Cyber Security: A New Front for Computational Science and Engineering
Securing cyber systems is paramount, but cyber defenders lack evidence-based techniques required to make high-consequence decisions. With lack of principled and rigorous measurements and models, cyber defenders resort to heuristics and expert intuitions. Cyber experimentation is commonly used in security, and we approach this problem as a new front in computation science and engineering. We use cyber emulation as a predictive capability and support this capability with uncertainty quantification (UQ) techniques to verify experiments, incorporate uncertainties in experiments for rigorous experimentation.
Finally, we use adversarial optimization to improve our defenses. The cyber systems are different from physics-based systems for which many UQ and optimization techniques were developed, and thus we need novel approaches. This talk will summarize our progress on this problem and point to the new challenges we are facing.