A HYBRID DEEP LEARNING ARCHITECTURE FOR ROBUST VULNERABILITY CLASSIFICATION

The increasing sophistication of cyber threats, particularly against critical infrastructure like healthcare Internet of Things (IoT) networks, demands advanced vulnerability management solutions. This paper presents a novel AI-driven Deep Packet Inspection (DPI) model for proactive vulnerability assessment and management. In the proposed study, a hybrid deep learning architecture, which combines a Convolutional Neural Network (CNN) to extract hierarchical features of network traffic with an ensemble of Feed-Forward Neural Networks to perform powerful vulnerability classification, was proposed. The model was trained and evaluated on an integrated dataset of more than 900,000 records, consisting of real-world vulnerability data of a healthcare IoT testbed and Common Vulnerabilities and Exposures (CVE) database. Our ensemble-based classifier showed outstanding results and the component models had test accuracy of up to 99.31%. The suggested system was implemented in such a way that it became a full-fledged Vulnerability Assessment and Management System with a rule-based decision support algorithm to score threats according to the National Vulnerability Database (NVD) standard and prioritize them. The results of the experiment show that the combined framework manages to automate the process of real-time packet capture and analysis to vulnerability reporting to action. The study offers a scalable, precise and holistic solution that can improve cybersecurity posture in critical network infrastructures that can effectively ensure that the disparity between AI-enabled threat detection and practical vulnerability management converges.