I'm interested in AI applications in nuclear fusion and plasma physics. During my undergraduate and graduate school, I majored in nuclear engineering and physics, focusing on the effect of RMP on electron heat transport in KSTAR. Recently, I studied plasma disruption prediction using Bayesian probabilistic deep learning and data-driven modeling of fusion plasma dynamics combined with autonomous control based on reinforcement learning. As a fusion AI researcher, I prioritize combining plasma physics and machine learning to achieve a physically consistent data-driven model. Several works are shared in my GitHub, so please see the repositories and share your opinions.
Feel free to contact me if you are interested in my research, work, or whatever you want to know from me.
- Notion page: [Jinsu Kim, profile]
- CV: [Jinsu Kim, CV]
- Linkedin:[Linkedin : zinzinbin]
- Disruption prediction using Deep Learning
- Disruption prediction using IVIS dataset(Video data) in KSTAR
- Disruption prediction using 0D data in KSTAR
- Multi-modal learning for disruption prediction
- Tokamak plasma operation control using Reinforcement Learning
- Development of a Transformer-based virtual KSTAR environment
- Development of PINN-based Grad-Shfranov solver
- 0D parameters / shape parameters control using RL algorithms(DDPG, SAC) under the virtual KSTAR environment
- Application of Multi-agent reinforcement learning for autonomous tokamak operation control
- Design optimization of a tokamak fusion reactor based on reinforcement learning
- Development of design computation code of virtual tokamak fusion reactor
- Single-step reinforcement learning for optimizing the design configuration of the tokamak reactor
- ML application on plasma etching process in Virtual Metrology
- Physics-based plasma etching process control
- WelfareForEveryone : https://play.google.com/store/apps/details?id=com.product.welfareapp&pli=1
- github link : https://github.com/21WelfareForEveryone/WelfareForEveryOne
- K-MolOCR(proceeding) : https://zinzinbin.notion.site/Project-K-MolOCR-783608fed0d443f38d3a42882d9bbf05