WarmPrior: Straightening Flow-Matching Policies with Temporal Priors
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866916011722145792 |
|---|---|
| author | Kang, Sinjae Kim, Chanyoung Wang, Kaixin Zhao, Li Lee, Kimin |
| author_facet | Kang, Sinjae Kim, Chanyoung Wang, Kaixin Zhao, Li Lee, Kimin |
| contents | Generative policies based on diffusion and flow matching have become a dominant paradigm for visuomotor robotic control. We show that replacing the standard Gaussian source distribution with WarmPrior, a simple temporally grounded prior constructed from readily available recent action history, consistently improves success rates on robotic manipulation tasks. We trace this gain to markedly straighter probability paths, echoing the effect of optimal-transport couplings in Rectified Flow. Beyond standard behavior cloning, WarmPrior also reshapes the exploration distribution in prior-space reinforcement learning, improving both sample efficiency and final performance. Collectively, these results identify the source distribution as an important and underexplored design axis in generative robot control. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_13959 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | WarmPrior: Straightening Flow-Matching Policies with Temporal Priors Kang, Sinjae Kim, Chanyoung Wang, Kaixin Zhao, Li Lee, Kimin Machine Learning Artificial Intelligence Robotics Generative policies based on diffusion and flow matching have become a dominant paradigm for visuomotor robotic control. We show that replacing the standard Gaussian source distribution with WarmPrior, a simple temporally grounded prior constructed from readily available recent action history, consistently improves success rates on robotic manipulation tasks. We trace this gain to markedly straighter probability paths, echoing the effect of optimal-transport couplings in Rectified Flow. Beyond standard behavior cloning, WarmPrior also reshapes the exploration distribution in prior-space reinforcement learning, improving both sample efficiency and final performance. Collectively, these results identify the source distribution as an important and underexplored design axis in generative robot control. |
| title | WarmPrior: Straightening Flow-Matching Policies with Temporal Priors |
| topic | Machine Learning Artificial Intelligence Robotics |
| url | https://arxiv.org/abs/2605.13959 |