A Machine Learning Approach to a Multidetector Array Response Function for Nuclear Search

During nuclear search operations, the localization of radioactive sources can be a time-consuming process that requires mapping relative radiation intensity in large area to determine position source. This article introduces use machine learning, specifically temporal convolutional network (TCN), estimate direction between detector array and static 137 Cs application learning provides directional vector $4\pi $ with 90% confidence 5.6° 99% within 11.2°. With low-cost NaI(Tl) detectors, effects self-shielding creates gamma-ray shadows depending on orientation Using convolved response function, we apply supervised neural predict unit points toward observed The is expected reduce times once implemented future work.