Application of quantum machine learning using variational quantum classifier in accelerator physics

TL;DR

This paper demonstrates that a variational quantum classifier can effectively evaluate the dynamic aperture in accelerator physics, outperforming classical neural networks in speed and accuracy, and remains robust under noise.

Contribution

First application of variational quantum classifier in accelerator physics, showing superior speed, accuracy, and noise robustness compared to classical neural networks.

Findings

Quantum classifier achieves faster training than neural networks.

Quantum classifier maintains higher accuracy, especially with limited data.

Performance remains robust under noisy conditions.

Abstract

Quantum machine learning algorithms aim to take advantage of quantum computing to improve classical machine learning algorithms. In this paper, we have applied a quantum machine learning algorithm, the variational quantum classifier for the first time in accelerator physics. Specifically, we utilized the variational quantum classifier to evaluate the dynamic aperture of a diffraction-limited storage ring. It has been demonstrated that the variational quantum classifier can achieve good accuracy much faster than the classical artificial neural network, with the statistics of training samples increasing. And the accuracy of the variational quantum classifier is always higher than that of an artificial neural network, although they are very close when the statistics of training samples reach high. Furthermore, we have investigated the impact of noise on the variational quantum classifier, and found that the variational quantum classifier maintains robust performance even in the presence of noise.