Po-Han (Kozak) Hou

I am a PhD student in Nuclear Fusion Research at the Graduate School of Frontier Sciences, University of Tokyo (UTokyo), under the supervision of Professor Naoto Tsujii in the Ejiri–Tsujii Laboratory. My research interests are in applied science via high performance computing, including fusion plasma modelling and machine learning surrogate. Recently, I have been working on accelerating Two-Dimensional Lower Hybrid Wave code via Physics-Informed Neural Operator on TST-2 spherical tokamak located at University of Tokyo, Kashiwa Campus.

Prior to my PhD, I was supervised by Professor Adriana Paluszny Rodriguez and coadvised by Professor Tsujii in the MSc Applied Computational Science and Engineering program (2024-2025) at Imperial College London. I also completed my bachelor’s degree in Space Science and Engineering with a credit program in Interdicipline Artificial Intilligence at National Central University in Taiwan (2020-2024). I was privileged to be advised by Dr. Chun-Yu Lin and Professor Jih-Hong Shue for their guidance in high-performance scientific computing and numerical methods in space physics.

Research Focus

• Scientific machine learning for fusion science and nuclear fusion
• Reg-PINNs, PINNs, neural operators, and Physics-Informed Neural Operator methods
• Fourier Neural Operator and geometry-embedded Fourier Neural Operator models
• Lower hybrid wave, plasma, and wave simulation for TST-2
• High-performance computing, parallel computing, and scalable scientific workflows