professor | Jeonggyu Huh

Department of Mathematics

#31313, Natural Science Building 1

Sungkyunkwan University, Suwon, Korea

Financial Mathematics & Machine Learning

Welcome! I’m interested in making financial mathematics work in high-dimensional settings. I bridge elegant theory and the reality of constrained, frictional, and uncertain markets. I develop structure-preserving learning methods that make dynamic portfolio choice deployable in practice.

PG-DPO Dynamic portfolio choice Deep hedging Asset pricing Stochastic control

Research

Overview of our methods and themes.

Papers

Published papers, preprints, and code.

People

Faculty, researchers, and students.

Join us

For prospective students.

selected publications

ICML

Beyond the Bellman Recursion: A Pontryagin-Guided Framework for Non-Exponential Discounting

Hojin Ko and Jeonggyu Huh

In International Conference on Machine Learning (ICML), 2026

arXiv Bib

@inproceedings{ko2026beyond,
  title = {{Beyond the Bellman Recursion: A Pontryagin-Guided Framework for Non-Exponential Discounting}},
  author = {Ko, Hojin and Huh, Jeonggyu},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2026},
}

Equity Premium Forecasting with Reliability-Screened Forward-Looking Signals

Jeonggyu Huh, Jaegi Jeon, and Seung-Won Jeong

PLoS ONE, 2026

Bib HTML

@article{huh2026equity,
  title = {{Equity Premium Forecasting with Reliability-Screened Forward-Looking Signals}},
  author = {Huh, Jeonggyu and Jeon, Jaegi and Jeong, Seung-Won},
  journal = {PLoS ONE},
  year = {2026},
}

JIMO

Pontryagin-Guided Direct Policy Optimization for Continuous-Time Portfolio Problem

Jeonggyu Huh, Jaegi Jeon, and Seung-Won Jeong

Journal of Industrial and Management Optimization, 2025

Bib HTML

@article{huh2025pontryagin,
  title = {{Pontryagin-Guided Direct Policy Optimization for Continuous-Time Portfolio Problem}},
  author = {Huh, Jeonggyu and Jeon, Jaegi and Jeong, Seung-Won},
  journal = {Journal of Industrial and Management Optimization},
  year = {2025},
}