Scholar
Xingchao Liu
Google Scholar ID: VOTVE0UAAAAJ
DeepSeek AI
Citations & Impact
Citations
10,896
H-index
25
i10-index
28
Publications
20
Co-authors
10
list available
Contact
Publications
Resume (English only)
Academic Achievements
- Involved in the development of the Janus series models (including Janus and JanusFlow), which aim to unify multimodal understanding and generation within the same model. Also developed the Rectified Flow framework, suitable for general distribution matching problems including generative modeling and unsupervised data transfer. Based on this, he and his collaborators have trained fast Stable Diffusion models such as InstaFlow and PeRFlow.
Research Experience
- Currently a researcher in the multimodal group at DeepSeek AI. Prior to that, he was pursuing his Ph.D. at the University of Texas at Austin.
Education
- Ph.D. from the University of Texas at Austin, advised by Prof. Qiang Liu; Undergraduate study at Beihang University, also worked with Prof. Hao Su at UCSD.
Background
- A machine learning researcher specializing in probabilistic inference and generative modeling, with a particular emphasis on their applications in multimodal intelligence.
Miscellany
- His personal website provides links to his research projects, GitHub repositories, and online demo platforms.
