Metraipolis › archive › dc_1bb41c2bb842

PiCo: Physics-Informed Concept Probing for Geo-Foundation Models — arXiv:2610.20973

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.20973
License (as archived)
CC0-1.0
Archived
2026-10-16T10:35:06.724Z
SHA-256 of the archived text
0b98bc149dc44705e9db3b84e682a96f7349d2ed816eb2f77e2d60e8b09e37af
Stable URL of this page
https://metraipolis.com/documento/dc_1bb41c2bb842

How to cite

PiCo: Physics-Informed Concept Probing for Geo-Foundation Models — arXiv:2610.20973 — arXiv (en), https://arxiv.org/abs/2610.20973 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_1bb41c2bb842 (Metraipolis archive dc_1bb41c2bb842, sha256 0b98bc149dc44705e9db3b84e682a96f7349d2ed816eb2f77e2d60e8b09e37af)
BibTeX
@misc{dc1bb41c2bb842,
  title = {PiCo: Physics-Informed Concept Probing for Geo-Foundation Models — arXiv:2610.20973},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.20973},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_1bb41c2bb842; sha256 0b98bc149dc44705e9db3b84e682a96f7349d2ed816eb2f77e2d60e8b09e37af},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.20973. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Geo-foundation models (GFMs) are typically evaluated by downstream task predictive performance, revealing little about which physical features they are sensitive to or use in prediction. This limits developers' ability to test hypotheses about model behaviour and users' ability to select models for new tasks. We propose Physics-Informed Concept Probing (PiCo), a post-hoc framework for explaining GFM predictions. PiCo identifies concept vectors in a GFM's embedding space by generating synthetic multispectral satellite observations from a 3D radiative transfer model coupled with a game engine. These vectors reveal physical sensitivities that inform domain-based hypotheses about downstream performance. Projecting embeddings onto these concept vectors produces low-dimensional representations for quantifying concept importance in making predictions. We evaluate PiCo on three GFMs with different input modalities using seven physical concepts covering forest leaf optics and canopy structure. In ground-cover estimation, RGB restriction reduces dry-matter sensitivity and disproportionately degrades predictions for dry vegetation and bare ground, supporting a hypothesis derived from sensitivity analysis. For hierarchical forest type and tree species classification, concept representations retain 70.0-93.8% of full-embedding accuracy. Models capture similar physical contrasts between classes, such as structural differences between broadleaf forest and cleared ground and leaf-water-content differences between oak and beech, yet differ in the strength of these contrasts and the concepts driving classification. PiCo thus provides a physical basis for hypothesis testing and model assessment, with potential applicability beyond Earth observation wherever forward models enable controlled simulation and pretrained embeddings are accessible for probing.

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.CV, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2610.20973 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-16T08:52:02.624Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).