Preprint
Oct 2026
World Action Learning via Interaction-Centric Spectral Latent Guidance
Learning general-purpose robot policies requires large-scale real-world interaction data, yet collecting robot demonstrations remains expensive and difficult to scale. Egocentric videos offer abundant human interaction experience with task-relevant semantics for robotic manipulation, but direct transfer is challenging...
Zhi-Ming Liu, Yi-Kun Miao, Ying Chen et al.
· 0 citations