Context-Guided Dynamic Retrieval for Improving Generation Quality in RAG Models

2025-04-28

AI Summary

This work addresses the weak contextual adaptability and rigid knowledge access of static RAG architectures in open-domain question answering and complex generative tasks. We propose a state-aware dynamic knowledge retrieval framework. Methodologically, we design multi-level perception-aware retrieval vectors and differentiable document-matching paths to enable end-to-end joint optimization of retrieval and generation modules. We introduce a novel context-guided dynamic retrieval paradigm that supports semantic ambiguity identification and collaborative multi-document fusion. Extensive cross-model evaluation on the Natural Questions dataset—using GPT-4, GPT-4o, and DeepSeek—demonstrates significant improvements in BLEU and ROUGE-L scores, enhanced generation consistency and robustness, and particularly strong performance in ambiguity resolution and multi-source information integration.

Technology Category

• Natural Language Processing: Question Answering
• Search and Optimization: Learning to Search
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

This paper focuses on the dynamic optimization of the Retrieval-Augmented Generation (RAG) architecture. It proposes a state-aware dynamic knowledge retrieval mechanism to enhance semantic understanding and knowledge scheduling efficiency in large language models for open-domain question answering and complex generation tasks. The method introduces a multi-level perceptive retrieval vector construction strategy and a differentiable document matching path. These components enable end-to-end joint training and collaborative optimization of the retrieval and generation modules. This effectively addresses the limitations of static RAG structures in context adaptation and knowledge access. Experiments are conducted on the Natural Questions dataset. The proposed structure is thoroughly evaluated across different large models, including GPT-4, GPT-4o, and DeepSeek. Comparative and ablation experiments from multiple perspectives confirm the significant improvements in BLEU and ROUGE-L scores. The approach also demonstrates stronger robustness and generation consistency in tasks involving semantic ambiguity and multi-document fusion. These results highlight its broad application potential and practical value in building high-quality language generation systems.

Problem

• Dynamic optimization of RAG architecture for better generation quality
• Enhancing semantic understanding in open-domain question answering tasks
• Improving context adaptation and knowledge access in static RAG models

Innovation

• State-aware dynamic knowledge retrieval mechanism
• Multi-level perceptive retrieval vector strategy
• Differentiable document matching path training