Beyond AlphaEarth: Toward Human-Centered Spatial Representation via POI-Guided Contrastive Learning

2025-10-10

AI Summary

Existing geospatial foundation models—such as AlphaEarth (AE)—effectively learn physical surface representations from global 10-meter-resolution remote sensing imagery but struggle to capture urban functional and socioeconomic semantics. To address this gap, we propose AETHER, the first framework that integrates Point-of-Interest (POI) textual semantics into geospatial foundation model pretraining via lightweight multimodal alignment. AETHER employs POI-guided contrastive learning to jointly model remote sensing features and human activity semantics. Its methodology unifies contrastive learning, cross-modal representation alignment, and spatial encoder fine-tuning—preserving model scalability while enabling joint encoding of surface morphology and urban functionality. Experiments over Greater London demonstrate significant improvements: +7.2% in land-use classification F1-score and −23.6% reduction in KL divergence for socioeconomic distribution estimation, substantially outperforming the AE baseline. AETHER thus bridges a critical semantic gap in purely remote sensing–driven geospatial understanding.

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Natural Language Processing: Language Grounding & Multi-modal NLP
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding

Abstract

General-purpose spatial representations are essential for building transferable geospatial foundation models (GFMs). Among them, the AlphaEarth Foundation (AE) represents a major step toward a global, unified representation of the Earth's surface, learning 10-meter embeddings from multi-source Earth Observation (EO) data that capture rich physical and environmental patterns across diverse landscapes. However, such EO-driven representations remain limited in capturing the functional and socioeconomic dimensions of cities, as they primarily encode physical and spectral patterns rather than human activities or spatial functions. We propose AETHER (AlphaEarth-POI Enriched Representation Learning), a lightweight framework that adapts AlphaEarth to human-centered urban analysis through multimodal alignment guided by Points of Interest (POIs). AETHER aligns AE embeddings with textual representations of POIs, enriching physically grounded EO features with semantic cues about urban functions and socioeconomic contexts. In Greater London, AETHER achieves consistent gains over the AE baseline, with a 7.2% relative improvement in land-use classification F1 and a 23.6% relative reduction in Kullback-Leibler divergence for socioeconomic mapping. Built upon pretrained AE, AETHER leverages a lightweight multimodal alignment to enrich it with human-centered semantics while remaining computationally efficient and scalable for urban applications. By coupling EO with human-centered semantics, it advances geospatial foundation models toward general-purpose urban representations that integrate both physical form and functional meaning.

Problem

• Enriching EO spatial representations with human-centered urban semantics
• Aligning physical landscape features with socioeconomic functional patterns
• Integrating POI-guided multimodal learning into geospatial foundation models

Innovation

• POI-guided contrastive learning enriches spatial representations
• Multimodal alignment integrates physical features with urban semantics
• Lightweight framework enhances geospatial models with socioeconomic contexts