Review-driven Personalized Preference Reasoning with Large Language Models for Recommendation

2024-08-12

arXiv.org

AI Summary

Existing large language models (LLMs) in recommender systems suffer from sparse input information and underutilized reasoning capabilities, leading to suboptimal rating prediction accuracy and limited interpretability. To address these limitations, we propose EXP3RT, a novel three-stage preference reasoning framework: (1) review-driven fine-grained user/item preference extraction; (2) structured profile construction; and (3) interpretable rating prediction via knowledge distillation and chain-of-thought reasoning. EXP3RT tightly integrates review semantics with structured preference representations, jointly optimizing causal interpretability and predictive performance. Extensive experiments on multiple benchmark datasets demonstrate that EXP3RT consistently outperforms state-of-the-art methods in both rating prediction and Top-K re-ranking tasks, achieving significant gains in prediction accuracy as well as explanation fidelity and plausibility.

Technology Category

• Machine Learning: Learning Preferences or Rankings
• Knowledge Representation and Reasoning: Preferences
• Data Mining & Knowledge Management: Recommender Systems

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• 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

Recent advancements in Large Language Models (LLMs) have demonstrated exceptional performance across a wide range of tasks, generating significant interest in their application to recommendation systems. However, existing methods have not fully capitalized on the potential of LLMs, often constrained by limited input information or failing to fully utilize their advanced reasoning capabilities. To address these limitations, we introduce EXP3RT, a novel LLM-based recommender designed to leverage rich preference information contained in user and item reviews. EXP3RT is basically fine-tuned through distillation from a teacher LLM to perform three key tasks in order: EXP3RT first extracts and encapsulates essential subjective preferences from raw reviews, aggregates and summarizes them according to specific criteria to create user and item profiles. It then generates detailed step-by-step reasoning followed by predicted rating, i.e., reasoning-enhanced rating prediction, by considering both subjective and objective information from user/item profiles and item descriptions. This personalized preference reasoning from EXP3RT enhances rating prediction accuracy and also provides faithful and reasonable explanations for recommendation. Extensive experiments show that EXP3RT outperforms existing methods on both rating prediction and candidate item reranking for top-k recommendation, while significantly enhancing the explainability of recommendation systems.

Problem

• Leveraging LLMs for personalized recommendation using reviews
• Enhancing rating prediction with reasoning from user-item profiles
• Improving recommendation explainability through step-by-step reasoning

Innovation

• Fine-tuned LLM for review-driven preference extraction
• Aggregates subjective preferences into user-item profiles
• Generates reasoning-enhanced rating predictions with explanations