key ethical considerations are explored, particu- cultural, or emotional barriers. However, significant challenges re-

larly concerning data privacy, patient consent, and accountability. main, particularly regarding ethical implications, technical hurdles,

The paper delves into how adopting specific ethical frameworks local customs, linguistic demands, and the critical aspect of patient-

or guidelines can help mitigate risks associated with bias, misin- doctor communication. This review examines current applications

formation, and patient autonomy. Additionally, the potential for of LLMs in healthcare, the associated challenges, ethical concerns,



LLMs to perpetuate biases or cultural misunderstandings is dis- and potential gaps, incorporating a range of recent research studies 1 in the field from 2023 and early 2024. cussed, emphasizing the need for fine-tuning these models to align

with clinical guidelines and patient needs across different geogra-

phies and medical customs. While the paper acknowledges the 1.1 Purpose and Scope

gaps in current research, such as the emotional sensitivity of LLMs This paper aims to critically examine the potential of integrating

and their ability to understand nuanced patient concerns, it also LLMs into healthcare settings, emphasizing their role in enhancing

proposes structured future research directions. This includes the de- communication. The scope of this review extends to a comparative

velopment of LLMs that are more contextually aware, emotionally evaluation of LLaMA, GPT-4, L2M3, and GatorTron, focusing on

intelligent, and capable of operating in diverse healthcare settings. their potential impact, ethical considerations, and future research

By synthesizing current studies and real-world applications, this needs. The goal is to provide healthcare practitioners, researchers,

paper aims to provide a transparent, reproducible framework for and developers with a transparent and reproducible framework for

evaluating the effectiveness of fine-tuned LLMs in transforming adopting LLMs.

healthcare communication, ultimately improving patient outcomes

and satisfaction. 2 METHODOLOGY

This review systematically analyzes the current applications of

KEYWORDS LLMs in healthcare by conducting a structured comparison of mod-

large language models, healthcare, patient-doctor communication, els such as GPT-4, Llama, GatorTron, and L2M3. The evaluation

ethics in AI, cultural sensitivity criteria include both qualitative and quantitative metrics, focusing

on model accuracy, error rates, patient satisfaction scores, and the

1 INTRODUCTION ability to generate culturally sensitive and clinically appropriate

Effective communication has always been at the forefront of success- responses. The sources for this review consist of peer-reviewed

ful interaction between physicians and patients. It has already been studies published between 2023 and 2024, covering applications

purported that integrating cultural and linguistic competency into of LLMs in real-time clinical settings, mental health, and patient-

doctor communication. Each model’s performance is assessed based



. 1https://doi.org/10.70314/is.2024.chtm.11



936

IS2024 - CHATGPT in MEDICINE,

Meeting Cultural and Linguistic Demands to Accommodate Fine-Tuned LLMs to Local Medical Customs and Patient Communication



on specific clinical tasks, adaptability to local medical customs, and decision-making and patient compliance[19]. Recent research ex-

ethical considerations. This methodology ensures transparency and plores how generative AI tools, including LLMs, have been applied

reproducibility, allowing future researchers to replicate and validate to reduce misunderstandings in clinical services. They show that pa-

the findings. tients often feel overwhelmed by medical language, and LLMs can

mitigate this issue by acting as interpreters, promoting clearer com-

3 CURRENT APPLICATION OF LLMS IN munication between patients and their healthcare providers[28].

HEALTHCARE A pictorial demonstration of the ways large language models can

LLMs are being adopted for various tasks within healthcare, with improve the communication between patients and their providers

their most prominent use in medical communication, clinical de- is given in Figure 2.

cision support, and facilitating patient-doctor interactions. Con-

versational agents like ChatGPT assist patients in understanding

medical terminology and provide immediate answers to healthcare-

related queries, often enhancing patient engagement with care

providers. LLMs also bridge communication gaps between patients

and healthcare providers, improving clarity and comprehension

in patient education[25]. In patient-doctor communication, LLMs

serve as intermediaries to help patients articulate their symptoms

and concerns more clearly. This enhances clinical encounters, as

doctors can quickly grasp patient issues without the noise of mis-

Figure 2: Patient-doctor communication facilitated by the

communication. They can improve the precision of patient-doctor

use of LLMs

dialogue, particularly when addressing complex conditions and

explaining treatment plans[21]. Similarly, mental health applica-

tions have seen LLMs being used to summarize counseling sessions, There is also research that addresses how LLMs can cater to mul-

allowing healthcare providers to focus on therapeutic interventions tilingual settings, facilitating better communication in regions with

rather than administrative tasks[1]. Beyond patient communication, diverse linguistic backgrounds. This ensures that patients from low-

LLMs assist clinicians with diagnoses and medical research. GPT-4 resource, multilingual regions receive equitable healthcare without

and LLaMA are increasingly used as clinical assistants, offering language barriers compromising the patient-doctor relationship[3].

support in diagnosis and treatment planning[29]. Large language However, the risks of over-reliance on LLMs in direct patient com-

models enhance decision-making in surgical care by answering munication must also be acknowledged. Some research argues that

patients’ questions, thereby improving pre-surgical communication cultural sensitivity and local medical customs play a crucial role

and reducing patient anxiety[18]. Figure in healthcare communication. LLMs that fail to adapt to these fac- ?? shows the areas of the

medical domain where LLMs are currently being applied. tors may risk undermining trust between patients and healthcare

providers. For instance, culturally inappropriate language or failure



to understand local medical norms could weaken the patient-doctor



bond and compromise care[11]. The diagram in Figure 3 shows a

basic workflow of the integration of LLMs in healthcare.





Figure 1: Applications of large language models in healthcare



4 IMPROVING PATIENT-DOCTOR

COMMUNICATION



Effective communication between patients and doctors is a corner- Figure 3: Workflow of LLM integration in healthcare

stone of healthcare. Miscommunication can lead to misunderstand-

ing, reduced adherence to treatment plans, and dissatisfaction with



care. LLMs offer a promising solution to enhance communication 5 COMPARATIVE ANALYSIS OF THE MOST by simplifying medical information into layman’s terms. The role of

LLMs in improving patient-doctor interactions is significant, as they PROMINENT LLMS IN HEALTHCARE

ensure the translation of medical jargon into easily understandable Our review would be incomplete if we failed to mention the most

language during consultations, ultimately leading to more informed prominent large language models that have been fine-tuned for



937

IS2024 - CHATGPT in MEDICINE,

Petrovska Dojchinovska et al.



Model Name Primary Applications Cultural and Linguistic Adaptations Current Limitations Future Research Needs

Diagnostic support, Multilingual capabilities, Domain-specific adaptations,

GPT-4 Bias, accuracy issues

patient communication cultural sensitivity ethical frameworks

Health equity, Designed for multiple languages, Research on impact

L2M3 Limited domain-specific fine-tuning

multilingual support cultural context integration in low-resource regions

Tailored for healthcare terminology, Enhanced cultural adaptation,

LlamaCare Healthcare knowledge sharing May lack emotional sensitivity

multilingual support emotional sensitivity

Clinical decision-making, Potential bias, Improving contextual understanding

GatorTron Adapted for diverse clinical contexts

patient interaction limited contextual awareness and bias reduction

Additional Models Specific clinical domains, Further development

Customizable for local languages and cultures Challenges with generalizability

(e.g., Me-LLaMA) patient interaction for specialized clinical needs

Table 1: Comparison of LLMs in Healthcare Applications



applications in the medical domain. Therefore, this section will has the potential to increase access to healthcare for marginalized

highlight the strong points and then perform a comparative analy- communities globally[8].

sis between GPT-4, LlamaCare, GatorTron, and L2M3. A graphic Cultural Sensitivity. L2M3 goes beyond just linguistic adapt-

display of these models and their most prominent features in the ability by embedding cultural knowledge into its model[23]. This

medical domain can be found in Figure 4, while Table 1 below sum- enables L2M3 to provide more contextually appropriate advice that

marizes the models’ strong points and current areas that can be aligns with local medical customs. For instance, it can recommend

improved, as well as areas where research can be focused in the treatments or healthcare guidelines that resonate with the cultural

future. practices of the patient’s region, something that more generalized

models like GPT-4 may fail to do effectively.

Consideration for Health Equity. L2M3 plays a pivotal role



in reducing healthcare disparities by making culturally and linguis-

tically appropriate care accessible in underdeveloped and diverse

regions. It is particularly adept at filling gaps left by monolingual

or culturally neutral models, such as GPT-4[23].



5.3 LlamaCare

Knowledge Sharing in Healthcare. LlamaCare was developed to

Figure 4: Overview of the key LLMs that are applied in health- facilitate knowledge sharing among healthcare professionals. Un-

care like GPT-4, which is designed for general applications, LlamaCare

is fine-tuned specifically for healthcare, giving it an edge in clinical

decision support[29]. LlamaCare’s training on specialized medical

datasets makes it particularly useful for knowledge-intensive tasks

5.1 such as diagnosing complex conditions or synthesizing information GPT-4

from clinical trials[12].

General Application in Medicine. As one of the most versatile

Fine-tuned for Medical Data. LlamaCare’s ability to provide

LLMs, GPT-4 has been widely studied for its role in clinical decision-

accurate, context-specific information gives it a significant advan-

making, medical education, and patient engagement. GPT-4’s broad

tage over more general models. Its precision comes from training

applications, from summarizing medical records to assisting with

on large volumes of healthcare-specific datasets, allowing it to out-

diagnosis and patient interaction. However, GPT-4’s lack of special-

perform models like GPT-4 when it comes to specialized clinical

ized medical training means it faces challenges when compared to

decision-making[26].

models like LlamaCare and GatorTron[23].

Patient Communication. Although LlamaCare’s primary role

Patient-Doctor Communication.GPT-4 has also demonstrated

is to assist healthcare providers, it can also be used to improve

potential in improving patient communication by generating empa-

patient communication by offering detailed and reliable medical

thetic, human-like responses. The model can generate emotionally

information. However, unlike GPT-4, which is more conversational,

intelligent text, potentially enhancing patient trust in AI-generated

LlamaCare’s focus remains on delivering precise medical knowledge

advice[17]. Despite this, GPT-4 struggles with more complex medi-

rather than generating empathetic dialogue[4].

cal queries where detailed clinical context is needed[24].



5.2 L2M3 5.4 GatorTron

Multilingual Healthcare Accessibility. One of L2M3’s greatest Focus on Medical Records. GatorTron is being specifically de-

strengths is its support for multilingual healthcare environments. signed for the U.S. healthcare system, where it excels in processing

L2M3 is a model designed specifically for low-resource healthcare vast amounts of electronic health records (EHRs). Its ability to

settings where language barriers and cultural diversity pose signifi- rapidly synthesize and analyze patient data has made it an essential

cant challenges. By offering support for multiple languages, L2M3 tool for improving diagnostic accuracy and reducing administrative



938

IS2024 - CHATGPT in MEDICINE,

Meeting Cultural and Linguistic Demands to Accommodate Fine-Tuned LLMs to Local Medical Customs and Patient Communication



burden[2]. GatorTron outperforms models like GPT-4 in this regard, 6.2 Ethical Implications of LLMs in Healthcare

owing to its fine-tuning on clinical records[7]. The ethical implications of deploying LLMs in healthcare extend

Enhancing Clinical Workflows. One of GatorTron’s key strengths beyond technical challenges, touching on critical issues such as

is its ability to streamline clinical workflows by accurately sum- data privacy, patient consent, and accountability. are substantial,

marizing medical histories, diagnoses, and treatment plans.[14]. particularly in terms of biases, transparency, and patient autonomy.

GatorTron enhances clinical efficiency by reducing the amount of While current studies briefly mention these concerns, this review

time doctors spend on administrative tasks like reviewing patient delves deeper into specific ethical frameworks that should guide

charts, allowing them to focus more on patient care.[16]. the development and deployment of LLMs. One such framework

Adaptation to U.S. Healthcare. GatorTron’s design is opti- is the "Data Protection by Design" principle, which emphasizes

mized for the U.S. healthcare system, making it particularly effective the need to integrate privacy safeguards into AI systems from the

in this context. However, this focus on English-language records outset. In addition, patient consent must be a core element when

means it may not be as adaptable in multilingual or international implementing LLMs in clinical settings, ensuring that patients are

healthcare environments. GatorTron’s success illustrates the need fully aware of how their data is used and stored. Accountability in

for more research on how to adapt models to diverse healthcare AI decision-making also needs to be addressed, particularly in high-

systems[27]. stakes scenarios like diagnostics and treatment planning, where

errors can have life-altering consequences. Finally, the risk of bias

5.5 and perpetuating healthcare inequities through LLMs necessitates Comparative Analysis

stricter ethical oversight, with transparent mechanisms for identi-

Local Medical Customs and Linguistic Demands. When com-

fying and mitigating bias. Racial and ethnic biases in GPT-4 were

paring L2M3 and GatorTron, L2M3’s strength in multilingual set-

explored for medical diagnosis and triage, uncovering disparities in

tings is highlighted, which makes it ideal for global healthcare

the model’s responses. These concerns underscore the need for eth-

applications. GatorTron, by contrast, is more specialized for English-

ical oversight when using LLMs in culturally diverse settings[10].

speaking environments. GPT-4 offers broader utility but lacks the

Furthermore, LLMs could reinforce healthcare inequities if not

cultural and linguistic specificity of L2M3 and the clinical precision

properly adapted to the specific needs and practices of various

of GatorTron and LlamaCare[23].

cultures[6]. Additionally, LLMs must account for linguistic diver-

Patient Communication. GPT-4 and LlamaCare both demon-

sity in global healthcare contexts. For example, healthcare systems

strate potential in patient-doctor communication, but they serve dif-

in multilingual nations, such as India, require models that can func-

ferent purposes. GPT-4 excels in generating empathetic responses,

tion across multiple languages while understanding the nuances of

while LlamaCare offers more medically precise information. L2M3’s

local medical customs. There is an innate link between culture, lan-

focus on multilingual communication gives it an edge in culturally

guage, and patient safety, making linguistic competency critical for

diverse settings, but GatorTron’s role remains more data-focused[17][9][26].

patient-centered care[11]. The LlamaCare model provides a frame-

Customization and Accuracy. Both GatorTron and LlamaCare

work for sharing healthcare knowledge across diverse linguistic

excel in accuracy due to their fine-tuning on medical data[30].

groups, showing how LLMs can facilitate cross-cultural knowledge

GPT-4, while highly versatile, does not have the same level of

sharing in healthcare[22]. Figure 5 below focuses on the challenges,

specialization[23]. L2M3, on the other hand, is a standout for global

as opposed to the benefits, of applying LLMs in the healthcare

healthcare, particularly in low-resource regions[3].

domain.





6 CHALLENGES



6.1 Technical Challenges of Implementing

Medical LLMs

Despite promising applications, adapting LLMs to specific medical

tasks presents technical challenges. A major issue is the need for

domain-specific training data. The models need considerable fine-

tuning for clinical natural language processing (NLP) tasks, making

zero-shot learning an emerging solution[20]. When it comes to the

application of zero-shot learning in preventive healthcare, it has an



ability to provide accurate decision support even in niche medical Figure 5: The benefits and challenges of applying large lan-

contexts[13]. In addition, fine-tuning LLMs to accommodate local

guage models in the medical domain

medical customs and linguistic variations is critical for effective

patient communication across diverse healthcare settings. That is



why multilingual models like L2M3, which cater to low-resource 7 EVALUATING THE EFFECTIVENESS OF regions where local dialects and cultural practices significantly

influence healthcare delivery, have such importance[3]. Without in- LLMS IN HEALTHCARE

corporating local linguistic demands and medical traditions, LLMs The evaluation of LLMs in healthcare requires a combination of

risk misinterpretation, leading to misdiagnosis or poor patient out- qualitative and quantitative metrics to ensure a balanced assess-

comes. ment of their performance. In addition to the qualitative analysis



939

IS2024 - CHATGPT in MEDICINE,

Petrovska Dojchinovska et al.



of their language generation capabilities, this review incorporates designed to recognize emotional cues in patient language and ad-

quantitative metrics such as accuracy rates, error rates in specific just their responses to provide more empathetic care. Another vital

tasks (e.g., diagnosis or patient communication), and patient satis- research direction is improving the contextual awareness of LLMs,

faction scores derived from recent studies. These metrics provide particularly in culturally diverse settings where understanding local

a robust framework[9] for comparing LLMs across various clini- customs, languages, and medical practices is essential. Furthermore,

cal applications. For example, models like GPT-4 and LlamaCare future research should explore ways to integrate LLMs with other

have been evaluated for their diagnostic accuracy, while GatorTron healthcare technologies, such as electronic health records (EHRs), to

has shown promise in reducing administrative burdens through streamline clinical workflows while ensuring data privacy. Finally,

improved EHR management. Quantitative evidence, such as the the creation of robust, standardized evaluation frameworks[19][9]

accuracy rates of LLMs in diagnosing rare conditions or the error will be crucial in assessing the long-term effectiveness of LLMs

rates in patient-facing applications, strengthens the argument for in healthcare, focusing on patient outcomes, model accuracy, and

their continued refinement and deployment in healthcare. ethical and regulatory compliance[6][15].

Evaluating LLMs in medical applications, focusing on adaptabil-

ity to local linguistic demands, accuracy, and transparency requires 10 CONCLUSION

a detailed framework[9]. There are METRICS in place that serve LLMs hold immense potential to revolutionize healthcare by en-

as a tool for standardizing LLM evaluations, ensuring they are tai- hancing diagnostic support, improving patient-doctor communica-

lored to local healthcare systems and relevant cultural needs[19]. tion, and facilitating equitable care. Their success, however, hinges

In multilingual and multicultural settings, the evaluation of LLMs on addressing challenges such as local medical customs, linguis-

should prioritize their ability to maintain cultural and linguistic ap- tic diversity, and ethical concerns. Proper adaptation of LLMs can

propriateness. The L2M3 model, designed for regions with limited significantly enhance patient-centered care by making communi-

healthcare resources, demonstrates how multilingual and cultur- cation between patients and healthcare providers more culturally

ally aware LLMs can improve healthcare outcomes in underserved sensitive and linguistically appropriate. Prioritizing inclusivity and

communities[3]. transparency is essential for improving healthcare outcomes and

equity. Future research must focus on addressing cultural and lin-



8 guistic diversity, improving emotional sensitivity in patient-doctor GAPS IN THE EXISTING RESEARCH

interactions, mitigating biases, and establishing ethical and legal

As LLMs become more prevalent in clinical environments, the frameworks for AI in healthcare. Specialized research is also needed

focus on improving patient-doctor communication and address- to tailor LLMs for specific clinical domains, such as mental health

ing cultural and linguistic considerations will be crucial. Over- and surgery, to ensure these tools are safe, reliable, and contextually

reliance on LLMs without adequate human oversight, particularly aware.

in sensitive clinical areas like diagnostics and treatment planning

is dangerous[29]. Additionally, there is a need for stringent ethi- 11 ACKNOWLEDGEMENT cal guidelines to ensure that LLMs do not perpetuate healthcare

This project is funded by the European Union under Horizon Europe

inequities[6]. In the papers that were evaluated for this review,

(project ChatMED grant agreement ID: 101159214).

we have identified the most notable gap in the limited focus on

multicultural and linguistic diversity. There is a need for more



comprehensive studies and models specifically designed for non-

Western and linguistically diverse healthcare settings. LLMs must

be fine-tuned to local medical practices and languages to truly

serve global healthcare needs. The second gap identified is the bias



that exists in LLM healthcare recommendations[10][6]. There is a REFERENCES lack of clear, actionable frameworks for systematically identifying

and reducing bias in LLMs. Addressing this will require a multi- [1] Prottay Kumar Adhikary, Aseem Srivastava, Shivani Kumar, Salam Michael

Singh, Puneet Manuja, Jini K Gopinath, Vijay Krishnan, Swati Kedia Gupta,

disciplinary approach combining AI ethics, clinical expertise, and Koushik Sinha Deb, and Tanmoy Chakraborty. Exploring the efficacy of large

sociocultural considerations. The third gap is the emotional sensi- language models in summarizing mental health counseling sessions: Benchmark

study. JMIR Mental Health, 11:e57306, 2024.

tivity in patient-doctor communication[1][21][18]. Finally, there is [2] Xi Chen. Large language models in the medical field: Principles and applications.

the danger of over-reliance on LLMs for decision support[6][29]. International Journal of Computer Science and Information Technology, 2(3):219–

Still, the field is quite promising, so exploring future directions for 224, 2024.

[3] Agasthya Gangavarapu. Introducing l2m3, a multilingual medical large language

research is worthwhile. model to advance health equity in low-resource regions, 2024.

[4] Marius Geantă, Daniel Bădescu, Narcis Chirca, Ovidiu Cătălin Nechita, Cos-

min George Radu, Stefan Rascu, Daniel Rădăvoi, Cristian Sima, Cristian Toma,

9 and Viorel Jinga. The potential impact of large language models on doctor–patient FUTURE DIRECTIONS FOR RESEARCH

communication: A case study in prostate cancer. In Healthcare, volume 12, page

The potential of LLMs in healthcare remains largely untapped, with 1548. MDPI, 2024.

[5] Tawara D Goode, M Claire Dunne, and Suzanne Bronheim. The evidence base

significant gaps in current research that need addressing. One criti- for cultural and linguistic competency in health care. Commonwealth Fund New

cal area for future investigation is the development of LLMs that York, NY, 2006.

are emotionally sensitive and capable of handling high-stakes, emo- [6] Joschka Haltaufderheide and Robert Ranisch. The ethics of chatgpt in medicine

and healthcare: a systematic review on large language models (llms). NPJ Digital

tionally charged patient interactions. For example, models could be Medicine, 7(1):183, 2024.



