Quantum-inspired federated learning framework for real time IoT cybersecurity

The evolution of IoT devices in Smart Cities, industrial infrastructures has raised the vulnerability to different types of cyberattacks, highlighting the deficiency of centralized detection models. In order to address this problem, this study suggests a Quantum-Inspired Federated Learning architecture that combines Quantum Generative Adversarial Networks, Quantum Support Vector Machines, and Quantum Neural Networks by means of a web-guided aggregation technique for safe and instantaneous edge-level threat detection. The QIFL design uses amplitude-weighted aggregation to boost union and strength across a variety of heterogeneous nodes, augments limited attack data via Q-GAN synthesis, and improves feature separability through quantum-based data mapping. In comparison to traditional federated systems, experiments on five yardstick datasets-NSL-KDD, CICIDS2017, UNSW-NB15, IoT-23, and Edge-IIoTset-in a hierarchical edge-fog-cloud environment show enhanced privacy (ε = 0.5), accuracy, and efficiency.