Fair Decision Making Using Privacy-Protected Data

Abstract

Differential privacy is a model of privacy protection being adopted by both commercial enterprises and government institutions. Its use protects individuals against privacy harms, but that protection comes at a cost to the accuracy of released data. Motivated by the Census Bureau's recent adoption of differential privacy, I will describe settings in which released data is used to decide who (i.e. which groups of people) will receive a desired resource or benefit. We show that if decisions are made using privacy-protected data, the noise added to achieve privacy may disproportionately impact some groups over others. Thus, while differential privacy offers equal privacy protection to participating individuals, it may result in disparities in utility across groups, with potentially serious consequences for affected individuals. The talk will explain how disparities in accuracy can be caused by commonly-used privacy mechanisms, highlight some of the social choices that arise in the design and configuration of privacy mechanisms, and discuss techniques for mitigating these affects.

Bio:

Gerome Miklau is a Professor of Computer Science at the University of Massachusetts, Amherst. His research focuses on private, secure, and equitable data management. He designs algorithms to accurately learn from data without disclosing sensitive facts about individuals, primarily in the model of differential privacy. He studies fair and responsible data management. He has also designed novel techniques for controlling access to data, limiting retention of data, and resisting forensic analysis.

He recently co-founded Tumult Labs, a start-up focused on commercializing privacy technology. Prior to that, he consulted for the U.S. Census Bureau on algorithms that will be deployed for the 2020 decennial census.

Professor Miklau received the ACM PODS Alberto O. Mendelzon Test-of- Time Award in both 2020 and 2012, the Best Paper Award at the International Conference of Database Theory in 2013, a Lilly Teaching Fellowship in 2011, an NSF CAREER Award in 2007, and he won the 2006 ACM SIGMOD Dissertation Award. He received his Ph.D. in Computer Science from the University of Washington in 2005. He earned Bachelor's degrees in Mathematics and in Rhetoric from the University of California, Berkeley, in 1995.

Please join via Zoom: https://tufts.zoom.us/j/97341864592

Password: see colloquium email

Dial-in not available.