VideoWorld 2: Learning Transferable Knowledge from Real-world Videos
Ren, Zhongwei, Wei, Yunchao, Yu, Xiao, Luo, Guixun, Zhao, Yao, Kang, Bingyi, Feng, Jiashi, Jin, Xiaojie
原文摘要
Learning transferable knowledge from unlabeled video data and applying it in new environments is a fundamental capability of intelligent agents. This work presents VideoWorld 2, which extends VideoWorld and provides the first investigation of learning transferable knowledge for complex, long-horizon tasks directly from raw real-world videos. At its core, VideoWorld 2 introduces a dynamics-enhanced Latent Dynamics Model (dLDM) that decouples action dynamics from visual appearance: a pretrained video diffusion model handles visual appearance modeling, enabling the dLDM to learn latent codes that focus on compact and meaningful task-related dynamics. These latent codes are then modeled autoregressively to learn task policies and support long-horizon reasoning. We evaluate VideoWorld 2 on challenging real-world handcraft making tasks, where prior video generation and latent-dynamics models struggle to operate reliably. Remarkably, VideoWorld 2 achieves up to 70% improvement in task success rate and produces coherent long execution videos. In robotics, we show that VideoWorld 2 can acquire transferable manipulation knowledge from the Open-X dataset, which substantially improves task performance on CALVIN, demonstrating strong cross-domain generalization. This study reveals the potential of learning transferable world knowledge directly from raw videos, with all code, data, and models open-sourced for further research.
cs.CV
Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition · 2026 · 正式发表