Pulse shape discrimination using a convolutional neural network for organic liquid scintillator signals

Abstract A convolutional neural network (CNN) architecture is developed to improve the pulse shape discrimination (PSD) power of gadolinium-loaded organic liquid scintillation detector reduce fast neutron background in inverse beta decay candidate events NEOS-II data. spectrum an event constructed using a Fourier transform time domain raw waveforms and put into CNN. An early data set evaluated by CNN after it trained low energy β α events. The signal-to-background ratio averaged over 1–10 MeV visible range enhanced more than 20% result method compared that existing conventional PSD method, improvement even higher region.