Machine Learning in Side Channel Analysis and Defenses

Abstract: The emerging, more sophisticated hardware-oriented side-channel attacks impact various computing platforms, from edge to cloud continuum, threatening computer systems owned by individuals, organizations, or governments. Compared with software-oriented malware, hardware-oriented attacks demand more dedicated detection and defense approaches beyond upgrading since they exploit hardware vulnerabilities. In this talk, I will talk about hardware-oriented side-channel vulnerabilities that exist in emerging workloads and computing platforms. Secondly, I will introduce our approach of defending critical computing system and sensitive application against side-channel attacks.

Bio: Han Wang is an assistant professor at Temple University, ECE department. She received her Ph.D. from the Department of Electrical and Computer Engineering, University of California Davis, advised by Prof. Houman Homayoun. Her research interests include side-channel attacks, machine learning, and embedded systems. In particular, she has been developing machine learning algorithms for effectively detecting microarchitectural attacks, proposing lightweight defense mechanisms to protect the computer system from attacks exploiting hardware vulnerability, and security assessment of emerging applications like deep learning in the face of hardware-oriented attacks. She participated in the Intel-sponsored Noyce research project for the security and privacy of AI-enabled IoT Eco-Systems. She has been published in top conferences and journals, including USENIX Security, NDSS, CCS, DATE, ICCAD, ICCD, and JETCAS. She received the DAC Young Student Fellowship in 2018 and Young Engineer of the Year in IEEE Philadelphia Section 2023.