BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Differential Privacy to Machine Learning\, and Back- Part II
DTSTART:20141127T100000
DTEND:20141127T120000
DTSTAMP:20260407T105544Z
UID:a55e22bdff04e98ba550cf1c1f1d5695ad8ae86f99777902e41129bf
CATEGORIES:Conferences - Seminars
DESCRIPTION:Abhradeep Guha Thakurta\, Yahoo! Labs\nIn this tutorial\, I wi
 ll outline some of the recent developments in the area of differentially p
 rivate machine learning\, which exhibit strong connections between statist
 ical data privacy and robust (stable) machine learning. The main theme of 
 this tutorial would be to demonstrate techniques to achieve one from the o
 ther.\n         In the first part\, I will first review some of th
 e high-profile privacy breaches (and also outline some the underlying prin
 ciples behind them) to motivate the need for a rigorous notion of statisti
 cal data  privacy. Subsequently\, I will move on to introduce the notion 
 of differential privacy\, and present some of the basic concepts commonly 
 used in the design of differentially private algorithms. I will conclude p
 art one by showing that  differential privacy is a strong form of regular
 ization (a common technique used in machine learning to control prediction
  error)\, and analyze the  Follow-the-perturbed-leader (FTPL) by Kalai an
 d Vempala' 2005 in this context.\n          In the second part\, 
 I will analyze the robustness (stability) properties of the some of the co
 mmonly used machine learning algorithms (e.g.\, gradient descent\, LASSO\,
  and\, Frank-Wolfe algorithm) and bootstrap the robustness properties into
  obtaining optimal differentially private algorithms for empirical risk mi
 nimization\, and model selection. Some of the robustness analyses are new\
 , and can be of independent interest irrespective of privacy.\nShort Bio:\
 nAbhradeep is currently a research scientist in the Systems research group
  at Yahoo! Labs Sunnyvale. Prior to that he  was a post-doctoral research
 er at Stanford University and Microsoft Research Silicon Valley Campus. He
  did his PhD from Pennsylvania State University. He is primarily intereste
 d in statistical data privacy and its relation to robust machine learning 
 and data mining. More precisely\, he studies the privacy and robustness im
 plications in problems spanning the areas of high-dimensional statistics\,
  risk minimization\, online learning and pattern mining. Recently\, he bee
 n also interested in studying the interrelation between various privacy no
 tions and their practical implications\, and robust property testing.
LOCATION:INM 200
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR