AgentDR Dynamic Recommendation with Implicit Item-Item Relations via LLM-based Agents

2025-10-07

AI Summary

Existing agent-based recommendation methods struggle to model implicit item relationships (e.g., substitutability or complementarity), tend to hallucinate items, fail to scale efficiently to full-catalog ranking, and inadequately leverage large language models’ (LLMs) commonsense reasoning capabilities. To address these limitations, we propose AgenDR—a novel LLM-agent-driven recommendation framework. First, it employs an LLM agent that explicitly infers semantic item relationships from user interaction history. Second, it introduces a personalized tool-selection mechanism that dynamically fuses LLM-generated rankings with collaborative filtering outputs. Third, it defines a new LLM evaluation metric jointly optimizing semantic alignment and ranking accuracy. Extensive experiments on three public grocery datasets demonstrate that AgenDR achieves, on average, a 2× improvement in full-catalog ranking performance over state-of-the-art baselines, while significantly enhancing recommendation relevance, interpretability, and scalability.

Technology Category

• Machine Learning: Learning Preferences or Rankings
• Data Mining & Knowledge Management: Recommender Systems
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

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

Abstract

Recent agent-based recommendation frameworks aim to simulate user behaviors by incorporating memory mechanisms and prompting strategies, but they struggle with hallucinating non-existent items and full-catalog ranking. Besides, a largely underexplored opportunity lies in leveraging LLMs'commonsense reasoning to capture user intent through substitute and complement relationships between items, which are usually implicit in datasets and difficult for traditional ID-based recommenders to capture. In this work, we propose a novel LLM-agent framework, AgenDR, which bridges LLM reasoning with scalable recommendation tools. Our approach delegates full-ranking tasks to traditional models while utilizing LLMs to (i) integrate multiple recommendation outputs based on personalized tool suitability and (ii) reason over substitute and complement relationships grounded in user history. This design mitigates hallucination, scales to large catalogs, and enhances recommendation relevance through relational reasoning. Through extensive experiments on three public grocery datasets, we show that our framework achieves superior full-ranking performance, yielding on average a twofold improvement over its underlying tools. We also introduce a new LLM-based evaluation metric that jointly measures semantic alignment and ranking correctness.

Problem

• Addresses hallucination and scalability in agent-based recommendation systems
• Captures implicit item relationships using LLM commonsense reasoning
• Combines traditional ranking models with LLM-based relational reasoning

Innovation

• LLM agents integrate multiple recommendation outputs
• LLM agents reason over substitute complement relationships
• Delegates full-ranking tasks to traditional scalable models