Multi-Type Context-Aware Conversational Recommender Systems via Mixture-of-Experts

2025-04-18

AI Summary

To address the challenge of effectively fusing heterogeneous multi-source contextual information—such as knowledge graphs, dialogue histories, and item reviews—in conversational recommendation, this paper proposes the Multi-type Context-aware Conversational Recommendation System (MCCRS). Methodologically, MCCRS introduces: (1) a domain-specific modeling mechanism that differentiates representation learning for structured (e.g., knowledge graphs) and unstructured (e.g., text-based) contexts; and (2) ChairBot, a dynamic coordination module integrating Mixture-of-Experts (MoE) with attention-guided gating to enable adaptive cross-modal fusion. The system jointly incorporates knowledge graph embedding, dialogue state tracking, and multimodal text encoding. Extensive experiments on multiple benchmark datasets demonstrate a 12.6% improvement in Recall@10, alongside significant gains in response relevance and user satisfaction—validating both the efficacy and generalizability of multi-source collaborative modeling.

Technology Category

• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Natural Language Processing: Conversational AI/Dialog Systems
• Machine Learning: Multimodal Learning

Application Category

• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

Conversational recommender systems enable natural language conversations and thus lead to a more engaging and effective recommendation scenario. As the conversations for recommender systems usually contain limited contextual information, many existing conversational recommender systems incorporate external sources to enrich the contextual information. However, how to combine different types of contextual information is still a challenge. In this paper, we propose a multi-type context-aware conversational recommender system, called MCCRS, effectively fusing multi-type contextual information via mixture-of-experts to improve conversational recommender systems. MCCRS incorporates both structured information and unstructured information, including the structured knowledge graph, unstructured conversation history, and unstructured item reviews. It consists of several experts, with each expert specialized in a particular domain (i.e., one specific contextual information). Multiple experts are then coordinated by a ChairBot to generate the final results. Our proposed MCCRS model takes advantage of different contextual information and the specialization of different experts followed by a ChairBot breaks the model bottleneck on a single contextual information. Experimental results demonstrate that our proposed MCCRS method achieves significantly higher performance compared to existing baselines.

Problem

• Combining multi-type contextual information in recommender systems
• Enhancing conversational recommendations via mixture-of-experts fusion
• Overcoming single-context limitations with structured and unstructured data

Innovation

• Mixture-of-experts fuses multi-type contextual information
• Combines structured and unstructured data sources
• ChairBot coordinates experts for final recommendations