CUE-RAG: Towards Accurate and Cost-Efficient Graph-Based RAG via Multi-Partite Graph and Query-Driven Iterative Retrieval

2025-07-11

AI Summary

Existing graph-structured RAG approaches for question answering suffer from incomplete knowledge extraction and insufficient query-aware information utilization, resulting in low-quality graph indices. This paper proposes Query-Driven Multi-Partite Graph RAG (QMPG-RAG). First, it introduces a multi-granularity knowledge extraction framework with hybrid extraction strategies to construct a multi-partite graph index integrating text chunks, knowledge units, and entities—significantly reducing LLM-based indexing overhead. Second, it designs a query-driven iterative retrieval mechanism (Q-Iter) that jointly leverages semantic matching and path constraints for precise graph traversal. Evaluated on three QA benchmarks, QMPG-RAG achieves up to 99.33% absolute accuracy gain and 113.51% F1-score improvement over baselines, reduces indexing cost by 72.58%, and requires no LLM involvement during index construction—matching or surpassing state-of-the-art methods in performance.

Technology Category

• Data Mining & Knowledge Management: Intelligent Query Processing
• Natural Language Processing: Question Answering
• Machine Learning: Graph-based Machine Learning

Application Category

• Graph Algorithms and Modeling for the Web: Querying, indexing, and retrieval in Web-related graphs
• 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

Abstract

Despite the remarkable progress of Large Language Models (LLMs), their performance in question answering (QA) remains limited by the lack of domain-specific and up-to-date knowledge. Retrieval-Augmented Generation (RAG) addresses this limitation by incorporating external information, often from graph-structured data. However, existing graph-based RAG methods suffer from poor graph quality due to incomplete extraction and insufficient utilization of query information during retrieval. To overcome these limitations, we propose CUE-RAG, a novel approach that introduces (1) a multi-partite graph index incorporates text Chunks, knowledge Units, and Entities to capture semantic content at multiple levels of granularity, (2) a hybrid extraction strategy that reduces LLM token usage while still producing accurate and disambiguated knowledge units, and (3) Q-Iter, a query-driven iterative retrieval strategy that enhances relevance through semantic search and constrained graph traversal. Experiments on three QA benchmarks show that CUE-RAG significantly outperforms state-of-the-art baselines, achieving up to 99.33% higher Accuracy and 113.51% higher F1 score while reducing indexing costs by 72.58%. Remarkably, CUE-RAG matches or outperforms baselines even without using an LLM for indexing. These results demonstrate the effectiveness and cost-efficiency of CUE-RAG in advancing graph-based RAG systems.

Problem

• Improves domain-specific QA accuracy via multi-partite graph indexing
• Reduces LLM token usage with hybrid knowledge extraction
• Enhances retrieval relevance through query-driven iterative strategy

Innovation

• Multi-partite graph index with Chunks, Units, Entities
• Hybrid extraction reduces LLM token usage
• Query-driven iterative retrieval enhances relevance