发表机构
Institute for Digital Transformation - University of Applied Sciences Ravensburg-Weingarten（拉芬斯堡-魏恩加腾应用科学大学数字转型研究所）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
该研究针对POI推荐的物品冷启动问题，提出LLM-MGCL模型，结合语义、地理与交互多图及对比学习，在Yelp数据集上显著优于基线模型，缓解了冷启动问题。
AI 中文摘要
基于图神经网络的兴趣点（POI）推荐模型通过在用户-物品交互中传播协同信号取得了良好性能，但面临冷启动问题，即仅有少量交互或无交互的物品无法得到有效表示。本文提出LLM增强型多图对比学习（LLM-MGCL），这是一种多图神经网络，利用物品的语义和空间信息扩展LightGCN主干，新增两个辅助物品-物品图：一个是由大语言模型（LLM）生成的照片摘要和关键词的句子嵌入构建的语义图，另一个是由商家位置间的半正矢（Haversine）距离得到的地理图。物品嵌入在三个图上并行传播、加性融合，并通过双向InfoNCE对比目标在视图间对齐，该目标连接同一物品的行为、语义和空间表示。在Yelp多模态推荐数据集上的实验表明，LLM-MGCL的性能优于经典协同过滤、矩阵分解及仅基于交互的图神经网络基线；与LightGCN相比，它将Recall@20提升了52.0%，NDCG@20提升了64.8%，同时与最强的对比基线自监督图学习（SGL）性能相当，但SGL也受冷启动问题影响。 ablation研究显示，跨视图对比对齐（CA）是这些性能提升的主要驱动因素，且当三个图全部结合时性能最佳。我们的结果表明，基于外部的、由LLM衍生的物品知识可有效弥补缺失的协同信号，缓解POI推荐中的物品冷启动问题。
英文摘要
Point-of-interest (POI) recommendation models based on graph neural networks achieve strong performance by propagating collaborative signals over user-item interactions, yet they struggle with the cold-start problem, where items with few or no interactions are not represented. In this paper, we propose LLM-augmented Multi-Graph Contrastive Learning (LLM-MGCL), a multi-graph neural network that uses semantic and spatial information about items to extend the LightGCN backbone with two auxiliary item-item graphs: a semantic graph constructed from sentence embeddings of LLM-generated photo summaries and keywords, and a geographic graph derived from Haversine distances between business locations. Item embeddings are propagated over all three graphs in parallel, fused additively, and aligned across views through a bidirectional InfoNCE contrastive objective that connects behavioral, semantic, and spatial representations of the same items. Experiments on the Yelp Multimodal Recommendation Dataset show that LLM-MGCL outperforms classical collaborative filtering, matrix factorization, and interaction-only graph neural network baselines. It improves Recall@20 by 52.0% and NDCG@20 by 64.8% over LightGCN while performing on par with the strongest contrastive baseline, Self-supervised Graph Learning (SGL), which is also affected by the cold-start problem. An ablation study reveals that the cross-view contrastive alignment (CA) is the primary driver of these gains, with the best performance achieved when all three graphs are combined. Our results suggest that externally grounded, LLM-derived item knowledge can effectively compensate for missing collaborative signal and mitigate the item cold-start problem in POI recommendation.
