Learning Generalizable Visuomotor Policy through Dynamics-Alignment
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866918180312580096 |
|---|---|
| author | Lee, Dohyeok Lee, Jung Min Kim, Munkyung Ju, Seokhun Koo, Jin Woo Lee, Kyungjae Kim, Dohyeong Cho, TaeHyun Lee, Jungwoo |
| author_facet | Lee, Dohyeok Lee, Jung Min Kim, Munkyung Ju, Seokhun Koo, Jin Woo Lee, Kyungjae Kim, Dohyeong Cho, TaeHyun Lee, Jungwoo |
| contents | Behavior cloning methods for robot learning suffer from poor generalization due to limited data support beyond expert demonstrations. Recent approaches leveraging video prediction models have shown promising results by learning rich spatiotemporal representations from large-scale datasets. However, these models learn action-agnostic dynamics that cannot distinguish between different control inputs, limiting their utility for precise manipulation tasks and requiring large pretraining datasets. We propose a Dynamics-Aligned Flow Matching Policy (DAP) that integrates dynamics prediction into policy learning. Our method introduces a novel architecture where policy and dynamics models provide mutual corrective feedback during action generation, enabling self-correction and improved generalization. Empirical validation demonstrates generalization performance superior to baseline methods on real-world robotic manipulation tasks, showing particular robustness in OOD scenarios including visual distractions and lighting variations. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_27114 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Learning Generalizable Visuomotor Policy through Dynamics-Alignment Lee, Dohyeok Lee, Jung Min Kim, Munkyung Ju, Seokhun Koo, Jin Woo Lee, Kyungjae Kim, Dohyeong Cho, TaeHyun Lee, Jungwoo Robotics Machine Learning Behavior cloning methods for robot learning suffer from poor generalization due to limited data support beyond expert demonstrations. Recent approaches leveraging video prediction models have shown promising results by learning rich spatiotemporal representations from large-scale datasets. However, these models learn action-agnostic dynamics that cannot distinguish between different control inputs, limiting their utility for precise manipulation tasks and requiring large pretraining datasets. We propose a Dynamics-Aligned Flow Matching Policy (DAP) that integrates dynamics prediction into policy learning. Our method introduces a novel architecture where policy and dynamics models provide mutual corrective feedback during action generation, enabling self-correction and improved generalization. Empirical validation demonstrates generalization performance superior to baseline methods on real-world robotic manipulation tasks, showing particular robustness in OOD scenarios including visual distractions and lighting variations. |
| title | Learning Generalizable Visuomotor Policy through Dynamics-Alignment |
| topic | Robotics Machine Learning |
| url | https://arxiv.org/abs/2510.27114 |