A Multi-Layered Hybrid Quantum-Classical Framework for Enhanced Feature Mapping in High-Dimensional Intrusion Detection

Abstract

Network intrusion detection systems are important in protecting the current computing infrastructure against more advanced cyber threats. The present paper suggests a new hybrid classical-quantum neural network architecture to intrusion de-tection that will have the representational capabilities of quantum computing and the strengths of classical deep learning systems. The model, which is going to be proposed combines variational quantum circuits and deep neural networks to optimize feature presentation and detection of anomalies in high-dimensional network traffic data. The quantum layer will use parameterized quantum gates to simulate complex and non-linear correlations and classical layers will process features and classify them finally. Comprehensive experiments on benchmark intrusion detection data sets show that the hybrid model has an accuracy of up to 96.2 to 98.7 percent in detection, which is better than the state-of-the-art classical models. Also, the suggested system has a very low rate of false positive and better inference performance. The best detection of advanced attack patterns such as distributed denial-of-service attacks, malware infiltration, and unauthorized access attempts is brought out by performance evaluation. These findings affirm the performance and realistic viability of hybrid classical-quantum learning models on future generation intrusion detection systems.