Security-Driven, Explainable Artificial Intelligence for Undergraduates

We propose to develop a new senior undergraduate class on security-driven explainable AI (XAI), with application to detecting backdoor attacks on deep learning in particular. The course will be available to a wide variety of students in science and engineering at Penn State, a school with a large number of Navy ROTC students. We will also run a set of ROTC research experience internships (both in summer and during school semesters), where the summerinternships will be available to ROTC students outside of Penn State. These internships will be based on different projects, some of which will be drawn from prior and on-going research of the PIs, all of which will be on security of AI or application of AI to cyber security, and all of which will provide exposure to working with real-world datasets. In extension years, the XAI course and security/AI projects will be revised and expanded. Also an additional course on secure and efficient cloud computing and a related set of ROTC research-experience projects on secure cloud computing will be developed.