- 'Shallow Diffuse: Robust and Invisible Watermarking through Low-Dimensional Subspaces in Diffusion Models' was accepted by NeurIPS 2025 and selected as Spotlight (top 3.2%);
- 'Model Collapse from MtoG Perspective' was accepted by NeurIPS 2025 and selected as Spotlight (top 3.2%);
- Published papers such as 'AlphaFlow: Understanding and Improving MeanFlow Models'.
Research Experience
- Research Intern at Snap Inc., Creative Vision Team, Supervisors: Ivan Skorokhodov, Aliaksandr Siarohin, and Sergey Tulyakov.
Education
- University of Michigan, Ann Arbor, Ph.D. student, Advisor: Prof. Qing Qu;
- University of Michigan, Ann Arbor, Master's degree, Advisor: Prof. Chad Jenkins;
- Huazhong University of Science and Technology, Bachelor's degree, Advisor: Prof. Zhigang Wu.
Background
Research Interests: Generative models, with a particular focus on developing a rigorous theoretical understanding of diffusion models and leveraging these insights to drive practical advancements. Professional Field: Mechanical Engineering, Electrical and Computer Engineering.