Retrieval-Augmented Conversational Recommendation with Prompt-based Semi-Structured Natural Language State Tracking

TL;DR

This paper introduces RA-Rec, a retrieval-augmented, LLM-driven system for conversational recommendation that leverages rich item reviews and semi-structured dialogue state tracking to improve understanding and recommendations.

Contribution

The paper presents RA-Rec, a novel retrieval-augmented LLM system that enhances conversational recommendation through semi-structured state tracking and review utilization.

Findings

RA-Rec improves recommendation relevance.

It effectively interprets complex user preferences.

The system is demonstrated via open-source tools.

Abstract

Conversational recommendation (ConvRec) systems must understand rich and diverse natural language (NL) expressions of user preferences and intents, often communicated in an indirect manner (e.g., "I'm watching my weight"). Such complex utterances make retrieving relevant items challenging, especially if only using often incomplete or out-of-date metadata. Fortunately, many domains feature rich item reviews that cover standard metadata categories and offer complex opinions that might match a user's interests (e.g., "classy joint for a date"). However, only recently have large language models (LLMs) let us unlock the commonsense connections between user preference utterances and complex language in user-generated reviews. Further, LLMs enable novel paradigms for semi-structured dialogue state tracking, complex intent and preference understanding, and generating recommendations, explanations, and question answers. We thus introduce a novel technology RA-Rec, a Retrieval-Augmented, LLM-driven dialogue state tracking system for ConvRec, showcased with a video, open source GitHub repository, and interactive Google Colab notebook.