Multi-Agent Collaborative Filtering: Orchestrating Users and Items for Agentic Recommendations

2025-11-23

AI Summary

Existing LLM-based recommender systems predominantly adopt generic single- or multi-agent paradigms, overlooking domain-specific characteristics of recommendation tasks and thus failing to effectively model collaborative signals inherent in user–item interactions. Method: We propose MACF (Multi-Agent Collaborative Filtering), a novel multi-agent framework that explicitly models users and items as distinct LLM agents endowed with personalized representations, coordinated by a central orchestrator agent. This orchestrator dynamically recruits agents and issues personalized collaboration instructions to enable adaptive, context-aware cooperation during recommendation. Contribution/Results: MACF is the first to formally instantiate collaborative filtering as a multi-agent coordination mechanism, integrating retrieval-augmented generation, structured multi-agent interaction, and orchestrator-driven control. Extensive experiments on three public benchmarks demonstrate statistically significant improvements over strong baselines, validating its superior capability in capturing and leveraging collaborative signals.

Technology Category

• Multiagent Systems: Coordination and Collaboration
• Machine Learning: Large Multimodal Models (LMMs)
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval

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: Agentic search

Abstract

Agentic recommendations cast recommenders as large language model (LLM) agents that can plan, reason, use tools, and interact with users of varying preferences in web applications. However, most existing agentic recommender systems focus on generic single-agent plan-execute workflows or multi-agent task decomposition pipelines. Without recommendation-oriented design, they often underuse the collaborative signals in the user-item interaction history, leading to unsatisfying recommendation results. To address this, we propose the Multi-Agent Collaborative Filtering (MACF) framework for agentic recommendations, drawing an analogy between traditional collaborative filtering algorithms and LLM-based multi-agent collaboration. Specifically, given a target user and query, we instantiate similar users and relevant items as LLM agents with unique profiles. Each agent is able to call retrieval tools, suggest candidate items, and interact with other agents. Different from the static preference aggregation in traditional collaborative filtering, MACF employs a central orchestrator agent to adaptively manage the collaboration between user and item agents via dynamic agent recruitment and personalized collaboration instruction. Experimental results on datasets from three different domains show the advantages of our MACF framework compared to strong agentic recommendation baselines.

Problem

• Existing agentic recommenders underuse collaborative signals from interaction history
• Current systems lack recommendation-oriented design for user-item collaboration
• Need to bridge traditional collaborative filtering with LLM-based multi-agent systems

Innovation

• Multi-agent framework simulates collaborative filtering dynamics
• LLM agents represent users and items with unique profiles
• Orchestrator manages dynamic agent recruitment and collaboration