In March, nuclear physicists using artificial intelligence (AI) technologies participated in a workshop at the Thomas Jefferson National Accelerator Facility (Jefferson Lab). The workshop brought together experimentalists and theorists representing a broad spectrum of nuclear physics research to produce a white paper on the current status and future direction of AI in nuclear science.
Recent work, including work done at the Facility for Rare Isotope Beams (FRIB), focuses on the potential of AI applications to further the priorities of the 2015 Long Range Plan on Research Opportunities and Directions. At FRIB, there are near and long-term prospects for scientific impact in both experimental and theoretical areas. Experimentally, the white paper proposes FRIB-related AI innovation for topics such as accelerator control optimization, experiment design, data acquisition, and data analysis. Theoretical prospects include improved nuclear predictions for astrophysical models and improved models of heavy nuclei. This work is now underway with a new NSF Cyberinfrastructure grant awarded to researchers at FRIB and collaborating institutions.
The workshop consisted of plenary talks as well as breakout sessions. Representatives from the U.S. Department of Energy (DOE) Office of Science and the DOE Artificial Intelligence and Technology Office presented their perspectives, along with leaders in AI applications for nuclear and high energy physics. Scientists involved in AI efforts gave presentations at the breakout sessions, which were organized by broad research topics. A subset of session topics of interest to the FRIB community include:
- Low-energy nuclear theory,
- Accelerator science and operations,
- Experimental methods, and
- Event generation and simulation.
For each workshop session, the white paper summarizes the current status of AI research and applications, case studies, and proposed future research directions.
The FRIB community had a strong presence at the workshop. Five of the working group conveners, six of the white paper co-authors, and two plenary speakers were from the FRIB community. This interchange between FRIB and the broader nuclear physics community built a foundation to drive innovation through AI to further FRIB’s scientific missions.
To drive innovation in nuclear physics research, the white paper looks to:
- Extend beyond the reach of low-energy nuclear science to build a community of scientists driving scientific innovation using AI;
- Build the language of AI methods into training and coursework for nuclear scientists;
- Form strong ties with AI scientists who can leverage the community’s challenges to drive AI research;
- Capitalize on the rapid technological advances to further scientific missions.
By Michelle P. Kuchera, Davidson College