Optimization & Machine Learning
I am a Ph.D. student in Machine Learning at the Georgia Institute of Technology, advised by Prof. Evangelos A. Theodorou in the Autonomous Control and Decision Systems (ACDS) Laboratory.
I develop machine learning methods for large-scale optimization through the lens of optimal control theory.
My research focuses on GPU-based optimization, reinforcement learning for solver acceleration, and robust generative models. I am interested in how learning and control can make computational methods more efficient and reliable.
Previously, I worked with Prof. Joonseok Lee at Seoul National University on video summarization and molecular representation learning, and with Prof. Yan Wang at Georgia Tech on quantum scientific computing. I received my B.S. degrees in Computer Science and Engineering and Mechanical Engineering from Seoul National University.
Large-scale optimization
GPU-based algorithms
Stochastic optimal control
Linear & mixed integer programming
ML for Optimization
Reinforcement learning
Schrödinger bridge
Diffusion-based generative models
Learning to accelerate optimization
I formulate solver acceleration as an optimal control problem and train feedback policies to coordinate solver parameters and restarts, while retaining GPU-friendly matrix–vector products and projections.
Read the preprintRobust generative modeling · DARPA AIQ
I develop distributionally robust fine-tuning methods for diffusion-based Schrödinger bridge models, studying resilience to noisy and shifted inputs without test-time retraining.
Read the preprintLarge-scale robust optimization
I formulate multi-item inventory problems as sparse robust linear programs and study solver performance and decomposition. Using Dantzig–Wolfe column generation, I solved an instance with 50,000 items.
Ph.D. in Machine Learning
ACDS Laboratory · Advisor: Prof. Evangelos A. Theodorou
Expected graduation: Spring 2029
B.S. in Computer Science and Engineering
B.S. in Mechanical Engineering
Doctoral Research · ACDS Laboratory
Advisor: Prof. Evangelos A. Theodorou
GPU-based optimization, reinforcement learning for solver acceleration, robust inventory optimization, and distributionally robust generative modeling.
Prior Research · Prof. Yan Wang
Quantum methods for differential equations and scientific computing; machine learning for semiconductor and additive manufacturing process monitoring.
Research Internship & Collaboration · Prof. Joonseok Lee
Generative video summarization, the Mr. HiSum dataset, and physically regularized molecular representation learning.
Intern · Seoul, Korea
Supported a long-range wireless charging project through rectifier-circuit and heat-transfer experiments and a review of wireless power transmission patents.
* Equal contribution. See the full publication list for earlier work.
Quantum Functional Expansion to Solve Stochastic Differential Equations
Invited talk · IEEE Quantum Week 2024 Review Workshop, The Institute of Semiconductor Engineers, Korea (online) · Nov 2024
Quantum Solver of Stochastic Differential Equations
Conference presentation · WCCM 2024
ICLR · 2027 cycle
NeurIPS Datasets and Benchmarks · 2023, 2024, 2026
WACV · 2024
Engineering Mathematics 2 · 2020–21
Foundation of Physics 2 · 2019
Seoul National University
Co-inventor · U.S. Provisional Patent Application No. 63/692,936
Filed September 10, 2024
For research conversations and professional inquiries, email is the best way to reach me.
jsul7@gatech.edu