Multimodal Graph Retrieval-Augmented Sequential Recommendation via Collaborative Filtering Paths

2026-10-08

AI Summary

This study addresses the limitations of existing sequential recommendation methods, which often neglect collaborative signals and incur substantial inference costs when processing long sequences. To overcome these challenges, this work proposes MGRASRec, a framework that leverages multimodal similarity to expand the coverage of user interaction graphs. By employing graph retrieval-augmented generation techniques, it extracts informative paths and injects them into the prompts of multimodal large language models. Integrated with parameter-efficient fine-tuning, the approach enables single-pass forward inference without repetitive summarization, significantly reducing computational overhead. Extensive experiments on three public datasets demonstrate that MGRASRec achieves state-of-the-art performance across all evaluated metrics, yielding substantial improvements in ranking quality.

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.

Problem

• Sequential Recommendation
• Multimodal Large Language Models
• Collaborative Filtering
• Computational Overhead

Innovation

• Multimodal Graph Retrieval-Augmented Generation
• Sequential Recommendation
• Collaborative Filtering Paths
• Multimodal Large Language Models
• Parameter-Efficient Fine-Tuning