Publication

Abstract : The greatest emerging threat to cybersecurity comes from the combination of quantum computing and artificial intelligence. Advances in quantum technology will compromise current encryption methods, and AI attacks are becoming in- creasingly complex to detect, frequently evading existing defenses. However, their combination also presents a promising solution to build a framework that would allow cybersecurity to be proactive rather than just reactive, as it is currently. This paper proposes a Quantum-Enhanced Adaptive Defense System for Network Threat Detection (QEAD), a quantum-classical model that enables the enhancement of network threat detection with a special focus on improved zero-day attack detection. QEAD consists of a Parameterized Quantum Circuit (PQC) to refine classical threat probabilities, combines them with anomaly scores to generate trust scores, and uses reinforcement learning to determine defense actions in real time. Using quantum computing principles, this model offers next-generation threat detection technology that demonstrates a significant improvement on a synthetic dataset, achieving a zero-day recall of 42.7 % compared to 3.3 % with Isolation Forest and 0 % with XGBoost. Due to the ability of QEAD to adapt to changing threats in real time, it represents a paradigm shift to improve cybersecurity against the evolving dynamics of zero-day attacks, positioning itself as an intelligent and active defense.