Title：Backdoor Attacks Are Fun and Expand the Defender’s Imagination
Date：2026/5/20 14:20-16:00
Location：R104, CSIE
Speaker：Prof. Khoa D Doan, VinUniversity
Host：Prof. Shao-Yuan Lo
Abstract:
Robust AI security in general, and backdoor defenses in specific, require understanding the strongest plausible threat models, not just average-case behavior. This lecture introduces the fundamentals of backdoor attacks in deep learning, tracing the evolution of the field from trigger-focused and adaptive attacks to emergence threat models such as quantization, multi-target, and natural backdoors. Beyond individual methods, the talk will also highlight a shift in perspective: from viewing backdoors as simple data artifacts to understanding them as hidden behaviors embedded within the training dynamics, model parameters, and learned representations. Finally, I will discuss potential future directions, including synthetic-data defenses and mechanistic interpretability.
Bio:
Khoa D Doan is currently an Assistant Professor in the College of Engineering and Computer Science (CECS) at VinUniversity, Vietnam and also the Associated Director of VinUni-Illinois Smart Health Center (VISHC), a joint initiative between VinUniversity and the University of Illinois Urbana-Champaign (UIUC). His research focuses on developing computational frameworks that enable the safe/secure and practical deployment of ML models in constrained and especially low-resource applications. Prior to his academic path, he held multiple industry positions, spanning from software developer to research scientist. He received his Ph.D. in Computer Science (machine learning) from Virginia Tech and his M.S. in Computer Science (high-performance distributed computing) from the University of Maryland, College Park.