Project Grant 2140175

This three-year, $425,000 Project Grant from the National Science Foundation's Division of Electrical, Communications and Cyber Systems, under the Engineering program (CFDA 47.041), will fund research at The Pennsylvania State University to develop provably secure, online learning-based adaptive cyber defenses for real-world servers. The research aims to provide mathematically rigorous guarantees by synthesizing new models integrating machine learning, game theory, and control theory. It will develop a co-design framework involving dynamic runtime environments, platforms, and networks to protect data centers against stochastic, strategic, and multi-stage attacks. The framework integrates novel mathematical models, online learning algorithms, and server retrofitting to bridge models to real-world systems. Hackathon events will engage students in machine learning and cybersecurity research. Results will be shared with stakeholders to improve adaptive and autonomous defense capabilities against zero-day attacks.