Saved in:
| Main Authors: | , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2605.15298 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916013822443520 |
|---|---|
| author | Lian, Shijie Yu, Bin Lin, Xiaopeng Wu, Changti Yuan, Hang Hu, Xiaolin Shen, Zhaolong Miao, Yuzhuo Liu, Haishan Tian, Yuxuan Shi, Yukun Huang, Cong Chen, Kai |
| author_facet | Lian, Shijie Yu, Bin Lin, Xiaopeng Wu, Changti Yuan, Hang Hu, Xiaolin Shen, Zhaolong Miao, Yuzhuo Liu, Haishan Tian, Yuxuan Shi, Yukun Huang, Cong Chen, Kai |
| contents | Vision-language-action models have advanced rapidly, but robot trajectories alone provide limited coverage for learning broad physical understanding. PhysBrain 1.0 studies a complementary route: converting large-scale human egocentric video into structured physical commonsense supervision before robot adaptation. Our data engine extracts scene elements, spatial dynamics, action execution, and depth-aware relations, then turns them into question-answer supervision for training PhysBrain VLMs. The resulting physical priors are further transferred to VLA policies through a capability-preserving and language-sensitive adaptation design. Across multimodal QA benchmarks and embodied control benchmarks, including ERQA, PhysBench, SimplerEnv-WidowX, LIBERO, and RoboCasa, PhysBrain 1.0 achieves SOTA results and shows especially strong out-of-domain performance on SimplerEnv. These results suggest that scaling physical commonsense from human interaction video can provide an effective bridge from multimodal understanding to robot action. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_15298 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | PhysBrain 1.0 Technical Report Lian, Shijie Yu, Bin Lin, Xiaopeng Wu, Changti Yuan, Hang Hu, Xiaolin Shen, Zhaolong Miao, Yuzhuo Liu, Haishan Tian, Yuxuan Shi, Yukun Huang, Cong Chen, Kai Robotics Artificial Intelligence Computation and Language Computer Vision and Pattern Recognition Vision-language-action models have advanced rapidly, but robot trajectories alone provide limited coverage for learning broad physical understanding. PhysBrain 1.0 studies a complementary route: converting large-scale human egocentric video into structured physical commonsense supervision before robot adaptation. Our data engine extracts scene elements, spatial dynamics, action execution, and depth-aware relations, then turns them into question-answer supervision for training PhysBrain VLMs. The resulting physical priors are further transferred to VLA policies through a capability-preserving and language-sensitive adaptation design. Across multimodal QA benchmarks and embodied control benchmarks, including ERQA, PhysBench, SimplerEnv-WidowX, LIBERO, and RoboCasa, PhysBrain 1.0 achieves SOTA results and shows especially strong out-of-domain performance on SimplerEnv. These results suggest that scaling physical commonsense from human interaction video can provide an effective bridge from multimodal understanding to robot action. |
| title | PhysBrain 1.0 Technical Report |
| topic | Robotics Artificial Intelligence Computation and Language Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2605.15298 |