1 paper
Zhenghao Xiao, Minting Pan, Nantian He +2
While action-conditioned video prediction provides an intuitive world model for robotics, purely data-driven predictors often suffer from compounding errors and physically implausi…