PhysBrain 1.0 Technical Report
摘要
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.
引用
@article{arxiv.2605.15298,
title = {PhysBrain 1.0 Technical Report},
author = {Shijie Lian and Bin Yu and Xiaopeng Lin and Changti Wu and Hang Yuan and Xiaolin Hu and Zhaolong Shen and Yuzhuo Miao and Haishan Liu and Yuxuan Tian and Yukun Shi and Cong Huang and Kai Chen},
journal= {arXiv preprint arXiv:2605.15298},
year = {2026}
}
备注
Project Page: https://phys-brain.github.io