An Ai-Integrated Real-Time Network Security Framework For Cyber Threat Mitigation In Smart Industry

Abstract

The transition of smart industry is showing tremendous efficiency in terms of operations but experiencing unexpected cyber security risks due to integration of operational and information technology. Present security models basically depending on signature-based detection mechanism and machine learning models. These security model may not suitable against sophistication and escalating volume of cyber-attacks. In this paper, we evaluated the proposed framework for AI integrated cyber security (AICSF) for real-time threat detection and mitigation in smart industry environments. This paper will compare with various baseline models Bayesian classification, SVM, Shallow DL on dataset which consists of network attack data and malware data. This framework operates in an AI-Mode, leveraging a recurrently refined DL architecture for real-time anomaly detection and adversarial learning. The comparative study demonstrated the superiority of proposed AICSF approach in terms of detection accuracy of 97%, F1-score of 98.75%, FPR of less than 1%, and detection rate of 96.12%, which adopting sophisticated AI driven solutions to safeguard high end physical and digital assets.