Open Call - Autonomous Decision Making for Cyber Defence

Under this open call, the Authority is seeking novel Artificial Intelligence (AI) and Machine Learning (ML) approaches for autonomous cyber defence decision making.

Specifically, the Authority is interested in research that aims to: • Develop AI and ML based approaches for autonomous response options planning.

This could include (but is not limited to) the application of reinforcement learning, adversarial machine learning, game theory etc.

Response options could include (but are not limited to): implementation of technical mitigation measures; initiating actions to increase information veracity or certainty before implementing a mitigation response; or initiating actions to identify the cause of system failure in order to recover from it. • Develop multi agent approaches and architectures for cyber defence decision making.

Key aspects include the trade-off between centralised and de-centralised agents, approaches for information sharing between agents, agent hierarchy and multi agent consensus.

Note that this should focus on the interaction of machine agents and not the interaction of humans with machine agents. • Develop methods and approaches to evaluating the decisions generated by the agents to determine their effectiveness and impact.