Principal Researcher, StepFun
1 paper at NeurIPS 2025
We develop a new framework that combines autoregressive modeling and diffusion for video data.