Advisor: Tian Li
Before starting my Phd, I completed my undergraduate from UChicago in 2026, finishing with a BS with honors in Maths and a BS with honors in Computer Science. I worked on developing a Hamiltonian perspective on Sherman’s Area-convexity algorithm; and on Asynchronous training under Heavy-tailed noise.
I was born and raised in Shanghai, China.
I am interested in theoretical machine learning. My current interests include and go beyond: (i) Trajectories, convergence, and generalizations of optimization algorithms, including non-traditional approaches like manifold and geometric learning. (ii)Developing more efficient optimizers for large-scale models, addressing heavy hyperparameter tuning, high resource requirements, etc. (iii) Understanding deep learning theory through statistical learning theory and high-dimensional statistics.