940



Zbornik 27. mednarodne multikonference



INFORMACIJSKA DRUŽBA – IS 2024

Zvezek L



Proceedings of the 27th International Multiconference



INFORMATION SOCIETY – IS 2024

Volume L



Digitalna preobrazba zdravstvene nege - SIZN 2024



Digital Transformation of Nursing - NIIG 2024



Uredniki / Editors



Jelena Ficzko, Ema Dornik, Uroš Rajkovič,

Tamara Štemberger Kolnik, Gregor Štiglic, Boštjan Žvanut



http://is.ijs.si



8. oktober 2024 / 8 October 2024

Ljubljana, Slovenia





PREDGOVOR




Pred vami je zbornik konference »Digitalna preobrazba zdravstvene nege (Sekcija za

informatiko v zdravstveni negi - SIZN)«, ki je potekala v okviru 27. mednarodne

multikonference Informacijska družba – IS 2024, 8. oktobra 2024 na Institutu Jožef Stefan v

Ljubljani.



Konferenco je organizirala SIZN pri Slovenskem združenju za medicinsko informatiko.

Partnerji konference so bili (v abecednem vrstnem redu):

• Fakulteta za vede o zdravju Univerze na Primorskem;

• Fakulteta za zdravstvene vede v Celju;

• Fakulteta za zdravstvene vede Univerze v Mariboru;

• Slovensko društvo za medicinsko informatiko;

• Zdravstvena fakulteta Univerze v Ljubljani;

• Zbornica zdravstvene in babiške nege Slovenije – Zveza strokovnih društev medicinskih

sester, babic in zdravstvenih tehnikov Slovenije



Konferenca predstavlja pomemben mejnik v razvoju informatike v zdravstveni negi, saj gre za

prvi dogodek SIZN na multikonferenci Informacijska družba. Predavatelji so predstavili

različne vidike digitalizacije zdravstvene nege in poudarili tako znanstvene kot praktične

implikacije digitalne preobrazbe, predvsem v povezavi z uporabo umetne inteligence.



Osrednja tema razprav je bila usmerjena v izboljšanje kakovosti oskrbe in učinkovitosti dela v

zdravstveni negi s pomočjo informacijsko komunikacijske tehnologije. Poleg tehnoloških

priložnosti so bili izpostavljeni tudi etični izzivi ter vpliv digitalizacije na zdravstvo. Večino

prispevkov so predstavile diplomirane medicinske sestre in zdravstveniki, kar poudarja

zavedanje stroke o pomenu digitalne transformacije za prihodnost zdravstvene nege.



Zahvaljujemo se vsem udeležencem za aktivno sodelovanje v razpravah. Posebna zahvala gre

avtorjem prispevkov, ki ste obogatili naše razumevanje digitalne transformacije v zdravstveni

negi in pripomogli k postavitvi sodobnih temeljev za nadaljnji razvoj informatike v zdravstveni

negi. Za konec bi se radi zahvalili še dr. Maximu Topazu, vrhunskemu strokovnjaku, ki

preučuje uporabo umetne inteligence v zdravstveni negi, za plenarno predavanje.



Programski odbor konference »Digitalna preobrazba zdravstvene nege (SIZN)«





FOREWORD




This is the conference proceedings of "Digital Transformation of Nursing (Slovenian Nursing

Informatics Interest Group - NIIG)," which took place within the 27th International

multiconference Information Society – IS 2024 on October 8, 2024, at the Jožef Stefan Institute

in Ljubljana, Slovenia.



The conference was organized by NIIG, interest group within the Slovenian Medical

Informatics Association. The conference partners were (in alphabetical order):

• • Faculty of Health Sciences, University of Primorska,

• • Faculty of Health Sciences, Celje,

• • Faculty of Health Sciences, University of Maribor,

• • Slovenian Association for Medical Informatics,

• • Faculty of Health Sciences, University of Ljubljana,

• • Nurses and Midwives Association of Slovenia.



The conference represents an important milestone in the development of nursing informatics,

as this is the first event organized by the NIIG at the Multiconference Information Society. The

authors addressed various aspects of the digitalization of nursing, emphasizing both the

scientific and practical implications of digital transformation in nursing, particularly in relation

to the use of artificial intelligence.



The central theme of the discussions was focused on improving the quality of care and the

efficiency of nursing work through information and communication technology. In addition to

technological opportunities, ethical challenges and the impact of digitalization on healthcare

were also highlighted. The majority of the contributions were presented by registered nurses

and health professionals, underscoring the profession's awareness of the importance of digital

transformation for the future of nursing.



We would like to thank all the participants for their active involvement in the discussions.

Special thanks go to the authors of the contributions, who enriched our understanding of digital

transformation in nursing and helped lay the modern foundations for further development of

nursing informatics. Finally, we would like to thank also Dr. Maxim Topaz, a leading expert in

the field of artificial intelligence in nursing, for delivering the plenary lecture.



Program Committee of the Conference "Digital Transformation of Nursing (NIIG)"



PROGRAMSKI ODBOR / PROGRAMME COMMITTEE



Jelena Ficzko



Zdravstvena fakulteta, Univerza v Ljubljani



Ema Dornik



Institut informacijskih znanosti



Uroš Rajkovič



Fakulteta za organizacijske vede, Univerza v Mariboru



Tamara Štemberger Kolnik



Fakulteta za zdravstvene vede v Celju



Gregor Štiglic



Fakulteta za zdravstvene vede, Univerza v Mariboru



Boštjan Žvanut



Fakulteta za vede o zdravju, Univerza na Primorskem



Ali lahko etična odobritev in soglasje pacienta sledita



pogovornim robotom na področju duševnega zdravja?



Can Ethical Approval and Patient Consent Keep Pace with



Conversational Agents in Mental Health?



Leona Cilar Budler Gregor Štiglic

Faculty of Health Sciences Faculty of Health Sciences

University of Maribor University of Maribor

Maribor, Slovenia Maribor, Slovenia

leona.cilar1@um.si gregor.stiglic@um.si



Povzetek insufficient research on how ethical issues are reported. Whether researchers routinely obtain ethical approval or informed consent Čeprav številne znanstvene raziskave uvajajo napredne remains unclear. This investigation seeks to determine how računalniške metode v raziskave na področju duševnega zdravja, mental health studies employing automated conversational ni dovolj raziskav o tem, kako se poroča o etičnih soglasjih. Ali agents either overlook or inadequately report patient consent and raziskovalci rutinsko pridobijo etično odobritev ali informirano ethical clearance from appropriate boards. Almost half of the soglasje, ostaja nejasno. Ta raziskava skuša ugotoviti, kako reviewed studies mentioned securing ethical approval and študije na področju duševnega zdravja, ki uporabljajo informed consent, reflecting varied opinions on adherence to avtomatizirane pogovorne robote, spregledajo ali neustrezno ethical standards. There is a link between the size of the study poročajo o soglasju pacienta in o etičnem dovoljenju ustreznih sample and obtaining ethical approval, especially in publications komisij. Le približno polovica pregledanih študij je omenila from journals with lower impact factors. Future research should zagotavljanje etične odobritve in informirane privolitve, kar explore how journal policies impact ethical standards in research. odraža različna mnenja o spoštovanju etičnih standardov. Additionally, educational programs could be established to Ugotovili smo tudi, da obstaja povezava med velikostjo vzorca emphasise the significance of ethical considerations, especially oz. številom udeležencev v raziskavi in pridobitvijo etične in studies involving smaller groups of participants. odobritve, zlasti pri objavah iz revij z nižjimi faktorji vpliva. V



prihodnosti bi bilo smiselno raziskati, kako politike revij vplivajo Keywords na etične standarde v raziskavah. Poleg tega bi lahko vzpostavili

izobraževalne programe, ki bi poudarjali in ozaveščali o pomenu Artificial intelligence; mental health; ethics; education; patient

etičnih vidikov, zlasti v študijah, ki vključujejo manjše skupine selection; automated conversational agents

udeležencev.



Ključne besede



Umetna inteligenca; duševno zdravje; etika; izobraževanje;

izbor pacientov; avtomatizirani pogovorni posredniki



Abstract

Although numerous scientific publications introduce advanced

computational methods in mental health research, there is



Permission to make digital or hard copies of part or all of this work for personal or Permission to make digital or hard copies of part or all of this work for personal or classroom use is granted without fee provided that copies are not made or distributed classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for third-party components of this work must citation on the first page. Copyrights for third-party components of this work must be honored. For all other uses, contact the owner/author(s). be honored. For all other uses, contact the owner/author(s). Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia © 2024 Copyright held by the owner/author(s). © 2024 Copyright held by the owner/author(s). http://doi.org/DOI_RECEIVED_AFTER_REVIEW http://doi.org/DOI_RECEIVED_AFTER_REVIEW



Uporaba umetne inteligence v procesu zdravstvene nege:



Analiza vpliva navodil na kakovost rezultatov



The use of artificial intelligence in the nursing process: An



analysis of the impact of instructions on quality of



outcomes



Andreja Hrovat Bukovšek Tamara Štemberger Kolnik Andreja Ljubič

Fakulteta za zdravstvene vede v Fakulteta za zdravstvene vede v Fakulteta za zdravstvene vede v

Celju Celju Celju

Celje, Slovenija Celje, Slovenija Celje, Slovenija

andreja.hrovat-bukovsek@fzvce.si tamara.stemberger-kolnik@fvze.si andreja.ljubic@fvzce.si



Povzetek Ključne besede

Prispevek predstavlja raziskavo, ki preučuje sposobnost

ChatGPT-ja za generiranje negovalnih diagnoz, ciljev in ChatGPT, negovalna diagnoza, cilji in intervencije zdravstvene

intervencij v negovalnih načrtih zdravstvene nege. Raziskava nege, natančnost navodil, NANDA klasifikacija



primerja rezultate ChatGPT-ja s standardnimi negovalnimi Abstract



učne negovalne dokumentacije - Študije primera. Analizirali smo ability of ChatGPT to formulate nursing diagnoses, goals and interventions in the nursing care plan. The research compares the besedilne odgovore, ki jih je ustvaril ChatGPT, in jih primerjali results of ChatGPT with standard nursing classifications and s standardi za negovalne diagnoze (NANDA), oblikovanje ciljev protocols, while investigating how students effectively use this učinkovito uporabljajo to orodje umetne inteligence za pripravo Introduction: The paper presents research that examines the klasifikacijami in protokoli ter hkrati preiskuje, kako študenti



od jasnosti in natančnosti podanih navodil. Študenti so dosegli SNOČ, and the creation of specific and measurable time frames for nursing interventions in the care plan. In addition, we različne rezultate kljub enotnim navodilom, kar poudarja vpliv examined the impact of different data entry methods and metode vnosa podatkov in interpelacije navodil. Predvsem so se instructions on the results generated by ChatGPT. Key findings razlike pokazale v kakovosti opredelitve/poimenovanja highlight that the effectiveness of ChatGPT depends on the negovalnih diagnoz, oblikovanju ciljev, saj le-ti niso bili dovolj clarity and precision of the instructions given. Students achieved specifični ali merljivi ter pri opredelitvi časovnih okvirjev different results despite uniform instructions, highlighting the oziroma frekvenc negovalnih intervencij. Umetna inteligenca influence of the data entry method and the interpellation of uspešno prepoznava negovalne probleme, vendar ne uporablja instructions. In particular, differences were evident in the quality uradne NANDA klasifikacije. ChatGPT je uporabno orodje za Poleg tega smo preučili vpliv različnih načinov vnosa podatkov documentation - Case Studies. We analyzed the text responses generated by ChatGPT and compared them to the Standards for in navodil na rezultate, ki jih ChatGPT generira. Ključne Nursing Diagnoses (NANDA), the creation of goals according to ugotovitve izpostavijo, da je učinkovitost ChatGPT-ja odvisna okvirov za intervencije zdravstvene nege v negovalnem načrtu. artificial intelligence tool to produce learning nursing po SNOČ-u ter oblikovanje specifičnih in merljivih časovnih



osnovno oblikovanje negovalnih načrtov zdravstvene nege, of the definition/naming of nursing diagnoses, the formulation of goals as these were not specific or measurable enough, and the vendar je njegova učinkovitost odvisna od natančnosti navodil in definition of timeframes or frequencies of nursing interventions. usposobljenosti uporabnika. Raziskava poudarja potrebo po Artificial intelligence successfully identifies nursing problems boljspecifičnih smernicah in prilagoditvah pri vključevanju but does not use the official NANDA classification. ChatGPT is ChatGPT-ja v izobraževalne procese zdravstvene nege a useful tool for the basic formulation of nursing care plans, but

its effectiveness depends on the accuracy of the instructions and

the competence of the user. The research highlights the need for

more specific guidelines and adaptations when integrating

Permission to make digital or hard copies of part or all of this work for personal or ChatGPT into nursing education processes.

classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full Keywords citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s). ChatGPT, nursing diagnosis, nursing goals and interventions,

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia instruction accuracy, NANDA classification © 2024 Copyright held by the owner/author(s).





Povezanost dimenzij informacijske varnostne kulture z




namero nepooblaščenega dostopa do zdravstvenih



podatkov



The relationship between the dimensions of information



security culture and the intention of unauthorised access



to health data



Samanta Mikuletič Boštjan Žvanut

Gregorčičeva cesta 8 Univerza na Primorskem

Fakulteta za vede o zdravju

6250 Ilirska Bistrica Polje 42, 6310 Izola, Slovenija

Slovenija bostjan.zvanut@fvz.upr.si

samanta.mikuletic@gmail.com



Povzetek Abstract

The world is experiencing an increase in information security

V svetu je zabeležena porast groženj in kršitev informacijske threats and breaches. While information about these breaches is

varnosti. Medtem ko so v tujini javno dostopne informacije o publicly available abroad, we have no official data on breaches



kršitvah varnosti in zasebnosti zdravstvenih podatkov. Zaposleni key role in protecting this data as they make up the majority of v zdravstveni negi igrajo ključno vlogo pri njihovi zaščiti, saj healthcare workers and have direct access to healthcare data due omenjenih kršitvah, pri nas ni na voljo uradnih podatkov o of health data security and privacy. Nursing employees play a



narave svojega dela omogočen neposreden dostop do to the nature of their work. The aim of this research was to predstavljajo večino zaposlenih v zdravstvu in imajo zaradi

zdravstvenih podatkov. Namen raziskave je s pomočjo Teorije explain information security breaches by nurses using the Theory

načrtovanega vedenja pojasniti kršitev informacijske varnosti s of Planned Behaviour by considering the impact of the

strani zaposlenih v zdravstveni negi in pri tem preveriti učinek dimensions of information security culture. A survey among

dimenzij informacijske varnostne kulture. V obdobju od aprila nursing employees in Slovenia (n = 527) was conducted between

2021, do marca 2022 je bila izvedena anketa med zaposlenimi v April 2021 and March 2022. The results showed that normative

zdravstveni negi v Sloveniji (n = 527). Rezultati so pokazali, da beliefs have the greatest impact on the behavioural intention to



namero nepooblaščenega dostopa do podatkov. Identificirali pa relationship was found between the dimensions of information smo tudi posredno povezavo med dimenzijami informacijske imajo normativna prepričanja največji učinek na vedenjsko access health data without authorisation. In addition, an indirect



varnostne kulture (usmerjenost k zagotavljanju zasebnosti; security culture (privacy oriented; security oriented; information security knowledge sharing; security education, training, and usmerjenost k varnosti podatkov; izmenjava znanja o awareness). and unauthorised access to health data. The results informacijski varnosti; varnostno izobraževanje, usposabljanje in

ozaveščanje) in nepooblaščenim dostopom do podatkov. Izsledki of this study contribute to the understanding of information

raziskave doprinašajo k razumevanju informacijske varnostne security culture and the relationship between its dimensions and

kulture in povezanosti njenih dimenzij na vedenjsko namero the behavioural intention of healthcare workers regarding

zaposlenih v zdravstveni negi do nepooblaščenega dostopa do unauthorised access to health data. zdravstvenih podatkov.



Ključne besede data privacy and security, health data, electronic health records, Keywords

zasebnost in varnost podatkov, zdravstveni podatki, elektronski data breaches, information security, nursing, healthcare

zdravstveni zapisi, kršenje varnosti podatkov, informacijska providers. varnost, zdravstvena nega, izvajalci zdravstvenih storitev.



∗Article Title Footnote needs to be captured as Title Note

†Author Footnote to be captured as Author Note



Permission to make digital or hard copies of part or all of this work for personal or

classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).





Odkrivanje znanja v zdravstvenih informacijskih sistemih




s pomočjo rudarjenja procesov



Knowledge discovery in healthcare information systems



with process mining



Gregor Polančič

Univerza v Mariboru,

Fakulteta za elektrotehniko,

računalništvo in informatiko,

Slovenija

gregor.polancic@um.si



POVZETEK

Rudarjenje procesov je sinonim za množico tehnik in tehnologij, 1 UVOD



neposredno, na osnovi podatkov, ki so pridobljeni iz vključuje ustvarjanje vrednosti za stranke, povečanje prihodkov, informacijskega sistema podjetja. V prispevku so predstavljene izboljšanje operativne učinkovitosti ter trajnostni razvoj. ki omogočajo avtomatsko odkrivanje znanja o procesih Vsaka organizacija stremi, k doseganju poslovnih ciljev, kar



naprednih algoritmov za odkrivanje, preverjanje skladnosti in Pomembno sredstvo za doseganje navedenega je ustrezna skrb osnove rudarjenja procesov, pomen dnevnikov dogodkov in

izboljšave procesov. Osrednji del prispevka je demonstracija za procese organizacije z metodologijo upravljanja poslovnih

zmožnosti sodobnih orodij za rudarjenje procesov, ki temeljijo procesov (angl. business process management), ki predstavlja

na dnevniku dogodkov, pridobljenem iz zdravstvenega skupek uveljavljenih praks, tehnik in tehnologij za kontinuirano

informacijskega sistema in so sledeče: avtomatizirano odkrivanje izboljševanje in prilagajanje poslovnih procesov.

procesov, rudarjenje uspešnosti procesov, preverjanje skladnosti Številne aktivnosti »tradicionalnega« upravljanja poslovnih

in analiza različic izvajanih procesov. procesov temeljijo na modelih procesov, ki služijo različnim



KLJUČNE BESEDE namenom kot so: analiziranje procesov, komuniciranje o

procesih, digitalizacija procesov in spremembe oz. izboljšave

Rudarjenje procesov, upravljanje poslovnih procesov, procesov. Ker v takšnih primerih odločitve upravljanja poslovnih

zdravstveni informacijski sistem procesov v veliki meri temeljijo na modelih procesov, je



ABSTRACT validity), kar pomeni, da modeli predstavljajo dejanske procese poglavitni izziv zagotavljanje njihove veljavnosti (angl.



Process mining is a synonym for a multitude of techniques and oziroma njihovo operativno izvajanje in da le te opisujejo v celoti

technologies that enable the automatic discovery of knowledge (angl. completeness) [1].

about processes based on data obtained from the company's V praksi se je izkazalo, da je ravno zagotavljanje veljavnih

information system. The paper presents the basics of process in pravilnih modelov procesov pogosto šibek člen njihovega

mining, the importance of event logs and advanced algorithms upravljanja. Zaradi nenehnih sprememb v poslovnem okolju se

for detection, compliance verification and process improvements. poslovni procesi kontinuirano spreminjajo in prilagajajo. Modeli

The central part of the paper is a demonstration of the capabilities poslovnih procesov tako hitro postanejo neskladni z dejansko

of modern tools for process mining, which are based on the event izvajanimi procesi kakor tudi s tehničnim in organizacijskim



follows: automated process discovery, process performance je prav tako podvrženo človeškim dejavnikom, saj je močno mining, compliance checking, and variant analysis of odvisno od spretnosti, znanj in razpoložljivosti analitika, kar log obtained from the healthcare information system and are as okoljem v katerem se izvajajo. Modeliranje poslovnih procesov

implemented processes.

vpliva na izdelane modele procesov. Tradicionalno odkrivanje in

modeliranje procesov je drago in časovno potratno tudi zaradi

KEYWORDS

vrzeli v poslovnem znanju deležnikov in pomanjkanja

Process mining, business process management, healthcare objektivnih validacij modelov [2]. Modeli so zato pogosto

information system nepopolni (ne opisujejo celotnega procesa), neskladni (ne

https://doi.org/10.70314/is.2024.sizn.4 predstavljajo dejanskega procesa) ali nepravilni (ne upoštevajo

Permission to make digital or hard copies of part or all of this work for personal or pravil diagramskega jezika, na primer BPMN, in so zato

classroom use is granted without fee provided that copies are not made or distributed nerazumljivi tako za ljudi kakor tudi za izvajalna okolja).

for profit or commercial advantage and that copies bear this notice and the full Poslovne odločitve, ki se sprejemajo na neustreznih modelih

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s). procesov, postanejo hitro neoptimalne ali celo napačne.

Information Society 2023, 9–13 October 2023, Ljubljana, Slovenia Z višanjem stopnje digitalizacije in avtomatizacije

© 2023 Copyright held by the owner/author(s).

