Zero-Shot Human Mobility Forecasting via Large Language Model with Hierarchical Reasoning

2025-09-20

AI Summary

Human mobility prediction faces a zero-shot generalization challenge: existing methods struggle to generalize to unseen users or locations due to scarce labeled data and difficulties in modeling dynamic user intent. This paper proposes NL-Move, the first framework to formulate mobility prediction as a natural language question-answering task, leveraging large language models’ (LLMs) semantic understanding for cross-user and cross-location zero-shot transfer. Its core innovation is a hierarchical reasoning architecture that jointly models activity-level long-term intent planning and location-level short-term preference selection, augmented by semantic-enhanced retrieval and reflective optimization to enable iterative, dynamic intent capture. On standard benchmarks, NL-Move significantly outperforms state-of-the-art methods. Ablation studies validate the contribution of each component, and case analyses demonstrate its capability to accurately infer user intent and adapt to diverse real-world scenarios.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Natural Language Processing: (Large) Language Models
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Human mobility forecasting is important for applications such as transportation planning, urban management, and personalized recommendations. However, existing methods often fail to generalize to unseen users or locations and struggle to capture dynamic intent due to limited labeled data and the complexity of mobility patterns. We propose ZHMF, a framework for zero-shot human mobility forecasting that combines a semantic enhanced retrieval and reflection mechanism with a hierarchical language model based reasoning system. The task is reformulated as a natural language question answering paradigm. Leveraging LLMs semantic understanding of user histories and context, our approach handles previously unseen prediction scenarios. We further introduce a hierarchical reflection mechanism for iterative reasoning and refinement by decomposing forecasting into an activity level planner and a location level selector, enabling collaborative modeling of long term user intentions and short term contextual preferences. Experiments on standard human mobility datasets show that our approach outperforms existing models. Ablation studies reveal the contribution of each module, and case studies illustrate how the method captures user intentions and adapts to diverse contextual scenarios.

Problem

• Forecasting human mobility without prior data on unseen users or locations
• Capturing dynamic user intentions amid complex mobility patterns and limited labels
• Generalizing predictions across diverse scenarios using semantic understanding of context

Innovation

• Semantic enhanced retrieval and reflection mechanism
• Hierarchical language model based reasoning system
• Activity level planner and location level selector