ALCHEMI: Attacker Learning in Cybernetworks using Heterogeneous Energy-guided Model Inference

The United States relies on networks of cyber-physical systems to conduct military and commercial operations, such as logistics, transportation, information sharing, energy production and distribution, financial transactions, elections, and infrastructure management. As the volume and diversity of cyber-attacks on these networks dramatically increase, there is a growing need for advanced tools and techniques to defend these networks. With this in mind, Aptima, Inc. and its academic partner at the University of South Florida, Prof. Xinming (Simon) Ou, propose to develop ALCHEMI: Attacker Learning in Cybernetworks using Heterogeneous Energy-guided Model Inference. ALCHEMI will provide operators supporting Defensive Cyber Operations (DCOs) with tools that automate the data collection and analysis of network observables. With a novel integration of machine learning, artificial intelligence, and dynamic attack graph technologies ALCHEMI will provide a suite of functionalities enabling visualization of both the current state of computer networks and the threats against it, as well as inferences regarding the goals of the attackers.