Bridging Semantic Gaps in RAG through Generated Context Knowledge Fusion

2026-09-29

AI Summary

This study addresses the semantic space misalignment between queries and documents in retrieval-augmented generation (RAG) by proposing the KASB framework. This framework innovatively integrates the strengths of generative and retrieval paradigms to construct a semantic bridging mechanism, achieving deep semantic alignment through a multi-stage knowledge fusion strategy that optimizes passage selection quality and enhances retrieval relevance. Leveraging large language model techniques, the proposed method is validated across three open-domain question answering datasets, demonstrating its effectiveness in significantly improving the overall accuracy of RAG systems.

Abstract

Retrieval-Augmented Generation has established itself as a fundamental framework in natural language processing, seamlessly integrating information retrieval with the generative capabilities of large language models. However, this process is fundamentally constrained by a critical challenge: semantic space mismatch between queries and retrieved contexts. We propose Knowledge-Aware Semantic Bridging (KASB), a novel framework that improves passage selection quality through semantic space alignment between queries and retrieved documents through intelligent knowledge fusion. Our approach leverages the complementary strengths of generative and retrieval-based knowledge through a multistage process that enhances both relevance and accuracy. We evaluate KASB on three popular open-domain Question Answering datasets to demonstrate the effectiveness of our approach.

Problem

• Retrieval-Augmented Generation
• Semantic Gap
• Semantic Space Mismatch
• Natural Language Processing

Innovation

• Retrieval-Augmented Generation
• Semantic Space Alignment
• Knowledge Fusion
• Knowledge-Aware Semantic Bridging
• Open-domain Question Answering