Second Workshop on AI for Accelerating Fusion and Plasma Science

Topics

Physics and Engineering Applications

PHY/ENG

Chair: Ryan McClarren (United States of America); Co-chair: Masayuki Yokoyama (Japan)
This session focuses on AI/ML applications to fusion and plasma physics and engineering. Speakers will discuss the current state of the art, demonstrated impacts, limitations, research opportunities, and cross-fertilization with other fields. The session will highlight how AI is currently being used in fusion research and development, stimulate interdisciplinary dialogue, and identify key challenges for AI researchers to address.

AI Methods

AI

Chair: Michael Churchill (United States of America); Co-chair: Alessandro Pau (Switzerland)
This session is dedicated to machine learning research that is broadly and specifically relevant to fusion and plasma science. Topics will include, but are not limited to: advanced AI techniques for time-series modelling and forecasting; self-supervised, unsupervised, and foundation model approaches applied to large experimental and simulation datasets; and reinforcement learning for control and scenario design. The session will also cover AI techniques to enhance simulation, including fast surrogate model development for digital twin applications and Bayesian inference for comparison with experiments.

Enabling Infrastructure and Data / ML Engineering

EI

Chair: Alejandra Gonzalez-Beltran (United Kingdom); Co-chair: Scott Klasky (United States of America)
This session focuses on the computing, data, and software infrastructure required to deploy AI/ML solutions at scale for fusion and plasma science. Topics include high-performance computing (HPC), cloud platforms, workflows, federated and reproducible data pipelines, and lifecycle management of large experimental and simulation datasets. The session will explicitly address data engineering methods for creating AI-ready data, the development and curation of benchmark datasets, metadata and interoperability standards, and practices that enable reproducibility, traceability, and long-term sustainability of AI workflows.

Special Track

ST

Chair: Cristina Rea (United States of America); Co-chair: Carlo Fiorina (United States of America)
This session is dedicated to lessons learned and best practices from AI applications in fields outside fusion and plasma science, including scientific computing, engineering, energy, materials science, and large-scale industrial AI deployment. The focus is on transferable approaches in validation, model credibility, workforce development, and collaboration models. The objective is to stimulate cross-disciplinary dialogue on how proven practices from mature AI communities can accelerate fusion research and innovation, while preserving Open Science principles and FAIR data practices and avoiding unnecessary duplication of effort.