analiza aiqbench
This research addresses a critical failure mode in diffusion-based robot learning where policies collapse to single modalities despite balanced training data. By introducing a label-free noise assignment strategy, the authors successfully preserve distinct action distributions without altering the core policy architecture or inference steps. The method demonstrates significant gains in recovering non-dominant behaviors across both simulated and real-world humanoid manipulation tasks.
Immiscible Diffusion Policy: Preserving Multimodal Robot Actions through Label-Free Noise Assignment
arXiv:2610.09369v1 Announce Type: cross 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.
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.
Robotics (cs.RO); Machine Learning (cs.LG)
arXiv:2610.09369 [cs.RO]
(or arXiv:2610.09369v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2610.09369 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 307 cuvinte
Subiecte conexe
FA-RDP: A Frequency-Adaptive Reactive Diffusion Policy for Contact-Rich Manipulation
arXiv · 31 iul. 2026
ReactVLA: Fast and Lightweight Reactive Robot Manipulation via Improved Mean Flow Action Generation
arXiv · 16 iun. 2026
CFG-Policy: Unified Diffusion Policy Learning Framework via Classifier-free Guidance
arXiv · 14 oct. 2026
Drift-Based Policy Optimization: Native One-Step Policy Learning for Online Robot Control
arXiv · 2 sept. 2026