|
Speaker: Adam Smith “Privacy in Machine Learning and Statistical Inference” Monday, September 30, 2024 3:30pm – Pre-talk meet and greet teatime – 219 Prospect Street, 13 floor, there will be light snacks and beverages in the kitchen area. |
Add To: Google Calendar | Outlook | iCal File
Zoom Link: https://yale.zoom.us/j/94223816617
Meeting ID: 942 2381 6617
Abstract: The results of learning and statistical inference reveal information about the data they use. This talk discusses the possibilities and limitations of fitting machine learning and statistical models while protecting the privacy of individual records.
I will begin by explaining what makes this problem difficult, using recent work on example memorization as an illustration. I will then present differential privacy, a rigorous definition of privacy in statistical databases that is now widely studied, and increasingly used to analyze and design deployed systems.
Finally, I will present recent algorithmic results on two fundamental problems: differentially private mean estimation and linear regression. We give time- and sample-efficient algorithms for “nicely” distributed (e.g. subgaussian) data. These algorithms adapt automatically to the geometry of the instance at hand, yielding instance-optimal accuracy guarantees.
The talk will be based on joint work with (among others) Gavin Brown, Mark Bun, Vitaly Feldman, Sam Hopkins, and Kunal Talwar (arxiv 2012.06421, 2301.12250, 2404.15409).
Bio: Adam Smith is a Professor of Computer Science and Engineering at Boston University. His research interests lie in data privacy and cryptography, and their connections to machine learning, statistics, information theory, and quantum computing. He obtained his Ph.D. from MIT in 2004 and has held visiting positions at the Weizmann Institute of Science, UCLA, and Harvard. He previously was a Professor of Computer Science and Engineering at Penn State. He received a Presidential Early Career Award for Scientists and Engineers (PECASE) in 2009; a 2016 Theory of Cryptography Test of Time award; the 2019 Eurocrypt Test of Time award; and the 2017 Gödel Prize.
Link to personal site: http://cs-people.bu.edu/ads22/
