ReproBridge：将可复现性内建于 ML 研究工件
ReproBridge: Building Reproducibility into ML Research Artifacts
AI 导读
ReproBridge 是一个用于构建精炼可复现工件（含代码库、环境配置、实验启动器与复现指南）的智能体工作流，配套 ReproBridge-Bench 提供 36 篇论文 387 个实验的 1,029 个任务，每任务含五个人工核验参考分数。
正文 · 原文
Abstract:Reproducing machine learning experiments requires reconstructing procedures from information scattered across papers, released code, and supporting artifacts. For coding agents, reproduction success can therefore depend on their ability to fill in missing details, rather than on the released artifacts alone. We ask whether preserving recovered procedures and decisions improves subsequent independent reproduction. We define a standardized artifact design comprising a refined codebase, environment setup, experiment launchers, and a reproduction guide, and introduce ReproBridge, an agent workflow for constructing these refined reproducible artifacts. ReproBridge-Bench provides 1,029 tasks from 387 experiments across 36 papers, anchored by five human-verified reference scores per task. Agents attempt tasks in fresh sessions with raw code or refined artifacts; a separate Auditor checks execution evidence. Across four evaluated models, refined artifacts raise mean reproduction success from 76.4% to 92.6% and reduce downstream API spending by 19.7% on average. The cross-model success-rate spread narrows from 21.6 to 3.9 percentage points, with all four models achieving mean success rates above 90%. Removing any component reduces success in the ablation study, with the largest losses from code and environment removal. These results suggest that preserving execution knowledge helps even models with lower raw-code success reproduce experiments more reliably and at lower API cost.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2611.08452 [cs.AI] |
| (or arXiv:2611.08452v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2611.08452 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yubo Wang [view email]
[v1]
Thu, 5 Nov 2026 18:45:36 UTC (3,856 KB)
来源：arXiv cs.AI · arxiv.org