pipette
ESEspañol

Towards Consistent Geospatial Representations: Aligning One Encoder to Many Foundation Models

Arthur Ouaknine, David Rolnick

Preprint

In the authors' words

Geospatial foundation models (GFMs) differ widely in pretraining objective, modality and resolution, and none leads across benchmarks. We ask whether their representations are complementary, and whether a single encoder can inherit that complementarity. We introduce Pangaea-Latents, an extension of the Pangaea benchmark providing frozen embeddings of GFMs over its datasets, along with reproduced baselines, and find GFM embeddings to be neither redundant nor disjoint. We then propose GeoMixture, a preliminary framework to align a ViT to several GFMs per dataset through a multi-representation alignment (mREPA) objective. Our encoder ties the best GFM in average rank across nine Pangaea datasets, trading peak performance for cross-task consistency.

Main resultLimitation the authors admit

Appeared: Wednesday, October 21. arXiv. Preprint, not yet peer-reviewed.

Authors' comment: Accepted at NeurIPS REO Workshop 2026