New Opportunities for Machine Learning in Nuclear Science

Artificial Intelligence (AI), based on machine learning (ML) approaches, pervades nearly all aspects of our lives today, including scientific discovery. ML is now used throughout the various subfields of nuclear science and across the stages of scientific discovery: from fast, online decision-making in experiments to providing theoretical predictions with uncertainty quantification. This talk will provide an overview of the core principles of AI and ML, followed by examples of notable work from our community in experiment design, real-time systems, data analysis, and theory. Then, we will look towards the near future at potential opportunities for ML in nuclear science from a technical perspective, along with a survey of emerging community norms surrounding data access, model dissemination, and publication. The talk will conclude by looking at recent conversations about AI ethics and regulation and how they intersect with AI/ML work in nuclear science.