Geography
GEO-Bench: Toward Foundation Models for Earth Monitoring
Document Type
Conference Paper
Abstract
Recent progress in self-supervision has shown that pre-training large neural networks on vast amounts of unsupervised data can lead to substantial increases in generalization to downstream tasks. Such models, recently coined foundation models, have been transformational to the field of natural language processing. Variants have also been proposed for image data, but their applicability to remote sensing tasks is limited. To stimulate the development of foundation models for Earth monitoring, we propose a benchmark comprised of six classification and six segmentation tasks, which were carefully curated and adapted to be both relevant to the field and well-suited for model evaluation. We accompany this benchmark with a robust methodology for evaluating models and reporting aggregated results to enable a reliable assessment of progress. Finally, we report results for 20 baselines to gain information about the performance of existing models. We believe that this benchmark will be a driver of progress across a variety of Earth monitoring tasks.
Publication Title
Advances in Neural Information Processing Systems
Publication Date
12-2023
Volume
36
First Page
51080
Last Page
51093
ISSN
1049-5258
DOI
10.48550/arXiv.2306.03831
Keywords
foundation models, remote sensing, Earth monitoring
Repository Citation
Lacoste, Alexandre; Lehmann, Nils; Rodriguez, Pau; Sherwin, Evan David; Kerner, Hannah; Lütjens, Björn; Irvin, Jeremy; Dao, David; Alemohammad, Hamed; Drouin, Alexandre; Gunturkun, Mehmet; and Huang, Gabriel, "GEO-Bench: Toward Foundation Models for Earth Monitoring" (2023). Geography. 986.
https://commons.clarku.edu/faculty_geography/986
