Adversarial ML-Based Secured Cloud Architecture for Consumer Internet of Things of Smart Healthcare

Abstract

The profound era of cloud computing (CC) is revolutionizing Industry 5.0in which users have online access to network services including better, transparent user management and the capacity to gather and analyze data. The services of the cloud paradigm have been adopted by academia, industries, healthcare, smart homes, and other areas due to cost-efficient and on-demand resources for IoT applications, but security and privacy of patient data is a major issue with cloud paradigm. For intrusion detection systems (IDS), adversarial machine learning (AML) has been promising to secure individual IoT devices from various modern threats. Network traffic anomalies brought on by both well-known assaults and recently discovered attacks are typically detected by adversarial machine learning-based intrusion detection systems (AML-IDS) in real time. This research proposes a novel technique in cloud security enhancement based on consumer IoT utilizing AML techniques in smart healthcare. The security of cloud networks is enhanced using trust-based encryption cryptographic analysis. The healthcare data is analyzed using structure-based Markov sparse Bayesian neural networks in smart healthcare. The experimental results are improved up to 9% in terms of data transmission ratio, specificity, training accuracy, validation accuracy, and security analysis.