(procesov) so se pojavile priložnosti za reševanje omenjenih

izzivov, z bolj neposrednim upravljanjem dejansko izvajanih Rudarjenje procesov omogoča poslovnim uporabnikom, da

procesov na osnovi podatkov, ki jih poslovne informacijske identificirajo ozka grla, nepotrebne ponovitve opravil,

rešitve generirajo v poslovnih operacijah (angl. business odstopanja in vire odpadkov v svojih procesih ter odkrijejo

operations). Te podatke je možno z usmerjenimi tehnikami priložnosti za optimizacijo delovanja in povečanje pozitivnih

rudarjenja podatkov (angl. data mining) pretvoriti v obliko, ki je poslovnih rezultatov.

uporabna za sprejemanje odločitev upravljanja procesov. V zadnjem desetletju je rudarjenje procesov dozorelo in na

Besedna zveza “rudarjenje procesov” (angl. process mining) trg poslalo širok nabor orodij za poslovno inteligenco in

označujemo družino podatkovno vodenih tehnik za analizo upravljanje poslovnih procesov. Rudarjenje procesov se danes

poslovnih procesov z uporabo podatkov o dogodkih, pridobljenih uporablja v skoraj vseh industrijskih sektorjih, vključno z

iz informacijskih sistemov, kot so ERP sistemi in sistemi za bančništvom in finančnimi storitvami, telekomunikacijami,

upravljanje odnosov s strankami (CRM). Rudarjenje procesov energetiko, zdravstvom, logistiko in proizvodnjo.

predstavlja uspešen primer prehoda akademskih zamisli in

rešitev v poslovna okolja [2]. 2.1 Dnevniki dogodkov

Za analizo poslovnega procesa s tehnikami rudarjenja procesov

je potrebno iz informacijskega sistema pridobiti dnevnik

2 OSNOVE RUDARJENJA PROCESOV dogodkov (Slika 2), ki beleži izvajanje procesa. Dnevnike

Rudarjenje procesov je krovni izraz ki predstavlja dogodkov je mogoče pridobiti iz skoraj katerega koli

kombinacijo podatkovnega rudarjenja in tehnik upravljanja informacijskega sistema podjetja, pa naj bo to iz sistemov ERP

procesov, ki uporabljajo napredne algoritme, strojno učenje in ali CRM, kot so SAP, Dynamics, Salesforce ali ServiceNow, ali

statistične metode za analizo podatkov o dogodkih, z namenom iz vertikalno specializiranih sistemov, kot so sistemi za izvajanje

analizirati poslovne operacije, vse od celovitih procesov do proizvodnje, sistemi za upravljanje zavarovanj, sistemi za

posameznih aktivnosti procesov. upravljanje bolnišnic, itd.

Slika 1 prikazuje rudarjenje procesov v širšem delovanju

organizacije. Poslovni procesi, med katere se uvrščajo tudi

zdravstveni procesi, preko interakcij med zdravniki,

medicinskimi sestrami, bolniki in drugimi deležniki, ustvarjajo

velike količine podatkov. Zdravstveni procesi so vedno bolj

podprti in nadzorovani z zdravstvenimi informacijskimi sistemi,

ki hranijo podatke o izvajanju procesov. Podatke o izvajanju

procesov je mogoče uporabiti za izdelavo dnevnikov dogodkov,

katerih ključni gradniki so ponazorjeni v Slika 2. Pridobljene

dnevnike dogodkov je mogoče uporabiti za tri osnovne namene

rudarjenja procesov: odkrivanje, skladnost in izboljšanje.

Odkrivanje se začne z dnevniki dogodkov in ustvari model

procesa. Preverjanje skladnosti primerja obstoječi model z

dnevnikom dogodkov in s tem izpostavi ujemanja in odstopanja

med modelom procesa in dnevnikom dogodkov. Tehnika

izboljšave se uporablja za razširitev, izboljšavo ali popravilo

modela procesa, npr. da ga uporabimo za izvajanje simulacij.





Slika 2: Preprosti dnevnik dogodkov





Dnevnik dogodkov, kot je ponazorjen na Slika 2, je niz zapisov

dogodkov, kjer je vsak zapis dogodka sestavljen iz naslednjih

atributov:

- identifikator primera ali ID primera (angl. case ID), ki

predstavlja enolični identifikator primera (instance), kot je

na primer: ID bolnika;

- časovni žig, ki predstavlja trenutek, ko je bil dogodek

zabeležen v sistemu;

- aktivnost; oznaka, ki se nanaša na dejavnost, ki je bila

izvedena. V tem primeru je šest različnih oznak dejavnosti,

in sicer: registracija, CT skeniranje, EEG test, posvetovanje,

urinski test in odpust;

- vrsta transakcije, ki predstavlja status aktivnosti. V tem

primeru dogodek bodisi predstavlja začetek ali zaključek

aktivnosti;

Slika 1: Delovanje rudarjenja procesov [3] - vir; član osebja ali medicinski pripomoček, povezan z

izvajanjem aktivnosti. V primeru je osem različnih

zaposlenih vključenih v dogodke.

Referenčni dnevniki dogodkov so dostopni na spletnem naslovu: 3 ZMOŽNOSTI ORODIJ ZA RUDARJENJE

https://processmining.org/event-data.html. PROCESOV



2.2 Sodobne rešitve za rudarjenje procesov omogočajo analitikom in Algoritmi poslovnim vodjem:

Kot je razvidno iz Slika 1 predstavlja dnevnik dogodkov vhod v - razumeti, kako se izvajajo poslovne operacije, na osnovi

algoritme rudarjenja procesov. Za potrebe odkrivanja, generiranih »AS-IS« diagramov procesov, ki temeljijo na

preverjanja skladnosti in izboljšav modelov procesov, uporablja podatkih o dogodkih, ki jih beleži informacijski sistem

rudarjenje procesov zapletene in napredne algoritme, ki lahko organizacije.

temeljijo na hevristikah, mehki logiki (angl. fuzzy logic), - analizirati podatke, z namenom, da se identificirajo točke

strojnem učenju (angl. maschine learning) in genetskih trenja v poslovnem procesu in jih poveže s ključnimi

algoritmih. kazalniki uspešnosti (angl. key performance indicators).

Med osnovne algoritme odkrivanja procesov spada algoritem - razumeti, kaj prispeva k zaželenim in nezaželenim

Alfa, ki analizira relacije urejenosti med pari dogodkov v sledeh rezultatom procesa, na primer različne aktivnosti, ki

(angl. traces) dnevnika dogodkov kot so: neposredni naslednik prispevajo k naročilom, ki so dostavljena pravočasno, v

(angl. direct successor), vzročnost (angl. causality), sočasnost primerjavi z naročili, ki so dostavljena z zamudo.

(angl. concurrency) in ekskluzivnost (angl. exclusiveness). - prepoznati neskladno vedenje, razumeti temeljne vzroke

Rezultat algoritma α je mreža delovnega toka, ki ohranja odstopanj in kvantificirati vplive teh odstopanj na uspešnost

omenjene relacije dnevnika dogodkov L, kar zapišemo kot: procesa.

𝛼(𝐿). [3] - napovedovati prihodnjo uspešnost procesa v različnih

scenarijih, tako da lahko ekipe sprejemajo boljše odločitve



in bolje razvrstijo prednostna prizadevanja za

avtomatizacijo in izboljšanje procesov.



Pridobivanje zgornjega znanja o procesih je možno s štirimi

osnovnimi zmožnostmi orodij za rudarjenja procesov (Slika 5):





Slika 3: Algoritem Alfa



Druga skupina algoritmov je namenjena preverjanju skladnosti,

ki kvantificirajo odstopanja med zapisi v dnevniku dogodkov (tj.

realno obnašanje procesa) in predpisanim modelom procesa. Na

primer, preverjanje skladnosti na osnovi ponovitve izvedbe

žetona (angl. token-based replay), natančno prešteje kateri

primerki procesa odstopajo od predpisane izvedbe in v kolikšni

meri.

566 566

971 971 1537 1537 461 461



1391 1391 b

1537 1537

examine g

thoroughly

pay

+443 compensation

a c -443 d e p4

start register p1 examine p2 check p3 decide end

request casually ticket h Slika 5: Zmožnosti rudarjenja procesov

problem reject request 930 f

443 tokens remain in place p2, reinitiate 930

because d did not occur although request 146 V nadaljevanju je predstavljena demonstracija zmožnosti the model expected d to happen

146

problem rudarjenje procesov z uporabo orodja Apromore

443 tokens were missing in place p2 during (www.apromore.com), ki temeljijo na referenčnem in

replay, because d happened even though

this was not possible according to the model anonimiziranem dnevniku dogodkov, pridobljenem iz

zdravstvenega informacijskega sistema.

Slika 4: Kvantifikacija veljavnih in neveljavnih korakov



izvedbe procesa [3] 3.1 Avtomatizirano odkrivanje procesov



Z uporabo funkcije prileganja (angl. fitness) se nato poda Kot je razvidno iz Slika 5, avtomatizirano odkrivanje procesov

vrednost, ki predstavlja stopnjo ujemanja izvedbe procesa z zajame dnevnik dogodkov in ustvari model »AS-IS«, ki natančno

modelom procesa v odstotkih oziroma na intervalu [0..1]. nakazuje, kako dejanski proces deluje, in omogoča primerjavo s

tem, kako podjetje želi, da proces deluje. Aplikacije za rudarjenje

procesov podjetjem omogočajo vizualizacijo procesa in



𝑓𝑖𝑡𝑛𝑒𝑠𝑠(𝜎, 𝑁) = 1 𝑚𝑁,𝜎 1 𝑟𝑁,𝜎 razumevanje, kje so težave in kako je mogoče proces izboljšati z (1 − ) + (1 − ) 2 𝑐 𝑁,𝜎 2 𝑝 𝑁,𝜎 avtomatizacijo ali drugimi prizadevanji za izboljšanje.

Avtomatizirano odkrivanje procesov prikazuje dejanski

potek procesa, točke, kjer se sprejemajo odločitve, kje se izvajajo

posegi, kdo jih izvaja, kje poteka predelava in odvečno delo ter

kje poteka predaja med zaposlenimi. Korak avtomatiziranega

odkrivanja procesov je ključnega pomena za doseganje

preglednosti procesa, kar je predpogoj za izdelavo načrta za

izboljšanje procesa.

Slika 6 prikazuje z rudarjenjem procesov odkrit model

procesa bolnišničnega zaračunavanja, ki je generiran na osnovi

100.000 primerkov izvedbe procesa.



Slika 8: Grafi uspešnosti procesa bolnišničnega



zaračunavanja



Z rudarjenjem uspešnosti lahko pridobimo odgovore na

vprašanja kot so:



- Kje v procesu so ozka grla (angl. bottleneck)? (1)

aktivnost je počasna zato predstavlja ozko grlo; (2) vse

vhodne povezave v aktivnost so počasne, zato je

najverjetneje ozko grlo vir, ki mu je dodeljena izvedba

aktivnosti; in (3) počasna je predaja dela (angl. handoff)

Slika 6: Model algoritmično pridobljenega procesa med dvema viroma.

- Kateri viri (zaposleni) so v procesu preobremenjeni in

Iz uporabniškega vmesnika in diagrama (Slika 6) so pod-obremenjeni? V kolikor dnevnik dogodkov beleži tudi

neposredno razvidni določeni statistični podatki o izvedbi vire, ki so bili zadolženi za izvedbo aktivnosti, lahko

procesa kot so: število variant izvedbe procesa (v danem primeru rudarjenje procesov generira mrežo odvisnosti med viri.

1000 variant), število različnih aktivnosti (v našem primeru 18), - Kje v procesu se izvajajo ponovitve opravil? Rudarjenje

minimalni, povprečni in najdaljši čas izvedbe primerka procesa, procesov lahko prepozna ponovno izvajanje opravila, parov

število ponovitev posameznih aktivnosti in kvantifikacija relacij opravil ali širših zank, ki potencialno predstavljajo

neposrednih naslednikov (tj. dveh zaporednih aktivnosti). redundantno ali jalovo delo.

Zaradi kompleksnosti realnih procesov (Slika 7), predvsem

iz vidika števila variacij izvedb enega procesa (varianta 3.3 Preverjanje skladnosti

predstavlja vse sledi procesa z istim zaporedjem aktivnosti), so Preverjanje skladnosti (angl. conformance checking) omogoča

običajno tudi procesne mape kompleksne, zato orodja običajno primerjavo izvajanega modela (oziroma dnevnika dogodkov) z

nudijo možnosti abstrakcije pridobljenih modelov in sicer se definiranimi poslovnimi pravili ali definiranim modelom procesa

najpogosteje omejuje prikaz glede na pogostost izvedbe (angl. prescribed process model) (Slika 5) in je s tem relevantno

aktivnosti ali povezav (na primer prikaz najpogosteje ali za usklajevanje poslovanja z zahtevami in za presojanje

najredkeje izvedenih variant procesa). poslovanja (angl. auditing). Primeri poslovnih pravil, ki se lahko

preverjajo so:





- omejitve kontrolnega toka, kot je analiza izvajanja



obveznih aktivnosti (na primer: odobritve zahtevkov ali

obvezna kontrola kakovosti);

- omejitve nivoja storitev oziroma SLA (angl. service level

agrement), kot je najdaljši dovoljen čas izvajanja aktivnosti,

pod-procesa ali procesa;

- omejitve virov kot je »ločevanje dolžnosti« (na primer: ista

oseba ne sme izvesti dveh zaporednih aktivnosti) in

- identifikacija redkih primerkov izvedbe, ki so

Slika 7: Dejanska kompleksnost realnega procesa

potencialno neskladni s poslovnimi pravili.



3.2 Rudarjenje uspešnosti procesov pravil ali definiranega modela procesa. Slika 9 prikazuje število Rezultat analize preverjanja skladnosti je seznam odstopanj od

Z rudarjenjem uspešnosti procesov (angl. performance mining) izvedb primerkov procesa glede na njihovo trajanje. Označeni so

pridobimo dodatne informacije o modelih procesov, ki lahko vsi primerki procesa bolnišničnega zaračunavanja, ki se izvedejo

vodijo v njihove izboljšave (Slika 5). Rezultat rudarjenja v manj kot 30 dneh (predpostavimo, da je to opredeljeno v SLA).

uspešnosti so grafi uspešnosti (Slika 8) in modeli procesov, ki so Primerki, ki ustrezajo SLA oziroma primerki, ki temu ne

dopolnjeni z informacijami kot so trajanja aktivnosti ali pogledi ustrezajo, se lahko nato še podrobneje analizirano kot je

na procese iz vidika določenega vira. predstavljeno v naslednjem podpoglavju.





Slika 11: Struktura počasnih primerkov procesa

Slika 9: Filtriran prikaz modela procesa bolnišničnega

zaračunavanja

4 ZAKLJUČEK

3.4 Analiza različic procesov

Digitalna preobrazba in optimizacija poslovanja je prepletena s

Analiza različic (angl. variant analysis) temelji na primerjavi tehnološkimi inovacijami, ki morajo zagotavljati hiter in

dveh ali več različic dnevnikov dogodkov istega procesa (Slika veljaven vpogled v delovanje organizacij, temelječ na realnih

5), ki tako predstavljajo različne variante procesa (na primer, podatkih, na način, ki je razumljiv vsem vpletenim. Med ključne

primerjava vseh primerkov procesa, ki so se uspešno zaključili tehnike za doseganje navedenega spada rudarjenje procesov, ki

glede na neuspešne). Primerjava variant procesov poda vpogled izkorišča močno povezanost med fizično in informacijsko

na vprašanja tipa »zakaj?«, na primer: realnostjo (digitalni dvojček) poslovanja, v kateri se poslovni

- zakaj se določeni primerki procesa (ki smo jih združili v eno dogodki beležijo v realnem času, le ti pa so uporabljeni za

izmed variant procesa) izvajajo hitreje kot drugi? usmerjanje, prilagajanje in nadzor poslovnih procesov.

- Zakaj se določeni primerki procesa uspešno zaključijo, Rudarjenje procesov postaja del rutine večjih podjetij razvitih

medtem ko se drugi neuspešno? držav, k čemu je pripomogel tudi širok nabor »enterprise-ready«

- Zakaj je vir, ki je vključen v izbrano varianto procesa manj orodij za rudarjenje procesov. Le ta so dostopna različnim vrstam

učinkovit kot drugi? in potrebam organizacij in omogočajo relativno nizek vstopni

Odgovore na zgornja vprašanja lahko pridobimo z enim prag v aktivnosti rudarjenja procesov. Slabost vpeljave orodij je

izmed naslednji pristopov analize različic procesov: (1) analizo pogosto, da so le ta v podjetjih implementirana v omejenem

metrik uspešnosti (angl. performance measures) različic obsegu in zato ne pokrivajo celotnega poslovanja [4].

procesov; (2) analizo različic procesov glede na čas izvedbe (na Poglavitna izziva širše vpeljave rudarjenja procesov ostajata

primer: analiza in primerjava različic procesov, ki so se izvajali kakovost podatkov in človeški dejavniki. Izkušnje kažejo, da je

pred in med pandemijo); (3) analiza različic procesov glede na okoli 80% časa potrebnega za lociranje, izbiranje, pridobivanje

lastnosti primerkov (na primer: vrsta izdelka, segment kupcev, in transformacijo podatkov, pogosto pa omenjene aktivnosti

geografsko področje, ipd.). odkrijejo tudi težave s kakovostjo podatkov, ki jih je potrebno

Na naslednjih dveh slikah (Slika 10 in Slika 11) je prikaz odpraviti neodvisno od rudarjenja procesov. Človeški dejavniki

strukture dela procesa bolnišničnega zaračunavanja, katerih so pogosto povezani z nepoznavanjem področja rudarjenja

primerki se izvedejo v manj kot 30 dnevih (31% vseh primerkov procesov in »strahom« pred odkritjem dejanskih procesov, ki bi

procesa) in strukture dela procesa bolnišničnega zaračunavanja, lahko izpostavili pomanjkljivo vodenje, neučinkovitosti ali

katerih primerki trajajo več kot leto dni (11% vseh primerkov neskladnosti med operativnih delovanjem in predpisi [4].

procesa).





5 LITERATURA IN VIRI



[1] M. Dumas, M. La Rosa, J. Mendling, and H. A. Reijers,

Fundamentals of Business Process Management. Berlin,

Heidelberg, Germany: Springer, 2018. doi: 10.1007/978-3-

662-56509-4.

[2] M. Kerremans, S. Searle, T. Srivastava, and K. Iijima,

“Market Guide for Process Mining,” Gartner, Sep. 2020.

Accessed: Sep. 08, 2021. [Online]. Available:

https://www.gartner.com/doc/reprints?id=1-

24ARMY34&ct=201002&st=sb

[3] E. De Roock and N. Martin, “Process mining in healthcare

– An updated perspective on the state of the art,” Journal of

Slika 10: Struktura hitrih primerkov procesa

Biomedical Informatics, vol. 127, p. 103995, Mar. 2022,

doi: 10.1016/j.jbi.2022.103995.

Iz primerjave slik je razvidno, da je v primeru počasnih izvedb [4] G. Polančič and M. Kocbek Bule, “Stanje in trendi na

procesa (Slika 11) veliko kratkih zank, ki predstavljajo področju rudarjenja procesov,” UI, vol. 30, no. 1, May

dolgotrajne, tudi več mesecev trajajoče ponovitve izvedbe 2022, doi: 10.31449/upinf.162.

opravil.



Ocena digitalnih kompetenc medicinskih sester v klinični



praksi: opisna raziskava



Assessing digital competencies of nurses in clinical



practice: a descriptive study



Anton Grmšek Svetlin Melita Peršolja

Katedra za zdravstveno nego, Fakulteta za vede o Katedra za zdravstveno nego, Fakulteta za vede o

zdravju/Univerza na Primorskem zdravju/Univerza na Primorskem

Izola, Slovenia Izola, Slovenia

97230368@student.upr.si melita.persolja@fvz.upr.si



Povzetek Abstract

Digitalizacija prinaša v zdravstveno nego novosti in inovativne The process of digitalization introduces innovations and novel

rešitve za medicinske sestre. Digitalne kompetence in uporaba solutions into healthcare, particularly benefiting nursing practice.

znanj s področja digitalne tehnologije predstavljajo ključen vidik Digital competencies and the application of knowledge in digital

dela medicinskih sester. Namen raziskave je bil opisati digitalno technologies are critical aspects of nurses' work. The aim of this

pismenost pri medicinskih sestrah. Uporabljena je bila study was to describe digital literacy among nurses. A

kvantitativna deskriptivna in eksplorativna neeksperimentalna quantitative, descriptive, and exploratory non-experimental

metoda empiričnega raziskovanja. Anketiranje je potekalo na empirical research method was employed. The survey was

vzorcu 51 medicinskih sester (36 žensk in 15 moških). conducted on a sample of 51 nurses (36 female and 15 male). The

Vprašalnik je bil sestavljen iz 25 trditev, povezanih z znanjem, questionnaire consisted of 25 statements related to knowledge,

sposobnostmi in odnosom do uporabe digitalnih tehnologij. skills, and attitudes towards the use of digital technologies. The

Upoštevana stopnja statistične značilnosti (p) je znašala ≤ 0,05. significance level (p) was set at ≤ 0.05. In addition to descriptive

Poleg deskriptivne statistike, smo uporabili Wilcoxonov test, statistics, we used the Wilcoxon test, Mann-Whitney U test,

Mann-Whitneyev U-test, Kruskal-Wallisov H-test, Hi-kvadrat Kruskal-Wallis H test, Chi-square test, and linear regression

test ter linearno regresijsko analizo. Medicinske sestre izkazujejo analysis. Nurses demonstrated a high level of knowledge and

visoko znanje in sposobnosti s področja uporabe digitalnih skills in the use of digital technologies and exhibited a positive

tehnologij in imajo do njih pozitiven odnos (Me = 122, p < 0,001). attitude towards them (Me = 122, p < 0.001). It was found that

Ugotovljeno je bilo, da obstajajo razlike med spoloma in sicer so gender differences exist, with women achieving higher scores (x̄

ženske dosegle višji rezultat (x̄ = 113,73, s = 15,80). Statistično = 113.73, s = 15.80). No statistically significant differences were

pomembnih razlik med različnimi ravnmi zdravstvenega varstva identified across different levels of healthcare. Most nurses are

nismo ugotovili. Medicinske sestre so v večini digitalno digitally competent, possessing both the knowledge and skills

kompetentne, kar pomeni, da imajo znanje in sposobnosti s required for the use of digital technologies and expressing a

področja uporabe digitalnih tehnologij in obenem izražajo positive attitude towards their application. It is essential for

pozitiven odnos do njihove uporabe. Pomembno je, da svoja nurses to continue enhancing and deepening their knowledge in

znanja s področja digitalnih tehnologij nadgrajujejo in digital technologies, as this contributes to the improvement of

poglabljajo, saj tako prispevajo k izboljšanju delovnih procesov work processes and better patient care. in boljši oskrbi pacientov.

