USB-Rec: An Effective Framework for Improving Conversational Recommendation Capability of Large Language Model
This work addresses the issue of training LLMs for conversational recommender systems, which is often overlooked in existing approaches, offering a model-level improvement for applications like personalized recommendations.
The authors tackled the problem of improving conversational recommendation in large language models by proposing an integrated training-inference framework called USB-Rec, which includes a preference optimization dataset for reinforcement learning and a self-enhancement strategy, resulting in consistent outperformance over previous state-of-the-art methods across various datasets.
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.