详情

原文传递 Generative Methods and Meta Learning for Cybersecurity.

题名：	Generative Methods and Meta Learning for Cybersecurity.
作者：	Chale, M. W.
摘要：	Cyberspace is the digital communications network that supports the internet ofbattlefield things (IoBT), the model by which defense-centric sensors, computers, ac-tuators and humans are digitally connected. A secure IoBT infrastructure facilitatesreal time implementation of the observe, orient, decide, act (OODA) loop acrossdistributed subsystems. Successful hacking efforts by cyber criminals and strategicadversaries suggest that cyber systems such as the IoBT are not secure. Three linesof effort demonstrate a path towards a more robust IoBT. First, a baseline data set ofenterprise cyber network traffic was collected and modelled with generative methodsallowing the generation of realistic, synthetic cyber data. Next, adversarial examplesof cyber packets were algorithmically crafted to fool network intrusion detection sys-tems while maintaining packet functionality. Finally, a framework is presented thatuses meta-learning to combine the predictive power of various weak models. This re-sulted in a meta-model that outperforms all baseline classifiers with respect to overallaccuracy of packets, and adversarial example detection rate. The National DefenseStrategy underscores cybersecurity as an imperative to defend the homeland andmaintain a military advantage in the information age. This research provides bothacademic perspective and applied techniques to further the cybersecurity posture ofthe Department of Defense into the information age.
总页数：	238 pages