Deep Learning for Polarization Analysis in Solid-State Targets via NMR

Nuclear Magnetic Resonance (NMR) has been an essential tool for solidstate polarized target experiments in Nuclear and High-energy physics. Phasesensitive detection with a Q meter enables monitoring of the polarization over the course of a scattering experiment, however, its performance is hindered by substantial noise and systematic uncertainties in the signals. In this study, we present a machine learning (ML)-based approach to enhance the signal-to-noise ratio and extend the reliability of NMR-based polarization measurements beyond the standard operational limits of Q-meters. Our initial findings demonstrate that the method enables more accurate real-time polarization tracking and offline data analysis, contributing to improved performance in scattering experiments.