QLFL: quantum-enabled lightweight federated learning-empowered threat detection model for Internet of Medical Things based healthcare

Abstract

The integration of Internet of Medical Things (IoMT) has significantly reshaped the landscape of modern healthcare, facilitating seamless communication between connected IoMT devices. Despite these advancements, the growing adoption of healthcare technologies has raised critical cybersecurity and data privacy concerns. Artificial Intelligence (AI)-powered solutions have gained traction as an effective means to strengthen Intrusion Detection Systems (IDS) against emerging cyber threats. Nevertheless, deploying AI-based IDS in healthcare environments presents obstacles, such as excessive computational overhead and the need for secure data handling. To address the above mentioned challenges, this article proposes Quantum Lightweight Federated Learning (QLFL), an innovative framework combining Quantum Computing (QC) and Federated Learning (FL) to deliver a high-performance, scalable, and privacy-centric intrusion detection system for IoMT networks. The proposed methodology adopts knowledge distillation techniques within FL to create an optimized detection model suitable for low-power devices. Moreover, Additive Homomorphic Encryption is implemented to safeguard model parameters during federated training, maintaining data confidentiality without compromising processing speed. Experimental evaluations highlight the system reliability, with accuracy rates of 98.90% on the IoT Healthcare Security dataset, demonstrating stable performance under varying network conditions. These findings suggest that QLFL could play a pivotal role in enhancing the cybersecurity framework for next-generation healthcare applications.