TOOL4POI: A Tool-Augmented LLM Framework for Next POI Recommendation

2025-11-09

AI Summary

Existing LLM-based next-POI recommendation methods face two critical bottlenecks: (1) excessive reliance on complete user historical sequences—leading to severe performance degradation in the out-of-history (OOH) scenario—and (2) inherent limitations in scaling to large candidate sets due to fixed context-window constraints. To address these, we propose a tool-augmented large language model framework that, for the first time, integrates external retrieval and sequential reasoning into open-set POI recommendation without requiring LLM fine-tuning. Our framework comprises three core modules—preference extraction, multi-turn retrieval, and re-ranking—that jointly enable long-term interest modeling and dynamic candidate generation via external knowledge integration. Extensive experiments on three real-world datasets demonstrate that our method achieves up to 40% absolute accuracy improvement in the OOH setting, with average gains of 20% and 30% in Acc@5 and Acc@10, respectively—significantly mitigating both historical sparsity and candidate-scale constraints.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• 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

Abstract

Next Point-of-Interest (POI) recommendation is a fundamental task in location-based services. While recent advances leverage Large Language Model (LLM) for sequential modeling, existing LLM-based approaches face two key limitations: (i) strong reliance on the contextual completeness of user histories, resulting in poor performance on out-of-history (OOH) scenarios; (ii) limited scalability, due to the restricted context window of LLMs, which limits their ability to access and process a large number of candidate POIs. To address these challenges, we propose Tool4POI, a novel tool-augmented framework that enables LLMs to perform open-set POI recommendation through external retrieval and reasoning. Tool4POI consists of three key modules: preference extraction module, multi-turn candidate retrieval module, and reranking module, which together summarize long-term user interests, interact with external tools to retrieve relevant POIs, and refine final recommendations based on recent behaviors. Unlike existing methods, Tool4POI requires no task-specific fine-tuning and is compatible with off-the-shelf LLMs in a plug-and-play manner. Extensive experiments on three real-world datasets show that Tool4POI substantially outperforms state-of-the-art baselines, achieving up to 40% accuracy on challenging OOH scenarios where existing methods fail, and delivering average improvements of 20% and 30% on Acc@5 and Acc@10, respectively.

Problem

• Addresses poor performance in out-of-history POI recommendation scenarios
• Solves limited scalability due to LLM context window constraints
• Enables open-set POI recommendation without task-specific fine-tuning

Innovation

• Tool-augmented LLM framework for POI recommendation
• External retrieval and reasoning for open-set recommendation
• Plug-and-play compatibility with off-the-shelf LLMs