AgentKGV: Agentic LLM-RAG Framework with Two-Stage Training for the Fact Verification of Knowledge Graphs
Quick Answer
AgentKGV is a novel LLM-RAG framework for verifying knowledge graphs, improving macro-F1 scores by 5.5% over single-turn RAG and 9.4% with two-stage training.
Quick Take
The method reduces search calls from 3.24 to 1.63 while maintaining accuracy, addressing the challenge of factual errors in KGs at scale.
Key Points
- AgentKGV integrates dynamic routing and iterative query rewriting for improved KG verification.
- Two-stage training includes turn-level distillation and trajectory-level GRPO for efficiency.
- The framework achieved a macro-F1 improvement of 9.4% on the T-REx benchmark.
- GRPO reduces average search calls significantly without compromising accuracy.
- Addresses critical challenges in verifying automatically constructed knowledge graphs.
Paper Resources
📖 Reader Mode
~2 min readAbstract:Knowledge graphs (KGs) are often automatically constructed from large-scale corpora, but they inevitably contain factual errors due to noisy sources and extraction failures, and verifying them reliably at industrial scale remains a critical challenge. To address this, we propose AgentKGV, the Agentic LLM-RAG framework for KG fact Verification, that integrates dynamic routing and iterative query rewriting, which handles surface-form mismatch in document-level retrieval. To make this framework more accurate and cost-efficient for industrial deployment, we further introduce a two-stage training strategy: turn-level distillation-based SFT that transfers reasoning ability from a large teacher model into a small model for stable query rewriting and reasoning, and trajectory-level GRPO that optimizes the search policy to reduce unnecessary retrieval at scale. On the long-tail-predicate split of the open-domain T-REx benchmark, our framework improves macro-F1 over single-turn RAG by 5.5 \%p, and two-stage training does it further by 9.4 \%p. GRPO also cuts the average number of search calls from 3.24 to 1.63 without lowering accuracy.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2607.09092 [cs.CL] |
| (or arXiv:2607.09092v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2607.09092 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yumin Heo [view email]
[v1]
Fri, 10 Jul 2026 04:22:34 UTC (585 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CL
See more →TriAgent: Divergence-Aware Committees for Cost-Efficient Financial Sentiment Analysis
TriAgent introduces a cost-efficient multi-agent system for financial sentiment analysis, combining VADER, FinBERT, and Qwen2.5. It achieves an F1 score of ~0.87 with significant savings of $9.3M/year at a 10M-user scale compared to GPT-4o-mini, while also detecting hallucinations with an AUC of 0.90.