■ 자기소개
현재 한양대학교 경상대학 경영학부의 조교수로 재직 중인 구재훈 교수는 아주대학교 산업공학과를 졸업하고, 한국과학기술원(KAIST)에서 산업공학 석사학위를, Northwestern University에서 산업공학 박사학위를 취득하였다. 한양대 교수로 부임하기 이전에 Argonne National Lab에서 Postdoc으로 재직하며 머신러닝을 이용해 High-Performance Computing과 Fusion Energy Science 분야의 응용연구를 수행하였다.
■ 학력
2020, Northwestern University, 산업공학 박사
2012, 한국과학기술원(KAIST), 산업공학 석사
2010, 아주대학교, 산업공학 학사
■ 경력
2022.09-현재, 한양대학교, 경상대학/경영학부, 조교수
2020-2022, Argonne National Lab, MCS, Postdoc
2011-2013, 현대모비스, 품질본부, 사원
■ 연구관심분야
Machine learning, deep learning, reinforcement learning, big data, artificial intelligence, optimization
■ 주요논문
- ytopt: Autotuning Scientific Applications for Energy Efficiency at Large Scales, CCPE, 2024
- EFIT-Prime: Probabilistic and physics-constrained reduced-order neural network model for equilibrium reconstruction in DIII-D, Physics of Plasmas, 2024
- Machine Learning Based Surrogate Models for Knock Prediction in Syngas (H2/CO) Added SI Engine Combustion, Journal of Energy & Climate Change, 2024
- Transfer-learning-based Autotuning using Gaussian Copula, ACM ICS, 2023
- An inverse classification framework with limited budget and maximum number of perturbed samples, Expert Systems with Applications, 2023
- Application of machine learning and artificial intelligence to extend EFIT equilibrium reconstruction, Plasma Physics and Controlled Fusion, 2022
- Customized Monte Carlo Tree Search for LLVM/Polly's Composable Loop Optimization Transformations, PMBS, 2021
- A Unified Defect Pattern Analysis of Wafer Maps Using Density-Based Clustering, IEEE ACCESS, 2021
- Combined Convolutional and Recurrent Neural Networks for Hierarchical Classification of Images, IEEE Big Data, 2020
- Improved Classification Based on Deep Belief Networks, ICANN, 2020
- Deep Reinforcement Learning for Network Slicing with Heterogeneous Resource Requirements and Time Varying Traffic Dynamics, CNSM, 2019
■ 프로젝트
2023-2026, 심층학습에 기반한 설명 및 해석 가능한 인공지능 분류 시스템 개발, 한국연구재단 기초연구사업 기본연구과제, 연구책임자
2021-2022, EFIT-AI: ML/AI Assisted Tokamak Equilibrium Reconstruction, DOE FES Project, Postdoc
2020-2022, PROTEAS-TUNE: Programming Toolchain for Emerging Architectures and Systems, DOE ASCR Exascale Computing Project, Postdoc