Plasma is a highly ionized state of matter governed by complex collective interactions among charged particles and electromagnetic fields. Its nonlinear, multiscale, and often high-dimensional dynamics pose formidable challenges for modeling, simulation, prediction, and control. Recent advances in scientific machine learning offer new opportunities to address these challenges by integrating data-driven methods with physical models and computational algorithms. We are excited to bring together in workshop researchers in plasma science, scientific computing, and machine learning to exchange ideas, identify foundational challenges, and develop new directions for scientific machine learning in plasma applications.

Left panel: landscape of electric energy for 1D1V Vlasov-Poisson as a function of two control variables. Middle and right panel: 2D2V Vlasov-Poisson with constant magnetic field in the Dory–Guest–Harris (DGH) instability regime. Pictures and movies are generated by Peiyi Chen, Martin Guerra and Ethan Hanold.

Funding institutions: National Science Foundation, American Family Insurance Data Science Institute,  Institute for Foundations of Data Science (IFDS)