arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.15922cs.ROcs.CVcs.LG
收藏

世界动作模型是零样本策略

World Action Models are Zero-shot Policies

NVIDIA
Seonghyeon Ye, Yunhao Ge, Kaiyuan Zheng, Shenyuan Gao, Sihyun Yu, George Kurian, Suneel Indupuru, You Liang Tan, Chuning Zhu, Jiannan Xiang, Ayaan Malik, Kyungm… 展开作者

Seonghyeon Ye, Yunhao Ge, Kaiyuan Zheng, Shenyuan Gao, Sihyun Yu, George Kurian, Suneel Indupuru, You Liang Tan, Chuning Zhu, Jiannan Xiang, Ayaan Malik, Kyungmin Lee, William Liang, Nadun Ranawaka, Jiasheng Gu, Yinzhen Xu, Guanzhi Wang, Fengyuan Hu, Avnish Narayan, Johan Bjorck, Jing Wang, Gwanghyun Kim, Dantong Niu, Ruijie Zheng, Yuqi Xie, Jimmy Wu, Qi Wang, Ryan Julian, Danfei Xu, Yilun Du, Yevgen Chebotar, Scott Reed, Jan Kautz, Yuke Zhu, Linxi "Jim" Fan, Joel Jang

AI总结:

DreamZero通过世界动作模型实现零样本策略,有效提升机器人在新任务和环境中的泛化能力,同时支持跨身体转移和少样本适应。

摘要:

最先进的视觉-语言-动作(VLA)模型在语义泛化方面表现出色,但在泛化到新环境中未见过的物理动作时遇到困难。我们介绍了DreamZero,一种基于预训练视频扩散骨干网络构建的世界动作模型(WAM)。与VLA不同,WAM通过预测未来世界状态和动作来学习物理动态,利用视频作为世界如何演变的密集表示。通过同时建模视频和动作,DreamZero能够从异构机器人数据中有效学习多种技能,而无需依赖重复的演示。这导致在真实机器人实验中,其在新任务和环境上的泛化能力比最先进的VLA提高了超过2倍。至关重要的是,通过模型和系统优化,我们使一个14B的自回归视频扩散模型能够以7Hz的速度进行实时闭环控制。最后,我们展示了两种跨身体转移形式:来自其他机器人或人类的视频-only演示在未见过的任务性能上带来了超过42%的相对改进,仅需10-20分钟的数据。更令人惊讶的是,DreamZero能够实现少样本身体适应,仅需30分钟的玩数据即可转移到新身体,同时保持零样本泛化能力。

英文摘要:

State-of-the-art Vision-Language-Action (VLA) models excel at semantic generalization but struggle to generalize to unseen physical motions in novel environments. We introduce DreamZero, a World Action Model (WAM) built upon a pretrained video diffusion backbone. Unlike VLAs, WAMs learn physical dynamics by predicting future world states and actions, using video as a dense representation of how the world evolves. By jointly modeling video and action, DreamZero learns diverse skills effectively from heterogeneous robot data without relying on repetitive demonstrations. This results in over 2x improvement in generalization to new tasks and environments compared to state-of-the-art VLAs in real robot experiments. Crucially, through model and system optimizations, we enable a 14B autoregressive video diffusion model to perform real-time closed-loop control at 7Hz. Finally, we demonstrate two forms of cross-embodiment transfer: video-only demonstrations from other robots or humans yield a relative improvement of over 42% on unseen task performance with just 10-20 minutes of data. More surprisingly, DreamZero enables few-shot embodiment adaptation, transferring to a new embodiment with only 30 minutes of play data while retaining zero-shot generalization.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