Mu2e, the muon-to-electron conversion experiment at Fermilab, will search for charged lepton flavor violation (CLFV) through the coherent conversion of a bound muon into an electron in the field of an aluminum nucleus. The experiment is expected to improve the current best limit on this process by four orders of magnitude. Detector commissioning with cosmic rays is currently underway, with the first physics data run anticipated in Fall 2027.
At the same time, advances in artificial intelligence (AI) are rapidly changing the landscape of scientific computing and analysis. While machine learning (ML) techniques have long been used in particle physics, recent developments -- including large language models and AI-assisted software tools -- are expanding the ways these methods can support experimental workflows.
In this talk, I will discuss early efforts to integrate modern AI tools and infrastructure within the Mu2e collaboration. As a concrete example of an ML algorithm applied directly to the experiment, I will present a hybrid magnetic field model that combines analytical expansions with a physics-informed neural network (PINN) to describe the Detector Solenoid magnetic field.
Cole Kampa, Postdoctoral Scholar, Caltech
Host: Susan Dittmer
Joan West
(847) 491-3645
Email