arXiv cs.AI· Haojin Wang, Yangzhen Wu, Zeyi Chen, Jian Xie, Jiaying Ye, Lili Meng, Shuyun Tang, Shangding Gu, Dawn Song·· 3 小时前AI 评分53
ParetoRSI-Gym 论文提出多领域递归自我改进评估框架
ParetoRSI-Gym: Can AI Agents Improve Language Models Across Multiple Domains?
AI 导读
论文提出 ParetoRSI-Gym，让智能体在数学、代码、金融、医学、科学五个领域中搜索并筛选数据、训练并评估自身后继模型，除数据外其余环节固定。
正文 · 原文
Abstract:Recursive self-improvement (RSI) requires an agentic system to design, train, and evaluate its own successor. Data-driven post-training is a central part of that loop, yet existing evaluations measure it on a single benchmark, where an agent selects from a static pool or distills a stronger model to push one score up in one run. Real post-training is multi-domain, and a gain in one domain has to be weighed against a regression in another. We propose ParetoRSI-Gym, in which an agent searches for data, curates it, trains, reads back per-domain results, and repeats, with everything but the data held fixed: the initial LLM, the five domains (math, code, finance, medicine, science), the toolset, and the SFT pipeline. Generalization benchmarks in each domain separate capability gains from overfitting to the scores the agent optimizes against. Across three agent systems, three base models, five rounds and multiple seeds, the strongest system, Claude Code with Opus-5, raises the generalization scores of its proxy-best checkpoint by 9.2% relative to the starting model (+2.3 absolute points), averaged with equal weight over three base models and five domains, with a single-domain relative gain of up to 63% (medicine on Ministral-3-3B). Further analysis shows that the harness matters as much as the model it drives, and that each agent keeps a unique data selection strategy. We see our work as a step towards evaluation of RSI systems in real-world machine learning engineering.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.34386 [cs.LG] |
| (or arXiv:2610.34386v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.34386 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: HaoJin Wang [view email]
[v1]
Sun, 25 Oct 2026 15:40:21 UTC (1,058 KB)
来源：arXiv cs.AI · arxiv.org