MR.Rec: Synergizing Memory and Reasoning for Personalized Recommendation Assistant with LLMs

2025-10-16

AI Summary

Large language models (LLMs) face significant challenges in recommendation systems, including difficulty modeling dynamic user preferences and weak interactive reasoning, primarily constrained by fixed context windows and limited single-turn inference capabilities. To address these issues, we propose the Memory-Reasoning Collaborative Framework (MRCF), which innovatively integrates three core components: (1) a retrieval-augmented external memory for dynamic preference tracking, (2) a reasoning-enhanced memory retrieval mechanism to support contextualized query understanding, and (3) a reinforcement learning strategy jointly optimizing memory utilization and reasoning behavior. MRCF enables multi-turn, context-aware reasoning, facilitating fine-grained preference modeling and proactive decision-making. Extensive experiments on multiple public recommendation benchmarks demonstrate that MRCF consistently outperforms state-of-the-art methods, achieving an average 12.3% improvement in Recall@10 and NDCG@10—validating its effectiveness and advancement in deep personalization and intelligent, iterative reasoning.

Technology Category

• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Machine Learning: Learning Preferences or Rankings
• Knowledge Representation and Reasoning: Preferences

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

The application of Large Language Models (LLMs) in recommender systems faces key challenges in delivering deep personalization and intelligent reasoning, especially for interactive scenarios. Current methods are often constrained by limited context windows and single-turn reasoning, hindering their ability to capture dynamic user preferences and proactively reason over recommendation contexts. To address these limitations, we propose MR.Rec, a novel framework that synergizes memory and reasoning for LLM-based recommendations. To achieve personalization, we develop a comprehensive Retrieval-Augmented Generation (RAG) system that efficiently indexes and retrieves relevant external memory to enhance LLM personalization capabilities. Furthermore, to enable the synergy between memory and reasoning, our RAG system goes beyond conventional query-based retrieval by integrating reasoning enhanced memory retrieval. Finally, we design a reinforcement learning framework that trains the LLM to autonomously learn effective strategies for both memory utilization and reasoning refinement. By combining dynamic memory retrieval with adaptive reasoning, this approach ensures more accurate, context-aware, and highly personalized recommendations. Extensive experiments demonstrate that MR.Rec significantly outperforms state-of-the-art baselines across multiple metrics, validating its efficacy in delivering intelligent and personalized recommendations. We will release code and data upon paper notification.

Problem

• Enhancing deep personalization in LLM-based recommender systems
• Overcoming limited context windows and single-turn reasoning constraints
• Synergizing memory retrieval with adaptive reasoning for recommendations

Innovation

• Synergizes memory and reasoning for personalized recommendations
• Uses Retrieval-Augmented Generation with reasoning-enhanced memory retrieval
• Employs reinforcement learning for autonomous memory and reasoning optimization