Xiaogang Xu

aka Theo
>

Ensure that artificial general intelligence benefits all of humanity.

— OpenAI Mission · the north star I work toward
~$
“A great man can bend in adversity and rise when his moment comes.”
Xiaogang Xu
Research Journey
Research journey timeline

Biography

I obtained my Ph.D. degree in the Department of Computer Science and Engineering at the Chinese University of Hong Kong, supervised by Prof. Jiaya Jia and Prof. Bei Yu. Before that, I obtained my B.E. degree in Information Engineering at the College of Information Science and Electronic Engineering, Zhejiang University.

During my Ph.D. life, I have spent wonderful times collaborating with, among others:
Academia
Prof. Chi-Wing FuCUHK Prof. Yingcong ChenHKUST Prof. Hengshuang ZhaoHKU Prof. Philip H. S. TorrOxford Prof. Antonio TorralbaMIT Prof. Christian TheobaltMPI-INF Dr. Thomas LeimkuehlerMPI-INF
Industry
Dr. Jiangbo LuSmartMore Dr. Nianjuan JiangSmartMore Dr. Ning XuAdobe Research Dr. Vibhav VineetMicrosoft Research

In the industrial community, I conduct research and development on AGI.

Zhejiang Lab
I primarily led the construction of a 10,000-card GPU cluster and spearheaded the training of domain-specific large science models, in close collaboration with Academician Jian Wang, Prof. Hujun Bao, and Prof. Zhe Liu.
Huawei
My research and development focused on training unified inference models for generation and understanding, as well as advancing AIGC models including text-to-image and text-to-video. I successfully drove the deployment of these models into core Huawei products, and established strong academic collaborations with Prof. Hanwang Zhang from NTU Singapore.
MiroMind
I dedicated myself to the post-training of agentic foundation models and the development of self-evolving harnesses, working in close partnership with Prof. Shuicheng Yan from NUS Singapore.
In parallel, I have been a visiting researcher at a number of academic institutions. The full record of these positions is in Experiences.
Several specific topics of our current research interests and focus:
01Multi-modality data (image, video, 3D, etc.) generation & manipulation via AIGC
02Multi-modality large model and agent → AGI
03Generative Computational Photography: large model and efficiency optimization
04Security and alignment for large models / AGI
05Embodied AI (VLA, VA, etc.) and World Model
I carry out development and research in the industrial and academic communities at the same time. If you are interested in collaborating with me, please feel free to contact me through the email.
Selected Collaborations
Zhejiang University
University of Oxford
CUHK
Adobe
MIT
Huawei
HKUST
Alibaba
Max Planck Institute for Informatics
Meta
Peking University
Microsoft
National University of Singapore
ByteDance
Tsinghua University
Tencent
The University of Hong Kong
Snapchat
Nanyang Technological University
SmartMore
Sun Yat-sen University
SenseTime
Zhejiang Lab
MiroMind
Collaborator

News

Research Summary

Multi-modality Generation
Multi-modality Restoration
Multi-modality Understanding