Adversarially Robust Industrial Control Systems Intrusion Detection Using A Generative Purification Pipeline

AI has become one of the leading changes in the world and with the growth, security concerns have been increasing linearly as well. Looking for methods that could improve robustness, and security will always be helpful with the direction the world is going. Investigating different adversarial attacks and ways to create countermeasures can be seen as a priority. Cybersecurity has also become a hot topic in recent years among the technology community. More and more attention is being paid towards improving security and defense techniques. This paper proposes a new approach that uses Generative Adversarial Networks to purify perturbations, thereby enhancing the security of the industrial control systems (ICS) against adversarial attacks.