Keywords

Ključne besede Digitalization, informatics, nursing care, skills

Digitalizacija, informatika, sposobnosti, zdravstvena nega



Permission to make digital or hard copies of part or all of this work for personal or

classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).



Prevod in validacija vprašalnika SHAIP za ocenjevanje



stališč zaposlenih v zdravstveni negi o umetni inteligenci*



Translation and validation of the SHAIP questionnaire for



assessment of nursing staff's perceptions of artificial



intelligence



Deja Praprotnik † Boštjan Žvanut

Dom upokojencev Idrija Univerza na Primorskem

Arkova ulica 4, 5280 Idrija Fakulteta za vede o zdravju

Slovenija Polje 42, 6310 Izola, Slovenija

deja.praprotnik@gmail.com bostjan.zvanut@fvz.upr.si



analysing internal consistency (Cronbach's α = 0.76) and the test-

Povzetek retest method (conducted on a sample of 48 respondents).

Validity was assessed using content validity (S-CVI/Ave = 0.91),



(SHAIP) je namenjen merjenju stališč zdravstvenih delavcev o concurrent validity and structural validity. Criterion validity was Vprašalnik »Shinners Artificial Intelligence Perception«

umetni inteligenci. Cilj študije je bil prevesti vprašalnik v assessed by using the construct attitudes toward use of artificial

slovenski jezik in ga validirati. Vprašalnik je izpolnilo 270 intelligence technology (rxy = 0.657, p < 0.001). The test-retest

zaposlenih v zdravstveni negi. Zanesljivost smo ocenili z analysis showed no statistically significant differences in the

analizo notranje skladnosti (Cronbachov α = 0,76) in metodo mean ranks between the test and retest phases, except for the last

ponovljenega testiranja (test-retest) na vzorcu 48 anketirancev. two items of the questionnaire. These items were later excluded

Veljavnost smo ocenili z analizo vsebinske (S-CVI/Ave = from the translated version following the analysis of structural



odgovorov test-retest je pokazala, da pri vseh, razen zadnjih umetne inteligence (rxy = 0,657, p < 0,001). Analiza constructs. The results suggest that the Slovenian version of the SHAIP questionnaire with the considered differences, is a dveh elementih vprašalnika, ni bilo statistično značilnih razlik v reliable and valid instrument for assessing perceptions of nursing povprečnih vrednostih rangov odgovorov med obema fazama staff on artificial intelligence. smo preverili s pomočjo konstrukta odnos do uporabe constructs, while the original instrument is based on two tehnologije 0,91), sočasne in strukturne veljavnosti. Kriterijsko veljavnost validity results. The final translated instrument consists of three

izpolnjevanja. Omenjena elementa vprašalnika sta bila po

analizi rezultatov strukturne veljavnosti izključena iz končnega

prevoda. Slovenska različica vprašalnika vsebuje tri konstrukte, Keywords

za razliko od izvirnega instrumenta, ki predvideva zgolj dva. Artificial intelligence, nursing, perceptions, validity, reliability

Rezultati nakazujejo, da je slovenska različica vprašalnika

SHAIP, ob upoštevanju razlik, zanesljiv in veljaven instrument

za ugotavljanje stališč zaposlenih v zdravstveni negi o umetni

inteligence.



Ključne besede

Umetna inteligenca, zdravstvena nega, stališča, veljavnost,

zanesljivost



Abstract

The Shinners Artificial Intelligence Perception (SHAIP)

questionnaire was developed to measure the attitudes of

healthcare professionals towards artificial intelligence. The aim

of the study was to translate the questionnaire in Slovenian

language and validate it. The questionnaire was completed by

270 employees in nursing care. Reliability was assessed by

∗Article Title Footnote needs to be captured as Title Note

†Author Footnote to be captured as Author Note



Permission to make digital or hard copies of part or all of this work for personal or

classroom use is granted without fee provided that copies are not made or distributed

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s).

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s).





Pridobivanje mikrodokazil z uporabo simulacij v




zdravstveni negi



Obtaining micro-credentials through simulation in nursing



Jakob Renko Patrik Pucer Igor Karnjuš

Fakulteta za vede o zdravju, Fakulteta za vede o zdravju, Fakulteta za vede o zdravju,

Univerza na Primorskem Univerza na Primorskem Univerza na Primorskem



Slovenija Slovenija Slovenija

jakob.renko@fvz.upr.si patrik.pucer@fvz.upr.si igor.karnjus@fvz.upr.si



Povzetek nursing education, it would make sense to use a simulation-based method of teaching for the acquisition of micro-credentials.

Uvod: Mikrodokazila predstavljajo nadgradnjo vseživljenjskega Purpose: To examine the introduction of micro evidence in

izobraževanja s katero bi tudi zdravstvena nega lahko zapolnila nursing through a simulation-based method of education.

vrzeli v okolju, kjer se potrebe po delovni sili hitro spreminjajo. Methods: An empirical qualitative research methodology was

Za reševanje potrebe po kontinuiranem vseživljenjskem razvoju used, which involved conducting semi-structured interviews and

in hkrati spremembe paradigme zdravstvenega izobraževanja, bi focus groups with nursing professionals. Data were analysed



pridobivanju mikrodokazil. Results: Ways in which simulations could contribute to the Namen : Preučiti vpeljevanje introduction of micro-evidence are also presented. Conclusion: mikrodokazil na področju zdravstvene nege z uporabo se bilo smiselno posluževati simulacijske metode poučevanja pri using content analysis to identify key themes and sub-themes.



simulacijske metode izobraževanja. Metode: Uporabljena je bila improve the competences of nurses and contribute to better The implementation of micro evidence and simulations could

empirična kvalitativna raziskovalna metodologija, ki je accessibility and standardisation of educational content. vključevala izvedbo pol-strukturiranih intervjujev in fokusnih

skupin s strokovnjaki na področju zdravstvene nege. Podatki so Keywords bili analizirani z uporabo vsebinske analize, da bi identificirali



ključne teme in podteme. Rezultati: Identificirana je tema education system. Micro-credentials, simulation training, lifelong learning, higher Simulacije kot metoda učenja za pridobivanje mikrodokazil in

podtemi z naslovoma Uporabnost simulacij v zdravstveni negi in

Pridobivanje mikrodokazil z uporabo simulacij. Predstavljeni so 1 UVOD tudi načini s katerimi bi simulacije doprinesle vpeljevanju

Hitrost napredka in sprememb v zdravstvu je sprožila potrebo po



izobraževanja v Sloveniji je ključna za karierni razvoj zaposlenih kontinuiranem vseživljenjskem izobraževanju in profesionalnem mikrodokazil. Zaključek: Nadgradnja sistema vseživljenjskega

v zdravstveni negi. Implementacija mikrodokazil in simulacij bi izpopolnjevanju vseh zdravstvenih profilov. Le s prilagajanjem



lahko izboljšala kompetence medicinskih sester ter prispevala k na potrebe zdravstvenih sistemov je možno nuditi varno in kakovostno zdravstveno oskrbo [1]. Nastajanje vrzeli med boljši dostopnosti in standardizaciji izobraževalnih vsebin. potrebami in formalno pridobljenimi kvalifikacijami, poskušajo



Ključne besede oblikovalci izobraževalnih politik po svetu, v Evropi in tudi

Sloveniji premostiti s pomočjo mikrodokazil [2]. Državna uprava

Mikrodokazila, simulacijsko usposabljanje, vseživljenjsko Republike Slovenije [3] definira mikrodokazila »kot zapise učnih

izobraževanje, visokošolski izobraževalni sistem. izidov, ki jih je posameznik dosegel z učenjem manjšega obsega.



Abstract Učni izidi so ovrednoteni s pomočjo Evropskega sistema prenašanja in zbiranja kreditnih točk (ECTS) na podlagi

Micro-credentials are an extension of lifelong learning that could pregledno in jasno definiranih standardov. Programi so

also help nursing to fill gaps in an environment where workforce pripravljeni tako, da opremijo posameznika s specifičnim

needs are changing rapidly. To address the need for continuous znanjem, spretnostmi in kompetencami, ki naslavljajo družbene,

lifelong development and at the same time a paradigm shift in osebne, kulturne potrebe oziroma potrebe trga dela.

Mikrodokazila so last posameznikov, ki jih lahko delijo z drugimi

∗Article Title Footnote needs to be captured as Title Note in prenašajo naprej. Lahko so samostojna ali se združujejo v

†Author Footnote to be captured as Author Note

večja. Podprta so s sistemi za zagotavljanje kakovosti in sledijo

Permission to make digital or hard copies of part or all of this work for personal or dogovorjenim standardom v sektorjih«. Mikrodokazila so na

classroom use is granted without fee provided that copies are not made or distributed globalni ravni še vedno v začetni fazi razvoja zato se soočamo z

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must veliko izzivi [4].

be honored. For all other uses, contact the owner/author(s). Izobraževanje današnjih študentov in strokovnjakov zdravstvene

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s). nege postaja vedno večji izziv za oblikovalce izobraževalnih

http://doi.org/10.70314/is.2024.sizn.7 politik, zdravstvene pedagoge, zdravstvene ustanove – učne baze

https://doi.org/10.70314/is.2024.sizn.7 https

://doi.org/10.70314/is.2024.sizn.7 in same udeležence, saj se potrebe in navade novih generacij zdravstvene nege [13]. Za namenski vzorec so bile uporabljene

korenito spreminjajo [5]. Avtorji v omenjeni raziskavi so tri demografske skupine:

dognali, da aktivno in konstruktivno učno okolje, ki temelji na • predstavniki visokošolskih učiteljev s področja

izzivih in učnih ciljih, pri njih spodbuja globlje učenje, ki zdravstvene nege, ki pri svojem delu uporabljajo

poudarja razumevanje in uporabo znanja ter se osredotoča na simulacije kot metodo učenja (6);

razumevanje namesto na pomnjenje. Takšno učno okolje je • predstavniki, ki sodelujejo pri oblikovanju politike

možno doseči tudi s simulacijsko metodo izobraževanja [6]. visokošolskega izobraževanja (4);

Simulacijska metoda izobraževanja vključuje uporabo različnih • predstavniki kliničnih mentorjev zdravstvene nege

modalnosti kot so simulatorji delnih nalog, simulatorji visoke oz. strokovnjaki iz prakse (5).

zvestobe, standardizirani pacienti, računalniško podprte Predstavniki visokošolskih učiteljev so zastopali Univerzo na

simulacije, virtualna resničnost in hibridne simulacije [7]. Primorskem, Univerzo v Mariboru in Univerzo v Ljubljani, s

Uporaba simulacijskih aktivnosti ima v povezavi z mikrodokazili čimer smo zajeli vse javne univerze v slovenskem prostoru.

dve prednosti. Ta pristop predstavlja inovativen način, s katerim Predstavniki kliničnih mentorjev so zastopali primarno,

lahko izobraževalne ustanove in izvajalci zdravstvenih storitev sekundarno in terciarno raven zdravstvenega varstva, s čimer

pridobijo kompetence potrebne za strokovni razvoj, izvajanje in smo si zagotovili celovit prispevek z vseh ravni. Predstavniki, ki

vrednotenje učinkovitih strategij poučevanja ter oblikovanje sodelujejo pri oblikovanju politike visokošolskega

učnih načrtov v spletnem in praktičnem okolju. Prav tako izobraževanja, so zastopali krovni organ za zdravstveno nego v

mikrodokazila podpirajo merjenje in nadgrajevanje kompetenc s slovenskem prostoru, in sicer Zbornico zdravstvene in babiške

pomočjo simulaciji, za doseganje optimalnih rezultatov nege Slovenije – Zvezo strokovnih društev medicinskih sester,

posameznika [8]. Mikrodokazila so pogosto tesno povezana tudi babic in zdravstvenih tehnikov Slovenije, Strateški svet za

z digitalnimi značkami. To so virtualni in digitalni simboli, ki zdravstvo Republike Slovenije ali pa so zasedali visoke položaje

predstavljajo dosežke, spretnosti ali znanja, pridobljene v na univerzitetni ravni.

različnih okoljih, kot so spletni tečaji ali druge vrste

izobraževanj. Pogosto se uporabljajo za spodbujanje učenja in 3.2 Raziskovalni instrument

priznavanje dosežkov v izobraževalnih platformah [9–11]. Pri izvedbi intervjujev in fokusnih skupin smo si pomagali z

Vključevanje simulacij v študijske programe zdravstvenih strok delno strukturiranim vodičem, ki je bil sestavljen iz treh sklopov

je pri študentih izboljšala pridobivanje znanja in veščin ter vprašanj. Pri oblikovanju vprašalnika smo si pomagali s

okrepila kakovost in varnost klinične prakse [12]. podobnim tipom raziskave, ki so jo izvedli Lok in sod. [1], in

Za reševanje potrebe po kontinuiranem vseživljenjskem razvoju pregledom relevantne literature.

in hkrati spremembe paradigme zdravstvenega izobraževanja, bi Sklopi delno strukturiranega vodiča so obsegali:

se bilo smiselno posluževati simulacijske metode poučevanja pri • sklop 1: Demografski podatki;

pridobivanju mikrodokazil [1, 8, 12]. • sklop 2: Stališča do mikrodokazil;

• sklop 3: Združevanje mikrodokazil, simulacijske



2 metode poučevanja in zdravstvene nege. NAMEN IN RAZISKOVALNO Vprašalnik je vseboval 6 vprašanj in kratko definicijo VPRAŠANJE obravnavanega področja, ker se je to izkazalo za potrebno zaradi

Namen raziskave je preučiti vpeljevanje mikrodokazil na specifike tematike. področju zdravstvene nege z uporabo simulacijske metode

izobraževanja. 3.3 Potek raziskave



Raziskovalno vprašanje: Kakšne so možnosti pridobivanja Prostovoljce za sodelovanje v raziskavi smo kontaktirali in mikrodokazil s pomočjo simulacijske metode izobraževanja na angažirali preko elektronske pošte. Uporabili smo elektronske področju zdravstvene nege? naslove dostopne preko spleta. Pridobivanje podatkov smo

izvedli preko platforme ZOOM ali v živo v mesecih septembru,



3 oktobru in novembru v letu 2023. Intervjuji so se snemali znotraj METODE omenjene platforme oz. s telefonom. Zvočni posnetki bodo



3.1 uničeni tri mesece po objavi raziskave. Od udeležencev smo pred Vzorec



deskriptivno interpretativni dizajn. Podatke smo zbrali s pomočjo sodelovanja, pravica do odstopa in nekaj osnovnih informacij dveh fokusnih skupin in 10 individualnih intervjujev. V prvi glede preučevane tematike. Sodelujoči so lahko raziskavo fokusni skupini sta sodelovala dva udeleženca, v drugi skupini zapustili med celotnim procesom izvajanja raziskave brez Uporabili smo empirično kvalitativno metodo dela in raziskavi, kjer so pojasnjeni namen raziskave, pravila izvedbo pridobili informirano soglasje za sodelovanje v

pa trije. Izvajanje fokusnih skupin se je izkazalo za pomanjkljivo obrazložitve. Na vprašanja jim ni bilo treba odgovarjati, če si

in neučinkovito zaradi specifike področja in težav z tega niso želeli. Raziskava je potekala v prostem času vseh



raziskave smo se odločili za pridobivanje informacij z Pridobljeni podatki so bili obdelani s pomočjo programa za individualnimi intervjuji. S procesom intervjuvanja smo kvalitativno analizo besedil NVivo. usklajevanjem prostih terminov udeležencev. V nadaljevanju sodelujočih posameznikov in je bila popolnoma anonimna.



prenehali, ko smo zaznali zasičenost podatkov. Skupno je v Uporabljena je bila fenomenološka metodologija, ki omogoča raziskavi sodelovalo 15 prostovoljcev, od tega 11 žensk in štirje kredibilnost in zanesljivost podatkov [14]. Po opravljenih moški. Razmerje v spolu se sklada z zastopanostjo na področju fokusnih skupinah in intervjujih je sledila transkripcija



posnetkov in slovnično popravljanje zapisov, pri čemer je pomen »Zelo so me navdušile, ker gre za poseben pedagoški pristop, ki

vsebine ostal nespremenjen. Za obdelavo podatkov smo ga lahko uporabljaš v kakršnihkoli razmerah, bodisi na cesti ali

uporabili proces po Colaizzijevi v sedmih stopnjah [15]: pa v visokotehnoloških okoljih. Predvsem je ta pristop tisti, v

1. seznanitev s transkripti: branje in večkratno ponovno katerem vidim prednost.« (VU5)

branje transkriptov; »Simulacije poznam in jih zelo podpiram, ker se mi zdi bistveno,

2. identifikacija pomembnih izjav: izvleček da določene stvari najprej poizkusimo na neživi stvari, preden se

pomembnih izjav; preselimo na realne paciente. To je dolžnost vsakega, da mora

3. formulacija pomenov: združevanje izjav v širše pred invazivnimi posegi najprej opraviti določeno število ur v

pomensko povezane skupine; simulacijskem centru. Vsekakor se mi zdi zelo dobrodošlo. Tudi

4. ustvarjanje tem: združevanje v smiselni sistem tem študenti bi se morali tega zavedati, da se najprej 100-odstotno

in podtem; naučiš na simulatorju in nato preideš na pacienta.« (KMZN5)

5. razvoj izčrpnega opisa: izpostavitev bistva nekega »Mislim, da je velika večina zaposlenih v zdravstvu tako

pojava in integracija vseh ugotovitev; naravnana, da so simulacije v sožitju s podajanjem teoretičnega

6. izdelava temeljne strukture: odstranitev vseh znanja tista metoda, s katero si bodo ljudje bistveno bolj

neustreznih vsebin in opisov in razkritje temeljne zapomnili pridobljeno znanje. Teorija pa vsekakor mora biti

strukture pojava; prisotna. Študenti si lažje predstavljajo, kako teče določen

7. verifikacija temeljne strukture: pregled in potrditev postopek in ga tudi sami izvajajo. Mislim, da so simulacije tisti

rezultatov s strani udeležencev z namenom aspekt, ki bi ga morali bistveno bolj vpeljati v prakso, pa nam ne

pridobivanja povratne informacije o pravilni gre najbolje.« (PP2)

interpretaciji.



4.2 Pridobivanje mikrodokazil z uporabo

4 REZULTATI simulacij

V sklopu raziskave smo prepoznali glavno temo, in sicer Iz intervjujev in fokusnih skupin smo razbrali, da so udeleženci

Simulacije kot metoda učenja za pridobivanje mikrodokazil. v raziskavi naklonjeni pridobivanju mikrodokazil s pomočjo

Poleg tem smo izpostavili tudi dve podtemi, ki sta razvrščeni k simulacijske metode učenja.

temi, kot je prikazano v Tabeli 1. »Jaz ta dva koncepta vidim kot zelo združljiva, zakaj pa ne. To je

seveda izvedljivo le s pravilnim načrtovanjem. Treba je razvijati

Tabela 1: Tema in podtemi vpeljevanja mikrodokazil v celovite vsebine in sistematične delavnice. Ne vidim, zakaj se ne

zdravstveno nego s pomočjo simulacijske metode učenja bi teh znanj pridobivalo s pomočjo simulacij. Vidim kvečjemu

veliko plusov.« (VU6)

»[Simulacije] so zelo pomemben način za učenje praktičnih

Tema Podtemi

veščin, manj dolgočasna metoda. Ljudje jo veliko hitreje

Simulacije kot Uporabnost simulacij v zdravstveni negi

sprejmejo kot neko klasično obliko učenja in poučevanja. Je tak

metoda učenja za Pridobivanje mikrodokazil z uporabo

način zelo dober zato, da se vzdržuje kondicija. Tudi za nekoga,

pridobivanje simulacij

ki mogoče ni več 100-odstoten v praksi.« (PP1)

mikrodokazil

»Jaz sem zelo nagnjena k pridobivanju mikrodokazil preko



katere demografske skupine in od katerega posameznika citat pridobiš s simulacijami, je to kot učenje vožnje z avtom. Lahko izhaja: poznaš vse predpise in pravila, avta pa ne spraviš s parkirišča. • VU 1 – 6 (predstavniki visokošolskih učiteljev s področja Zagotovo to podpiram in mislim, da bi se kakovost obravnave ne Citati so označeni s kraticami in številkami, ki ponazarjajo, iz zelo pomembna, toda če nimaš različnih spretnosti, ki jih simulacijske metode, ker ni vse samo v teoriji. Seveda je [teorija]



zdravstvene nege, ki pri svojem delu uporabljajo simulacije kot glede na področje dvignila.« (KMZN2) metodo učenja); Poleg izrazite podpore pa so bili izraženi tudi pomisleki glede • PP 1 – 4 (predstavniki, ki sodelujejo pri oblikovanju politike dejstva, da bi se mikrodokazila pridobivala izključno s pomočjo visokošolskega izobraževanja); simulacijske metode učenja. Sodelujoči so bili mnenja, da se • KMZN 1 – 5 (predstavniki kliničnih mentorjev zdravstvene simulacije lahko uporabljajo kot metoda preverjanja znanja in da nege oz. strokovnjaki iz prakse). je ob simulacijah še vedno prisotna potreba po kliničnem

praktičnem usposabljanju in podajanju teoretičnega znanja preko



4.1 Uporabnost simulacij v zdravstveni negi » predavanj.

Simulacijski scenariji so lahko tudi tehnika preverjanja znanja.

