Towards Self-cognitive Exploration: Metacognitive Knowledge Graph Retrieval Augmented Generation

TL;DR

This paper introduces MetaKGRAG, a novel metacognitive framework for knowledge graph retrieval that enables self-assessment and path-aware refinement, significantly improving reasoning accuracy in large language models across multiple domains.

Contribution

It proposes a new closed-loop, path-aware refinement framework inspired by human metacognition, addressing cognitive blindness in KG-RAG systems.

Findings

MetaKGRAG outperforms existing KG-RAG methods on five datasets.

The framework effectively identifies and corrects exploration deficiencies.

Results highlight the importance of path-aware self-refinement in structured knowledge retrieval.

Abstract

Knowledge Graph-based Retrieval-Augmented Generation (KG-RAG) significantly enhances the reasoning capabilities of LargeLanguage Models by leveraging structured knowledge. However, existing KG-RAG frameworks typically operate as open-loop systems, suffering from cognitive blindness, an inability to recognize their exploration deficiencies. This leads to relevance drift and incomplete evidence, which existing self-refinement methods, designed for unstructured text-based RAG, cannot effectively resolve due to the path-dependent nature of graph exploration. To address this challenge, we propose Metacognitive Knowledge Graph Retrieval Augmented Generation (MetaKGRAG), a novel framework inspired by the human metacognition process, which introduces a Perceive-Evaluate-Adjust cycle to enable path-aware, closed-loop refinement. This cycle empowers the system to self-assess exploration quality, identify deficiencies in coverage or relevance, and perform trajectory-connected corrections from precise pivot points. Extensive experiments across five datasets in the medical, legal, and commonsense reasoning domains demonstrate that MetaKGRAG consistently outperforms strong KG-RAG and self-refinement baselines. Our results validate the superiority of our approach and highlight the critical need for path-aware refinement in structured knowledge retrieval.