模仿学习策略学会本体感知锚点:感知噪声的预测定律及其闭环代价
Imitation Policies Learn a Proprioceptive Anchor: A Predictive Law for Perception Noise and Its Closed Loop Cost
AI 导读
一项 arXiv 研究揭示,机器人操作的模仿策略会将带噪物体位置读数按权重 α = V/(V+s²) 折扣,其中 V 是本体感知锚点的误差方差。该预测一次测得的 V 可跨 2 个 robomimic 任务、2 个学习器和 4 个新噪声水平预测 10 个留出敏感度,平均绝对误差 0.025,而经典 EIV 回归为 0.37。
正文 · 原文
Abstract:Imitation policies for robot manipulation are often trained on noisy object positions, and they learn to discount the noisy reading. We show how much they discount it, and why. Because demonstrators steer toward the object, the robot's own state carries information about where the object is. A trained policy combines this proprioceptive anchor with the reading and weights the reading by $\alpha = V/(V + s^2)$, where $s^2$ is the reading noise and $V$ the anchor's error variance. $V$ belongs to the learner and is far smaller than a linear model of the demonstrations implies, which is why classical errors-in-variables (EIV) regression overestimates $\alpha$. Measured once, $V$ predicted ten held-out sensitivities across two robomimic tasks, two learners and four new noise levels, with mean absolute error 0.025 against 0.37 for classical EIV, and a parameter-free bias test rules out the simpler explanation that the policies just overestimate the noise. The discount does not carry over to deployment, where the policy produces its own state: closed-loop sensitivity to a persistent reading offset exceeded $\alpha$ in all 15 policies that passed the registered gates, although the stricter registered prediction, that all would follow the offset fully, held for 8 of them. Post hoc, lower $\alpha$ went with lower success, steeply for diffusion, and a registered prospective test confirmed this for diffusion on Can at two new noise levels: the measured $\alpha$ predicted success before any rollout, within 0.03 (95% intervals $\pm 0.08$), although training noise predicts it equally well. Every test was registered before its data, including the classical benchmark that failed, and all other analyses are labelled post hoc; all results are in simulation with synthetic Gaussian noise.
| Comments: | 18 pages, 4 figures |
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2611.17658 [cs.RO] |
| (or arXiv:2611.17658v1 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2611.17658 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Christopher Malili [view email]
[v1]
Fri, 13 Nov 2026 02:09:51 UTC (143 KB)
来源:arXiv cs.AI · arxiv.org