arXiv：cs.CL· Haishuo Fang, Yue Feng, Iryna Gurevych·· 16 小时前
ProReviewer：面向科学同行评审的证据驱动审查智能体
ProReviewer: Evidence-Driven Investigation for Scientific Peer Review
AI 导读
ProReviewer 将同行评审建模为马尔可夫决策过程，以结构化评审日志作为状态核心，智能体依据累积发现决定下一步检查什么、何时结束，最终每条评审意见均引用日志条目以追溯证据。在 1,000 篇 ICLR 2026 留出论文上，训练后的 8B ProReviewer 平均得分超过所有对比方法，包括基于 397B 模型的方法，人工评估在四个评审质量维度中的三个上更偏好其评审。
正文
Abstract:Effective peer review requires adaptive investigation that responds to evidence emerging from each paper. Existing LLM-based reviewers largely organize the review process around predefined workflows, leaving limited scope for emerging evidence to reshape investigation. We propose ProReviewer, an evidence-driven review agent that formulates reviewing as a Markov Decision Process, with a structured review log forming the core of its state. At each step, the agent uses accumulated findings to decide what to inspect next and when to conclude. As new evidence emerges, it updates the log by revising claim statuses, resolving open questions, and recording new findings. Each point in the final review cites supporting log entries, making its judgments traceable to the recorded evidence. On 1,000 held-out ICLR 2026 papers, the trained 8B ProReviewer achieves the highest average score among all compared methods, including those built on a 397B model; human evaluators prefer its reviews over all baselines on three of four review-quality dimensions.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2606.13349 [cs.CL] |
| (or arXiv:2606.13349v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2606.13349 arXiv-issued DOI via DataCite |
Submission history
From: Haishuo Fang [view email]
[v1]
Thu, 11 Jun 2026 13:38:23 UTC (1,866 KB)
[v2]
Wed, 14 Oct 2026 13:33:49 UTC (904 KB)
来源：arXiv：cs.CL · arxiv.org