Electrical and Computer Engineering Faculty Candidate Seminar
Hanshen Xiao
PhD Candidate, Massachusetts Institute of Technology
Monday, April 8, 2024, 10:00-11:00 am
B02 CSL Auditorium or Online via Zoom
Title: When is Automatic Privacy Proof Possible for Black-Box Processing?
Abstract: Can we automatically and provably quantify the information leakage from a black-box processing? From a statistical inference standpoint, in this talk, I will start with a unified framework to summarize existing privacy definitions based on input-independent indistinguishability and unravel the fundamental challenges in crafting privacy proof for general data processing. Yet, the landscape shifts when we gain access to the secret generation. By carefully leveraging the secret entropy, we unlock the black-box analysis. This breakthrough enables us to automatically “learn" the underlying inference hardness for an adversary to recover arbitrary sensitive features related to the secret, and our privacy analysis is fully based on end-to-end simulations by treating the processing mechanism causing the leakage as a black box. Meanwhile, a set of new information-theoretical tools will be introduced to minimize noise perturbation and I will unveil the win-win relationship between privacy and stability for simultaneous algorithm improvements. Concrete applications will be given in diverse domains, including privacy-preserving machine learning, side-channel leakage mitigation and formalizing long-standing heuristic data obfuscations.
Hanshen Xiao is a PhD student in MIT. His research interests lie at the intersection of the fundamentals of information security and privacy, robust statistics and applied cryptography. He received the B.S. degree in Mathematics from Tsinghua University and is the recipient of several awards, including Mathwork Fellowship (2021-2023) and Tsinghua Future Scholar Fellowship (2015-2017). His work has also been supported by DSTA Singapore, Capital One and Cisco.
