GraphSearch: An Agentic Deep Searching Workflow for Graph Retrieval-Augmented Generation

TL;DR

GraphSearch addresses key weaknesses in GraphRAG—shallow retrieval and underutilization of graph structure—by introducing an agentic, six-module deep searching workflow that orchestrates iterative reasoning. A dual-channel retrieval system combines semantic text from chunk-level data and relational graph signals, enabling more comprehensive evidence gathering and reasoning. Empirical results across six multi-hop benchmarks show GraphSearch consistently improves answer accuracy and generation quality, with strong plug-and-play compatibility across graph KBs and robustness under smaller models and tighter retrieval budgets. This approach advances fact-grounded graph-based generation, offering a practical, scalable path for complex knowledge-intensive tasks in retrieval-augmented generation systems.

Abstract

Graph Retrieval-Augmented Generation (GraphRAG) enhances factual reasoning in LLMs by structurally modeling knowledge through graph-based representations. However, existing GraphRAG approaches face two core limitations: shallow retrieval that fails to surface all critical evidence, and inefficient utilization of pre-constructed structural graph data, which hinders effective reasoning from complex queries. To address these challenges, we propose \textsc{GraphSearch}, a novel agentic deep searching workflow with dual-channel retrieval for GraphRAG. \textsc{GraphSearch} organizes the retrieval process into a modular framework comprising six modules, enabling multi-turn interactions and iterative reasoning. Furthermore, \textsc{GraphSearch} adopts a dual-channel retrieval strategy that issues semantic queries over chunk-based text data and relational queries over structural graph data, enabling comprehensive utilization of both modalities and their complementary strengths. Experimental results across six multi-hop RAG benchmarks demonstrate that \textsc{GraphSearch} consistently improves answer accuracy and generation quality over the traditional strategy, confirming \textsc{GraphSearch} as a promising direction for advancing graph retrieval-augmented generation.

Figures (13)

• Figure 1: Shallow retrieval.
• Figure 2: Comparison of using graph data only, text data only, or all data as commonly adopted in GraphRAG approaches. The metric is SubEM. The contribution of retrieved graph data is marginal.
• Figure 3: Overview of our GraphSearch framework.
• Figure 4: Judge results across eight metrics on A-Score and E-Score.
• Figure 5: Comparisons between dual-channel and single-channel retrieval in GraphSearch, integrated with the graph KB retrievers built upon LightRAG, PathRAG and HyperGraphRAG.