arXiv：cs.LG· Jason Marcell Setiadi, Xin Cao, Lina Yao·· 3 小时前
MGRASRec：基于协同过滤路径的多模态图检索增强序列推荐
Multimodal Graph Retrieval-Augmented Sequential Recommendation via Collaborative Filtering Paths
AI 导读
MGRASRec 提出一种多模态图检索增强的序列推荐框架，通过从用户-物品交互图中检索结构化路径，将协同过滤信号直接注入 MLLM 提示词，并借助多模态相似度扩展覆盖范围。该方法无需循环摘要，每个候选物品仅需一次前向推理，所有组件统一进增强提示词以支持 MLLM 的参数高效微调。在三个公开数据集上，MGRASRec 在所有指标上取得最佳表现，排名质量提升尤为显著。
正文
Abstract:Multimodal Large Language Models (MLLMs) have demonstrated strong potential for sequential recommendation through their ability to reason over complex multimodal data. However, existing approaches either rely solely on the target user's own interaction history, neglecting collaborative signals from neighboring users, or incur substantial computational overhead through repeated MLLM inference over long interaction histories. To address these challenges, we propose MGRASRec, a multimodal graph retrieval-augmented framework for sequential recommendation. MGRASRec injects collaborative filtering signals conditioned on the candidate item directly into the MLLM prompt by retrieving structured paths from a user-item interaction graph, extended via multimodal similarity to increase coverage beyond exact co-interaction overlap. This retrieval also surfaces the history items most relevant to the candidate at no additional cost, removing the need for recurrent summarization and keeping inference to a single forward pass per candidate. All components are unified into an augmented prompt for parameter-efficient fine-tuning of an MLLM. Extensive evaluations across three publicly available datasets validate the effectiveness of MGRASRec, achieving the best performance on all metrics with particularly strong gains in ranking quality.
| Comments: | Accepted to AJCAI 2026 |
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.11228 [cs.LG] |
| (or arXiv:2610.11228v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.11228 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jason Marcell Setiadi [view email]
[v1]
Thu, 8 Oct 2026 04:27:52 UTC (767 KB)
来源：arXiv：cs.LG · arxiv.org