BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:IC Colloquium: From Differential Privacy to Generative Adversarial
  Privacy
DTSTART:20180312T101500
DTEND:20180312T113000
DTSTAMP:20260416T151626Z
UID:0c60e15b89b0f8358a9f7eeab47a9edbfc63ff8896938b01f750f2ab
CATEGORIES:Conferences - Seminars
DESCRIPTION:By: Peter Kairouz - Stanford University\nIC Faculty candidate\
 n\nAbstract:\nThe explosive growth in connectivity and data collection is 
 accelerating the use of machine learning to guide consumers through a myri
 ad of choices and decisions. While this vision is expected to generate man
 y disruptive businesses and social opportunities\, it presents one of the 
 biggest threats to privacy in recent history. In response to this threat\,
  differential privacy (DP) has recently surfaced as a context-free\, robus
 t\, and mathematically rigorous notion of privacy. \n \nThe first part o
 f my talk will focus on understanding the fundamental tradeoff between DP 
 and utility for a variety of unsupervised learning applications. Surprisin
 gly\, our results show the universal optimality of a family of extremal pr
 ivacy mechanisms called staircase mechanisms. While the vast majority of w
 orks on DP have focused on using the Laplace mechanism\, our results indic
 ate that it is strictly suboptimal and can be replaced by a staircase mech
 anism to improve utility. Our results also show that the strong privacy gu
 arantees of DP often come at a significant loss in utility. \n \nThe sec
 ond part of my talk is motivated by the following question: can we exploit
  data statistics to achieve a better privacy-utility tradeoff? To address 
 this question\, I will present a novel context-aware notion of privacy cal
 led generative adversarial privacy (GAP). GAP leverages recent advancement
 s in generative adversarial networks (GANs) to arrive to a unified framewo
 rk for data-driven privacy that has deep game-theoretic and information-th
 eoretic roots. I will conclude my talk by showcasing the performance of GA
 P on real life datasets.\n\nBio:\nPeter Kairouz is a postdoctoral scholar 
 at Stanford University. He received his PhD in ECE from the University of 
 Illinois at Urbana-Champaign (UIUC). He interned twice at Qualcomm and mor
 e recently at Google where he designed privacy-aware machine learning algo
 rithms. He is the recipient of the 2015 ACM SIGMETRICS Best Paper Award\, 
 the 2012 Roberto Padovani Scholarship from Qualcomm's Research Center\, an
 d the 2016 Harold L. Olesen Award for Excellence in Undergraduate Teaching
  from UIUC. His research interests are interdisciplinary and span the area
 s of data and network sciences\, privacy-preserving data analysis\, machin
 e learning\, and information theory.\n\nMore information\n 
LOCATION:BC 420 https://plan.epfl.ch/?room==BC%20420
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR