ASTERICS
Full Title
Adversarial Strategies for counTERing Intelligent CyberattackSDescription
ASTERICS aims to enhance the cybersecurity of Operational Technology (OT)-based Cyber-Physical Systems (CPS) by developing innovative defence strategies against AI-driven cyberattacks. The project explores the novel use of Adversarial Machine Learning (AML), particularly data poisoning techniques, to disrupt the learning processes of intelligent attacks. These techniques are not intended for deployment in real systems in operation, to avoid any functional impact. Instead, they will be applied in two scenarios: in datasets that may be made accessible to attackers; and in honeynets designed to lure attackers by mimicking real systems. By researching both the offensive and defensive aspects of intelligent attacks, the project aims to enhance the security of industrial systems and critical infrastructures. Its impact goes beyond a better understanding of how AI techniques can be used for cyberattacking purposes. It advances to the novel use of AML as a defence mechanism against intelligent attacks, unfolding possibilities of new studies in the field beyond the duration of the project.