Convolutional Neural Networks for Shower Energy Prediction in Liquid Argon Time Projection Chambers

TL;DR

This paper explores the use of convolutional neural networks to improve energy reconstruction of electromagnetic showers in liquid argon TPCs, outperforming traditional linear calibration especially with imperfect data.

Contribution

It introduces CNN-based models for shower energy prediction in LArTPCs and demonstrates their superior performance over linear methods in simulated data.

Findings

CNN models outperform linear calibration in energy accuracy

CNNs handle unresponsive wires more effectively

Reconstruction within 5% accuracy for more events

Abstract

When electrons with energies of $O(100)$ MeV pass through a liquid argon time projection chamber (LArTPC), they deposit energy in the form of electromagnetic showers. Methods to reconstruct the energy of these showers in LArTPCs often rely on the combination of a clustering algorithm and a linear calibration between the shower energy and charge contained in the cluster. This reconstruction process could be improved through the use of a convolutional neural network (CNN). Here we discuss the performance of various CNN-based models on simulated LArTPC images, and then compare the best performing models to a typical linear calibration algorithm. We show that the CNN method is able to address inefficiencies caused by unresponsive wires in LArTPCs and reconstruct a larger fraction of imperfect events to within 5% accuracy compared with the linear algorithm.