arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2512.01809cs.ROcs.LG
收藏

许多关于噪声的争论:消解生成机器人控制的神话

Much Ado About Noising: Dispelling the Myths of Generative Robotic Control

Chaoyi Pan, Giri Anantharaman, Nai-Chieh Huang, Claire Jin, Daniel Pfrommer, Chenyang Yuan, Frank Permenter, Guannan Qu, Nicholas Boffi, Guanya Shi, Max Simchowitz

AI总结:

本文通过评估生成控制策略在行为克隆基准上的表现,发现其成功主要源于迭代计算和监督训练,而非多模态捕捉或复杂映射表达。

摘要:

生成模型,如流和扩散,最近在机器人学中作为流行且有效的策略参数化方法出现。人们对于其成功因素有很多猜测,从捕捉多模态动作分布到表达更复杂的行为。在本工作中,我们对流行的生成控制策略(GCPs)在常见行为克隆(BC)基准上进行了全面评估。我们发现,GCPs的成功并非源于其捕捉多模态的能力或表达更复杂观察到动作映射的能力。相反,我们发现其优势来自于迭代计算,只要在训练过程中中间步骤受到监督,并且这种监督与适当的随机性相结合。作为我们发现的验证,我们展示了最小迭代策略(MIP),一种轻量级的两步回归策略,实际上与流GCPs的性能相当,经常优于提炼的快捷模型。我们的结果表明,GCPs的分布拟合组件比人们普遍认为的要不显眼,并指向仅专注于控制性能的新设计空间。项目页面:https://simchowitzlabpublic.github.io/much-ado-about-noising-project/

英文摘要:

Generative models, like flows and diffusions, have recently emerged as popular and efficacious policy parameterizations in robotics. There has been much speculation as to the factors underlying their successes, ranging from capturing multi-modal action distribution to expressing more complex behaviors. In this work, we perform a comprehensive evaluation of popular generative control policies (GCPs) on common behavior cloning (BC) benchmarks. We find that GCPs do not owe their success to their ability to capture multi-modality or to express more complex observation-to-action mappings. Instead, we find that their advantage stems from iterative computation, as long as intermediate steps are supervised during training and this supervision is paired with a suitable level of stochasticity. As a validation of our findings, we show that a minimum iterative policy (MIP), a lightweight two-step regression-based policy, essentially matches the performance of flow GCPs, and often outperforms distilled shortcut models. Our results suggest that the distribution-fitting component of GCPs is less salient than commonly believed, and point toward new design spaces focusing solely on control performance. Project page: https://simchowitzlabpublic.github.io/much-ado-about-noising-project/

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