Hang Gao

I am a final-year Ph.D. at UC Berkeley, advised by Angjoo Kanazawa. My research revolves around neural rendering and deep generative models.

I studied at Jiao Tong University and Columbia. I have spent time working as a student researcher at Stability, Adobe and Microsoft.

I am on the industrial job market this spring (2025).

I work on high-quality 3D rendering in real-world settings. My research focuses on scaling and accelerating neural rendering and diffusion models, with an emphasis on post-training for camera control in video generation.

We recover complete human avatars from self-occluded videos in the wild.

We propose a method for joint 4D reconstruction and 3D tracking that works reasonably well in-the-wild.

We build and release a toolbox for accelerating all kinds of NeRFs by efficient sampling.

We show that existing 4D reconstruction methods are not working well in-the-wild.

We predict long-term, diverse human motion in 3D by understanding scene context from an image.

We propose a new convolutional operator that does attention in the kernel space.

We find relational graph useful for action recognition.

We make a system that propagates seen parts and generates unseen ones.

We learn feature augmentation for low-shot classifiers.

We propose a weakly-supervised method for temporal action localization.

We build an audio-based mobile app for inattentive driving detection.

Thanks Jon!
Last updated Jan 2025.