Hume Seminar Series: Future of Automated Penetration Testing with Reinforcement Learning

About the Event

Nation-states increasingly target national security and domestic prosperity by way of small and medium-sized businesses (SMBs). Automated approaches to cyber test are the only affordable option for fortifying the cyber defense of SMBs. Automated penetration testing is an important aspect of automated cyber test and re-engineering. In recent years, there has been a rapid increase in the use of reinforcement learning (RL) that takes lessons learned from applying RL to Starcraft, DOTA, and other video games and applies them to penetration testing. In this talk, the concept of whole campaign emulation will be proposed as a benchmark and roadmap for the development of future automated penetration testing capabilities with reinforcement learning

About the Speaker

Tyler Cody, Ph.D.

Tyler Cody, Ph.D., is an assistant research professor in the Intelligent System Division of the Virginia Tech National Security Institute. Tyler's research interests are developing first principles for the systems engineering of artificial intelligence. Recently, he has developed various reinforcement learning algorithms for penetration testing, including for crown jewel analysis (CJA), exfiltration (Exfil), surveillance detection routes (SDR), and command and control (C2). More broadly, Tyler works in systems and test engineering for machine learning, with applications to cyber, prognostics and health management, cognitive communications, and other areas where machine learning is applied to engineered systems. Tyler received his Ph.D. in systems engineering from the University of Virginia in 2021 for his work on a systems theory of transfer learning.

If you are an individual with a disability and desire an accommodation, please contact Sarah Downer at humecenter@vt.edu during regular business hours at least 10 business days prior to the event.