Jaeyeon Kim

Jaeyeon Kim

Ph.D. Student

Harvard University

Biography

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 began in junior high school, with multiple awards in Math Olympiads. Since then, I’ve been driven by a passion for solving challenging problems through mathematics. Afterward, I earned a B.S. in Mathematics from Seoul National University, where I enjoyed organizing mathematical concepts in my own words.

During my undergraduate years, I had the privilege of working with Prof. Ernest Ryu in Optimization Theory-leading to the discovery of H-duality-a duality between first-order algorithms. Unlike conventional dualities, H-duality characterizes the relationship between algorithms themselves. My research has since extended to various setups. (Mirror Descent, Fixed-point problems).

At Harvard, my research has shifted toward Diffusion Models. With incredible advisors, my recent work on Masked Diffusion Models sheds light on their training and inference processes. My broader goal is to develop efficient generative models for discrete data while deepening our mathematical understanding of Diffusion Models.

I’m always happy to chat about Machine Learning, Optimization, and beyond—feel free to reach out!

Interests
  • Optimization Theory
  • Science of Deep Learning
Education
  • Ph.D. in Computer Science, 2024

    Harvard University

  • B.S. in Mathematics, 2020

    Seoul National University

Recent News

  2024.09   I’m starting my Ph.D. at Harvard University, prospectively advised by Prof. Sitan Chen and Sham KaKade. I’m really thrilled to pursue my research career at Harvard University!

  2024.08   I’m honored to be selected as Ilju Foundation scholarship, which supports graduate students studying abroad.

  2024.04   Excited to announce my new paper, Optimal Acceleration for Minimax and Fixed-Point Problems is Not Unique (ICML 2024, Spotlight, Top 3.5%). By proposing novel algorithms, we suggested that the optimal acceleration mechanism in minimax optimization and fixed-point problems is not unique. Surprisingly, our new algorithms are H-dual to the prior anchor-based accelerated methods: We discover H-duality in another setups!

  2023.12   I attended NeurIPS 2023 and gave a poster presentation.

Experience

 
 
 
 
 
Ph.D. Student at Harvard
Under Professor Sitan Chen and Sham KaKade
September 2024 – Present
Diffusion Models
 
 
 
 
 
Research Intern at MIT
Under Professor Asuman Ozdaglar
July 2023 – August 2023
Research on Optimization Theory
 
 
 
 
 
Research Intern at Seoul National University
Under Professor Ernest Ryu
September 2022 – August 2024
Research on Optimization Theory