| title | Magic-W0: A Structured World-Action Foundation Model for Physical Intelligence |
|---|---|
| authors | Xuhua Chen, Zhenhan Yin, Yuan Zhang, Lingfeng Zhang, He Zheng, Tong Mu, Shun Zuo, Dian Zhou, Di Wu, Xuan Zhou, Shaojie Wan, Rongtian Shen, Qiulong Xu, Yiduo Li, Yinglong Wang, Yanqian Wang, Kun Wang, Tao Zhang |
| date | 2026-10-03 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2609.39870 |
| upvotes | 2 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf code |
| abstract | World-action models (WAMs) augment robot policies with action-conditioned environment dynamics, yet existing approaches largely rely on future observation reconstruction or generic latent prediction and lack structured, control-oriented world representations tightly coupled with action generation. We introduce Magic-W0, a world-action foundation model that jointly models structured physical state evolution and continuous actions. Magic-W0 represents interaction as a Structured World Transition consisting of Current State, Transition, and Future State. Current State combines vision-language context with Current 3D Geometry; Transition is represented by 3D Motion capturing action-induced three-dimensional changes; and Future State is represented by Future Semantics describing task-relevant outcomes. To couple prediction and control, we propose a layer-aligned world-action interaction architecture in which evolving action hypotheses condition world-transition prediction, while predicted world representations continuously inform action generation. Magic-W0 is pre-trained on large-scale egocentric human manipulation, UMI, real-robot, and simulation data, with latent supervision for geometry, 3D motion, and future semantics from pre-trained visual models. Inference-time interventions show that structured world representations respond systematically to changes in candidate actions and that action-related information propagates through shared 3D representations into future semantic predictions. On RoboDojo-Sim, Magic-W0 achieves an average Score of 27.10, the highest among the compared WAMs. Across multiple real-robot tasks, it also demonstrates strong downstream performance after fine-tuning with limited downstream data, supporting generalization and rapid adaptation. |
cite
@misc{chen2026magic,
title={Magic-W0: A Structured World-Action Foundation Model for Physical Intelligence},
author={Xuhua Chen and Zhenhan Yin and Yuan Zhang and Lingfeng Zhang and He Zheng and Tong Mu and Shun Zuo and Dian Zhou and Di Wu and Xuan Zhou and Shaojie Wan and Rongtian Shen and Qiulong Xu and Yiduo Li and Yinglong Wang and Yanqian Wang and Kun Wang and Tao Zhang},
year={2026},
eprint={2609.39870},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2609.39870}
}