SPARTA
Secure and Privacy-Aware Reinforcement Learning Agents. Cluster of Excellence CASA, with Veelasha Moonsamy (2026–2029).
Reinforcement learning agents increasingly make decisions in settings where attackers may manipulate what they observe and where the data they learn from is sensitive. SPARTA develops reinforcement learning agents that stay safe and secure under such adversarial conditions and that respect privacy, for instance by being able to remove (unlearn) information they should no longer use, without retraining from scratch.
SPARTA is a fundamental research project of the Cluster of Excellence CASA (Cyber Security in the Age of Large-Scale Adversaries) at Ruhr University Bochum, led jointly with Veelasha Moonsamy.