Adversarial Machine Learning for Detecting Advanced Threats Inspired by StuxNet in Critical Infrastructure Networks
Hafiz Bilal Ahmad, Haichang Gao, Naila Latif, Abdul Aziiz, Muhammad Auraangzeb, Muhammad Tanveer Khan · 2024
Inspired by the notorious StuxNet malware, this paper presents a unique and resilient architecture to address the imminent issue of identifying advanced threats within critical infrastructure networks. We use adversarial machine learning techniques to keep ahead of ever-changing threats. Our defense system's fundamental component is a Multi-Layer Perceptron (MLP) architecture that we've dubbed “StuxNet.” This innovative method combines threat intelligence, anomaly detection, and deep learning to forecast and counter sophisticated attacks. The effectiveness of our method in bolstering the security of critical infrastructure networks is supported by empirical evaluations and real-world case studies. Importantly, the StuxNet-MLP architecture has a 98 percent detection accuracy for sophisticated threats. These results highlight the paramount significance of preventative cybersecurity measures in protecting mission-critical systems and provide useful insights for fending off sophisticated attacks. To ensure a safer and more resilient digital future, this study highlights the need to stay ahead in the continuous battle to secure key infrastructure.