HuRo: Robotizing Human Videos for Scalable VLA Pretraining
Human video datasets offer an abundant and diverse source of interaction data that can complement expensive real-robot data. To bridge the human-to-robot embodiment gap, existing approaches either robotize videos in task-matched settings or address observation and action alignment separately at scale. In this work, we systematically examine whether robotized human videos can se
Reading in Plain English mode — toggle below to switch
Human video datasets offer an abundant and diverse source of interaction data that can complement expensive real-robot data. To bridge the human-to-robot embodiment gap, existing approaches either robotize videos in task-matched settings or address observation and action alignment separately at scale. In this work, we systematically examine whether robotized human videos can se
This work addresses a critical challenge in robotics: leveraging abundant human video data for training Vision-Language-Action (VLA) models. By robotizing human videos, it offers a scalable and cost-effective alternative to expensive real-robot data, accelerating the development of more capable robotic agents.