QCNN-ID: A Quantum-Classical Hybrid Model for IoT Intrusion Detection
Marwen Amara, Marwen Amara, Marwa Amara, Marwa Amara, Sami Mnasri, Thierry Val · Procedia Computer Science · 2025
Security threats in Internet of Things (IoT) networks, especially in smart cities, are currently increasing due to the fast spread of IoT devices, which highlights the need for reliable intrusion detection systems. Classical machine learning models, such as convolutional neural networks (CNNs), were effectively utilized for detecting intrusions. However, they encountered scalability challenges. This paper proposes a hybrid quantum CNN-based intrusion detection system for securing healthcare IoT networks, as part of smart city (IoTSMC) infrastructures. The introduced model integrates the computational capacity of quantum computing into the CNN architecture to optimize the feature extraction and classification process, for a better detection accuracy. Compared with standard CNN models using reliable IoT attack datasets, Quantum Convolutional Neural Network (QCNN) gives better efficiency in terms of accuracy, precision, recall, and computational efficiency. These findings indicate the potential of using QCNN in detecting IoT cyber-security threats.