1 paper · 1 filter
Byungjun Kim, Taeksoo Kim, Hyunsoo Cha +1
Action-conditioned video world models predict future observations from an initial observation and an action signal. In robotics, actions influence future observations through two d…