Variational Quantum Convolutional Neural Network Based on Particle Swarm Optimization

With the rapid development of the Internet and digitalization, cybersecurity threats have become increasingly severe, and traditional intrusion detection techniques struggle to address the growing complexity of network attacks. Machine learning-based intrusion detection models face limitations in detection accuracy. To overcome the shortcomings of traditional methods in identifying intrusions with high precision, this paper proposes a Variational Quantum Convolutional Neural Network (VQCNN) model optimized using Particle Swarm Optimization (PSO) to improve detection accuracy and generalization capability. By leveraging PSO to optimize the parameters of variational quantum circuits, the proposed model integrates quantum properties with intelligent optimization algorithms, enabling more efficient processing of complex data features. The UNSW-NB15 network intrusion detection dataset was used for experimental validation, and the results demonstrate that the proposed PSO-VQCNN model achieves an accuracy of 96.25%, outperforming other traditional intrusion detection models such as Artificial Neural Networks (ANN), Support Vector Machines (SVM), k-Nearest Neighbors (k-NN), Naive Bayes, and Decision Trees. This highlights the potential and challenges of combining variational quantum circuits with intelligent algorithms.