Physics-Informed Video Generation via Mixture-of-Experts Latent Alignment

Cong Wang, Hanxin Zhu, Jiayi Luo +6

Large-scale video generation models have made remarkable progress in semantic consistency and visual quality, producing videos that are increasingly coherent and visually convincin…