Asking Clarifying Questions for Preference Elicitation With Large Language Models

2025-10-13

AI Summary

Large language model (LLM)-based recommender systems struggle to infer deep user preferences under sparse historical interaction data. Method: We propose a two-stage clarification question generation framework grounded in diffusion modeling: the forward process injects noise into an ideal question sequence to model uncertainty, while the reverse process learns iterative denoising to generate a funnel-shaped, semantically progressive question sequence; user profile enhancement and question-answer sequence joint modeling further enable dynamic preference refinement. Contribution/Results: This work is the first to introduce the diffusion paradigm into interactive recommendation question generation, eliminating reliance on extensive dialogue annotations required by conventional reinforcement learning or supervised fine-tuning approaches. Experiments demonstrate significant improvements in question-guided efficiency and recommendation accuracy across domains, with particularly strong performance in cold-start scenarios.

Technology Category

• Natural Language Processing: Question Answering
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Large Language Models (LLMs) have made it possible for recommendation systems to interact with users in open-ended conversational interfaces. In order to personalize LLM responses, it is crucial to elicit user preferences, especially when there is limited user history. One way to get more information is to present clarifying questions to the user. However, generating effective sequential clarifying questions across various domains remains a challenge. To address this, we introduce a novel approach for training LLMs to ask sequential questions that reveal user preferences. Our method follows a two-stage process inspired by diffusion models. Starting from a user profile, the forward process generates clarifying questions to obtain answers and then removes those answers step by step, serving as a way to add ``noise'' to the user profile. The reverse process involves training a model to ``denoise'' the user profile by learning to ask effective clarifying questions. Our results show that our method significantly improves the LLM's proficiency in asking funnel questions and eliciting user preferences effectively.

Problem

• Eliciting user preferences with limited history
• Generating effective sequential clarifying questions
• Training LLMs to ask funnel questions effectively

Innovation

• Training LLMs to ask sequential clarifying questions
• Using a two-stage process inspired by diffusion models
• Forward process adds noise, reverse process denoises profiles