Yiyuan Zhang

Generation-first training eliminates latent collapse, achieving unprecedented stability and performance in latent generative modeling.

CameraAnything enables filmmakers to reshoot videos with arbitrary camera angles and focal lengths in a single generation, revolutionizing video editing workflows.

Accurately simulate complex underwater robots, including soft and underactuated designs, without tedious manual tuning, by jointly optimizing structural and hydrodynamic parameters.

Control hybrid rigid-soft robots with the ease of AR teleoperation, thanks to a new pipeline that accurately models the soft robot's real-world behavior in simulation.