My research focuses on developing scalable and robust [geospatial] machine learning solutions for environmental conservation, sustainability, and disaster response. I work on methods such as out-of-distribution detection for planetary scale deployment and interpretable ML for influential sample selection.
I am a PhD student and Research Associate at the University of the Bundeswehr Munich, Germany, supervised by Michael Schmitt. Previously, I was a research assistant at Koç University and received my Master’s degree from Istanbul Technical University under the supervision of Elif Sertel.
News
[July, 2025] I am starting my Student Researcher position with the Climate and Sustainability team at Google Research in London!
[December, 2024] Our paper Distribution Shifts at Scale: Out-of-distribution Detection in Earth Observation was accepted to the EarthVision workshop at CVPR 2025.
[November, 2024] After six amazing months, I completed my Research Residency at Microsoft’s AI for Good Lab in Nairobi, Kenya.
[September, 2024] Our paper Deep Occlusion Framework for Multimodal Earth Observation Data has been published in IEEE GRSL.
[May, 2024] Presented our paper Mapping Land Naturalness from Sentinel-2 using Deep Contextual and Geographical Priors at the ICLR 2024 Tackling Climate Change with Machine Learning Workshop.
[June, 2023] Participated in a tutorial on Ethics in AI for Earth Observation at the International Conference on Geosciences and Remote Sensing. I presented the ethical challenges and considerations of conducting geospatial machine-learning research with a focus on using satellite data for environmental monitoring and biodiversity conservation. You can find the LinkedIn post here.
[March, 2023] Our paper Explaining Multi Modal Data Fusion: Occlusion Analysis for Wilderness Mapping was accepted at the ICLR 2023 Machine Learning for Remote Sensing Workshop. Additionally, I received a travel grant to attend the conference.
[October, 2022] Visited Ribana Roscher’s Remote Sensing Group at Bonn University, Germany.
[October, 2022] Took over a demo station at the Geo for Good 2022 Summit organized by Google. I discussed our Naturalness Index and how it complements our main project on wilderness mapping. See the LinkedIn post here.
Publications
Burak Ekim, Michael Schmitt, 2024. Deep Occlusion Framework for Multimodal Earth Observation Data. Paper.
Burak Ekim, Michael Schmitt, 2024. Mapping Land Naturalness from Sentinel-2 Using Deep Contextual and Geographical Priors. ICLR Tackling Climate Change with Machine Learning Workshop. Paper.
Tobias Landmann, Michael Schmitt, Burak Ekim, Jandouwe Villinger, Faith Ashiono, Jan C. Habel, Henri E. Z. Tonnang, 2023. Insect Diversity is a Good Indicator of Biodiversity Status in Africa. Communications Earth & Environment. Paper.
Burak Ekim, Michael Schmitt, 2023. Explaining Multimodal Data Fusion: Occlusion Analysis for Wilderness Mapping. Oral presentation at ICLR 2023 Machine Learning for Remote Sensing Workshop. Preprint.
Burak Ekim, Michael Schmitt, 2023. Explaining Multimodal Data Fusion: Occlusion Analysis for Wilderness Mapping. IGARSS 2023. Paper.
Burak Ekim, Timo T. Stomberg, Ribana Roscher, Michael Schmitt, 2022. MapInWild: A Remote Sensing Dataset to Address the Question “What Makes Nature Wild”. IEEE Geoscience and Remote Sensing Magazine. Paper, Code, Dataset.
Burak Ekim, Michael Schmitt, 2022. MapInWild: A Dataset for Global Wilderness Mapping. IGARSS 2022. Paper, Code, Dataset.
Elif Sertel, Burak Ekim, Paria Ettehadi Osgouei, M. Erdem Kabadayi, 2022. Land Use and Land Cover Mapping Using Deep Learning-Based Segmentation Approaches and VHR Worldview-3 Images. Remote Sensing, 14(18):4558. Paper, Code and Dataset.
Burak Ekim, Zeyu Dong, Dmitry Rashkovetsky, Michael Schmitt, 2021. The Naturalness Index for the Identification of Natural Areas on a Regional Scale. International Journal of Applied Earth Observation and Geoinformation. Paper.
Burak Ekim, Elif Sertel, 2021. Deep Neural Network Ensembles for Remote Sensing Land Cover and Land Use Classification. International Journal of Digital Earth. Paper.
Burak Ekim, Elif Sertel, 2021. A Multi-Task Deep Learning Framework for Building Footprint Segmentation. IGARSS 2021, Brussels. Paper, Code.
Burak Ekim, Elif Sertel, M. Erdem Kabadayı, 2021. Automatic Road Extraction from Historical Maps Using Deep Learning Techniques: A Regional Case Study of Turkey in a German World War II Map. ISPRS International Journal of Geo-Information. Paper, Code, Dataset and Weights.