Locatability-Guided Adaptive Reasoning for Image Geo-Localization with Vision-Language Models

2026-03-13

AI Summary

Existing vision-language models for image geolocation are often constrained by fixed reasoning depths or low-quality retrieval corpora, leading to hallucinations and insufficient localization accuracy. To address these limitations, this work proposes Geo-ADAPT, an adaptive reasoning framework that introduces a novel localizability scoring mechanism to dynamically adjust reasoning depth. The authors construct Geo-ADAPT-51K, a hierarchical reasoning dataset, and integrate retrieval-augmented generation with group relative policy optimization (GRPO) guided by tailored rewards. This approach achieves state-of-the-art performance across multiple geolocation benchmarks, significantly reducing hallucination rates while simultaneously improving both localization accuracy and reasoning efficiency.

Technology Category

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

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

The emergence of Vision-Language Models (VLMs) has introduced new paradigms for global image geo-localization through retrieval-augmented generation (RAG) and reasoning-driven inference. However, RAG methods are constrained by retrieval database quality, while reasoning-driven approaches fail to internalize image locatability, relying on inefficient, fixed-depth reasoning paths that increase hallucinations and degrade accuracy. To overcome these limitations, we introduce an Optimized Locatability Score that quantifies an image's suitability for deep reasoning in geo-localization. Using this metric, we curate Geo-ADAPT-51K, a locatability-stratified reasoning dataset enriched with augmented reasoning trajectories for complex visual scenes. Building on this foundation, we propose a two-stage Group Relative Policy Optimization (GRPO) curriculum with customized reward functions that regulate adaptive reasoning depth, visual grounding, and hierarchical geographical accuracy. Our framework, Geo-ADAPT, learns an adaptive reasoning policy, achieves state-of-the-art performance across multiple geo-localization benchmarks, and substantially reduces hallucinations by reasoning both adaptively and efficiently.

Problem

• image geo-localization
• vision-language models
• locatability
• reasoning depth
• hallucination

Innovation

• locatability
• adaptive reasoning
• vision-language models
• geo-localization
• policy optimization