外观
论文arXiv 2608.075423 个月前
不漂移的 AI Scientist：四足导航研究循环中的品味、结构与可证伪发现
原文链接：An AI Scientist that Doesn't Drift: Taste, Structure, and Falsifiable Findings in a Quadruped Navigation Research Loop（arXiv 2608.07542 · 发布日期：2026-07-30）
An AI Scientist that Doesn't Drift: Taste, Structure, and Falsifiable Findings in a Quadruped Navigation Research Loop
提出结构化AI Scientist研究仿真四足导航泛化，用实验卡、子代理和kkanbu在十一条研究流验证。
核心要点
- 动机是避免大语言模型驱动的自主研究循环偏向优化指标的局部改进，而非检验原始假设。
- 方法加入不可变实验卡、受限机械角色的专门子代理，以及以类型化知识图保存用户研究品味的kkanbu偏好预言机。
- 在十一条研究流中分别运行有无kkanbu的相同循环；两臂都未漂移，并各自证伪约四分之三假设。
- kkanbu改变研究方向而非分数：它探索测试时适应，并在其领先时写出获胜设计、跨研究流迁移经验。
验证范围与局限：验证范围为仿真中的四足机器人导航策略泛化研究，并比较有无kkanbu的十一条研究流。
英文摘要 (Abstract)
Autonomous research loops driven by large language models can run machine-learning experiments at scale but tend to drift toward local refinements of whichever metric they optimise rather than testing the hypotheses that motivate the experiments. We address this structurally and present an AI Scientist for studying generalisation in quadruped robot navigation policies in simulation. Building on the autoresearch paradigm of Karpathy, our loop adds three components: an immutable experiment card that pairs each iteration's prediction with its outcome under a fixed schema, so a falsified hypothesis cannot be retconned; specialised subagents restricted to mechanical roles; and kkanbu, a preference oracle that holds the user's research taste as a typed knowledge graph and is the only component permitted to make subjective judgements. To isolate the oracle we run the identical loop twice across eleven research streams, with and without kkanbu. Neither arm drifts: both falsify roughly three quarters of their own hypotheses, and the best trained policy comes from the oracle-less arm. What the oracle changes is direction, not score: it alone explores test-time adaptation, it authored the winning designs where its arm led, and it carried lessons across streams that the other arm repeatedly re-derived. The scaffold keeps the loop honest; kkanbu decides where it looks.
作者：Yiwen Zhang, Eloise Zeng, Jaeha Lee, Tony Yue Yu
来源
本页事实对应的公开页面，以原文为准。