1 paper
Yixin Zheng, Jiangran Lyu, Yuntian Deng +6
World Action Models (WAMs) jointly generate robot actions and predict future world states, transferring priors from video pretraining to robot control. However, future visual predi…