“Artificial intelligence and scientific machine learning are transforming the way fusion and plasma research is conducted. These awards will advance a broad set of capabilities across the Fusion Energy Sciences (FES) program, making essential capabilities available for all stakeholders,” said Jean Paul Allain, DOE associate director of science for FES. “The U.S. is leveraging every tool in its pursuit of an aggressive program that will bring fusion energy to the grid on the most rapid timescale.”
The projects were selected by competitive peer review under the DOE’s funding opportunity announcement for Machine Learning, Artificial Intelligence, and Data Resources for Fusion Energy Sciences. Approaches that could support the deployment of a fusion pilot plant on a decadal timescale were preferred for their “high programmatic importance,” according to the FOA, which was opened to applications in December 2022.
Making fusion data easier to access and use: The recipients will pursue research in science discovery, diagnostic data analysis, model extraction and reduction, plasma control, analysis of extreme-scale simulation data, and data-enhanced prediction, according to the DOE. Multiple awards will focus on establishing new systems for managing, formatting, curating, and accessing experimental and simulation data, and the products of that work will be available in public databases.
Total funding is $29 million for projects lasting up to three years in duration, with $11 million for fiscal year 2023, and outyear funding contingent on congressional appropriations.
The projects: A list of the seven projects chosen for funding is available on the FES program homepage. The participants, and their project titles, are as follows:
- University of Rochester and Hewlett Packard Enterprise: “Applications of machine learning and data science to predict, design, and improve laser-fusion implosions for inertial fusion energy.”
- Oak Ridge National Laboratory: “Enabling tokamak pulse simulation by machine learning of core-pedestal-boundary physics.”
- University of California–Berkeley, Princeton University, and Princeton Plasma Physics Laboratory: “Active learning-guided discovery and data-enabled active control of plasma synthesis of nanomaterials.”
- Massachusetts Institute of Technology, University of Wisconsin–Madison, the HDF Group, Auburn University, and the College of William and Mary: “Open and FAIR fusion for machine learning applications.”
- General Atomics, Hewlett Packard Enterprise, SapientAI, and the University of California–San Diego: “A fusion machine learning data science platform to support the design and safe operation of a fusion pilot plant.”
- Los Alamos National Laboratory, University of Texas–Austin, University of Florida, and Pennsylvania State University: “DeepFusion accelerator for fusion energy sciences in disruption mitigation.”
- Purdue University: “Accelerating discovery and diagnostics of plasma-wall interactions using machine learning.”
