Yuyang Hu
Research Scientist @ Google. Generative models for image and video.
I am a Research Scientist in the Computational Imaging team at Google, working on generative models for images and video, and a core contributor to Gemini Omni, Google’s multimodal video generation and editing model. I received my Ph.D. from Washington University in St. Louis, advised by Prof. Ulugbek Kamilov, and previously worked at Google as a Student Researcher and at Mitsubishi Electric Research Laboratories (MERL) as a Research Intern.
I am particularly interested in the training and distillation of multimodal generative models, and in sampling methods that make them fast and reliable on real-world data. My work spans video and image generation, editing, restoration, and computational imaging.
news
| Aug 21, 2026 | Our paper “Generative Manifold Distillation: Aligning Restoration Trajectories with Natural Image Prior” was accepted to ECCV 2026. |
|---|---|
| Jun 04, 2026 | Gave an invited talk, “Towards Robust Computational Photography: From Hallucination Reduction to Domain Adaptation”, at the 1st Workshop on Low-Level Vision Frontiers (LoViF) at CVPR 2026 in Denver. |
| May 19, 2026 | Gemini Omni, Google’s multimodal video generation and editing model, was announced at Google I/O. I am a core contributor. |
| Apr 13, 2026 | Successfully defended my Ph.D. dissertation at Washington University in St. Louis. Looking forward to the next chapter! |
| Mar 08, 2026 | Our paper “ADOBI: Adaptive Diffusion Bridge for Blind Inverse Problems with Application to MRI Reconstruction” was accepted to IEEE Transactions on Computational Imaging (TCI). |