具身开发者社区

外观

论文arXiv 2610.0936921 天前

Immiscible Diffusion Policy：通过无标签噪声分配保留多模态机器人动作 ​

针对扩散策略动作模态坍缩，提出无标签动作-噪声分配，并在仿真与真实人形操作任务中验证。

AI 整理
  • 本体 SDK 与控制
  • 数据与采集
  • 具身模型

原文链接：Immiscible Diffusion Policy: Preserving Multimodal Robot Actions through Label-Free Noise Assignment（arXiv 2610.09369 · 发布日期：2026-10-07）

Immiscible Diffusion Policy: Preserving Multimodal Robot Actions through Label-Free Noise Assignment

针对扩散策略动作模态坍缩，提出无标签动作-噪声分配，并在仿真与真实人形操作任务中验证。

核心要点 ​

  • 扩散策略即使在模态平衡和批内对称时，也可能坍缩到单一动作模态。
  • 分析认为独立动作-噪声配对会增加路径混合与交叉，抑制模态特异行为。
  • 该训练期附加方法分配动作与噪声，不改变策略架构或推理流程。
  • 在五个仿真和两个真实任务中，提高动作模态保留并保持较强任务性能。

验证范围与局限：验证范围为五个仿真和两个真实世界人形操作任务，观测包括状态、RGB和点云。

英文摘要 (Abstract) ​

When diffusion policies were first introduced, they were expected to recover multi-modal action distributions. However, we find this expectation does not always hold, as diffusion policies often collapse to a single modality even when we guarantee the balance of dataset modalities and exact within-batch symmetry. Our analysis indicates that independent action-noise pairing contributes to this failure by increasing mixing and crossing among diffusion paths, which can produce averaged denoising responses and suppress modality-specific behavior. This issue is especially severe in robot planning, where action spaces are dense and low-dimensional, significantly increasing such mixing and crossing. To alleviate this problem, we propose Immiscible Diffusion Policy, a label-free training-time add-on to diffusion policy that uses action-noise assignment to preserve relatively distinct noise-to-action routes without modifying the policy architecture or inference procedure. Across five simulated and two real-world humanoid manipulation tasks spanning state, RGB, and point-cloud observations, our method significantly improves the policy's preservation of action modalities while maintaining strong task performance. It increases the proportion of the non-dominant modality by 6.0x-14.6x across three two-modality tasks and recovers demonstrated modalities that are entirely absent from vanilla policy rollouts on both four-modality tasks. These results demonstrate that Immiscible Diffusion Policy provides a simple yet robust approach to preserving action multi-modality in general robot learning tasks.

作者：Xiao Zhang, Yuxin Chen, Zhixuan Liang, Guojian Zhan, Chenran Li, Chenfeng Xu, Masayoshi Tomizuka, Yiheng Li

来源

本页事实对应的公开页面，以原文为准。

  1. arXiv 2610.09369：Immiscible Diffusion Policy: Preserving Multimodal Robot Actions through Label-Free Noise Assignment访问 2026-10-21