"An AI-Driven Framework for Ransomware Detection and Defense in Smart Healthcare IoT Systems"

Abstract

The growth of Internet of MedicalThings (IoMT) has made real-time patient monitoring and clinical decision-making much easier, but at the same time, security threats havealso surged sharply owing to the constraints of the resources used, different communications protocols, and the increasing numberofcyber-attacks.Tosolve the identified issue, in this work, we present asmart and explainable Software-as-a-Service (SaaS) solution Ransomware detection system (IDS) specifically designed for IoMT applications that are constrained in terms of resources. For feature selection in the proposed framework, the PSO method has been adopted, while ML and DL methods were used to distinguish between benign and malicious traffic. Doctors could rely on explanations in case of suspicious decisions thanksto SHAP being used as the explanation approach. Our IDS has been evaluated using IoMT synthetic data and WUSTLEHMS-2020 dataset, where we achieved outstanding performance in terms of accuracy proving our model’s ability to work inreal-time and scale up via cloud computing.