Simulacije kot metoda učenja so se izkazale za zaželen pristop Veliko naših izobraževanj je zastavljenih tako, da se tega niti ne

za pridobivanje novega znanja. Intervjuvanci so jih identificirali izvaja. Pri simulacijski metodi pa preverjaš tako teoretično

za najprimernejši način za povezavo teoretičnega znanja s znanje kot tudi praktične veščine in na koncu oceniš celoto. To je

praktičnim. odlično, sploh če želimo podeliti ECTS točke, kjer potrebujemo

»Odlična metoda za implementacijo teorije v prakso. Veliko se o tudi neko metodo preverjanja znanja.« (VU3)

tem govori. Slovenija se v tem počasi prebuja, tujina ima že leta »Pridobivanje mikrodokazil, ki se nanašajo na praktične veščine

dodelane programe, kjer so simulacije vpete v izobraževalni v zdravstveni negi, se nikakor ne more dogajati izključno v

program.« (KMZN4)



simuliranem okolju. Del bi moral biti izveden tudi v kliničnem spretnosti in utrjujejo znanje kot tudi spodbujajo kritično

okolju.« (PP3) razmišljanje [19]. Zato je pri novih generacijah ključnega

»Mikrodokazil samo preko simulacijske metode jaz osebno ne bi pomena, da se opredeli kombinacija metod, ki presega okvir

uvajala. Za samo pridobitev dokazila je potrebno izobraževanje običajnih učnih pristopov in se vključi tudi uporabo bolj

intervencij v kliničnem okolju. Brez dela na realnih pacientih se inovativnih in prožnih metod učenja kot so simulacije [19, 20].

mi izdajanje certifikatov o znanju ne zdi legitimno.« (KMZN5) Rezultati, ki smo jih pridobili lahko služijo kot vodilo pri

»Jaz vedno vidim situacijo tako, da vedno potrebuješ neko vzpostavljanju sistema pridobivanja mikrodokazil v zdravstveni

teoretično ozadje, potem pa sledi praksa. Teoretično naj bi bile negi.

samo osnovne stvari povedane, s prakso pa utrjuješ in povezuješ Izvedena raziskava ima nekaj omejitev. V vzorec intervjuvancev

znanje.« (KMZN1) bi bilo smotrno vključiti tudi predstavnike akreditacijskih

»Želim si, da bi bil nek manjši, krajši del namenjen teoriji. Ne v organov in predstavnike državnih sistemov, ki so zadolženi za

smislu, da izvajalec pride in predava, ampak da se udeleženci na uvajanje mikrodokazil v slovenski prostor. Prav tako je omejitev

usposabljanje doma pripravijo, na lokaciji pa se snov samo relativna novost področja mikrodokazil in slabo poznavanje

ponovi.« (VU2) omenjene tematike med zaposlenimi v kliničnih okoljih. Ker je

prenova vseživljenjskega učenja v visokošolskih sistemih še v

zelo zgodnji fazi, je bilo iskanje kandidatov s primernim znanjem

5 RAZPRAVA na tem področju zahtevno.

Rezultati naše raziskave so nam omogočili vpogled v vpeljevanje V nadaljnjih raziskavah na obravnavanem področju bi se bilo

mikrodokazil na področju zdravstvene nege s pomočjo zanimivo usmeriti v študije primerov vpeljevanja mikrodokazil.

simulacijske metode izobraževanja. S pomočjo intervjujev in Smiselno bi bilo izvajati kvantitativne raziskave med

fokusnih skupin smo podrobneje raziskali možnosti vpeljave zdravstvenim osebjem glede želenih področij izobraževanja in

sistema mikrodokazil s pomočjo simulacij kot metode pridobivanja mikrodokazil. Menimo, da je treba nadaljevati

izobraževanja. kvalitativno raziskovanje tega področja, pri čemer bi informacije

Vsi intervjuvanci so izrazili podporo združevanju pridobivanja zbirali preko vseh deležnikov ekosistema mikrodokazil, ki ga

mikrodokazil s simulacijsko metodo učenja. Pomanjkanje skušamo vzpostaviti. Digitalne značke v samostojni obliki lahko

sorodnih raziskav, ki so se ukvarjale s to tematiko, je izrazito. Če tudi predstavljajo primerno tematiko za raziskovanje na področju

primerjamo sestavne dele kvalitetnega simulacijskega zdravstvene nege. Prav tako smo mnenja, da bi bila boljša

usposabljanja, ki jih določajo Standardi najboljših praks za komunikacija med deležniki ključnega pomena za kakovosten

simulacijo v zdravstveni negi [16], in sestavne dele zagon nadgradnje vseživljenjskega učenja. Glede na to, da je

mikrodokazil, ki jih izpostavljata Bideau in Kearns [17], lahko vizija Evropske unije [21] vpeljevanje mikrodokazil na vseh

vidimo, da sta si načrta podobna in posledično ne moremo podati delovnih področjih, menimo, da je prisotno občutno

smiselnih argumentov za večje težave pri soobstajanju. Oba pomanjkanje raziskav, ki bi se osredotočale na izzive in prednosti

načrta potrebujeta jasno določene cilje in rezultate, ki jih bodo specifičnih področij.

udeleženci dosegali, za oba je potrebna predpriprava na vsebine,

ki so jasno določene, v obeh primerih gre za specifično področje

pridobivanja kompetenc, oba načrta spodbujata visoko stopnjo 6 ZAKLJUČEK

lastne angažiranosti udeležencev, oba načrta podpirata uporabo Nadgradnja sistema vseživljenjskega izobraževanja v Sloveniji

evalvacije, oba načrta sta jasno strukturirana in ju lahko hitro je izredno pomembna. Zaposleni v zdravstveni negi se pogosto

prilagajamo potrebam učečih se in delovnega okolja. Ugotovili znajdejo v situacijah, kjer je karierni razvoj otežen in pravi

smo, da bi se simulacije lahko uporabljale tudi kot metoda napredek z dodatnimi kompetencami skoraj nemogoč.

preverjanja znanja. Prav tako je še vedno prisotna potreba po Mikrodokazila predstavljajo novost na področju nadgrajevanja in

kliničnem praktičnem usposabljanju in podajanju teoretičnega posodabljanja znanja ter priznavanja spretnosti, ki bi zaposlenim

znanja na druge že ustaljene načine, kot so predavanja. Druge na vseh ravneh zdravstva koristila. Simulacije pri tem

raziskave podpirajo simulacije kot metodo za dokazovanje predstavljajo metodo učenja, s katero je mogoče izvajati celo

znanja in kompetenc pri profesionalnem razvoju, če jih je vrsto izobraževanj v kontroliranem okolju, pri čemer pacienti

mogoče uskladiti s kompetencami in cilji izobraževalnih niso ogroženi. Raziskali smo vpeljevanje mikrodokazil na

programov. To predstavlja pomembne priložnosti za ocenjevanje področje zdravstvene nege s pomočjo simulacij, kjer smo

spretnosti, ki ne vključujejo zgolj praktičnih veščin [18]. ugotovili, da je nadgradnja vseživljenjskega učenja z

Peisachovich in sod. [8] navajajo, da ima uporaba simulacij kot mikrodokazili zaželena ter prinaša mnogo prednosti, kot so

metode za evalvacijo in dopolnjevanje vsebin mikrodokazil boljša dostopnost izobraževanj, izboljšanje samih kompetenc

nekaj prednosti. Navajajo, da mikrodokazila lahko podpirajo strokovnjakov v zdravstveni negi, nadgradnja vseživljenjskega

merjenje in nadgrajevanje kompetenc preko simulacijskih učenja, osredotočenost na učečega in standardizacija

aktivnosti, preko česar spodbujajo tudi optimalne učne izide. izobraževalnih vsebin. Dognali smo tudi, da se intervjuvancem

Prav tako navajajo, da združevanje mikrodokazil s simulacijami zdi izvajanje mikrodokazil s pomočjo simulacij smiselno in

omogoča inovativen način pridobivanja kompetenc za razvoj, pogosto tudi nujno v kontekstu pridobivanja novih kliničnih

izvajanje in oceno učinkovitih poučevalnih strategij, oblikovanje veščin ter spretnosti. Ekosistem mikrodokazil v slovenskem

učnih načrtov in dokazovanje znanja v obliki spletnega portfolia. prostoru je še vedno v začetnih fazah razvoja. Menimo, da bi s

Različne simulacijske metode bi lahko v okviru pridobivanja kvalitetno implementacijo tega novega pristopa k

mikrodokazil dopolnjevale tradicionalne metode poučevanja, saj vseživljenjskemu učenju pripomogli k nadgradnji

slednje pri učečem izboljšujejo tako tehnične kot netehnične izobraževalnega sistema. Mnogo strokovnih področij se bo



spoprijemalo z velikimi težavami, če ne bodo temelji ekosistema

in zakonodaje postavljeni enotno in jasno. Menimo, da je v

našem prostoru mogoče z bolj kakovostnim in konstantnim

prenosom informacij med deležniki ustvariti sistem, ki bo

deloval v dobro uporabnikov. Prav tako bi bilo smiselno postaviti

enega od deležnikov v vodilno vlogo, ker bi tako proces imel

motivatorja in iniciatorja, ki bi spodbujal ter koordiniral vse

druge vključene. Glede na velikost prostora Slovenije to vsaj v

teoriji ne bi smelo predstavljati prevelikih težav.



ACKNOWLEDGMENTS / ZAHVALA

Zahvalili bi se intervjuvancem, da so si vzeli čas za sodelovanje

v raziskavi. Brez njihove udeležbe bi bilo doseganje namena

nemogoče.



REFERENCES

[1] Lok P, Beyene K, Awaisu A, et al. Microcredentials training in

pharmacy practice and education: an exploratory study of its viability and

pharmacists’ professional needs. BMC Med Educ 2022; 22: 332.

[2] Wheelahan L, Moodie G. Analysing micro-credentials in higher

education: a Bernsteinian analysis. J Curric Stud 2021; 53: 212–228.

[3] Marentič U. Mikrodokazila za vseživljenjsko učenje in zaposljivost,

https://www.gov.si/assets/ministrstva/MIZS/Dokumenti/Visoko-

solstvo/Strokovni-sveti/SVS/Zapisniki-2021/SVS_7e-seja_priloga.pdf (2022).

[4] Grant S. Building Collective Belief in Badges: Designing Trust

Networks. In: Ifenthaler D, Bellin-Mularski N, Mah D-K (eds) Foundation of

Digital Badges and Micro-Credentials. Cham: Springer International Publishing,

pp. 97–114.

[5] Shin S, Park J-H, Kim J-H. Effectiveness of patient simulation in

nursing education: meta-analysis. Nurse Educ Today 2015; 35: 176–182.

[6] Bradley P. The history of simulation in medical education and possible

future directions. Med Educ 2006; 40: 254–262.

[7] Lioce L. Healthcare Simulation Dictionary. Second. Agency for

Healthcare Research and Quality. Epub ahead of print 15 January 2020. DOI:

10.23970/simulationv2.

[8] Peisachovich EH, Dubrowski A, Da Silva C, et al. Using Simulation-

Based Methods to Support Demonstration of Competencies Required by Micro-

Credential Courses. Cureus 2021; 13: e16908.

[9] Noyes JA, Welch PM, Johnson JW, et al. A systematic review of

digital badges in health care education. Med Educ 2020; 54: 600–615.

[10] Ellis LE, Nunn SG, Avella JT. Digital Badges and Micro-credentials:

Historical Overview, Motivational Aspects, Issues, and Challenges. In: Ifenthaler

D, Bellin-Mularski N, Mah D-K (eds) Foundation of Digital Badges and Micro-

Credentials. Cham: Springer International Publishing, pp. 3–21.

[11] Garnett T, Button D. The use of digital badges by undergraduate

nursing students: A three-year study. Nurse Educ Pract 2018; 32: 1–8.

[12] Padilha JM, Machado PP, Ribeiro A, et al. Clinical Virtual Simulation

in Nursing Education: Randomized Controlled Trial. J Med Internet Res 2019; 21:

e11529.

[13] Nurse journal. Male Nurse Statistics: A Look At The Numbers,

https://nursejournal.org/articles/male-nurse-statistics/ (2023, accessed 5 March

2024).

[14] Neubauer BE, Witkop CT, Varpio L. How phenomenology can help

us learn from the experiences of others. Perspect Med Educ 2019; 8: 90–97.

[15] Wirihana L, Welch A, Williamson M, et al. Using Colaizzi’s method

of data analysis to explore the experiences of nurse academics teaching on satellite

campuses. Nurse Res 2018; 25: 30–34.

[16] Watts PI, McDermott DS, Alinier G, et al. Healthcare Simulation

Standards of Best PracticeTM Simulation Design. Clin Simul Nurs 2021; 58: 14–

21.

[17] Bideau Y-M, Kearns T. A European Approach to Micro-credentials for

Lifelong Learning and Employability. J Eur CME 2022; 11: 2147288.

[18] Dove A, Borland J, Wiley CRH, et al. The Potential of Simulation

Assessments in Professional Development. J Educ Technol Syst 2023; 51: 340–371.

[19] Nestel D, Bearman M. Theory and Simulation-Based Education:

Definitions, Worldviews and Applications. Clin Simul Nurs 2015; 11: 349–354.

[20] MacLean S, Geddes F, Kelly M, et al. Realism and Presence in

Simulation: Nursing Student Perceptions and Learning Outcomes. J Nurs Educ

2019; 58: 330–338.

[21] European Commision. A European approach to micro-credentials |

European Education Area, https://education.ec.europa.eu/education-levels/higher-

education/micro-credentials (2023, accessed 12 September 2023).



Uporaba navidezne in obogatene resničnosti



v zdravstveni negi



Using virtual and augmented reality in nursing



Mojca Peterlin Jelena Ficzko Danica Dolničar

Zdravstvena fakulteta Zdravstvena fakulteta Naravoslovnotehniška fakulteta

Univerza v Ljubljani Univerza v Ljubljani Univerza v Ljubljani

Ljubljana, Slovenija Ljubljana, Slovenija Ljubljana, Slovenija

mojca.peterlinp@gmail.com jelena.ficzko@zf.uni-lj.si danica.dolnicar@ntf.uni-lj.si



POVZETEK

S hitrim razvojem informacijske tehnologije se v zadnjem času 1 UVOD

močno povečuje uporaba oblik razširjene resničnosti, kar se Zmanjševanje razkoraka med teorijo in prakso med

odraža tudi v zdravstveni negi. V prispevku najprej izobraževalnim procesom je nujno. Uporaba tehnologij, kot je

predstavljamo povzetek stanja raziskav razvoja in uporabe navidezna resničnost, podpira teorijo situacijskega učenja, ki

navidezne resničnosti in obogatene resničnosti na tem pravi, da lahko simulacija zagotavlja alternativna okolja za

strokovnem področju, nato navajamo nekaj na tržišču dostopnih učenje na kraju samem prek različnih virtualnih kontekstov, ki

aplikacij, primernih za uporabo v izobraževanju, usposabljanju uporabnikom dajejo občutek prisotnosti ali "biti tam" in s tem

ali klinični praksi. Podrobneje sta predstavljeni dve aplikaciji, možnost učenja v verodostojnem in edinstvenem kontekstu. [1]

SimX in AccuVein. Uporaba tovrstnih aplikacij prinaša številne Tudi za usposabljanje zaposlenih v zdravstveni negi (ZN) ima

prednosti, ima pa tudi nekaj slabosti. V prihodnje je pričakovati uporaba novih tehnologij potencial, prinaša pa tudi izzive.

še več raziskav o njihovi primernosti in učinkovitosti ter razvoj

novih in/ali izboljšanih aplikacij za vse več področij zdravstvene 1.1 Navidezna in obogatena resničnost

nege. Razširjena resničnost (eXtended Reality, XR) je pogosto



ABSTRACT uporabljen krovni izraz za več podvrst resničnosti, vključno z



The rapid development of information technology has recently resničnostjo ( navidezno resničnostjo (Virtual Reality, VR), obogateno

led to a significant increase in the use of types of extended reality, (Mixed Reality, MR). VR je tehnologija, ki zagotavlja skoraj Augmented Reality, AR) in mešano resničnostjo



a summary of the state of research on the development and use način [2]. Omejili se bomo na tip potopitvene VR, pri kateri of virtual reality and augmented reality in this professional field, dobimo which is also reflected in nursing. In this paper, we first present resnično in/ali verodostojno izkušnjo na sintetičen oz. virtualen

and then list some commercially available applications suitable svetu. To lahko dosežemo občutek, da se nahajamo v računalniško ustvarjenem



applications, SimX and AccuVein, are presented in more detail. projektorjev. Pogosto VR uporabniku omogoča interakcijo z The use of such applications brings many advantages, but also ustvarjenim 3D okoljem [3]. Za razliko od VR je AR tehnologija, some disadvantages. More research on their suitability and katere cilj je digitalna povezava in razširitev fizičnega okolja ali for use in education, training or clinical practice. Two (Head-Mounted Display, HMD) ali z uporabo sistema s tehnologijo naglavnega zaslona



effectiveness and the development of new or improved uporabnikovega sveta v realnem času z dodajanjem plasti applications for more and more aspects of nursing are expected digitalnih informacij. [4]. Prikaz poteka na različne načine, npr. in the future. z uporabo kamere in markerjev v realnem svetu, na lokaciji,

zaznani npr. z GPS, s projekcijo na objekte ali s prekrivanjem



KLJUČNE BESEDE prenosni napravi. objektov. Rezultat si ogledujemo bodisi s pametnimi očali ali na

zdravstvena nega, navidezna resničnost, obogatena resničnost,

simulacija 1.2 Ravni uporabe XR



KEYWORD Aplikacije XR nastopajo na številnih strokovnih področjih, med

katerimi ima zdravstvo zelo vidno mesto, saj se uporabljajo tako

nursing, virtual reality, augmented reality, simulation med potekom študija/šolanja, za usposabljanje zaposlenih ali v

klinični praksi. Nekatere rabe so namenjene neposredno

pacientom (npr. za zdravljenje različnih fobij, depresije ipd.), a

Permission to make digital or hard copies of part or all of this work for personal or jih tu ne bomo posebej obravnavali. V ZN so prisotne v

classroom use is granted without fee provided that copies are not made or distributed izobraževanju bodočih medicinskih sester, za dodatno

for profit or commercial advantage and that copies bear this notice and the full usposabljanje obstoječega kadra in v klinični praksi.

citation on the first page. Copyrights for third-party components of this work must

be honored. For all other uses, contact the owner/author(s). Izobraževalni vidik sicer v raziskavah prevladuje. Med glavnimi



Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia cilji uvedbe novih pristopov poučevanja je zmanjševanje © 20 24 Copyright held by the owner/author(s). stroškov in tveganja https://doi.org/10.70314/is.2024.sizn.8 s posnemanjem avtentičnih kliničnih

scenarijev brez neposrednega stika s pacienti. [5]. S simulacijo



imajo študenti različne praktične priložnosti za ponovitev VR v ZN je ugotovil izboljšave v znanju, ne pa tudi v spretnostih,

kliničnih scenarijev ter takojšnje rezultate in samorefleksije. [6] zadovoljstvu ali samozavesti. [6] Pri uporabi 360° VR videa za

V medicini in ZN se pojem simulacije sicer pogosto uporablja za mehansko ventilacijo pa je prišlo celo do nižjega zadovoljstva,

preigravanje scenarijev, ki ni nujno računalniško podprto. Prav brez razlik v znanju ali spretnostih v primerjavi s pristopom v

tako gre pri izrazu virtualne simulacije v literaturi večkrat le za živo. [11] Ena od preglednih študij je ugotovila same pozitivne

prikaz virtualnega sveta na računalniškem zaslonu. V prispevku učinke (na znanje, spretnosti, samozavest in zadovoljstvo) pri

se bomo omejili na uporabe, ki vključujejo 3D potopitveno uporabi XR v izobraževanju ZN. [12]

izkušnjo, torej vizualno izključujejo resnični svet okoli nas - Kljub porastu simulacij VR v ZN je še vedno premalo študij o

simulacije VR. Izobraževalne simulacije z uporabo VR lahko potopitvenih scenarijih VR, temelječih na HMD. Za večjo

razdelimo v štiri skupine glede na cilje [7]: učenje postopkovnih učinkovitost VR bi morali snovalci aplikacij razmisliti o

veščin za izboljšanje tehničnega znanja in spretnosti; vključitvi haptičnih naprav za vadbo psihomotoričnih veščin ter

usposabljanje za ukrepanje v sili, ki se osredotoča na zaupanje; vključiti socialno interakcijo za poučevanje mehkih veščin. [7]

učenje mehkih veščin, ki uči empatije; in končno, vadba

psihomotoričnih sposobnosti. 2.3 Prednosti in slabosti uporabe XR

Pregledne študije vključevanja VR v izobraževanje za poklice v



2 zdravstveni dejavnosti med prednostmi navajajo dostopnost, RAZISKAVE XR

ponovljivost in stroškovno učinkovitost pristopa, med slabostmi



2.1 pa visok začetni strošek, tehnične omejitve, in možnost nelagodja Rast števila raziskav

za uporabnika [13]. Za tovrstno nelagodje se pogosto uporablja

Raziskave XR z omembo ZN so v zadnjem času v vzponu, kar izraz kibernetska bolezen. Pod tehnične omejitve bi lahko šteli

vidimo iz rezultatov poizvedbe v bibliografski zbirki podatkov tehnološke težave in pomanjkanje realizma v virtualnih svetovih,

Scopus, ki nam za iskalni profil (virtual-reality OR augmented- s katerimi so se srečevali tako študentje ZN kot zaposleni. [1].

reality OR mixed-reality) AND nursing-education vrne 493 Pri aplikaciji z 360° VR videom so uporabniki v nasprotju z

objav, pri čemer število objav v zadnjih letih močno narašča prejšnjo ugotovitvijo med drugim pohvalili tudi stopnjo

(Slika 1). Večina objav (445) vsebuje temo VR, manj (83) pa AR. realizma. Stopnja sprejemanja VR je lahko odvisna tudi od

Obe temi vsebuje 42 objav. Pri VR gre v določeni meri tudi za značilnosti posameznega uporabnika ter od njegovega dojemanja

objave brez potopitvene izkušnje. in odnosa do VR. Realnost, s katero se srečujemo v praksi, se

120 lahko razlikuje od naših pričakovanj. [14] Ena od skrbi ob

100 uporabi virtualnih okolij brez dejanskih pacientov je tudi

nevarnost, da lahko pride do pomanjkanja sočutnosti. [15]

v 80

b 60 ja

. o 3 APLIKACIJE VR IN AR

Št

40

20 V tabeli 1 je prikazan pregled obstoječih aplikacij VR in AR s

0 poudarkom na uporabi v ZN. Za vsako aplikacijo je navedeno

1995 njeno ime ter opis področja uporabe. V nadaljevanju sta malo 2000 2005 2010 2015 2020 2025

Leto podrobneje predstavljeni komercialno najbolj poznani aplikaciji

SimX (VR) in Accuvein (AR).



