SBIR/STTR Award attributes
The topic of vulnerabilities in machine learning system utilized in the cyber defense domain has not been sufficiently explored. Compared to the counterparts in other domains, the attacks to machine learning systems in the cyber defense domain are more complicated, dynamic and associated with higher cost. With recent deployment of machine learning systems for the network security applications, it is thus in critical need to conduct extensive investigation and evaluation of adversarial machine learning in the cyber defense domain. To address this need, Intelligent Automation, Inc. (IAI) proposes the AdversarIal Machine/Deep Learning (AI-Mining) toolset that develops adversarial machine learning in the cyber defense domain, including the evaluation of machine learning concepts, methods, attack methodology, the impact in the cyber security, and the countermeasures to enhance the defense. The results of this effort will aid in better understanding of machine learning based network security solutions as well as enhanced penetration testing and cyber defensive techniques.
