Graph Counselor: Adaptive Graph Exploration via Multi-Agent Synergy to Enhance LLM Reasoning

2025-06-04

AI Summary

Existing GraphRAG methods suffer from two key limitations: (1) inefficient information aggregation—relying on single-agent, fixed-iteration paradigms that struggle to jointly leverage textual, structural, and centrality cues; and (2) rigid reasoning mechanisms—employing predefined inference paths without dynamic depth control or semantic self-correction. To address these, we propose a multi-agent collaborative GraphRAG framework featuring: (i) a novel “Plan–Think–Execute” tri-agent architecture; (ii) an Adaptive Graph Information Extraction Module (AGIEM) for multi-granularity dependency modeling; and (iii) a Multi-perspective Self-Reflection (SR) mechanism enabling dynamic reasoning depth adjustment and semantic error correction. Our approach synergistically integrates graph neural networks, abductive reasoning, and retrieval-augmented generation. Extensive experiments across diverse graph reasoning benchmarks demonstrate substantial improvements in accuracy and generalization, consistently outperforming state-of-the-art GraphRAG methods.

Technology Category

• Knowledge Representation and Reasoning: Computational Complexity of Reasoning
• Machine Learning: Graph-based Machine Learning
• Reasoning under Uncertainty: Graphical Models

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Graph Retrieval Augmented Generation (GraphRAG) effectively enhances external knowledge integration capabilities by explicitly modeling knowledge relationships, thereby improving the factual accuracy and generation quality of Large Language Models (LLMs) in specialized domains. However, existing methods suffer from two inherent limitations: 1) Inefficient Information Aggregation: They rely on a single agent and fixed iterative patterns, making it difficult to adaptively capture multi-level textual, structural, and degree information within graph data. 2) Rigid Reasoning Mechanism: They employ preset reasoning schemes, which cannot dynamically adjust reasoning depth nor achieve precise semantic correction. To overcome these limitations, we propose Graph Counselor, an GraphRAG method based on multi-agent collaboration. This method uses the Adaptive Graph Information Extraction Module (AGIEM), where Planning, Thought, and Execution Agents work together to precisely model complex graph structures and dynamically adjust information extraction strategies, addressing the challenges of multi-level dependency modeling and adaptive reasoning depth. Additionally, the Self-Reflection with Multiple Perspectives (SR) module improves the accuracy and semantic consistency of reasoning results through self-reflection and backward reasoning mechanisms. Experiments demonstrate that Graph Counselor outperforms existing methods in multiple graph reasoning tasks, exhibiting higher reasoning accuracy and generalization ability. Our code is available at https://github.com/gjq100/Graph-Counselor.git.

Problem

• Inefficient graph information aggregation via single-agent methods
• Rigid reasoning mechanisms lacking dynamic depth adjustment
• Poor semantic consistency in graph-augmented LLM reasoning

Innovation

• Multi-agent collaboration for adaptive graph exploration
• Dynamic reasoning depth adjustment via AGIEM module
• Self-reflection mechanism enhances semantic consistency