AEF-Econ: Toward Plug-and-Play Socioeconomic Foundation Embeddings from AlphaEarth for Urban Remote Sensing

2026-06-15

AI Summary

This study addresses the limitation of existing remote sensing foundation embeddings—such as AlphaEarth—which primarily capture physical surface signals and are ill-suited for socioeconomic tasks. To bridge this gap, the authors integrate seven heterogeneous data sources across 36 Chinese cities over eight years to construct CHN-Econ, the first remote sensing benchmark tailored for socioeconomic analysis. They propose a Capacity-Adaptive Reconstruction (CAR) mechanism, which leverages multi-axis controlled experiments to optimize fusion architectures and self-supervised objectives, effectively mitigating information suppression between high- and low-dimensional semantic streams during shared reconstruction. The resulting plug-and-play socioeconomic embedding achieves R² scores of 0.848 and 0.693 in cross-regional and cross-hierarchical evaluations, respectively, and is released as the AEF-Econ dataset covering 14.4 million pixels.

Abstract

AlphaEarth Foundations (AEF) unify global remote sensing foundation embeddings through multimodal self-supervised learning, but their pretraining focuses on physical land-surface signals, limiting plug-and-play use in socioeconomic tasks. We integrate seven heterogeneous data streams across 36 Chinese cities over eight years - AEF embeddings, population, nighttime lights, remote sensing indices, points of interest (POIs), urban morphology, and cross-lingual text - and construct CHN-Econ, a socioeconomic benchmark with 16 labels in three categories. We conduct 31 controlled experiments along five axes: fusion architecture, self-supervised objective, text integration, embedding dimensionality, and normalization. Used alone as a linear probe, AEF achieves R2 values of only 0.301 for cross-region and 0.160 for cross-tier evaluation. The five-axis ablated backbone improves these scores to 0.832 and 0.671, respectively, but reveals that low-dimensional semantic streams are consistently suppressed by high-dimensional streams under shared reconstruction. To address this bottleneck, we propose Capacity-Adaptive Reconstruction (CAR), replacing shared reconstruction with per-stream decoders and stream-level losses to mitigate inter-stream capacity competition. CAR further raises cross-region and cross-tier R2 to 0.848 and 0.693, and restores collapsed labels from negative R2 to a stable range. Using CAR, we infer 14.4 million pixels across 36 cities and eight years and release AEF-Econ, including 128d and 64d compressed versions. Self-diagnostics and case studies show that AEF-Econ captures cross-city hierarchies and intra-urban spatial organization under unsupervised settings, providing a socioeconomic remote sensing foundation embedding complementary to AEF physical embeddings.

Problem

• socioeconomic embedding
• remote sensing
• multimodal fusion
• foundation model
• urban analytics

Innovation

• Capacity-Adaptive Reconstruction
• socioeconomic embeddings
• multimodal fusion
• remote sensing foundation model
• plug-and-play representation