VR AR

Tabela 1: Aplikacije VR in AR v ZN



Slika 1: Rast števila objav v zbirki Scopus na temo VR in

AR v izobraževanju Tip: navidezna resničnost (VR) ZN

IME ZNAČILNOSTI



2.2 APLIKACIJE SPLETNA STRAN Učinkovitost uporab e XR

SimX simulacije ZN za študente/zaposlene, virtualni

Poleg temeljnih raziskav je bilo izdelanih več preglednih študij o pacienti, sodelovanje

uporabi VR v ZN. Nekatere temeljne raziskave se dotikajo https://www.simxvr.com/virtual-reality-

področja uspešnosti simulation-for-nurses/ pristopov (vpliv na znanje in spretnosti),

druge npr UbiSim simulacije ZN za študente/zaposlene, ZN, . na uporabniško izkušnjo (primernost aplikacij,

zadovoljstvo) ali psihološki vidik (npr. samozavest, motivacija, https://www.ubisimvr.com/ odločanje in sodelovanje



samoučinkovitost). Courseta simulacije ZN, interakcija z AI pacienti, igra

V več primerih raziskav uporabe AR v izobraževanju se je ta Nursing vlog, podpora odločanju

pokazala kot uspešna, npr. pri učenju mehanskega predihavanja https://coursetanursing.com/

(vpliv na znanje in motivacijo) [8] in namestitve intravenskega 3D Organon 3D anatomija, vizualizacija medicinskih slik,

katetra (vpliv na samozavest) [9]. simulacija ultrazvoka

Pri uporabi VR pa vsi rezultati niso bili vzpodbudni. Virtualna https://www.3dorganon.com/3d-organon-

intravenozna kateterizacija med študenti in solutions/ zaposlenimi sicer ni

bila učinkovitejša od klasičnega pristopa učenja, je pa izkazovala izkušnjo v težave staranja Embodied Labs usposabljanje na področju ZN s potopitveno

višjo samozavest udeležencev. [10] Pregled več študij o uporabi https://www.embodiedlabs.com/





Tip: obogatena resničnost (AR)

IME ZNAČILNOSTI

APLIKACIJE SPLETNA STRAN

AccuVein projekcija površinskih ven na telo pacienta

https://www.accuvein.com/

Complete prikaz modelov organov in okostja v

Anatomy izobraževanju

https://3d4medical.com/support/complete-

anatomy/ar

EchoPixel interaktivni hologram srca pacienta na osnovi

True 3D, slikanja MRI in CT

HTG* https://echopixeltech.com/true3d

izvedba z ultrazvokom v realnem času

https://echopixeltech.com/htg

Insight Heart interaktivni hologram srca v izobraževanju Slika 2: Simulacija politravme v aplikaciji SimX

https://animares.com/portfolio/insight-heart

HoloPatient simulacije kliničnih scenarijev s hologramskimi 3.2 APLIKACIJA ACCUVEIN (AR)

videoposnetki pacientov

https://www.gigxr.com/holopatient/ Aplikacija AccuVein [17] uporabnikom omogoča projekcijo

ARtery 3D projekcija pacientovih arterij na telo, osnovana slike vaskulature na površino kože (Slika 3). Slika se kaže na

na slikanja MRI zaslonu prenosne naprave v realnem času, s pomočjo

https://artery3d.com/ kombinacije laserskega optičnega bralnika, sistema za obdelavo

*Holographic Therapy Guidance in digitalne laserske projekcije. Zdravstveni delavci pri uporabi



3.1 APLIKACIJA SIMX (VR) niso vidne, kar vodi k boljšim odločitvam, boljši namestitvi igle lahko vidijo žile, zaklopke in bifurkacije, ki s prostim očesom

Aplikacija SimX za področje ZN [16] ponuja dinamično in s tem boljši oskrbi pacienta. Na zaslonu prikazana

platformo za realistične simulacije scenarijev klinične prakse v vaskularnost jim pomaga najti veno prave velikosti in položaj za

VR-okolju. Služi kot varen in nadzorovan poligon za vensko punkcijo ter druge medicinske postopke, ki zahtevajo

usposabljanje, kjer lahko zaposleni ali študenti izpopolnjujejo lokacijo površinskih ven. Aplikacijo uporabljamo tako, da

svoje klinične spretnosti, vadijo ocenjevanje stanja pacientov in zaslon naprave premaknemo nad pacientovo kožo. To lahko

izvajajo postopke brez kakršnihkoli tveganj v resničnem svetu. storimo pred palpacijo žil ali po tem, ko nam je palpacija potrdila

Ponuja možnost timskega dela in sodelovanja med zdravstvenimi lokacijo in primernost žile. Na zaslonu se prikaže le površinska

ekipami prek funkcionalnosti za več igralcev, kar spodbuja vaskularnost. Največja globina, pri kateri so vene prikazane, se

komunikacijo in usklajevanje. Omogočeno je prilagajanje razlikuje glede na pacienta. Poleg tega vene nekaterih pacientov

scenarijev glede na posameznikove potrebe in izkušnje, kar ali del njihovih ven morda ne bo dobro prikazan ali sploh ne bo

omogoča personalizirano učenje in vodi v izboljšanje spretnosti. prikazan. Možni vzroki so lahko globina ven, kožne bolezni (npr.

Poleg tega aplikacija omogoča tudi interaktivno komunikacijo s ekcem), tetovaže, dlake, brazgotine ali druge izrazite spremembe

pacienti, kjer lahko zdravstveni delavci razvijajo svoje površine kože ter maščobno tkivo. Za uporabo aplikacije

sposobnosti poslušanja, postavljanja vprašanj in razlage AccuVein ni potrebno posebno usposabljanje. Gre za preprosto

postopkov na razumljiv način. Po vsaki simulaciji uporabniki rešitev, ki lahko poveča zadovoljstvo pacientov, zmanjša zaplete

prejmejo povratne informacije in meritve uspešnosti, kar pri periferni infuziji in izboljša učinkovitost postopka.

omogoča nenehno izboljševanje in sledenje napredka. Dostop do



aplikacije SimX je mogoč z različnih lokacij, zato je vsestransko

uporabna za sodelovanje in učenje tudi takrat, ko je fizični dostop

omejen. Orodje ponuja več kot 300 vnaprej pripravljenih

simulacijskih scenarijev, ki pokrivajo širok spekter kliničnih

področij, npr. nujna medicinska pomoč, pediatrija, geriatrija,

psihiatrija in kirurgija. Eden od primerov je simulacija

politravme v urgentnem centru (Slika 2). Aplikacija predstavlja

most med učenjem v učilnici in osebnimi kliničnimi izkušnjami.



Slika 3: Prikaz uporabe aplikacije Accuvein



4 ZAKLJUČEK [15] Asma Al Yahyaei. 2024. Reserving the human touch in nursing education while integrating virtual reality and simulation. Journal of Professional

Z nadaljnjim razvojem tehnologij lahko tudi v ZN pričakujemo Nursing 54 (Sep, 2024), 36–8. DOI:

večje število https://doi.org/10.1016/j.profnurs.2024.06.005 komercialnih dostopnih aplikacij VR in AR, kot se

[16] SimX. 2024. SimX Virtual Reality Simulation for Nurses. Available from:

trenutno kaže, predvsem za področje izobraževanja in https://www.simxvr.com/virtual-reality-simulation-for-nurses/

usposabljanja. Razvoj tovrstnih aplikacij je zahteven in drag, [17] AccuVein. 2024. Available from: https://www.accuvein.com/

posledično zato tudi nakup aplikacij, skupaj s potrebno strojno

opremo, predstavlja za institucije velik finančni zalogaj.

Kljub kontroverznim izsledkom nekaterih raziskovalcev o

uporabi teh tehnologij lahko pričakujemo porast njene uporabe

za izobraževanje in usposabljanje, saj je dobrodošel pripomoček

učečim, predno se znajdejo v realnih situacijah. Zlasti v okoljih,

kjer je potrebno hitro ukrepanje, bodo morda podobno kot danes

piloti, lahko zaposleni tudi osveževali svoje znanje in spretnosti

s podporo VR in AR tehnologij.



LITERATURA

[1] Shefaly Shorey, Esperanza Debby Ng. 2021. The use of virtual reality

simulation among nursing students and registered nurses: A systematic

review. Nurse Education Today 98 (Mar, 2021), 104662. DOI:

https://doi.org/10.1016/j.nedt.2020.104662

[2] Borko Furht (ed.). 2008. Immersive Virtual Reality. In: Encyclopedia of

Multimedia Boston, MA: Springer US; p. 345–6. DOI:

https://doi.org/10.1007/978-0-387-78414-4_85

[3] Ronak Dipakkumar Gandhi, Dipam S Patel. 2018. Virtual Reality –

Opportunities and Challenges. International Research Journal of

Engineering and Technology 5, 1 (Jan, 2018), 482–90.

[4] Fabio Arena, Mario Collotta, Giovanni Pau, Francesco Termine. 2022. An

Overview of Augmented Reality. Computers 11, 2 (Feb, 2022), 28. DOI:

https://doi.org/10.3390/computers11020028

[5] Kai Liu, Weiwei Zhang, Wei Li, Ting Wang, Yanxue Zheng. 2023.

Effectiveness of virtual reality in nursing education: a systematic review and

meta-analysis. BMC Med Educ 23 (Sep, 2023), 710. DOI:

https://doi.org/10.1186/s12909-023-04662-x

[6] Feng-Qin Chen, Yu-Fei Leng, Jian-Feng Ge, Dan-Wen Wang, Cheng Li, Bin

Chen, Zhi-Ling Sun. 2020. Effectiveness of Virtual Reality in Nursing

Education: Meta-Analysis. Journal of Medical Internet Research 22, 9 (Sep,

2020), e18290. DOI: https://doi.org/10.2196/18290

[7] Christian Plotzky, Ulrike Lindwedel, Michaela Sorber, Barbara Loessl, Peter

König, Christophe Kunze, Christiane Kugler, Michael Meng. 2021. Virtual

reality simulations in nurse education: A systematic mapping review. Nurse

Education Today 101 (Jun, 2021), 104868. DOI:

https://doi.org/10.1016/j.nedt.2021.104868

[8] Sahar Younes Othman, Eman Ghallab, Sameh Eltaybani, Alaa Mostafa

Mohamed. 2024. Effect of using gamification and augmented reality in

mechanical ventilation unit of critical care nursing on nurse students’

knowledge, motivation, and self-efficacy: A randomized controlled trial.

Nurse Education Today 142 (Nov, 2024), 106329. DOI:

https://doi.org/10.1016/j.nedt.2024.106329

[9] Miyase Avci, Serap Parlar Kilic. 2024. The Effect of Augmented Reality

Applications on Intravenous Catheter Placement Skill in Nursing Students:

A Randomized Controlled Study. Clinical Simulation in Nursing 90 (May,

2024), 101524. DOI: https://doi.org/10.1016/j.ecns.2024.101524

[10] Sümeyye Arslan, Nevin Kuzu Kurban, Şenay Takmak, Arife Şanlialp

Zeyrek, Sinem Öztik, Hande Şenol. 2022. Effectiveness of simulation-based

peripheral intravenous catheterization training for nursing students and

hospital nurses: A systematic review and meta-analysis. Journal of Clinical

Nursing 31, 5–6 (Mar, 2022), 483–96. DOI:

https://doi.org/10.1111/jocn.15960

[11] Abdulfatai Olamilekan Babaita, Mayumi Kako, Chie Teramoto, Miho

Okamoto, Yoko Hayashi, Shinichiro Ohshimo, Takuma Sadamori, Minoru

Hattori, Michiko Moriyama. 2024. Face-to-face versus 360° VR video: a

comparative study of two teaching methods in nursing education. BMC Nurs

23, 1 (Mar, 2024), 199. DOI: https://doi.org/10.1186/s12912-024-01866-4

[12] Yuma Ota, Gen Aikawa, Ayako Nishimura, Tetsuharu Kawashima, Ryota

Imanaka, Hideaki Sakuramoto. 2024. Effects of educational methods using

extended reality on pre-registration nursing students’ knowledge, skill,

confidence, and satisfaction: A systematic review and meta-analysis. Nurse

Education Today 141 (Oct, 2024), 106313. DOI:

https://doi.org/10.1016/j.nedt.2024.106313

[13] Marvin Mergen, Marcel Meyerheim, Norbert Graf. 2023. Reviewing the

current state of virtual reality integration in medical education – a scoping

review protocol. Systematic Reviews 12 (Jun, 2023), 97. DOI:

https://doi.org/10.1186/s13643-023-02266-6

[14] Hyojin Son, Alyson Ross, Elizabeth Mendoza-Tirado, Lena Jumin Lee.

2022. Virtual Reality in Clinical Practice and Research: Viewpoint on Novel

Applications for Nursing. JMIR Nursing 5, 1 (Jan, 2022), e34036. DOI:

https://doi.org/10.2196/34036



Implementacija digitalnih rešitev v specialističnih



ambulantah in bolniških oddelkih Splošne bolnišnice



Jesenice ter njihov vpliv na kakovostno zdravstveno



obravnavo pacientov



Implementation of digital solutions in specialist clinics



and inpatient wards of the Jesenice General Hospital and



their impact on the quality of patient care



Anton Justin Alenka Bijol

Splošna bolnišnica Jesenice Splošna bolnišnica Jesenice

Jesenice, Slovenija Jesenice, Slovenija

anton.justin@sb-je.si alenka.bijol@sb-je.si



Povzetek Kljub uspešnemu uvajanju digitalnih rešitev se bolnišnica sooča z izzivi, kot so odpor zaposlenih do sprememb, potreba po Digitalizacija v zdravstvu postaja ključna za izboljšanje dodatnem izobraževanju in usposabljanju ter vprašanja glede kakovosti obravnave pacientov, kar se odraža tudi v Splošni varstva in varovanja pacientovih podatkov. Potrebna je tudi bolnišnici Jesenice. Prispevek analizira uvedbo digitalnih rešitev dodatna zakonodaja, ki bi še bolj zaščitila občutljive podatke v specialističnih ambulantah in bolnišničnih oddelkih ter njihov pacientov in zagotovila varno uporabo informacijskih tehnologij. vpliv na diagnostično-terapevtske postopke. Pravilna uporaba

digitalnih orodij omogoča bolj varno, transparentno in Prispevek poudarja, da je ključ do uspešne digitalne

učinkovito obravnavo pacientov, zmanjšuje možnost napak, transformacije v podpori vodstva, aktivnem vključevanju vseh

zmanjšuje administrativne obremenitve izvajalcev in izboljšuje deležnikov in nenehni motivaciji zaposlenih. V prihodnosti se

zadovoljstvo pacientov. načrtuje razširitev uporabe e-kartonov na vse specialistične

ambulante, nadgradnja zdravstveno-vzgojnih video vsebin z

Članek predstavlja štiri ključne implementacije digitalnih uporabo QR kod ter integracija aplikacij, ki bodo omogočale

rešitev: uvedbo e-kartona v diabetološki ambulanti, vzpostavitev prenos podatkov iz domačega okolja v bolnišnične informacijske

povezljivosti informacijskih sistemov v enoti za srčno-žilno sisteme. diagnostiko, informiranje pacientov z video vsebinami v

otorinolaringološki ambulanti ter elektronsko beleženje Zaključujemo, da digitalizacija in transformacija zdravstvenih

preventive ter kurative padcev pacientov. Uvedba e-kartona storitev ponujata priložnost za bolj kakovostno, varno in k

omogoča kontinuirano in poenoteno obravnavo pacientov, pacientu usmerjeno zdravstveno obravnavo, hkrati pa

zmanjšuje podvajanje preiskav ter znižuje stroške zdravljenja. izboljšujeta delovne procese izvajalcev. Za uspešno prihodnost

Povezljivost informacijskih sistemov v enoti za srčno-žilno zdravstva bo potrebno medsebojno sodelovanje vseh ravni

diagnostiko izboljšuje sledljivost podatkov, zmanjšuje možnost organizacije pri iskanju rešitev za prihodnje izzive.

napak in omogoča hitrejše odzive pri obravnavi pacientov.

Uporaba informativnih video vsebin v otorinolaringološki Ključne besede



ambulanti pacientom olajša razumevanje diagnostičnih posegov, zdravstvena nega, elektronski zdravstveni zapisi, digitalna povečuje njihovo sodelovanje in pripravljenost na zdravljenje. transformacija, elektronska dokumentacija, kakovost, varnost, Elektronsko beleženje padcev pacientov omogoča zaposlenim izboljšani delovni procesi

hitrejšo zabeležko o oceni stopnje tveganja za padec ter dosledno

beleženje vseh podatkov, ki jih zahteva kurativa padca pacienta. Abstract

Digitalization in healthcare is essential for improving the quality

Permission to make digital or hard copies of part or all of this work for personal or of patient care, as demonstrated by the General Hospital Jesenice.

classroom use is granted without fee provided that copies are not made or distributed This article explores the implementation of digital solutions in

for profit or commercial advantage and that copies bear this notice and the full

citation on the first page. Copyrights for third-party components of this work must specialist outpatient clinics and hospital departments and their

be honored. For all other uses, contact the owner/author(s). impact on diagnostic and therapeutic procedures. The use of

Information Society 2024, 7–11 October 2024, Ljubljana, Slovenia

© 2024 Copyright held by the owner/author(s). digital tools enables safer, more transparent, and efficient patient

https://doi.org/10.70314/is.2024.sizn.9 care, reducing errors and administrative burdens while increasing

patient satisfaction. The article highlights four key digital



implementations: the introduction of electronic health records za pojav varnostnih odklonov, ki velikokrat nastanejo kot

(EHR) in the diabetic clinic, the integration of information posledica procesa zdravljenja in ne kot posledica bolezni.

systems in the cardiovascular diagnostics unit, patient education

through video content in the otorhinolaryngology clinic, and Približno 84 % vseh varnostnih odklonov, ki se zgodijo v

electronic documentation of patient falls. The EHR enables bolnišnicah je povezanih s padci, le ti so najpogosteje zabeleženi

continuous and standardized care, reducing redundant tests and neželeni dogodek v bolnišničnem okolju [4]. Vseh padcev ni

lowering treatment costs. The integration of information systems mogoče preprečiti, a tudi neizogibni niso. Po podatkih in

enhances data traceability and speeds up responses in patient care, izkušnjah različnih zdravstvenih okolij je 20–30 % padcev

while video content improves patients' understanding of mogoče preprečiti z ustreznimi ukrepi usmerjenimi v zmanjšanje

diagnostic procedures and their engagement in treatment. prepoznanih tveganj [5]. V raziskavi [6] je bilo ugotovljeno, da

Electronic documentation of falls allows quicker risk assessment ima kritična presoja izvajalcev zdravstvene nege podobne izide

and consistent recording of all required data. Despite the success kot rutinska uporaba standardiziranih lestvic za oceno tveganja

in digitalization, the hospital faces challenges such as employee za padce, saj obstoječe lestvice niso dovolj specifične in ne

resistance to change, the need for additional training, and pokrivajo vseh dejavnikov tveganja.

concerns about data protection. Additional legislation is needed

to safeguard sensitive patient data and ensure the secure use of

information technology. The article emphasizes that successful

digital transformation relies on leadership support, active 2 Vpeljava digitalnih rešitev v specialističnih

stakeholder involvement, and continuous staff motivation. ambulantah Splošne bolnišnice Jesenice

Future plans include expanding the use of EHRs across all clinics,

upgrading video content with QR codes, and integrating Uvajanje e-kartona v diabetološki ambulanti Splošne bolnišnice

applications that allow data transfer from home environments to Jesenice (v nadaljevanju SBJ) je bilo uvedeno z namenom

hospital systems. Digitalization and the transformation of izboljšanja zdravstvene obravnave pacientov. E-karton omogoča

healthcare services offer opportunities for higher-quality, safer, zdravstvenim delavcem vpogled v celotno zdravstveno

and more patient-centered care, while also improving workflows zgodovino pacienta na primarni in sekundarni ravni, kar

for healthcare providers. zagotavlja neprekinjeno obravnavo. S tem sistemom se zbirajo in

enotno beležijo podatki o pacientovih boleznih ter morebitnih

Keywords zapletih, kar omogoča takojšen dostop do vseh relevantnih

nursing, electronic health records, digital transformation, informacij in olajšuje analizo podatkov. Uvedba e-kartona

electronic documentation, quality, safety, improved work preprečuje podvajanje laboratorijskih preiskav in nepotrebne

processes napotitve, kar zmanjšuje stroške zdravljenja sladkorne bolezni.

