Machine Beats Machine: Machine Learning Models to Defend Against Adversarial Attacks

We propose using a two-layered deployment of machine learning Artificial Intelligence (AI) solutions have penetrated the Industry models to prevent adversarial attacks. The first layer determines 4.0 domain by revolutionizing the rigid production lines enabling whether the data was tampered, while the second layer solves a innovative functionalities like mass customization, predictive maintenance, domain-specific problem. We explore three sets of features and zero defect manufacturing, and digital twins. However, three dataset variations to train machine learning models. Our results AI-fuelled manufacturing floors involve many interactions between show clustering algorithms achieved promising results. In the AI systems and other legacy Information and Communications particular, we consider the best results were obtained by applying Technology (ICT) systems, generating a new territory for malevolent the DBSCAN algorithm to the structured structural similarity index actors to conquer. Hence, the threat landscape of Industry 4.0 is measure computed between the images and a white reference expanded unpredictably if we also consider the emergence of adversary image.