Beyond the Bellman Recursion: A Pontryagin-Guided Framework for Non-Exponential Discounting, Hojin Ko*, Jeonggyu Huh*, International Conference on Machine Learning (ICML), 2026
Sobolev-Stable Regime-Conditioned DeepONet for Monthly Implied-Volatility Surface Forecasting, submitted
Finite-Horizon Portfolio Choice with Stochastic Income: From Analytical Approximation to Reinforcement Learning, submitted [Link]
Papers in Progress
High-Dimensional Portfolio Optimization with Transaction Costs via Neural Continuation Gradients
Deep Hedging through Bellman-Guided Direct Policy Optimization
A Computational Framework for Decision-Aligned Conditional Betas in Cost-Aware Portfolio Optimization
Project in Progress
ELS Hedging Decision-Making System Using Neural Optimal Control, Jul. 2026–Jun. 2027, RiskX-led project, 2026 Seoul FinTech Technology Commercialization Support Program
Grants
Research Scholarship, Growth Pathways Support Program for Outstanding STEM Talent, Sungkyunkwan University, 2026 [Program]
Awards
Outstanding Poster Award, Korean Society for Industrial and Applied Mathematics (KSIAM), Fall 2026 [News]
Outstanding Poster Award, Korean Mathematical Society (KMS), Fall 2026 [News]
Presentations
Knowing When Not to Trade: Latent No-Trade Region Recovery Hidden in Neural Control, Korean Society for Industrial and Applied Mathematics (KSIAM), Poster Presentation, Fall 2026
Knowing when not to act: Latent no-action region recovery hidden in neural control, Korean Mathematical Society (KMS), Poster Presentation, Fall 2026
Beyond the Bellman Recursion: A Pontryagin-Guided Framework for Non-Exponential Discounting, The 2nd Sookmyung-TMU Mathematical Finance Workshop with Young Researchers, 2026
Breaking the Dimensional Barrier in Dynamic Portfolio Choice with Transaction Costs, The Korean Operations Research and Management Science Society (KORMS), Oral Presentation, Spring 2026
Beyond the Bellman Recursion: A Pontryagin-Guided Framework for Non-Exponential Discounting, International Conference on Machine Learning (ICML), Poster, 2026