Discover the mysteries of the Maya : Selected contributions from the Machine Learning Challenge & the Discovery Challenge Workshop, ECML PKDD 2021

english

2022

1 spletni vir (1 datoteka PDF (IX, 27 str.))

Narodna in univerzitetna knjižnica
(Obvezni izvod spletne publikacije)

This book contains the proceedings for the competition 'Discover the mysteries of Maya' held within the Discovery Challenge track of the 2021 European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD 2021, https://2021.ecmlpkdd.org/). The goal of the competition was locating and identifing ancient Maya architectures (buildings, aguadas and platforms) by performing integrated image segmentation of different types of satellite imagery (from Sentinel-1 and Sentinel-2) data and aerial laser scanning data (lidar data). Details for the competition as well as the final leader board are available at https://biasvariancelabs.github.io/maya_challenge/.Driven by the great interest in the competition as well as its uniqueness in terms of area of application of machine learning, we invited the the top 7 solutions to a workshop, to further discuss motivations, challenges and experience during the challenge. The workshop was organized as part of the Discovery Challenge Track of ECML PKDD 2021. Four out of the seven invitees responded positively and prepared reports discussing their solutions. These proceedings include the reports on the solutions provided by the following teams:- Chapter 1: Team \textbf{ArchAI} -- Matthew Painter and Iris Kramer, 'Discover the Mysteries of the Maya';- Chapter 2: Team \textbf{German Computer Archaeologists} -- J\"{u}rgen Landauer, Burkhard Hoppenstedt, and Johannes Allgaier, 'Image Segmentation To Locate Ancient Maya Architectures Using Deep Learning';- Chapter 3: Team \textbf{The Sentinels} -- Thorben Hellweg, Stefan Oehmcke, Ankit Kariryaa, Fabian Gieseke, and Christian Igel, 'Ensemble Learning for Semantic Segmentation of Ancient Maya Architectures';- Chapter 4: Team \textbf{cayala} -- Christian Ayala, Carlos Aranda, and Mikel Galar, 'A Deep Learning Approach to Ancient Maya Architectures Detection Using Aerial Laser Scanning Data'.

06.09.2022