Machine Learning for Event Identification in the Nab Experiment

Despite the predictive success of the standard model (SM), it is known that it is still incomplete. For example, the three-sigma discrepancy found in the unitary test of the Cabibbo-Kobayashi-Maskawa (CKM) matrix is suggestive of physics beyond the SM. The Nab collaboration aims to measure the beta decay correlation coefficients to a sufficiently high precision to sensitively test CKM unitarity. In the Nab experiment, two detectors are used to detect protons and electron energies coming from the beta decay of cold neutrons. In this work we attempt to use machine learning models to provide a filter for data quality with the goal of classifying proton and electron signals, and noise signals that can be discarded or ambiguous signals that need further analysis. These models will be compared to the standard filter techniques used in the Nab experiment, ie sinlge/double trapezoidal and cusp filters.