Through the development of powerful algorithms and design tools, machine learning, especially deep neural network (DNN), is becoming state-of-the-art in various fields. However, due to the ubiquity and complexity of DNN models, recent works have shown that they are quite vulnerable to many categories of adversarial attacks, thereby posing serious challenges and security concerns on the practical deployment of artificial intelligence technologies. This talk will discuss machine learning security and privacy from both the algorithmic and hardware perspectives. I will share our recent results on the backdoor attack, model watermarking, and privacy-preserving machine learning.

Yingjie Lao is currently an assistant professor in the Department of Electrical and Computer Engineering at Clemson University. He received the B.S. degree from Zhejiang University, China, in 2009, and the Ph.D. degree from the Department of Electrical and Computer Engineering at University of Minnesota, Twin Cities, in 2015. Prior to joining Clemson, he spent one year at Broadcom Corporation. His research has been recognized with an NSF CAREER Award, an IEEE Circuits and Systems Society Very Large Scale Integration Prize Paper Award, and an ISLPED Best Paper Award. His research interests include trusted AI, hardware security, VLSI architectures for machine learning and emerging cryptographic systems, cybersecurity, and robotics.