Towards Self-cognitive Exploration: Metacognitive Knowledge Graph Retrieval Augmented Generation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918124338544640 |
|---|---|
| author | Yuan, Xujie Di, Shimin Tang, Jielong Zheng, Libin Yin, Jian |
| author_facet | Yuan, Xujie Di, Shimin Tang, Jielong Zheng, Libin Yin, Jian |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_09460 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Towards Self-cognitive Exploration: Metacognitive Knowledge Graph Retrieval Augmented Generation Yuan, Xujie Di, Shimin Tang, Jielong Zheng, Libin Yin, Jian Information Retrieval 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. |
| title | Towards Self-cognitive Exploration: Metacognitive Knowledge Graph Retrieval Augmented Generation |
| topic | Information Retrieval |
| url | https://arxiv.org/abs/2508.09460 |