Hello, my name is Jaeyeon Kim! I’m a first-year Ph.D. student at Harvard CS, prospectively advised by Sitan Chen and Sham Kakade.

My academic journey started in junior high school, with multiple awards in Math Olympiads. Since then, I’ve always been driven by solving challenging problems with mathematics. Afterward, I studied Mathematics at Seoul National University, where I received B.S. in Mathematics. I loved to organize mathematical concepts in my own words–also I got 4.3/4.3 in every math course that I’ve taken!

Fortunately, I worked with Prof. Ernest Ryu in Optimization Theory. I discovered H-duality, a duality between first-order algorithms. H-duality is the duality between algorithms, thus distinct from any previously known duality, and further extended on various setups by my works (Mirror Descent, Fixed-point problems).

At Harvard University, I turned my attention to Diffusion Models. With amazing and supportive professors, my recent work on Masked Diffusion Models demystifies its training and inference. My broad goal is to develop an efficient generative model for discrete data, as well as mathematically identify scientific phenomena on Diffusion Models.

Feel free to reach out to me! I’d be happy to chat on various topics on Machine Learning.