Zdravstveno osebje lahko tako bolje sledi obravnavi pacienta in

izvajanju zdravstveno-vzgojnih aktivnosti. V enoti za srčno-

1 Uvod žilno diagnostiko smo zagotovili povezljivost informacijskih

sistemov za učinkovitejši prenos podatkov med različnimi

Pravilna uporaba ogromnega obsega podatkov o pacientih bolnišničnimi informacijskimi sistemi. Nova pooblastila za

zahteva natančno zbiranje in analizo podatkov, vključno z dostop do podatkov zagotavljajo varnost in kakovost zdravstvene

zdravstvenimi zapisi, genomiko in informacijami, pridobljenimi obravnave ter zmanjšujejo možnost napak, kot je npr. zamenjava

iz različnih aplikacij. Kljub temu, da je nabor pacientovih identitete pacienta v specialistični ambulantni obravnavi. Cilj je

podatkov velik analiza v digitalnem sistemu omogoča zagotoviti sledljivost podatkov in kakovostno obravnavo

prilagojeno individualno oskrbo in napovedne modele za velike pacienta skozi vse faze bolnišnične obravnave. Učinkovita

skupine prebivalstva. V informacijskih sistemih, ki se organizacija diagnostično-terapevtskih aktivnosti v enoti za

uporabljajo v zdravstvu pogosto uporabljen e-karton. E-karton srčno-žilno diagnostiko temelji na zmanjševanju

ali elektronska kartoteka je pacientova zdravstvena kartoteka, ki administrativnih obremenitev izvajalcev, kar omogoča večjo

funkcionalno zagotavlja pacientovo anamnezo, demografske osredotočenost na zdravljenje, zdravstveno-vzgojne nasvete in

podatke in podatke o pregledih za uporabo strokovnjakov, ki razvoj stroke. Povezljivost podatkov omogoča enkraten vnos, ki

pogosto temeljijo na telemedicinskih pristopih [1]. se avtomatično posodobi v bolnišničnem informacijskem

Raziskovalca [2] sta analizirala uporabo informacijske sistemu, če je vsebina podatka povezana z drugimi podatki. V

tehnologije in ugotovila pozitivno korelacijo med intervencijami otorinolaringološki ambulanti smo izdelali informativne video

in zadovoljstvom pacientov, nekatere raziskave pa navajajo da se vsebine ki se predvajajo na zaslonih pred ambulantami, kjer

je zaradi uporabe informacijskih tehnologij in boljše pacientom predstavljajo potek obravnave in diagnostičnih

produktivnosti izvajalcev zdravstvene oskrbe zniža ležalna doba preiskav. Namen teh vsebin je boljša informiranost pacientov

pacientov za 34 %. Izhajamo iz dejstva, da je eden izmed pred preiskavami, kar pripomore k njihovemu boljšemu

vzrokov, ki predstavlja nujno potrebo po strategijah za sodelovanju med postopki. To omogoča hitrejšo izvedbo

preprečevanje, diagnozo in optimizacijo zdravstvene obravnave preiskav in bolj natančne izvide. Če bo ta način informiranja

tudi veliko število smrti zaradi kroničnih nenalezljivih bolezni učinkovit, bodo video vsebine razširili na druge specialistične

(leta 2015 so predstavljale 70 % vseh primerov na svetu) [3]. ambulante in področja.

Hkrati pa na bolniških oddelkih nastaja visoka stopnja tveganja



3 za padec: nizka (0-24 točk), zmerna (25-50 točk) in visoka (51 Vpeljava digitalnih rešitev na bolniških

oddelkih Splošni bolnišnici Jesenice točk in več) [7].



Na bolniških oddelkih v SBJ je bilo vpeljano elektronsko 4 Razprava beleženje padcev pacientov. Izbran je bil program, ki omogoča

Implementacija programa je v zdravstveni sistem prinesla

zaposlenim sistematično beleženje in obravnavo padcev, saj

številne pozitivne spremembe, ki so znatno izboljšale kakovost

omogoča tako preventivne kot kurativne ukrepe, kar prispeva k

oskrbe pacientov in zmanjšale tveganja, povezana s padci.



2 % pacientov pade najmanj enkrat med svojim bivanjem v prihodnjih zdravstvenih storitev zahteva kreativno in inovativno razmišljanje. Izhodišče za digitalizacijo in transformacijo bi bolnišnici. Vsak četrti padec se konča s poškodbo, od katerih je moralo biti kolektivno razumevanje izzivov, s katerimi se sooča 10 % v obliki težjih poškodb [6]. Poleg fizičnih (počasnejše v primeru padcev. Izhajali smo iz pregleda literature, da približno Raziskovanje možnih načinov za zagotavljanje in organizacijo izboljšanju varnosti pacientov in k bolj učinkovitemu ukrepanju

organizacija, kar kaže na potrebo srednjih menedžerjev oz.



posledic za pacienta (slabša kvaliteta življenja, strah pred operativnega menedžmenta, da olajšajo komunikacijo in znanje okrevanje, višja stopnja obolevnosti in smrtnosti) in psihičnih

na različnih ravneh v organizacijski hierarhiji s ciljem

ponovnim padcem, nezadovoljstvo z zdravstveno oskrbo) je

premostitve vrzeli v znanju in pretoku informacij. Znanje o

nezanemarljiv tudi dodaten zdravstveni izdatek, ki ga predstavlja

vprašanjih, povezanih z delom, načinih zagotavljanja storitev ter

obravnava pacienta po padcu (daljše trajanje hospitalizacije, višji

zaznanih težavah in izzivih med zaposlenimi je treba sporočiti

stroški zdravljenja, večje obremenitev zdravstvenega osebja) [4].

navzgor do najvišjega vodstva, da jim zagotovijo bolj niansirane



Cilj implementacije je bil izboljšati nadzor nad dejavniki repertoarje kategorij za smiselnost in alternative ukrepanja, potrebnih za odločanje. Po drugi strani pa strategije, ukazi in tveganja in optimizirati odziv zdravstvenega sistema ob smernice, zlasti razloge, ki so za njimi, je treba posredovati prvim dogodkih padcev. Sistem beleženja omogoča: zaposlenim na način, ki jim pomaga razumeti. Posedovanje



• kolektivnega razumevanja in čustvene izkušnje v zvezi s

evidentiranje dejavnikov tveganja pri pacientih, kot so

posebnimi situacijami bo dobra podlaga za učenje in koristen

omejena mobilnost, uporaba zdravil in starost;

pristop k soustvarjanju digitalizacije in transformacije prihodnjih

• identifikacijo pacientov z visokim tveganjem ter

zdravstvenih storitev [8].

uvedbo ustreznih preventivnih ukrepov, kot so



• prilagoditev okolja, fizična pomoč ali prilagoditev Ena izmed večjih prednosti programa je hitrost in učinkovitost terapije; odziva na že zabeležene padce. Pred uvedbo sistema je bil odziv beleženje dogodkov padcev, kurativnih ukrepov in na padce pogosto počasnejši, zaradi česar so se lahko posledice spremljanje pacientovega stanja po padcu; poškodb poslabšale. Z uvedbo digitalnega sistema beleženja se • povezavo s centralnim sistemom zdravstvene lahko zdravstveno osebje takoj odzove in začne z ustreznimi dokumentacije, kar omogoča učinkovito izmenjavo kurativnimi ukrepi, kar vodi k hitrejšemu okrevanju pacientov in podatkov med oddelki in zdravstvenimi ustanovami. zmanjšanju resnosti poškodb. Ta izboljšava se odraža tudi v

primerjavi z obdobjem pred uvedbo programa, kjer se je jasno

pokazalo, da je bil čas, potreben za obravnavo padcev, bistveno

V procesu implementacije smo upoštevali ključne elemente za daljši. Ta vidik izboljšuje kakovost zdravstvene obravnave in

uspešno delovanje sistema, kot so usposabljanje osebja, zadovoljstvo pacientov, saj se počutijo varnejše in bolje

prilagoditev obstoječih protokolov in uvedba rednih posodobitev oskrbljene. za spremljanje podatkov. Velik del časa smo posvetili pripravi



modula za preventivno oceno in kurativno beleženje padcev. V Druga pomembna prednost programa je izboljšana sposobnost

sodelovanju z aplikativno inženirko smo opredelili ključne prepoznavanja pacientov, ki so bolj izpostavljeni tveganju za

kazalnike tveganja za padec, v sistem vključili preventivne padec. Z natančnim beleženjem dejavnikov tveganja, kot so

ukrepe ter določili redno spremljanje in ocenjevanje pacientov omejena mobilnost, uporaba zdravil, ki vplivajo na ravnotežje in

glede na njihovo tveganje za padec. Podoben pristop smo kognitivne funkcije, ter starost, je mogoče prepoznati paciente z

uporabili tudi pri oblikovanju kazalnikov za kurativno beleženje večjim tveganjem in pravočasno uvesti ustrezne preventivne

padcev. V program smo vključili opomnike za spremljanje ukrepe. Ti ukrepi vključujejo prilagoditev bolnišničnega okolja,

zdravstvenega stanja po padcu in za nadaljnje ukrepe, kot so kot je namestitev dodatnih ročajev, premestitev nevarnih

oskrba morebitnih poškodb, zdravniška poročila in podobno. predmetov, zagotavljanje ustrezne osvetlitve in sodelovanje

fizioterapevtov ter kliničnih farmacevtov pri načrtovanju in

V SBJ je že dolgo v uporabi protokol za preprečevanje padcev, spremljanju terapije. Tako je mogoče ne le zmanjšati število

kjer se vsi pacienti ob sprejemu ocenijo po Morsejevi lestvici padcev, ampak tudi izboljšati kakovost življenja pacientov.

padcev. Ta ocenjevalni pripomoček preverja šest kategorij

dejavnikov tveganja in tako pomaga prepoznati paciente z večjim Program omogoča natančno in strukturirano beleženje vseh

tveganjem za padec. Pri ocenjevanju sodelujeta diplomirana podatkov, povezanih s padci, kar predstavlja pomembno

medicinska sestra in tehnik zdravstvene nege, ki skupaj s izboljšavo glede na prejšnji sistem papirnatih obrazcev.

pacientom in njegovimi svojci, če so prisotni, opravita oceno. Na Digitalizacija procesov omogoča centraliziran dostop do

podlagi seštevka točk (od 0 do 125) se določi stopnja ogroženosti podatkov, kar zdravstvenemu osebju omogoča natančne analize

vzorcev padcev in učinkovitost preventivnih ukrepov. Možnost



izgube ali napačne obdelave obrazcev se je z digitalizacijo nenalezljivih bolezni in večji nadzor pacientov v domačem

praktično izničila, kar izboljšuje zanesljivost podatkov in okolju.

omogoča boljše sledenje preventivnim in kurativnim ukrepom. Z

možnostjo analize podatkov v realnem času je mogoče hitreje

prepoznati trende in vzorce, kar omogoča nenehno prilagajanje 5 Zaključek

strategij za preprečevanje padcev.

Uvedba programa je v bolnišnični sistem prinesla številne Digitalne rešitve v zdravstvu praviloma prinašajo bolj

pozitivne spremembe, ki so pomembno prispevale k izboljšanju kakovostno, varno in k pacientu osredotočeno obravnavo, hkrati

varnosti in kakovosti zdravstvene obravnave. Hitrejši odziv na pa izvajalcem olajšajo delovni proces in preprečijo podvajanje

padce, boljše prepoznavanje tveganj in natančnejše beleženje dokumentacije, ki nastaja ob papirni dokumentaciji. podatkov so ključni dejavniki, ki izboljšujejo oskrbo pacientov

in zmanjšujejo tveganje za resne poškodbe. S sistematičnim Izvajalce je potrebno tudi primerno izobraziti in jim omogočiti

spremljanjem in analizo podatkov je mogoče nenehno podporo pri uporabi digitalnih rešitev. Ker živimo v digitalni

izboljševati protokole in prilagajati ukrepe, kar predstavlja dobi imajo različne gospodarske dejavnosti že v celoti

pomemben napredek v zdravstveni oskrbi. digitaliziran proces, zato bi bilo potrebno razmišljati tudi o

digitalizaciji celotnega procesa v zdravstvu.

Za nadaljnjo optimizacijo je ključnega pomena redno

izobraževanje zaposlenih za učinkovitejšo uporabo sistema. Digitalizacija in transformacija zdravstva bosta vplivali na

Nadgradnja sistema bi bila smiselna z vidika boljše integracije z izvajalce v zdravstvu in na srednji in najvišji menedžment, zato

drugimi informacijskimi sistemi in poenostavitvijo komunikacije zahtevata vključevanje in izvajanje vseh deležnikov. Vodje na

med zdravstvenimi ustanovami. V primeru ponovnih različnih ravneh in vključeni zaposleni bodo pridobivali in

hospitalizacij ter oskrbe v domačem okolju bi bila dobrodošla soproizvajali znanje o negotovostih, zaželenih prihodnostih in

širitev digitalnih orodij na področje telemedicine, kar bi njihovih protislovjih. Predstavljali bodo izzive in spodbude drug

omogočilo spremljanje tveganj za padce tudi izven bolnišničnega drugemu in drugim relevantnim deležnikom, zato je potrebno

okolja. medsebojno in timsko sodelovanje ter izobraževanje izvajalcev.



Pri uvajanju e-kartona v nekaterih specialističnih ambulantah se

je zmanjšala papirna oblika dokumentacije. S tem smo References / Literatura zagotovili, da je komunikacija med izvajalci v ambulanti

dejavnosti potekala preko informacijskega sistema, kar

zagotavlja večjo kakovost, varnost in posledično hitrejšo [1] Chakravorty, T., Jha K. & Barthwal, K., 2018. Digital technologies as

obravnavo, predvsem pa transparentnost obravnave. Pri enablers of care-quality and performance: A conceptual review of hospital

ambulantnih obravnavah se je izkazalo, da so tudi pacienti bolj supply chain network IUP. Journal of Supply Chain Management. 15 (3),

pp. 7-25.

zadovoljni. Podobno so ugotovili tudi v raziskavah [2, 9] in sicer [2] Hong, K. & Lee, D., 2017. Impact of operational innovations on customer

je bila ugotovljena pozitivna korelacija med operativnimi loyalty in the healthcare sector. Service Business, 12 (3), pp. 575-600.

inovacijami in zadovoljstvom pacientov, kar je vplivalo na DOI: https://doi.org/10.1007/s11628-017-0355-4 [3] World Health Organization, 2018. NCD mortality and morbidity [online].

zmanjšanje stroškov in izboljšanje poteka dela. Trenutno je Avaliable at: http://www.who.int/gho/ncd/mortality_morbidity/en/.

uvedba e kartona, integracija povezljivosti sistemov v [Accessed 22 March 2023].

[4] Wang, J., Chen, B., Xu, F., Chen, Q., Yue, J., Wen, J… Zhang, Y., (2023).

bolnišnični informacijski sistem in predvajanje video vsebin Clinical study of falls among inpatients with hematological diseases and

potekala samo v nekaterih specialističnih dejavnostih. V exploration of risk prediction models. Frontiers in Public Health; 11: 1–9.

DOI: https://doi.org/10.3389/fpubh.2023.1150333

prihodnosti bi želeli imeti povezljiv informacijski sistem, ki bi [5] Morris, R., & O'Riordan, S., (2017). Prevention of falls in hospital. Journal



nam pospešil obravnavo pacientov v specialističnih of Clinical Medicine; 17(4): 360–362. DOI: https://doi.org/10.7861/clinmedicine.17-4-360 diagnostičnih ambulantah. Uporabo e-kartona nameravamo [6] LeLaurin, J. H., & Shorr, R. I., (2019). Preventing falls in hospitalized

uspešno razširiti na vse nivoje specialističnih ambulant, s ciljem patients: state of the science. Clinics in Geriatric Medicine; 35(2): 273–

posredovanja različnih podatkovnih zbirk zbranih tekom 283. DOI: https://doi.org/10.1016/j.cger.2019.01.007

[7] Morse,J.M. (2006).Thesafetyofsafetyresearch: Thecaseofpatient fall

obravnave pacientov na različne ravni zdravstvenega varstva. research. Canadian Journalof Nursing Research,38(2),74–88.



Zdravstveno vzgojni nasveti v elektronski obliki nam pomagajo, [8] Gjellebæk, C., Svensson, A., Bjørkquist, C., Fladeby, N. & Grundén, K., 2020. Management challenges for future digitalization of healthcare da so pacienti, ki prihajajo v specialistično ambulanto pred services. Futures. 124, pp.1-10. DOI:

obravnavo ustrezno informirani in razumejo potek obravnave saj https://doi.org/10.1016/j.futures.2020.102636

nam to omogoča učinkovito zdravljenje. V nadaljevanju želimo [9] Rubbio, I., Bruccoleri, M., Pietrosi, A. & Ragonese, B., 2019. Digital

health technology enhances resilient behaviour: Evidence from the ward2.

video vsebine nadgradi s QR kodo, ki jo bodo pacienti lahko International Journal of Operations & Production Management, 39 (4), pp.

odčitali na zaslonu. Preko dostopa bodo v domačem okolju lahko 594-627. DOI: https://doi.org/10.1108/IJOPM-02-2018-0057

dostopali do zdravstveno vzgojnih vsebin v digitalni obliki s

katerim bomo omogočili opolnomočenje pacienta in njihovih

svojcev za obvladovanje kroničnih bolezni, zmanjševanje

zapletov in pogostosti hospitalizacije. Ob tem pa bi bila smiselna

uporaba nekaterih aplikacij, ki jih pacienti uporabljajo v

domačem okolju in bi omogočale prenos podatkov direktno v

bolnišnični informacijski sistem ambulantne dejavnosti, s tem bi

zagotovili bolj nadzorovano vodenje nekaterih kroničnih



Indeks avtorjev / Author index



Abkari M. Wahib........................................................................................................................................................................ 221

Ahlin Doljak Sara ....................................................................................................................................................................... 850

Albreht Jaka ............................................................................................................................................................................... 557

Amiel Tel ................................................................................................................................................................................... 217

Andova Andrejaana ...................................................................................................................................................................... 51

Andrenšek Luka ......................................................................................................................................................................... 237

Anžur Zoja ................................................................................................................................................................................... 31

Avdić Elma ................................................................................................................................................................................... 15

Aydin Aleyna ............................................................................................................................................................. 420, 424, 428

Aykut Yalçın .............................................................................................................................................................. 420, 424, 428

Ayvaz Emrah .............................................................................................................................................................. 420, 424, 428

Babič Davorin .................................................................................................................................................................... 560, 566

Bajec Marko ............................................................................................................................................................................... 880

Balas Marius ............................................................................................................................................................................... 812

Balas Valentina .......................................................................................................................................................................... 812

Baruca Arne ............................................................................................................................................................................... 733

Baş Seda ..................................................................................................................................................................................... 424

Batagelj Vladimir ............................................................................................................................................................... 106, 209

Bavdaž Mojca ............................................................................................................................................................................. 543

Behrami Sami ............................................................................................................................................................................. 501

Bengeri Katja ............................................................................................................................................................................... 11

Bernik Igor ................................................................................................................................................................................. 878

Bežan Simona ............................................................................................................................................................................. 571

Bijol Alenka ............................................................................................................................................................................... 966

Binić Iva ..................................................................................................................................................................................... 911

Bizjak Blanka ............................................................................................................................................................................. 575

Bobnarič Roman ......................................................................................................................................................................... 579

Bohanec Borut ............................................................................................................................................................................ 393

Bohanec Marko ............................................................................................................................................................................ 59

Bolješić Luka ............................................................................................................................................................................. 901

Bosnar Darinka ........................................................................................................................................................................... 584

Božak Tomi .................................................................................................................................................................................. 83

Bratko Ivan ......................................................................................................................................................................... 145, 150

Breznik Andreja ......................................................................................................................................................................... 588

Brežnik Dornik Maruša .............................................................................................................................................................. 164

Britchkovski Viatcheslav ........................................................................................................................................................... 432

Buben Adam ............................................................................................................................................................................... 867

Čakš Peter................................................................................................................................................................................... 801

Calcina Erik ................................................................................................................................................................................ 275

Can Tuncer ................................................................................................................................................................................. 839

Candellari Zdenka ...................................................................................................................................................................... 592

Candia Vieira Joao Paulo ........................................................................................................................................................... 259

Carlos Neves José....................................................................................................................................................................... 846

Carmona Natalija........................................................................................................................................................................ 595

Čepar Drago ............................................................................................................................................................................... 507

Cerovec Bogdan ......................................................................................................................................................................... 820

Cestnik Bojan ............................................................................................................................................................................. 130

Cherakaoui Manal ...................................................................................................................................................................... 221

Čibej Jaka ................................................................................................................................................................................... 205

Cigoj Primož .................................................................................................................................................................................. 7

Cilar Budler Leona ..................................................................................................................................................................... 947

Čipin Ivan ................................................................................................................................................................................... 514

Cizelj Leon ................................................................................................................................................................................. 384

Cork Jordan .................................................................................................................................................................................. 51

Correia Sergio .................................................................................................................................................................... 812, 828



Costa Luiz .................................................................................................................................................................................. 259

Cunha Ana .................................................................................................................................................................................. 828

Dajčman Silvo ............................................................................................................................................................................ 733

Debevc Matjaž ........................................................................................................................................................................... 824

Dečman Klara ............................................................................................................................................................................. 102

Değermenci Beril ....................................................................................................................................................... 420, 424, 428

Demšar Jani ................................................................................................................................................................................ 805

Dernovšček Hafner Nataša ......................................................................................................................................................... 600

Divjak Saša ................................................................................................................................................................................. 605

Dodevski Zlate ........................................................................................................................................................................... 905

Đoković Lazar .............................................................................................................................................................................. 19

Dolinar Lenart ............................................................................................................................................................................ 275

Dolničar Danica.......................................................................................................................................................................... 962

Dolza Enrico ............................................................................................................................................................................... 842

Dončev Aleksandar .................................................................................................................................................................... 606

Doratiotto Silvia ......................................................................................................................................................................... 846

Došenović Bonča Petra .............................................................................................................................................. 510, 539, 543

