Adaptation and Deception in Adversarial Cyber Operations
George V. Cybenko · 2020
This chapter outlines a variety of challenges related to adaptation in adversarial operations. Such issues arise in autonomous systems that operate in hostile environments, where the adversaries can adapt and shape the environment as well. This includes national security systems like autonomous cyber operations, battlefield Internet of Things, unmanned vehicles, electronic warfare, and social media information operations. We note that recent progress in reinforcement learning has resulted in remarkable advances in game playing technology, to the extent that machines are now demonstrating superior play against human experts in games such as Go and some versions of Texas Hold 'Em poker. Those advances use a variety of combinations of computational game theory and deep reinforcement learning. However, there are several fundamental reasons that such successes do not translate immediately to more complex adversarial interactions such as arise in national security operations. This chapter reviews some of those reasons and how they relate to deceptions that are possible by opponents in adaptive, adversarial systems.