Context-Informed Neural ODEs Unexpectedly Identify Broken Symmetries: Insights from the Poincaré-Hopf Theorem
In Huh et al., International Conference on Machine Learning (ICML), 2025
Curriculum Vitae
Education
- Ph.D. in Electrical and Computer Engineering, Purdue University (Aug. 2024–Present)
- Supervised by Prof. Muhammad Ashraful Alam
- Specializing in AI for science
- Supervised by Prof. Muhammad Ashraful Alam
- M.S. in Electronic Engineering, Sogang University (Mar. 2015–Feb. 2017)
- Supervised by Prof. Woo Young Choi (now at Seoul National University)
- Specializing in semiconductor device modeling and simulation
- Supervised by Prof. Woo Young Choi (now at Seoul National University)
- B.S. in Electronic Engineering, Sogang University (Mar. 2009–Feb. 2015)
- Including two years of mandatory military service
Work Experience
- Staff Research Engineer, Advanced Simulation Group, CSE Team, Samsung Electronics (Mar. 2018–Present)
- Developed and optimized machine learning algorithms for semiconductor device design and analysis. Key achievements include:
- Developed symmetry- and geometry-aware representation learning methods for high-dimensional data such as wafer map images.
- Proposed a novel physics-informed neural differential equation model for learning dynamical systems.
- Enabled real-time device and process simulations using deep neural networks.
- Automated DRAM design test vector generation using deep reinforcement learning.
- Developed and optimized machine learning algorithms for semiconductor device design and analysis. Key achievements include:
- Research Engineer, Logic TCAD Group, CAE Team, Samsung Electronics (Mar. 2017–Feb. 2018)
- Analyzed gate stack structures of 7-nm fin transistors using theoretical approaches and numerical simulations.
Publications
Design-Aware Full-Chip Warpage Modeling for STCO: Bridging Reliability and Design for a New Era of Advanced Systems
Hyunjae Jang, ..., In Huh, ..., et al., Symposium on VLSI Technology and Circuits (VLSI), 2025
ML-Driven Compact Models for RRAMs: Addressing Variability and Simulation Efficiency
Giyong Hong, In Huh, ..., et al., IEEE Electron Device Letters (IEEE EDL), 2025
Isometric Quotient Variational Auto-Encoders for Structure-Preserving Representation Learning
In Huh et al., Neural Information Processing Systems (NeurIPS), 2023
Deep Learning for Semiconductor Materials and Devices Design
Changwook Jeong, ..., In Huh, ..., et al., IEEE Electron Devices Technology & Manufacturing Conference (EDTM), 2023
Comprehensive Studies on Deep Learning Applicable to TCAD
Sanghoon Myung, ..., In Huh, ..., et al., Japanese Journal of Applied Physics (JJAP), 2023
PAC-Net: A Model Pruning Approach to Inductive Transfer Learning
Sanghoon Myung, In Huh, ..., et al., International Conference on Machine Learning (ICML), 2022
Application of Deep Reinforcement Learning to Dynamic Verification of DRAM Designs
Hyojin Choi, In Huh, ..., et al., ACM/IEEE Design Automation Conference (DAC), 2021
Bridging TCAD and AI: Its Application to Semiconductor Design
Changwook Jeong, Sanghoon Myung, In Huh, ..., et al., IEEE Transactions on Electron Devices (IEEE TED), 2021
Time-Reversal Symmetric ODE Network
In Huh et al., Neural Information Processing Systems (NeurIPS), 2020
Real-time TCAD: A New Paradigm for TCAD in the Artificial Intelligence Era
Sanghoon Myung, ..., In Huh, ..., et al., International Conference on Simulation of Semiconductor Processes and Devices (SISPAD), 2020
An Accurate Drain Current Model of Monolayer Transition-Metal Dichalcogenide Tunnel FETs
In Huh et al., IEEE Transactions on Electron Devices (IEEE TED), 2017
Subthreshold-Swing-Adjustable Tunneling-Field-Effect-Transistor-Based Random-Access Memory for Nonvolatile Operation
In Huh et al., Applied Physics Letters (APL), 2016
Academic Activities
- Reviewer of NeurIPS (2021–Present)
- Reviewer of ICML (2021–Present)
- Reviewer of ICLR (2022–Present)
- Reviewer of AISTATS (2024–Present)
Honors and Awards
- Full sponsorship for overseas Ph.D. study, Samsung Electronics, 2024
- SEC Annual Award (Business Innovation Section, Bronze Prize), Samsung Electronics, 2021
- DIT Center Best Paper Award, Samsung Electronics, 2021
- DIT Center Best Paper Award, Samsung Electronics, 2020
- Samsung Best Paper Award (Silver Prize), Samsung Electronics, 2019
- SEC Annual Award (Creativity Section, Gold Prize), Samsung Electronics, 2018
- Memory Business Innovation Challenge (Gold Prize), Samsung Electronics, 2018
- Data Science Festival (Special Prize), Samsung Electronics, 2018
- Qualcomm Best Paper Award, Qualcomm Korea, 2016
- EDISON Challenge (Special Prize), National Research Foundation of Korea, 2016
Patents
- Method and apparatus with system verification based on reinforcement learning, US Patent 12175177, 2024
- Device for generating verification vector for circuit design verification, circuit design system, and reinforcement learning method of the device and the circuit design system, US Patent 11861280, 2024
- Simulation system for semiconductor process and simulation method there of, US Patent 11886783, 2024
- Method and apparatus for selecting model of machine learning based on meta-learning, US Patent 11681947, 2023
- Electronic devices generating verification vector for verifying semiconductor circuit and methods of operating the same, US Patent 11669773, 2023
- Method for detecting defects in semiconductor device, US Patent 11507801, 2022