An Adversarial ML Approach to Parameter Learning in Quantum Chromodynamics
Katherine Keegan, Emory University
Parameter fitting of equations in quantum chromodynamics (QCD) based on empirical data suffers from the error inherent in the measurement of quantum particles. To combat this, we introduce an adversarial machine learning framework in which parameters are iteratively learned through learning to generate fake samples through simulated physics measurement which match true data. This project is part of the QuantOm SCIDAC collaboration and was completed with collaborators at RIKEN, Virginia Tech, Jefferson Laboratory, and Argonne National Laboratory.