Mixture-of-Experts Knowledge Graph Retrieval-Augmented Generation for Multi-Agent LLM-based Recommendation

2026-05-27

AI Summary

This work addresses critical limitations in existing large language model (LLM)-based recommender systems, which rely on parametric knowledge prone to obsolescence, and current knowledge graph retrieval-augmented generation (KG-RAG) approaches that employ uniform retrieval strategies ill-suited for heterogeneous query complexities and suffer from information loss during graph-to-text conversion due to the absence of end-to-end supervision. To overcome these challenges, the authors propose MixRAGRec, a novel framework featuring a Mixture-of-Experts Multi-Agent Policy Optimization (MMAPO) mechanism. MMAPO enables query-aware, fine-grained knowledge retrieval and alignment through collaborative multi-agent coordination: a mixture-of-experts retrieval agent adaptively selects knowledge granularity, a knowledge preference alignment agent losslessly linearizes graph structures into natural language, and a contrastive learning-enhanced recommendation agent jointly optimizes recommendation outcomes. Extensive experiments demonstrate that MixRAGRec significantly outperforms state-of-the-art methods across multiple real-world datasets, achieving superior recommendation accuracy and knowledge utilization efficiency.

Abstract

Large language models (LLMs) have recently been adopted for recommendations due to their ability to understand user intent and item semantics. However, LLM-based recommender systems often rely on parametric knowledge and suffer from outdated knowledge, motivating knowledge graph retrieval-augmented generation (KG-RAG) to ground recommendations on structured, up-to-date KGs. Despite this promise, effective KG-RAG in recommendations faces great challenges. First, users' queries vary in complexity and require KG knowledge at different granularities, whereas existing methods adopt a one-size-fits-all retrieval strategy, leading to over-retrieval for simple queries and under-retrieval for complex ones. In addition, augmenting LLMs with KG knowledge requires translating graph-structured data into linear text, which may introduce noise and cause structural information loss. Moreover, the selection of retrieval granularity lacks direct supervision and must be inferred from the final recommendation after alignment and downstream utilization, making query-aware retrieval hard to learn end-to-end. To address these issues, we propose MixRAGRec, a cooperative multi-agent framework for KG-RAG recommendations. MixRAGRec integrates a Mixture-of-Experts Retrieval Agent that routes each query to a KG retrieval expert with different granularities, a Knowledge Preference Alignment Agent that converts structured knowledge into LLM-friendly natural language, and a Contrastive Learning-reinforced Recommendation Agent trained with contrastive preference feedback. Notably, we introduce Mixture-of-Experts Multi-Agent Policy Optimization (MMAPO) to train three agents under a unified objective. Extensive experiments on real-world datasets demonstrate the effectiveness of our framework.

Problem

• Knowledge Graph Retrieval-Augmented Generation
• Multi-Agent Recommendation
• Retrieval Granularity
• Structural Information Loss
• Query-Aware Retrieval

Innovation

• Mixture-of-Experts
• Knowledge Graph Retrieval-Augmented Generation
• Multi-Agent Framework
• Contrastive Learning
• LLM-based Recommendation