让大模型学会地理推理，提升地点推荐准确率。
Reasoning Over Space: Enabling Geographic Reasoning for LLM-Based Generative Next POI Recommendation
- 用分层地理语义编码，把位置信息转为可推理的文本令牌。
- 在三个数据集上，推荐命中率比最强基线高10%以上。
- 适合做跨城市地点推荐，且用更小模型就达到好效果。
基于大语言模型的生成式推荐将预测重构为序列生成任务，但现有方法在利用移动与本地服务场景中至关重要的地理信号方面仍受限。本文提出地理推理框架Reasoning Over Space（ROS），将地理信息作为推理过程中的关键决策变量。ROS引入层次化空间语义标识符（SID），将粗粒度到细粒度的区域与兴趣点（POI）语义分解为组合式令牌，并赋予大模型三阶段移动思维链（CoT）机制：建模用户个性、构建意图对齐候选空间、进行基于地理信息的剪枝。此外，通过空间引导的强化学习，使模型与真实地理环境对齐。在三个广泛使用的基于位置社交网络（LBSN）数据集上的实验表明，尽管采用较小的主干模型，ROS相较最强的LLM基线在命中率上实现超过10%的相对提升，并显著改善跨城市迁移性能。
原文摘要 · Abstract (English)
Generative recommendation with large language models (LLMs) reframes prediction as sequence generation, yet existing LLM-based recommenders remain limited in leveraging geographic signals that are crucial in mobility and local-services scenarios. Here, we present Reasoning Over Space (ROS), a framework that utilizes geography as a vital decision variable within the reasoning process. ROS introduces a Hierarchical Spatial Semantic ID (SID) that discretizes coarse-to-fine locality and POI semantics into compositional tokens, and endows LLM with a three-stage Mobility Chain-of-Thought (CoT) paradigm that models user personality, constructs an intent-aligned candidate space, and performs locality informed pruning. We further align the model with real world geography via spatial-guided Reinforcement Learning (RL). Experiments on three widely used location-based social network (LBSN) datasets show that ROS achieves over 10% relative gains in hit rate over strongest LLM-based baselines and improves cross-city transfer, despite using a smaller backbone model.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。