1 paper
Tenny Yin, Zhiting Mei, Zhonghe Zheng +8
Action-conditioned video models offer a promising path to building general-purpose robot simulators that can improve directly from data. Yet, despite training on large-scale robot…