Enhancing Retrieval and Managing Retrieval: A Four-Module Synergy for Improved Quality and Efficiency in RAG Systems
Yunxiao Shi, Xing Zi, Zijing Shi, Haimin Zhang, Qiang Wu, Min Xu
TL;DR
The paper tackles reliability and efficiency gaps in retrieval-augmented generation by introducing a four-module, modular RAG design. It combines Query Rewriter+ (multi-query generation and intent clarification) with Knowledge Filter (NLI-based relevance), Memory Knowledge Reservoir (non-parametric caching), and Retrieval Trigger (calibration-based retrieval pacing), all built on the Gemma-2B model via LoRA fine-tuning. Empirical results across six open-domain QA datasets show consistent improvements in accuracy (5–10% higher target-hit rates) and substantial efficiency gains (up to ~46% faster on recurring questions) when modules are combined, with ablations confirming their complementary effects. The work advances practical RAG deployments by reducing noise, mitigating information plateaus, and dialing in retrieval costs, offering a scalable path for robust, efficient external-knowledge grounded QA.
Abstract
Retrieval-augmented generation (RAG) techniques leverage the in-context learning capabilities of large language models (LLMs) to produce more accurate and relevant responses. Originating from the simple 'retrieve-then-read' approach, the RAG framework has evolved into a highly flexible and modular paradigm. A critical component, the Query Rewriter module, enhances knowledge retrieval by generating a search-friendly query. This method aligns input questions more closely with the knowledge base. Our research identifies opportunities to enhance the Query Rewriter module to Query Rewriter+ by generating multiple queries to overcome the Information Plateaus associated with a single query and by rewriting questions to eliminate Ambiguity, thereby clarifying the underlying intent. We also find that current RAG systems exhibit issues with Irrelevant Knowledge; to overcome this, we propose the Knowledge Filter. These two modules are both based on the instruction-tuned Gemma-2B model, which together enhance response quality. The final identified issue is Redundant Retrieval; we introduce the Memory Knowledge Reservoir and the Retriever Trigger to solve this. The former supports the dynamic expansion of the RAG system's knowledge base in a parameter-free manner, while the latter optimizes the cost for accessing external knowledge, thereby improving resource utilization and response efficiency. These four RAG modules synergistically improve the response quality and efficiency of the RAG system. The effectiveness of these modules has been validated through experiments and ablation studies across six common QA datasets. The source code can be accessed at https://github.com/Ancientshi/ERM4.