RANGE: Retrieval Augmented Neural Fields for Multi-Resolution Geo-Embeddings

TL;DR

RANGE introduces a retrieval-augmented approach to improve geospatial embeddings by combining visual features from similar locations, significantly enhancing performance across classification and regression tasks.

Contribution

The paper proposes RANGE, a novel retrieval-augmented method that better captures visual features for geospatial representations, outperforming existing contrastive alignment techniques.

Findings

Up to 13.1% improvement in classification accuracy

0.145 R^2 increase in regression tasks

Outperforms state-of-the-art models across various geospatial tasks

Abstract

The choice of representation for geographic location significantly impacts the accuracy of models for a broad range of geospatial tasks, including fine-grained species classification, population density estimation, and biome classification. Recent works like SatCLIP and GeoCLIP learn such representations by contrastively aligning geolocation with co-located images. While these methods work exceptionally well, in this paper, we posit that the current training strategies fail to fully capture the important visual features. We provide an information-theoretic perspective on why the resulting embeddings from these methods discard crucial visual information that is important for many downstream tasks. To solve this problem, we propose a novel retrieval-augmented strategy called RANGE. We build our method on the intuition that the visual features of a location can be estimated by combining the visual features from multiple similar-looking locations. We evaluate our method across a wide variety of tasks. Our results show that RANGE outperforms the existing state-of-the-art models with significant margins in most tasks. We show gains of up to 13.1% on classification tasks and 0.145 $R^2$ on regression tasks. All our code and models will be made available at: https://github.com/mvrl/RANGE.