papers in progress

Preprints

• PG-DPOBreaking the Dimensional Barrier: A Pontryagin-Guided Direct Policy Optimization for Continuous-Time Multi-Asset Portfolio Jeonggyu Huh*, Jaegi Jeon, Hyeng-Keun Koo, Byung-Hwa Lim*
  PaperCodeSlides
  + Supplementary proofs: Detailed Proof of Theorem 3 — BPTT–BSDE Equivalence PDF · Detailed Proof of Theorem 4 — Policy-Gap Bounds PDF
• PG-DPOBreaking the Dimensional Barrier: Dynamic Portfolio Choice with Parameter Uncertainty via Pontryagin Projection Jeonggyu Huh*, Hyeng-Keun Koo
  PaperSlides

Papers in Progress

» PG-DPO

• Unverified Optimism in World Models: Pontryagin-Guided Control Verification
• Pontryagin-Guided Policy Iteration for Discrete-Time Infinite-Horizon Portfolio Choicewith Seungwon Jeong
• Finite Horizon and Optimal Portfolio Choice with Stochastic Income: A Reinforcement Learning Approachwith Seyoung Park, Hojin Ko
• Pontryagin-Guided Deep Hedging for Nonsmooth Payoffs under Entropic Riskwith Seungho Na
• Breaking the Dimensional Barrier in Dynamic Portfolio Choice under High-Dimensional Jump Riskwith Dongwan Shin
• Boundary-Layer Stable Pontryagin-Guided Deep Hedging for Barrier Optionswith Seungho Na
• Breaking the Dimensional Barrier for Dynamic Portfolio Choice with Optimal Stoppingwith Hojin Ko
• Breaking the Dimensional Barrier in Non-Markovian Dynamic Portfolio Choicewith Seungwon Jeong
• Utility Maximization under Shortfall Tail Risk: A CRRA–CVaR Formulationwith Jung Min Lee

» DeepONet

• Deep Operator Learning for Option Pricing with Functional Coefficients: A MIONet Approachwith Myeongsik Kim, Woo-Chul Choi
• Real-Time Pricing of Equity-Linked Securities Using Deep Operator Networkswith Yoonyoung Byun
• Macro-Based Forecasting of Implied Volatility Surface Dynamicswith Hojin Ko, Ho-Jun Lee, Wonwoo Choi

» Asset Pricing

• Scalable Dynamic Portfolio Allocation via Physics-Informed Neural Networkswith Seungwon Jeong, Yeoneung Kim
• Adversarial Time-Series Domain Adaptation for Early-Stage IPO Price Predictionwith Youngwoo Lee, Seungwon Jeong
• Discounted Alpha: A Machine Learning Framework for Equity Valuationwith Dongwan Shin, Thummim Cho

Future directions

Beyond its current scope, PG-DPO admits natural extensions to partial equilibrium settings—including robust optimization, rough volatility, regime, taxation, Epstein-Zin utility, smooth ambiguity and belief-state dynamics—as well as general equilibrium frameworks such as mean-field control and multi-agent games.