The Year in AI Papers
The year/Independent research

Paper 2512.16793

PhysBrain: Human Egocentric Data as a Bridge from Vision Language Models to Physical Intelligence

Xiaopeng Lin, Shijie Lian, Bin Yu, Ruoqi Yang, Zhaolong Shen, Changti Wu, Yuzhuo Miao, Yurun Jin, Yukun Shi, Jiyan He, Cong Huang, Bojun Cheng, et al.

Published
Dec 2025
Research lab
Independent
Citations
8
GitHub
Not linked

01 In brief

Summary

The paper introduces PhysBrain, a model that bridges vision-language models (VLMs) and physical intelligence for robotics by using human egocentric videos.

The authors propose the Egocentric2Embodiment (E2E) Translation Pipeline, which converts raw egocentric videos into multi-level, schema-driven VQA supervision with evidence grounding and temporal consistency, creating the E2E-3M dataset (nearly 3 million instances) from Ego4D, BuildAI, and EgoDex.

PhysBrain is obtained by fine-tuning Qwen3-VL (4B and 8B) on E2E-3M, improving egocentric understanding and planning.

When used as the VLM backbone in a VLA system (PhysVLA), it achieves high success rates on SimplerEnv (67.4% for 8B) and RoboCasa (55.25% for 8B), outperforming baselines trained on large-scale robot data, without robot-data pretraining.

Ablations show consistent gains across architectures and scales, and that scaling egocentric supervision improves performance.

The work demonstrates that human egocentric data can serve as effective supervision for embodied brains, enhancing VLA sample efficiency and generalization.

02 From the paper

Abstract

Robotic generalization relies on physical intelligence: the ability to reason about state changes, contact-rich interactions, and long-horizon planning under egocentric perception and action. Vision Language Models (VLMs) are essential to Vision-Language-Action (VLA) systems, but the reliance on third-person training data creates a viewpoint gap for humanoid robots. Collecting massive robot-centric data is an ideal but impractical solution due to cost and diversity constraints. Conversely, human egocentric videos offer a highly scalable data source with rich interaction context, yet the embodiment mismatch prevents the direct application. To bridge this gap, we propose an Egocentric2Embodiment Translation Pipeline that transforms raw human egocentric videos into multi-level, schema-driven embodiment supervision with enforced evidence grounding and temporal consistency, enabling the construction of the Egocentric2Embodiment dataset (E2E-3M) at scale. An egocentric-aware embodied brain, termed PhysBrain, is obtained by training on the E2E-3M dataset. PhysBrain exhibits substantially improved egocentric understanding, particularly for planning. It provides an egocentric-aware initialization that enables more sample-efficient VLA fine-tuning and higher success rates, demonstrating effective transfer from human egocentric supervision to downstream robot control.