Events Calendar

The Fundamental Limits of Data and Metadata Privacy

Speaker: Peter Kairouz, University of Illinois at Urbana-Champaign

Talk Title: The Fundamental Limits of Data and Metadata Privacy

Abstract: With the ability to surf the web efficiently comes the danger of being monitored. There is an increasing tension between the need to share data and the need to preserve the privacy of Internet users. The need for privacy appears in two main contexts: the data privacy context, as in when individuals want to share their personal data with a potentially malicious service provider or when a trusted service provider wants to release sensitive information about individuals, and the metadata privacy context, as in when individuals want to broadcast information on a social network without the fear of being judged by friends, the public or authorities.

In the metadata privacy context, anonymity is achieved by controlling the way information spreads over a network. In the first half of my talk, I will introduce a novel anonymous messaging protocol (called adaptive diffusion) and show that it spreads a message quickly over a network while "perfectly" hiding authorship information from a powerful adversary with global access to metadata.

In the data privacy context, privacy is achieved by randomizing the data before releasing it. This leads to a fundamental trade-off between privacy and utility. In the second half of my talk, I will present a new class of privacy mechanisms (called staircase mechanisms) and show that they achieve the optimal privacy-utility trade-off under various settings of interest.

Biography: Peter Kairouz is a PHD student at the University of Illinois at Urbana-Champaign. For his masters, he was mainly interested in signal processing and digital communications. He interned twice at Qualcomm (in 2012 and 2013), and was awarded The 2012 Roberto Padovani Scholarship from Qualcomm's Research Center. For his PhD, he chose to work on data and metadata privacy, winning the Best Paper Award at ACM SIGMETRICS 2015. He recently interned at Google, where he designed privacy-aware machine learning algorithms. His primary research interests include privacy enhancing technologies, machine learning, and wireless communications.


Host: Professor Rahul Jain

Location: Hughes Aircraft Electrical Engineering Center (EEB) - 248

Audiences: Everyone Is Invited

Contact: Suzanne Wong