Saved in:
| Main Authors: | , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2602.11075 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908996655382528 |
|---|---|
| author | Yang, Jiazhi Lin, Kunyang Li, Jinwei Zhang, Wencong Lin, Tianwei Wu, Longyan Su, Zhizhong Zhao, Hao Zhang, Ya-Qin Chen, Li Luo, Ping Yue, Xiangyu Li, Hongyang |
| author_facet | Yang, Jiazhi Lin, Kunyang Li, Jinwei Zhang, Wencong Lin, Tianwei Wu, Longyan Su, Zhizhong Zhao, Hao Zhang, Ya-Qin Chen, Li Luo, Ping Yue, Xiangyu Li, Hongyang |
| contents | Despite the sustained scaling on model capacity and data acquisition, Vision-Language-Action (VLA) models remain brittle in contact-rich and dynamic manipulation tasks, where minor execution deviations can compound into failures. While reinforcement learning (RL) offers a principled path to robustness, on-policy RL in the physical world is constrained by safety risk, hardware cost, and environment reset. To bridge this gap, we present RISE, a scalable framework of robotic reinforcement learning via imagination. At its core is a Compositional World Model that (i) predicts multi-view future via a controllable dynamics model, and (ii) evaluates imagined outcomes with a progress value model, producing informative advantages for the policy improvement. Such compositional design allows state and value to be tailored by best-suited yet distinct architectures and objectives. These components are integrated into a closed-loop self-improving pipeline that continuously generates imaginary rollouts, estimates advantages, and updates the policy in imaginary space without costly physical interaction. Across three challenging real-world tasks, RISE yields significant improvement over prior art, with more than +35% absolute performance increase in dynamic brick sorting, +45% for backpack packing, and +35% for box closing, respectively. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_11075 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | RISE: Self-Improving Robot Policy with Compositional World Model Yang, Jiazhi Lin, Kunyang Li, Jinwei Zhang, Wencong Lin, Tianwei Wu, Longyan Su, Zhizhong Zhao, Hao Zhang, Ya-Qin Chen, Li Luo, Ping Yue, Xiangyu Li, Hongyang Robotics Despite the sustained scaling on model capacity and data acquisition, Vision-Language-Action (VLA) models remain brittle in contact-rich and dynamic manipulation tasks, where minor execution deviations can compound into failures. While reinforcement learning (RL) offers a principled path to robustness, on-policy RL in the physical world is constrained by safety risk, hardware cost, and environment reset. To bridge this gap, we present RISE, a scalable framework of robotic reinforcement learning via imagination. At its core is a Compositional World Model that (i) predicts multi-view future via a controllable dynamics model, and (ii) evaluates imagined outcomes with a progress value model, producing informative advantages for the policy improvement. Such compositional design allows state and value to be tailored by best-suited yet distinct architectures and objectives. These components are integrated into a closed-loop self-improving pipeline that continuously generates imaginary rollouts, estimates advantages, and updates the policy in imaginary space without costly physical interaction. Across three challenging real-world tasks, RISE yields significant improvement over prior art, with more than +35% absolute performance increase in dynamic brick sorting, +45% for backpack packing, and +35% for box closing, respectively. |
| title | RISE: Self-Improving Robot Policy with Compositional World Model |
| topic | Robotics |
| url | https://arxiv.org/abs/2602.11075 |