Intrusion Detection System based on Quantum Generative Adversarial Network and Blockchain-based Federated Learning

Quantum Generative Adversarial Networks (QGANs) represent a promising solution for anomaly detection in network traffic, with the overarching goal of tackling the increasing complexity and sophistication of cyber threats in Next-Generation Internet infrastructures. Given the inherent noise and instability of current quantum devices, as well as the massive workload of current network flow traces, developing robust and scalable Quantum Machine Learning (QML) models that maintain performance under noisy conditions is a critical challenge. A solution focuses on designing and evaluating QGAN architectures for federated learning (F-QGAN) to address scalability and robustness. This paper presents a solution to integrate blockchain into this federated approach, enabling asynchronous implementation and improving coordination among local learners who cannot exchange weights with the centralized aggregator synchronously.