Speaker
Description
With the Main Injector Neutrino Oscillation Search (MINOS) experiment decommissioned, muon and hadron monitors became an important diagnostic tool for the NuMI Off-axis $v_\mu$ Appearance (NOvA) experiment at Fermilab to monitor the Neutrinos at the Main Injector (NuMI) beam. The goal of this study is to maintain the quality of the monitor signals and to establish correlations with the neutrino beam profile. By combining individual pixel information from muon monitors and pattern recognition algorithms, we use simulation results and measurement data to build machine learning-based predictions of the muon monitor response and neutrino flux. A new improved simulation allows us to generate high statistics data samples as the training material for machine learning (ML). The model is trained using simulation results with different beam configurations. ML predictions can be used to monitor beamline issues in the future. The plan is to implement the ML model predictions for daily NuMI beamline data monitoring and catching common failure modes.
