AI for Fusion

AI for Fusion

Accelerating fusion development .

Machine learning methods are advancing rapidly, and their potential to accelerate fusion research and development is beginning to take shape. Fusion research relies on a close interplay between theory, simulation, and experiment, providing a rich setting for exploring how AI can contribute to scientific progress. Our collaboration develops and evaluates methods for fusion science and engineering, with an emphasis on advanced optimization, neural network surrogates, generative models, and large language model agents.

AI + Fusion

Our group connects plasma physics and AI researchers around open problems in fusion science and engineering—bringing together domain expertise and emerging AI methods to develop practical solutions.

  • Plasma physics
  • Nuclear engineering
  • Scientific computing
  • Machine learning
  • Applied mathematics
  • Materials science

Kickoff event

Oct. 27, 15:30-17:00 @ MH 7560

Join talks, working sessions, and discussions on AI for fusion. Speakers:

Help set the scientific direction.

We’re looking for UW–Madison organizers who want to contribute to event planning and the collaboration’s scientific direction. Join the group and introduce yourself—we’d love to hear what you want to build.

Current organizing committee

Misha Khodak

Department of Computer Sciences
UW–Madison

Rogerio Jorge

Department of Physics
UW–Madison

Yijing Zhang

Department of Computer Sciences
UW–Madison

Raheem Hashmani

Department of Physics
UW–Madison