Machine Learning for Cybersecurity

4 credits

Application of machine learning algorithms to cybersecurity. Anomaly detection, spam detection and IP blacklisting, use of natural language processing to improve performance in architecture identification. Optimization methods for determining adversarial inputs to bias the detection and classification outputs of deep neural networks (DNN). Backdoor trigger embedding in DNN. Generative Adversarial Networks. Hands-on, practical course. Prerequisite: either CSE 163 or E E 241; either AMATH 352, MATH 208, MATH 308, or MATH 136; and either IND E 315 , MATH 394/STAT 394, or STAT 390; recommended: E E 445.