Dupuis Aymeric ......................................................................................................................................................................... 213

Dvoraček Ana ............................................................................................................................................................................. 610

Džeroski Sašo ....................................................................................................................................................................... 67, 213

E. Wachowicz Marta .................................................................................................................................................................. 454

El-Zoheiry Abdelhamid .............................................................................................................................................................. 436

Evkoski Bojan ............................................................................................................................................................................ 201

Falkowski Michal J. ................................................................................................................................................................... 462

Farčnik Daša ....................................................................................................................................................................... 510, 539

Farič Ana .................................................................................................................................................................................... 150

Ficzko Jelena .............................................................................................................................................................................. 962

Fijavž Zoran ............................................................................................................................................................................... 233

Filipič Bogdan .............................................................................................................................................................................. 51

Fink Laura .................................................................................................................................................................................. 130

Fir Jakob ..................................................................................................................................................................................... 287

Fir Nejc....................................................................................................................................................................................... 733

Fischer Evelyn ............................................................................................................................................................................ 113

Florjančič Urška ......................................................................................................................................................................... 441

Forte Mateja ............................................................................................................................................................................... 809

Fortun Novak Maja .................................................................................................................................................................... 413

Fric Urška ................................................................................................................................................................................... 441

Gallo Florian .............................................................................................................................................................................. 842

Gams Matjaž ................................................................................................................................ 27, 158, 313, 349, 518, 901, 920

Gašparič Lea ................................................................................................................................................................................. 67

Gaurina Marija ........................................................................................................................................................................... 171

Gilliani Khasa ............................................................................................................................................................................. 295

Gjoreski Hristijan ......................................................................................................................................................................... 35

Gjoreski Martin ............................................................................................................................................................................ 35

Gladović Karen .......................................................................................................................................................................... 436

Godoy Oliveira Cristina ............................................................................................................................................................. 259

Göksu Hüseyin ........................................................................................................................................................................... 839

Golob Luka ................................................................................................................................................................................. 229

Gourari Kamal ............................................................................................................................................................................ 221

Gradišek Anton ...................................................................................................................................................................... 23, 79

Grigor Patricia-Carla .................................................................................................................................................................. 201

Grmšek Svetlin Anton ................................................................................................................................................................ 955

Grobelnik Marko ........................................................................................................................................................ 225, 263, 283

Guček Alenka ..................................................................................................................................................................... 263, 267

Gültekin Güler Tuğba ................................................................................................................................................. 420, 424, 428

Guštin Mojca .............................................................................................................................................................................. 614

Hachimi Hanaa ........................................................................................................................................................................... 221

Hafner Ana ................................................................................................................................................................................. 416

Hafner Izidor .............................................................................................................................................................................. 620

Hafner Miha ................................................................................................................................................................................. 59



Halbwachs Helena ........................................................................................................................................................................ 23

Hartmans Anouk ........................................................................................................................................................................ 167

Hočevar Domen.......................................................................................................................................................................... 189

Horvat Laura .............................................................................................................................................................................. 801

Hrib Ivo ...................................................................................................................................................................................... 249

Hribar Andraž ............................................................................................................................................................................. 343

Hrovat Bukovšek Andreja .......................................................................................................................................................... 948

Infante Moro Neves Alfonso ...................................................................................................................................................... 846

İskender Balaban Dilek .............................................................................................................................................. 420, 424, 428

Istenič Tanja ............................................................................................................................................................................... 539

Ivanuša Kline Darja .................................................................................................................................................................... 846

Jakšić Danijela ........................................................................................................................................................................... 667

Janković Isidora ......................................................................................................................................................................... 926

Janković Sonja ........................................................................................................................................................................... 926

Javornik Branka ......................................................................................................................................................................... 861

Jelen Polona ............................................................................................................................................................................... 622

Jelenc Matej ................................................................................................................................................................................. 79

Jereb Eva .................................................................................................................................................................................... 606

Jermol Mitja ............................................................................................................................................................................... 217

Jonnagaddala Jitenndra ................................................................................................................................................................ 79

Jordan Marko ............................................................................................................................................................................... 71

Jurgele Anže ............................................................................................................................................................................... 625

Justin Anton ............................................................................................................................................................................... 966

Juvan Žana ................................................................................................................................................................................. 816

Kalan Živčec Gordana ................................................................................................................................................................ 863

Kalin Jan....................................................................................................................................................................................... 27

Kalyoncu Sedanur ...................................................................................................................................................... 420, 424, 428

Kaminski Jaroslaw ..................................................................................................................................................................... 462

Karnelutti Lucija ........................................................................................................................................................................ 167

Karnjuš Igor ............................................................................................................................................................................... 957

Katavić Ivana ............................................................................................................................................................................. 171

Keček Vaupotič Lenka ............................................................................................................................................................... 579

Kenda Klemen .................................................................................................................................................... 189, 193, 255, 295

Kholmska Ganna ........................................................................................................................................................................ 255

King Melissa .............................................................................................................................................................................. 865

Kirn Urša .................................................................................................................................................................................... 628

Kizhevska Emilija ........................................................................................................................................................................ 75

Klančič Rok ................................................................................................................................................................................ 193

Koç Ayhan ......................................................................................................................................................................... 424, 428

Kocuvan Primož ................................................................................................................................................................. 901, 920

Kokalj Anton ................................................................................................................................................................................ 67

Kokec Barbara ............................................................................................................................................................................ 633

Kokelj Martina ........................................................................................................................................................................... 636

Kolar Žiga .................................................................................................................................................................................... 27

Koloski Boshko .......................................................................................................................................................................... 213

Konečnik Martin .......................................................................................................................................................................... 27

Konečnik Simon ......................................................................................................................................................................... 694

Kopinč Gregor ............................................................................................................................................................................ 639

Kopinč Rok ................................................................................................................................................................................ 864

Kordeš Urban ..................................................................................................................................................................... 175, 179

Korlat Alex ................................................................................................................................................................................. 901

Koščak Veronika ........................................................................................................................................................................ 643

Kosi Jelka ................................................................................................................................................................................... 647

Košir Andrej ............................................................................................................................................................................... 832

Košir Janja .................................................................................................................................................................................. 832

Košir Katja ................................................................................................................................................................................. 824

Košmrlj Lea ................................................................................................................................................................................ 145

Kous Katja .................................................................................................................................................................................. 842

Kovač Klaudija ........................................................................................................................................................................... 651

Koželj Ksenija ............................................................................................................................................................................ 654



Kožuh Ines ......................................................................................................................................................................... 767, 801

Kralj Novak Petra ....................................................................................................................................................................... 201

Kramar Andreja .......................................................................................................................................................................... 661

Kramar Sebastjan ................................................................................................................................................................... 35, 71

Kregar Aleš ................................................................................................................................................................................ 388

Krisper Marjan ........................................................................................................................................................................... 885

Krstevska Ana .............................................................................................................................................................................. 35

Kuhar Marjeta ............................................................................................................................................................................ 657

Kukar Matjaž ................................................................................................................................................................................ 39

Kulauzović Bajko ......................................................................................................................................................................... 27

Kuzman Taja .................................................................................................................................................................................. 7

Lachheb Hatim ........................................................................................................................................................................... 221

Lasič Nastja ................................................................................................................................................................................ 661

Lazore Courtney ........................................................................................................................................................................... 98

Leandro Isabel ............................................................................................................................................................................ 836

Leban Gregor ..................................................................................................................................................................... 245, 271

Letonja Stanislava ...................................................................................................................................................................... 665

Lipič Karel ................................................................................................................................................................................. 402

Ljubič Andreja ........................................................................................................................................................................... 948

Longar Mark David ............................................................................................................................................................ 283, 287

Lovrenčič Držanič Irena ............................................................................................................................................................. 801

Lovrić Marina ............................................................................................................................................................................. 805

Lukan Junoš ........................................................................................................................................................................... 11, 35

Lukić Stevo ........................................................................................................................................................................ 911, 926

Luštrek Mitja .................................................................................................................................................. 11, 31, 35, 71, 75, 83

Lutman Tomaž ........................................................................................................................................................................... 441

Malačič Janez ............................................................................................................................................................................. 533

Mali Franc .................................................................................................................................................................................. 125

Margan Erik ............................................................................................................................................................................... 321

Marmol Marta ............................................................................................................................................................................ 836

Marolt Marjeta ........................................................................................................................................................................... 667

Martinc Matej ............................................................................................................................................................................. 213

Massri M. Besher ....................................................................................................................................................................... 263

Mattová Veronika ......................................................................................................................................................................... 98

Maver Nuša ................................................................................................................................................................................ 824

Mehanović Dželila ....................................................................................................................................................................... 15

Meira Silva Rafael ...................................................................................................................................................................... 259

Mejač Anja ................................................................................................................................................................................. 672

Mernik Marjan ........................................................................................................................................................................... 875

Mihalič Rafael ............................................................................................................................................................................ 329

Mikuletič Samanta...................................................................................................................................................................... 949

Mishev Kostadin ........................................................................................................................................................................ 936

Mladenić Dunja .................................................................................................................................. 245, 263, 267, 271, 279, 295

Mlakar Tina ................................................................................................................................................................................ 678

Mlinar Renata ............................................................................................................................................................................. 683

Močnik Alenka ........................................................................................................................................................................... 688

Mores Neto Antonio J. ............................................................................................................................................................... 217

Motamedi Elham ........................................................................................................................................................................ 241

Možina Feliks ............................................................................................................................................................................. 733

Mrgole Urška ............................................................................................................................................................................. 445

Mrkela Vesna ............................................................................................................................................................................. 692

Muha Simon ............................................................................................................................................................................... 694

Muhvič Šumandl Romana .......................................................................................................................................................... 524

N. Brečko Barbara ...................................................................................................................................................................... 449

Nančovska Šerbec Irena ............................................................................................................................................................. 672

Nedić Mila .................................................................................................................................................................................... 63

Nica Iulia .................................................................................................................................................................................... 930

Novak Erik ................................................................................................................................................................. 275, 283, 295

Novalija Inna .............................................................................................................................................................................. 241

Obrecht Matevž .......................................................................................................................................................................... 809



Odić Duško ................................................................................................................................................................................. 445

Ogrin Tomaž .............................................................................................................................................................................. 375

Ojsteršek Milan .......................................................................................................................................................................... 891

Osredkar Mari Jože .................................................................................................................................................................... 496

Pahor Marko ............................................................................................................................................................................... 510

Pajmon Sabina .................................................................................................................................................................... 121, 167

Pajnik Tina ................................................................................................................................................................................. 698

Palčič Devid ............................................................................................................................................................................ 888

Pangeršič Bor ............................................................................................................................................................................. 287

Paravlić Armin ........................................................................................................................................................................... 510

Pavleska Tanja ....................................................................................................................................................................... 7, 905

Pečnik Klemen ........................................................................................................................................................................... 816

Pejanovič Nosaka Tomo ............................................................................................................................................................... 27

Perko Alexander ................................................................................................................................................................. 914, 930

Peršolja Melita ........................................................................................................................................................................... 955

Pestotnik Stres Luna ................................................................................................................................................................... 703

Peterlin Mojca ............................................................................................................................................................................ 962

Petrović Jovana .......................................................................................................................................................................... 911

Petrovska Dojchinovska Gordana .............................................................................................................................................. 936

Piciga Aleksander ......................................................................................................................................................................... 39

Pisanski Jan ................................................................................................................................................................ 106, 110, 209

Pisanski Tomaž .................................................................................................................................................................. 106, 209

Pita Costa Joao ................................................................................................................................................... 217, 221, 225, 259

Plaskan Jure ................................................................................................................................................................................ 449

Polajnar Anja .............................................................................................................................................................................. 217

Polančič Gregor .......................................................................................................................................................................... 950

Poljak Lukek Saša ................................................................................................................................................................ 47, 529

Poljšak Kus Maša ....................................................................................................................................................................... 175

Pollak Senja ................................................................................................................................................................................ 237

Ponikvar Nina ............................................................................................................................................................................. 543

Povh Dušan ................................................................................................................................................................................ 399

Povšič Miha ................................................................................................................................................................................ 708

Praprotnik Deja .......................................................................................................................................................................... 956

Prestor Domen .............................................................................................................................................................................. 27

Pruš Daša .................................................................................................................................................................................... 510

Prybylski Maxim ........................................................................................................................................................................ 466

Pucer Patrik ................................................................................................................................................................................ 957

Purver Matthew .......................................................................................................................................................................... 237

Pušnik Maja ................................................................................................................................................................ 828, 842, 846

Rad Dana .................................................................................................................................................................................... 812

Rad Gavril .................................................................................................................................................................................. 812

Radonjič Gregor ......................................................................................................................................................................... 336

Rajher Polona ..................................................................................................................................................................... 482, 711

Rajhman Tina ............................................................................................................................................................................. 715

Rajkovic Vladislav ..................................................................................................................................................................... 718

Ratajec Mariša .............................................................................................................................................................................. 23

Rebec Suzana ............................................................................................................................................................................. 719

Redek Tjaša ........................................................................................................................................................................ 539, 543

Rei Luis ...................................................................................................................................................................................... 241

Rener Roman .............................................................................................................................................................................. 805

Renko Jakob ............................................................................................................................................................................... 957

Repič Slavič Tanja ..................................................................................................................................................................... 491

Reščič Nina .................................................................................................................................................................................. 71

Ribič Marko ............................................................................................................................................................................... 723

Robič Borut ................................................................................................................................................................................ 883

Robnik-Šikonja Marko ................................................................................................................................................................. 19

Rodman Grega ........................................................................................................................................................................... 155

Rolih Eva ............................................................................................................................................................................ 726, 730

Romih Dejan .............................................................................................................................................................................. 733

Rotnicki Michał .......................................................................................................................................................................... 458



Rožanec Jože M. ........................................................................................................................................................ 245, 255, 271

Rupnik Urban ................................................................................................................................................................................. 7

Sabir Hülya ................................................................................................................................................................. 420, 424, 428

Sadikov Aleksander...................................................................................................................................................................... 43

Sağlam Gözde ............................................................................................................................................................ 420, 424, 428

Sambt Jože ................................................................................................................................................................................. 547

Schenk Richard .......................................................................................................................................................................... 317

Šebenik Tina ............................................................................................................................................................................... 746

Šebjanič Oražem Maja ....................................................................................................................................................... 486, 749

Seme Barbi ................................................................................................................................................................................. 179

Senegačnik Andrej ..................................................................................................................................................................... 395

Šercar Tvrtko-Matija .................................................................................................................................................................. 524

Shulajkovska Miljana ................................................................................................................................................................... 79

Simjanoska Misheva Monika ..................................................................................................................................................... 936

Simonič Barbara ......................................................................................................................................................................... 477

Simsek Irfan ....................................................................................................................................................................... 839, 846

Šinik Bogdan ...................................................................................................................................................................... 197, 309

Sirk Maruša .......................................................................................................................................................................... 93, 179

Sitar Šuštar Katarina ................................................................................................................................................................... 237

Sittar Abdul ........................................................................................................................................................................ 229, 267

Šker Tesia ................................................................................................................................................................................... 271

Skobir Matjaž ............................................................................................................................................................................... 27

Škrjanc Maja .............................................................................................................................................................................. 249

Škrlj Gregor ................................................................................................................................................................................ 754

Slapničar Gašper .............................................................................................................................................................. 31, 35, 83

Slavuj Vanja ............................................................................................................................................................................... 667

Slomšek Šlamberger Bojana ...................................................................................................................................................... 809

Smerkol Maj ................................................................................................................................................................................. 23

Smrekar Erika ............................................................................................................................................................................. 654

Šoln Kristjan ................................................................................................................................................................................. 55

Sönmez Kerim ............................................................................................................................................................ 420, 424, 428

Sousa Carla ................................................................................................................................................................................. 846

Špiler Jure................................................................................................................................................................................... 356

Šprah Lilijana ............................................................................................................................................................................. 600

Stavrov Filip ............................................................................................................................................................................... 291

Stegnar Jernej ............................................................................................................................................................................. 245

Stepanić Josip ............................................................................................................................................................................. 171

Sterle Luka ................................................................................................................................................................................. 737

Štibi Ivana .................................................................................................................................................................................. 171

Stiglic Gregor ............................................................................................................................................................................. 947

Stopar Luka ................................................................................................................................................................................ 291

Strgar Sonja ................................................................................................................................................................................ 698

Strle Toma .......................................................................................................................................................................... 121, 167

Strniša Iva................................................................................................................................................................................... 741

Šturm Jan .................................................................................................................................................................................... 249

Šumak Boštjan ........................................................................................................................................................... 828, 842, 846

Susič David .................................................................................................................................................................................. 27

Susič Rok ..................................................................................................................................................................................... 23

Šuštaršič Urška ........................................................................................................................................................................... 757

Svetozarević Mihailo .................................................................................................................................................................. 926

Swati ........................................................................................................................................................................................... 279

Szabo Sandor .............................................................................................................................................................................. 299

Teklu Henok ............................................................................................................................................................................... 313

Tietz - Latza Alexander .............................................................................................................................................................. 866

Tomat Nastja .............................................................................................................................................................................. 136

Topal Oleksandra ....................................................................................................................................................................... 249

Tori Lea ...................................................................................................................................................................................... 762

Torre Cambón Silvia .................................................................................................................................................................. 767

Tošić Aleksander ................................................................................................................................................................ 197, 309

Tounsi El Azzoiani Jad .............................................................................................................................................................. 221



Trajkovik Vladimir ..................................................................................................................................................................... 905

Traven Eva ................................................................................................................................................................................. 775

Trkov Andrej .............................................................................................................................................................................. 381

Trobec Marjeta ........................................................................................................................................................................... 445

Trojer Sebastijan .................................................................................................................................................................... 31, 35

Trost Marina ............................................................................................................................................................................... 771

Tušar Tea ................................................................................................................................................................................ 51, 63

Ünver Müslüm Serhat ................................................................................................................................................ 420, 424, 428

Urbanč Luka ............................................................................................................................................................................... 225

Uspenski Aliaksei ....................................................................................................................................................................... 466

Uspenskiy Alexander ................................................................................................................................................................. 466

Vacev Maša ................................................................................................................................................................................ 911

Vake Domen ............................................................................................................................................................................... 197

Valenčič Matjaž.................................................................................................................................................................. 364, 369

Valentinčič Lara ......................................................................................................................................................................... 778

Vavpotič Mitja ........................................................................................................................................................................... 781

Verbič Jože ................................................................................................................................................................................. 339

Vičić Jernej ................................................................................................................................................................................. 197

Vidmar Eva ................................................................................................................................................................................ 117

Vladić Ervin ................................................................................................................................................................................. 15

Vodičar Janez ............................................................................................................................................................................. 510

Voglar Rafaela ........................................................................................................................................................................... 692

Wachowicz Marta ...................................................................................................................................................................... 462

Werber Borut .............................................................................................................................................................................. 784

Wielgosz Maciej ......................................................................................................................................................................... 313

Wotawa Franz .................................................................................................................................................................... 914, 930

Yildiz İslam ................................................................................................................................................................ 420, 424, 428

Yildiz Oktay ............................................................................................................................................................... 420, 424, 428

Yilmaz Eren ............................................................................................................................................................... 420, 424, 428

Yüksel Harun ............................................................................................................................................................................. 420

Zadobovšek Matic ...................................................................................................................................................................... 920

Zakrajšek Miha ........................................................................................................................................................................... 791

Zakrajšek Srečo .......................................................................................................................................................................... 791

Zalaznik Maja ............................................................................................................................................................................. 510

Zaletelj Janez .............................................................................................................................................................................. 832

Založnik Marcel ..................................................................................................................................................................... 55, 71

Zaouini Mustafa ......................................................................................................................................................................... 221

Zavalnij Bogdan ......................................................................................................................................................................... 299

Zibrek Katja ............................................................................................................................................................................... 141

Zirkelbach Maj ............................................................................................................................................................................. 43

Žnidaršič Anja ............................................................................................................................................................................ 667

Žužek Leon ................................................................................................................................................................................. 167

Žvanut Boštjan ................................................................................................................................................................... 949, 956



Slovenska konferenca o umetni inteligenci



Slovenian Conference on Artificial Intelligence



Kognitivna znanost



Cognitive Science



Odkrivanje znanja in podatkovna skladišča-SiKDD



Data Mining and Data Warehouses-SiKDD



Miti in resnice o varovanju okolja



Myths and Truths about Environmental Protection



17. Mednarodna konferenca o prenosu tehnologij



17th International Technology Transfer Conference



Demografske in družinske analize



Demographic and Family Analyses



Vzgoja in izobraževanje v informacijski družbi



Education in Information Society



Digitalna vključenost v informacijski družbi



Digital Inclusion in Information Society



Konferenca o zdravi dolgoživosti



Conference on Healthy Longevity



Legende računalništva in informatike



Legends of Computing and Informatics



ChatGPT v zdravstvu



ChatGPT in Medicine



Digitalna preobrazba zdravstvene nege-SIZN 2024



Digital Transformation of Nursing-NIIG 2024



Uredniki in urednice > Editors:

Mitja Luštrek, Matjaž Gams, Rok Piltaver, Anka Slana Ozimič, Borut Trpin, Toma Strle,

Olga Markič, Dunja Mladenić, Marko Grobelnik, Rafael Mihalič, Tomaž Ogrin, Erik Margan,

Urška Florjančič, Robert Blatnik, Špela Stres, Janez Malačič, Mari Jože Osredkar,

Uroš Rajkovič, Borut Batagelj, Matjaž Debevc, Ines Kožuh, Martin Lipovšek, Boštjan Petrič,

Tone Stanovnik, Monika Simjanoska Misheva, Stevo Lukić, Franz Wotawa, Jelena Ficzko,

Ema Dornik, Uroš Rajkovič, Tamara Štemberger Kolnik, Gregor Štiglic, Boštjan Žvanut