Conversational Recommendation Systems

The field of conversational recommendation systems is moving towards more sophisticated and human-like interactions. Researchers are exploring the use of large language models and multimodal generative retrieval to improve the accuracy and relevance of recommendations. One notable direction is the development of test-time scaling strategies to refine recommendations based on user feedback and evolving intent. Another area of focus is the creation of preference-aligned user simulators that can mimic human decision-making processes and provide more insightful signals for recommendation systems. Additionally, there is a growing interest in autonomous deliberative reasoning capabilities for recommender systems, which can lead to more accurate and personalized recommendations. Noteworthy papers in this area include: STARec, which introduces a slow-thinking augmented agent framework for recommender systems via autonomous deliberate reasoning. Mirroring Users, which proposes a novel framework for building preference-aligned user simulators with user feedback in recommendation systems.

Sources

Test-Time Scaling Strategies for Generative Retrieval in Multimodal Conversational Recommendations

Mirroring Users: Towards Building Preference-aligned User Simulator with User Feedback in Recommendation

STARec: An Efficient Agent Framework for Recommender Systems via Autonomous Deliberate Reasoning

Refining Text Generation for Realistic Conversational Recommendation via Direct Preference Optimization

Built with on top of