GeoResponder: Towards Building Geospatial LLMs for Time-Critical Disaster Response

arXiv cs.CL / 3/27/2026

💬 OpinionSignals & Early TrendsIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper argues that current LLMs need “inner” geospatial reasoning to support time-critical disaster response scenarios involving road networks, coordinates, and access to key infrastructure like hospitals, shelters, and pharmacies.
It presents GeoResponder, a framework that improves spatial reasoning via a scaffolded instruction-tuning curriculum that maps semantic knowledge onto a continuous coordinate space and helps enforce spatial axioms.
The authors evaluate the approach across four topologically distinct cities and multiple geospatial tasks, reporting performance gains over both foundation-model baselines and domain-specific alternatives.
The results indicate that LLMs can internalize and generalize geospatial structures, suggesting a pathway toward future geospatial LLMs tailored for emergency response workflows.

Abstract

LLMs excel at linguistic tasks but lack the inner geospatial capabilities needed for time-critical disaster response, where reasoning about road networks, coordinates, and access to essential infrastructure such as hospitals, shelters, and pharmacies is vital. We introduce GeoResponder, a framework that instills robust spatial reasoning through a scaffolded instruction-tuning curriculum. By stratifying geospatial learning into different cognitive layers, we anchor semantic knowledge to the continuous coordinate manifold and enforce the internalization of spatial axioms. Extensive evaluations across four topologically distinct cities and diverse tasks demonstrate that GeoResponder significantly outperforms both state-of-the-art foundation models and domain-specific baselines. These results suggest that LLMs can begin to internalize and generalize geospatial structures, pointing toward the future development of language models capable of supporting disaster response needs.