Suppression of Neutron Background using Deep Neural Network and Fourier Frequency Analysis at the KOTO Experiment

TL;DR

This paper introduces two novel analysis techniques using deep neural networks and Fourier analysis to significantly reduce neutron background noise in the KOTO experiment's search for rare particle decays.

Contribution

The study develops and applies deep learning and Fourier frequency analysis methods to effectively discriminate neutron background events from photon signals in a particle physics experiment.

Findings

Neutron background was suppressed by a factor of 5.6×10^5.

Photon signal efficiency was maintained at 70%.

The techniques improved background rejection without sacrificing signal detection.

Abstract

We present two analysis techniques for distinguishing background events induced by neutrons from photon signal events in the search for the rare $K^0_L\rightarrow\pi^0\nu\bar{\nu}$ decay at the J-PARC KOTO experiment. These techniques employed a deep convolutional neural network and Fourier frequency analysis to discriminate neutrons from photons, based on their variations in cluster shape and pulse shape, in the electromagnetic calorimeter made of undoped CsI. The results effectively suppressed the neutron background by a factor of $5.6\times10^5$, while maintaining the efficiency of $K^0_L\rightarrow\pi^0\nu\bar{\nu}$ at $70\%$.