Zhiqi Li

I'm a first-year Ph.D. student in Computer Science at Georgia Tech, supervised by Prof. Bo Zhu and Prof. Gred Turk, focusing on physical-based simulation in computer graphics. I'm also interested in deep learning theory and seek to apply deep learning technique to physical-based simulation. Previously, I was a research intern in Internet Graphics (IG) Group in Microsoft Research-Asian (MSRA) for one year supervised by Dr. Xing Tong and Dr. Yizhong Zhang, and I was also an research intern in Dartmouth College supervised by Prof. Bo Zhu since March 2021.

Email  /  Download CV  /  Scholar  /  Github

Research Interests

Currently, my research focuses on flow matching for image and geometry generation, as well as EditFlow and discrete flow matching/masked diffusion for large language models and geometry generation. Previously, I worked on differentiable physics-based simulation in computer graphics, as well as deep learning and virtual reality.

One of my current research directions is Flow Map, which connects my prior work in physics simulation with my recent work on one-step generation. In physics simulation, the primary role of flow maps is to reduce numerical dissipation and enable more faithful simulations. In generative modeling, flow maps help achieve fast one-step or few-step generation.

Awards

SIGGRAPH Asia Best Paper Award (first author), 2024

SIGGRAPH Honorable Mention for Best Paper Award (first author), 2025

Publications

Education

2019 - 2023, Zhejiang University, BS in Mathematics and Applied Mathematics, BE in Computer Science

2023 - 2025, Georgia Tech, Master in Computer Science

2025 - 20xx, Georgia Tech, PhD in Computer Science