Gamma-hadron Separation in Imaging Atmospheric Cherenkov Telescopes using Quantum Classifiers

TL;DR

This paper introduces a novel quantum machine learning approach for gamma-hadron separation in Imaging Atmospheric Cherenkov Telescopes, demonstrating comparable performance to classical methods and proposing new architectures for large datasets.

Contribution

It presents the first application of Quantum Support Vector Classifier and Variational Quantum Classifier for gamma-hadron separation in IACTs, with improved accuracy through hyperparameter tuning and clustering.

Findings

Quantum classifiers achieve performance comparable to classical methods.

Hyperparameter tuning improves classification accuracy.

Clustering enhances quantum classifier performance on large datasets.

Abstract

In this paper we have introduced a novel method for gamma hadron separation in Imaging Atmospheric Cherenkov Telescopes (IACT) using Quantum Machine Learning. IACTs captures images of Extensive Air Showers (EAS) produced from very high energy gamma rays. We have used the QML Algorithms, Quantum Support Vector Classifier (QSVC) and Variational Quantum Classifier (VQC) for binary classification of the events into signals (Gamma) and background(hadron) using the image parameters. MAGIC Gamma Telescope dataset is used for this study which was generated from Monte Carlo Software Coriska. These quantum algorithms achieve performance comparable to standard multivariate classification techniques and can be used to solve variety of real-world problems. The classification accuracy is improved by hyper parameter tuning. We propose a new architecture for using QSVC efficiently on large datasets and found that clustering enhance the overall performance.