Research Project Title:
Synthesizing Adversarial Attacks Against CPU Schedulers With Reinforcement Learning
abstract:Modern CPU schedulers have a tough job: given the many processes running on a computer at any given time, schedulers need to decide how to equitably divide CPU time between these processes. We propose a framework for attacking CPU schedulers with reinforcement learning. A reinforcement learning algorithm can learn how to exploit weaknesses in CPU scheduling algorithms to generate an attacker program that uses a very small compute budget to effectively cripple a processor. We hope that in building this framework to generate adversarial attacks against processors, we will be able to more effectively identify weaknesses in scheduling algorithms and build defenses to mitigate attacks like ours.
I am participating in SuperUROP to continue doing meaningful research that I enjoy. My classes at MIT have given me many potential questions and ideas to investigate, and I hope to use this opportunity to further my knowledge in computer security. I also recently finished a paper with my lab group and am excited to investigate follow-up work and related research questions.