Dynamic Retrieval Strategy Optimization in Retrieval-Augmented Generation Based on User Feedback
Lulu Lin, Zhu Qiang Xiao · 2025
In this paper, we propose a novel approach for dynamically optimizing the retrieval strategy in Retrieval- Augmented Generation (RAG) models based on user feedback. Unlike traditional RAG methods that rely on fixed retrieval. Our Methodology Adapts Retrieval Process in Real Mechanism. Duration that derives from the actions of the user like clicks, ratings, or selected responses. By utilizing user history and feedback, the model learn that retrieved a result that best meets the user’s needs. Which results in more accurate content that is more tailored to audiences. Experimental results performed show the effectiveness of the retrieval relevance in a method proposed in enhancing but the quality of the generative process render RAG more adaptive to user preferences.