Skip to main navigation Skip to search Skip to main content

USB-Rec: An Effective Framework for Improving Conversational Recommendation Capability of Large Language Model

  • Jianyu Wen*
  • , Jingyun Wang
  • , Cilin Yan
  • , Jiayin Cai
  • , Xiaolong Jiang
  • , Ying Zhang*
  • *Corresponding author for this work
  • Harbin Institute of Technology Shenzhen
  • Beihang University
  • Xiaohongshu

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Recently, Large Language Models (LLMs) have been widely employed in Conversational Recommender Systems (CRSs). Unlike traditional language model approaches that focus on training, all existing LLMs-based approaches are mainly centered around how to leverage the summarization and analysis capabilities of LLMs while ignoring the issue of training. Therefore, in this work, we propose an integrated training-inference framework, User-Simulator-Based framework (USB-Rec), for improving the performance of LLMs in conversational recommendation at the model level. Firstly, we design a LLM-based Preference Optimization (PO) dataset construction strategy for RL training, which helps the LLMs understand the strategies and methods in conversational recommendation. Secondly, we propose a Self-Enhancement Strategy (SES) at the inference stage to further exploit the conversational recommendation potential obtained from RL training. Extensive experiments on various datasets demonstrate that our method consistently outperforms previous state-of-the-art methods. Codes are available at https://github.com/John-Wendell/USB_Rec.

Original languageEnglish
Title of host publicationRecSys2025 - Proceedings of the 19th ACM Conference on Recommender Systems
PublisherAssociation for Computing Machinery, Inc
Pages472-481
Number of pages10
ISBN (Electronic)9798400713644
DOIs
StatePublished - 7 Aug 2025
Externally publishedYes
Event19th ACM Conference on Recommender Systems, RecSys 2025 - Prague, Czech Republic
Duration: 22 Sep 202526 Sep 2025

Publication series

NameRecSys2025 - Proceedings of the 19th ACM Conference on Recommender Systems

Conference

Conference19th ACM Conference on Recommender Systems, RecSys 2025
Country/TerritoryCzech Republic
CityPrague
Period22/09/2526/09/25

Keywords

  • Conversational Recommendation
  • Large Language Model
  • Reinforcement Learning

Fingerprint

Dive into the research topics of 'USB-Rec: An Effective Framework for Improving Conversational Recommendation Capability of Large Language Model'. Together they form a unique fingerprint.

Cite this