Speaker
Joachim Tsakanikas
(University of Virginia)
Description
Modern particle experiments require large, sophisticated machines and the expertise of many scientists and dedicated personnel. Coordinating machine performance with different, possibly competing goals for simultaneous experiments is a major challenge. To address this challenge, Jefferson Lab's Coupling Experiment to Accelerator Control (CEAC) project members will train a deep reinforcement learning model to optimize accelerator parameters, subject to experiment-specific constraints. My research in CEAC focuses on minimizing uncertainties about beam polarization for the MOLLER experiment and thereby maximizing MOLLER's discovery potential.
Author
Joachim Tsakanikas
(University of Virginia)