Pulse Shape Simulation and Discrimination using Machine-Learning Techniques

Shubham Dutta,Sayan Ghosh,Satyaki Bhattacharya,Satyajit Saha

DOI: https://doi.org/10.1088/1748-0221/18/03/P03038

2024-05-16

Abstract:An essential metric for the quality of a particle-identification experiment is its statistical power to discriminate between signal and background. Pulse shape discrimination (PSD) is a basic method for this purpose in many nuclear, high-energy and rare-event search experiments where scintillation detectors are used. Conventional techniques exploit the difference between decay-times of the pulses from signal and background events or pulse signals caused by different types of radiation quanta to achieve good discrimination. However, such techniques are efficient only when the total light-emission is sufficient to get a proper pulse profile. This is only possible when adequate amount of energy is deposited from recoil of the electrons or the nuclei of the scintillator materials caused by the incident particle on the detector. But, rare-event search experiments like direct search for dark matter do not always satisfy these conditions. Hence, it becomes imperative to have a method that can deliver a very efficient discrimination in these scenarios. Neural network based machine-learning algorithms have been used for classification problems in many areas of physics especially in high-energy experiments and have given better results compared to conventional techniques. We present the results of our investigations of two network based methods \viz Dense Neural Network and Recurrent Neural Network, for pulse shape discrimination and compare the same with conventional methods.

Instrumentation and Detectors,Machine Learning

What problem does this paper attempt to address?

The problem that this paper attempts to solve is that under low - energy deposition conditions (such as in direct dark matter detection experiments), the traditional Pulse Shape Discrimination (PSD) technique has a low efficiency in distinguishing signal from background events. Specifically, when the energy received by the detector is very low, the traditional PSD techniques based on time - or frequency - domain information are difficult to effectively distinguish the signals generated by different particles (such as electron scintillation caused by gamma - ray - induced electron recoil and nuclear recoil scintillation caused by neutrons). The paper explores the use of machine - learning (ML) techniques, especially Dense Neural Network (DNN) and Recurrent Neural Network (RNN), to improve the ability to distinguish signal from background events in this low - energy range. The research results show that, compared with the traditional PSD methods, the ML - based methods can provide better discrimination results.

Pulse Shape Simulation and Discrimination using Machine-Learning Techniques

Pulse Shape Discrimination and Exploration of Scintillation Signals Using Convolutional Neural Networks

Model-Based Deep Learning Algorithm for Detection and Classification at High Event Rates

Neutron-Gamma Pulse Shape Discrimination for Organic Scintillation Detector using 2D CNN based Image Classification

Near threshold pulse shape discrimination techniques in scintillating CsI(Tl) crystals

Development of Indigenous Pulse-Shape Discrimination Algorithm for Organic Scintillation detectors

Detection of Dispersed Radio Pulses: A machine learning approach to candidate identification and classification

Fast neutron-gamma discrimination in organic scintillators via convolution neural network

Optimization of digital neutron-gamma discrimination methods in time and frequency domain

Pulse shape discrimination technique for diffuse supernova neutrino background search with JUNO

Neutron-gamma pulse shape discrimination for EJ301 liquid scintillator based on machine learning

A novel single-pulse search approach to detection of dispersed radio pulses using clustering and supervised machine learning

Pulse shape discrimination in an organic scintillation phoswich detector using machine learning techniques

Pulse Shape Discrimination of Bulk and Very Bulk Events within HPGe Detectors

Deep Learning-based Alpha Particles Spectroscopy with Solid-state Nuclear Track Detector CR-39

Using deep neural networks to improve the precision of fast-sampled particle timing detectors

Machine learning-based analysis of experimental electron beams and gamma energy distributions

Pulse Shape Discrimination Based on the Tempotron: A Powerful Classifier on GPU

Special Topic: Applications of the Machine Learning in Nuclear Physics

Searching for Exotic Particles in High-Energy Physics with Deep Learning