Yan Li (李延)

PhD student, ISyE, Georgia Tech

I study simple problems. Hopefully they can be useful.

Greetings! I am a final-year PhD student at ISyE, Georgia Tech. I obtained M.S. in Statistics from the same institute. Before that, I obtained B.S. in Mathematics and B.B.A. in Accounting from Nankai University.

I work with George Lan on the design and analysis of data-driven methods for dynamic decision-making under uncertainty, and with Tuo Zhao on understanding and improving gradient-based methods in ML practice.

At Georgia Tech, I have also been fortunate enough to collaborate with Anton Kleywegt and Alexander Shapiro.

Kindly find below some of my personal favorites:

• A Novel Catalyst Scheme for Stochastic Minimax Optimization
• Rectangularity and Duality of Distributionally Robust Markov Decision Processes
• First-order Policy Optimization for Robust Policy Evaluation
• Policy Mirror Descent Inherently Explores Action Space
• First-order Policy Optimization for Robust Markov Decision Process
• Homotopic Policy Mirror Descent
• Implicit Bias of Gradient Descent based Adversarial Training

Of course, a bit more detail on things I have been working on can be found here.