FactReview：基于执行验证与证据溯源的论文审稿框架
FactReview: Evidence-Grounded Peer Review with Execution-Based Claim Verification
FactReview 框架通过对照文献核查论文主张、检查推导过程并执行论文代码验证实证结论，在 300 篇论文评测中于 7 项指标中的 5 项排名第一，错误支持率 6.8%、分类准确率 79.2%。在 36 篇论文的用户研究中，使用 FactReview 的审稿人每篇发现 9.4 个问题，独立审稿仅 5.6 个，且耗时减少 74.1%。代码已公开。
Authors:Ling Yue, Chaoqian Ouyang, Ruijun Huang, Hang Xu, Lu Li, Libin Zheng, Yuchen Liu, Wei Liu, Shaowu Pan, Shimin Di, Min-Ling Zhang
Abstract:AI is accelerating scientific research, but providing timely and trustworthy feedback remains challenging. Such feedback requires evaluating specific claims against verified evidence, helping researchers and autonomous research systems determine which findings can guide subsequent work. Thorough verification is time-consuming, and reviewers under time pressure may overlook or misjudge critical issues. Existing AI reviewers have yet to fill this gap because they often assess claims without adequately verifying the supporting evidence. We introduce FactReview, a framework that verifies claims in papers to provide more trustworthy feedback. It checks claims against the literature and examines derivations, while also executing the paper's code to test empirical claims. Across 300 papers, FactReview ranks best among all evaluated systems on five of seven metrics, achieving 6.8% false support and 79.2% classification accuracy. In a study of 36 papers, reviewers using FactReview detect 9.4 issues per paper compared with 5.6 independently, while spending 74.1% less time. FactReview thus supports more reliable peer review, with the potential to accelerate research iteration. The code is public at this https URL.
| Subjects: | Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2604.04074 [cs.AI] |
| (or arXiv:2604.04074v5 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2604.04074 arXiv-issued DOI via DataCite |
Submission history
From: Chaoqian Ouyang [view email]
[v1]
Sun, 5 Apr 2026 11:45:22 UTC (977 KB)
[v2]
Tue, 7 Apr 2026 17:20:55 UTC (971 KB)
[v3]
Wed, 27 May 2026 02:42:30 UTC (10,129 KB)
[v4]
Sun, 16 Aug 2026 07:16:01 UTC (3,515 KB)
[v5]
Fri, 16 Oct 2026 11:06:34 UTC (1,487 KB)
来源：arXiv cs.AI · arxiv.org