Two-Steps Diffusion Policy for Robotic Manipulation via Genetic Denoising
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908611406462976 |
|---|---|
| author | Clemente, Mateo Brunswic, Leo Yang, Rui Heng Zhao, Xuan Khalil, Yasser Lei, Haoyu Rasouli, Amir Li, Yinchuan |
| author_facet | Clemente, Mateo Brunswic, Leo Yang, Rui Heng Zhao, Xuan Khalil, Yasser Lei, Haoyu Rasouli, Amir Li, Yinchuan |
| contents | Diffusion models, such as diffusion policy, have achieved state-of-the-art results in robotic manipulation by imitating expert demonstrations. While diffusion models were originally developed for vision tasks like image and video generation, many of their inference strategies have been directly transferred to control domains without adaptation. In this work, we show that by tailoring the denoising process to the specific characteristics of embodied AI tasks -- particularly structured, low-dimensional nature of action distributions -- diffusion policies can operate effectively with as few as 5 neural function evaluations (NFE).
Building on this insight, we propose a population-based sampling strategy, genetic denoising, which enhances both performance and stability by selecting denoising trajectories with low out-of-distribution risk. Our method solves challenging tasks with only 2 NFE while improving or matching performance. We evaluate our approach across 14 robotic manipulation tasks from D4RL and Robomimic, spanning multiple action horizons and inference budgets. In over 2 million evaluations, our method consistently outperforms standard diffusion-based policies, achieving up to 20\% performance gains with significantly fewer inference steps. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_21991 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Two-Steps Diffusion Policy for Robotic Manipulation via Genetic Denoising Clemente, Mateo Brunswic, Leo Yang, Rui Heng Zhao, Xuan Khalil, Yasser Lei, Haoyu Rasouli, Amir Li, Yinchuan Robotics Artificial Intelligence 68T40, 93C85, 68T07, 68U35 Diffusion models, such as diffusion policy, have achieved state-of-the-art results in robotic manipulation by imitating expert demonstrations. While diffusion models were originally developed for vision tasks like image and video generation, many of their inference strategies have been directly transferred to control domains without adaptation. In this work, we show that by tailoring the denoising process to the specific characteristics of embodied AI tasks -- particularly structured, low-dimensional nature of action distributions -- diffusion policies can operate effectively with as few as 5 neural function evaluations (NFE). Building on this insight, we propose a population-based sampling strategy, genetic denoising, which enhances both performance and stability by selecting denoising trajectories with low out-of-distribution risk. Our method solves challenging tasks with only 2 NFE while improving or matching performance. We evaluate our approach across 14 robotic manipulation tasks from D4RL and Robomimic, spanning multiple action horizons and inference budgets. In over 2 million evaluations, our method consistently outperforms standard diffusion-based policies, achieving up to 20\% performance gains with significantly fewer inference steps. |
| title | Two-Steps Diffusion Policy for Robotic Manipulation via Genetic Denoising |
| topic | Robotics Artificial Intelligence 68T40, 93C85, 68T07, 68U35 |
| url | https://arxiv.org/abs/2510.21991 |