Nuclear Theory at the Intelligence Frontier

A new generation of nuclear and particle physics experiments will probe the structure of hadrons and nuclei and confront the Standard Model with unprecedented precision. Hadron structure quantities are key to both efforts, as both direct science targets and critical inputs to the search for new physics. In this talk, I will highlight recent results and ongoing work to advance our theoretical control of these quantities, including: (1) recent advances in lattice QCD calculations of the gravitational structure of hadrons including the nucleon and scalar glueball; (2) a new data analysis framework based on applying the Lanczos algorithm to the Hilbert space of QCD, providing more rigorous spectroscopic analyses and a simple, reliable approach to lattice hadron structure calculations at new scales of complexity; (3) a fully data-driven approach to neutrino-nucleus interaction modeling at DUNE, which uses ML to circumvent the need for difficult-to-obtain ab initio nuclear theory inputs; (4) ongoing development of ML approaches based on normalizing flows to accelerate lattice QCD calculations, including early demonstrations of computational advantage in QCD; and (5) how rapidly emerging agentic AI will drive progress in computational nuclear theory and science more broadly.