Parameter-Efficient Conversational Recommender System as a Language Processing Task

TL;DR

This work reframes conversational recommender systems as a unified language-processing task and introduces PECRS, a decoder-only, parameter-efficient framework that jointly handles retrieval, re-ranking, and response generation using textual item metadata. By employing LoRA-style adapters and a single-stage training objective, PECRS avoids external knowledge graphs and heavy multi-component pipelines, achieving competitive results on ReDial and INSPIRED with only a frozen backbone LM plus small trainable plugins. Key contributions include a retrieval via noise-contrastive estimation, a context-aware re-ranking scheme, and an end-to-end training regime that scales with larger backbones while maintaining efficiency. The study also offers critical insights into evaluation practices, showing that reference-based metrics are more reliable than Dist@K for assessing generated conversations, and compares PECRS against large language models to highlight the strengths of integrated retrieval and generation components for CRS tasks.

Abstract

Conversational recommender systems (CRS) aim to recommend relevant items to users by eliciting user preference through natural language conversation. Prior work often utilizes external knowledge graphs for items' semantic information, a language model for dialogue generation, and a recommendation module for ranking relevant items. This combination of multiple components suffers from a cumbersome training process, and leads to semantic misalignment issues between dialogue generation and item recommendation. In this paper, we represent items in natural language and formulate CRS as a natural language processing task. Accordingly, we leverage the power of pre-trained language models to encode items, understand user intent via conversation, perform item recommendation through semantic matching, and generate dialogues. As a unified model, our PECRS (Parameter-Efficient CRS), can be optimized in a single stage, without relying on non-textual metadata such as a knowledge graph. Experiments on two benchmark CRS datasets, ReDial and INSPIRED, demonstrate the effectiveness of PECRS on recommendation and conversation. Our code is available at: https://github.com/Ravoxsg/efficient_unified_crs.

Figures (5)

• Figure 1: An example of dialogue from ReDial redial, where blue color denotes the movie items.
• Figure 2: The overall architecture of the proposed Parameter-efficient Conversation Recommendation System (PECRS), where the PEFT denotes the parameter-efficient fine-tuning. Instead of fine-tuning backbone model, we inject PEFT plugins into backbone model and fine-tune the PEFT weights (see the figure in the right).
• Figure 3: The R@50 results of PECRS-small using the different $M_{\text{train}}$ and $M_{\text{inference}}$ pairs on ReDial dataset.
• Figure 4: R@50 of PECRS on ReDial per number of conversation turns prior to the CRS response.
• Figure 5: An example of dialogue from INSPIRED inspired, where blue color denotes the movie items.