Back to AI News
arxiv

Awaken-HI: Scaling Humanoid Intelligence with Whole-Body Human Experience

Researchers introduce Awaken-Human, a dataset pairing roughly 50,000 hours of egocentric stereo video with synchronized whole-body human motion across more than 100 skills. Building on it, they present Awaken-HI, a vision-language-action model for humanoid loco-manipulation that jointly predicts human motion and retargeted robot actions through a shared action expert. The work reports that direct human trajectory supervision improves transfer, with an empirical scaling law linking more human pretraining data to lower held-out robot action prediction error and higher downstream success. After full-corpus pretraining and T800 adaptation, Awaken-HI reaches a mean success rate of 93.75% across eight tasks and is said to outperform GR00T N1.7 and π_0.5 in individual and long-horizon evaluations.

AI Brief

It offers a large human-motion dataset and a humanoid VLA model with reported real-robot gains over prominent baselines, useful for readers tracking robot learning progress.

arXiv:2611.26420v1 Announce Type: cross Abstract: A central goal of humanoid robotics is for robots to perform the everyday work people do in homes and workplaces. Existing egocentric video datasets cover many of these activities but provide little synchronized whole-body motion data for learning loco-manipulation at scale. We build Awaken-Human, which pairs approximately 50,000 hours of egocentric stereo video with synchronized whole-body human motion and covers more than 100 skills. Building on this, we also introduce Awaken-HI (Humanoid Intelligence), a vision-language-action model for humanoid loco-manipulation that jointly predicts original human motion and retargeted robot actions through a shared action expert. We show that loco-manipulation skills transfer more effectively from people to humanoids when human trajectories directly supervise the shared action expert. Increasing human pretraining data reduces held-out robot action prediction error according to an empirical scaling law and raises the mean success rate on downstream real-robot tasks. After full-corpus pretraining and T800 adaptation, Awaken-HI achieves a mean success rate of 93.75% across eight individual tasks and outperforms GR00T N1.7 and $\pi_{0.5}$ in both individual and long-horizon evaluations. The same policy generalizes to objects unseen during robot post-training and to new pairings of familiar objects and destinations.
Be the first to rate this article.

Comments & Questions (0)

Captcha
Please be respectful — let's keep the conversation friendly.

No comments yet

Be the first to comment!