AMEM4Rec: Leveraging Cross-User Similarity for Memory Evolution in Agentic LLM Recommenders

2026-02-09

AI Summary

This work addresses the limitations of existing large language model (LLM)-based recommender agents, which struggle to effectively model collaborative filtering signals and are constrained by parameter inefficiency, limited context length, and hallucination risks. To overcome these challenges, the authors propose a memory-guided end-to-end framework that explicitly stores user behavioral patterns in a global memory pool. By introducing cross-user memory linking and an evolution mechanism, the framework dynamically captures implicit user preferences and integrates semantic and collaborative signals—without requiring a pre-trained collaborative filtering model. Notably, this approach is the first to enable memory evolution within LLM-based recommender systems, achieving significant performance gains over state-of-the-art LLM recommendation methods on both the Amazon and MIND datasets.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Data Mining & Knowledge Management: Recommender Systems
• Natural Language Processing: (Large) Language Models

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information

Abstract

Agentic systems powered by Large Language Models (LLMs) have shown strong potential in recommender systems but remain hindered by several challenges. Fine-tuning LLMs is parameter-inefficient, and prompt-based agentic reasoning is limited by context length and hallucination risk. Moreover, existing agentic recommendation systems predominantly leverages semantic knowledge while neglecting the collaborative filtering (CF) signals essential for implicit preference modeling. To address these limitations, we propose AMEM4Rec, an agentic LLM-based recommender that learns collaborative signals in an end-to-end manner through cross-user memory evolution. AMEM4Rec stores abstract user behavior patterns from user histories in a global memory pool. Within this pool, memories are linked to similar existing ones and iteratively evolved to reinforce shared cross-user patterns, enabling the system to become aware of CF signals without relying on a pre-trained CF model. Extensive experiments on Amazon and MIND datasets show that AMEM4Rec consistently outperforms state-of-the-art LLM-based recommenders, demonstrating the effectiveness of evolving memory-guided collaborative filtering.

Problem

• agentic recommender
• collaborative filtering
• LLM-based recommendation
• memory evolution
• cross-user similarity

Innovation

• memory evolution
• cross-user similarity
• collaborative filtering
• agentic LLM
• global memory pool