The Optimal Mechanism in Differential Privacy

Professor Pramod Viswanath
Professor, University of Illinois at Urbana-Champaign
Given on: May 9, 2013


Differential Privacy is a framework to quantify the level of privacy of released data. In this talk we study the fundamental tradeoff between utility and privacy. The optimal privacy mechanism involves adding noise with a staircase shaped probability density function, in a departure from the state of the art Laplacian and geometric noise mechanisms. Several important utility measures (eg: variance of added noise) with the optimal staircase mechanism behave exponentially better than with the Laplacian mechanism as the privacy requirement is reduced. We conclude that the gains of the optimal privacy mechanism over state of the art are pronounced in the moderate and low privacy regimes.


Pramod Viswanath received the PhD degree in EECS from the University of California at Berkeley in 2000. He was a member of technical staff at Flarion Technologies until August 2001 before joining the ECE department at the University of Illinois, Urbana-Champaign. He is a recipient of the Xerox Award for Faculty Research from the College of Engineering at UIUC (2010), the Eliahu Jury Award from the EECS department of UC Berkeley (2000), the Bernard Friedman Award from the Mathematics department of UC Berkeley (2000), and the NSF CAREER Award (2003). He was an associate editor of the IEEE Transactions on Information Theory for the period 2006-2008.