BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//wp-events-plugin.com//7.4.0.1//EN
TZID:America/New_York
X-WR-TIMEZONE:America/New_York
BEGIN:VEVENT
UID:630@fds.yale.edu
DTSTART;TZID=America/New_York:20240911T113000
DTEND;TZID=America/New_York:20240911T130000
DTSTAMP:20250916T142137Z
URL:https://fds.yale.edu/events/fds-colloquium-sofya-raskhodnikova-boston-
 privately-evaluating-untrusted-black-box-functions/
SUMMARY:FDS Colloquium: Sofya Raskhodnikova (Boston)\, "Privately evaluatin
 g untrusted black-box functions"
DESCRIPTION:Abstract: \n\n\n\nWe provide tools for sharing sensitive data 
 in situations when the data curator does not know in advance what question
 s an (untrusted) analyst might want to ask about the data. The analyst can
  specify a program that they want the curator to run on the dataset. We mo
 del the program as a black-box function $f$. We study differentially priva
 te algorithms\, called privacy wrappers\, that\, given black-box access to
  a real-valued function $f$ and a sensitive dataset $x$\, output an accura
 te approximation to $f(x)$. The dataset $x$ is modeled as a finite subset 
 of a possibly infinite universe\, in which each entry $x$ represents the d
 ata of one individual. A privacy wrapper calls $f$ on the dataset $x$ and 
 on some subsets of $x$ and returns either an approximation to $f(x)$ or a 
 nonresponse symbol $perp$. The wrapper may also use additional information
  (that is\, parameters) provided by the analyst\, but differential privacy
  is required for all values of these parameters. Correct setting of these 
 parameters will ensure better accuracy of the privacy wrapper. The bottlen
 eck in the running time of our privacy wrappers is the number of calls to 
 $f$\, which we refer to as queries. Our goal is to design privacy wrappers
  with high accuracy and small query complexity. \n\n\n\nWe consider two s
 ettings: in the automated sensitivity detection setting\, the analyst supp
 lies only the black-box function $f$ and the intended (finite) range of $f
 $\; in the provided sensitivity bound setting\, the analyst also supplies 
 additional parameters that describe the sensitivity of $f$. We define accu
 racy for both settings. We present the first privacy wrapper for the autom
 ated sensitivity detection setting. For the setting where a sensitivity bo
 und is provided by the analyst\, we design privacy wrappers with simultane
 ously optimal accuracy and query complexity\, improving on the constructio
 ns provided (or implied) by previous work. We also prove tight lower bound
 s for both settings. In addition to addressing the black-box privacy probl
 em\, our private mechanisms provide feasibility results for the differenti
 ally private release of general classes of functions. \n\n\n\nJoint work 
 with Ephraim Linder\, Adam Smith\, and Thomas Steinke\n\n\n\nBio: Sofya 
 Raskhodnikova is a professor of Computer Science at Boston University. He
 r office is located in the newly constructed Center for Computing and Data
  Sciences\, colloquially known as the Jenga building. She received all her
  degrees (S.B.\, S.M.\, and Ph.D.) from MIT\, and then was a postdoctoral 
 fellow at the Hebrew University of Jerusalem and the Weizmann Institute of
  Science. She was a Professor of Computer Science and Engineering at Penn 
 State and held visiting positions at the Institute for Pure and Applied Ma
 thematics at UCLA\, Boston University\, Harvard University\, and at the Si
 mons Institute for the Theory of Computing at Berkeley. Sofya works in t
 he areas of randomized and approximation algorithms. Her main interest is 
 the design and analysis of sublinear-time algorithms for combinatorial pro
 blems. She has also made important contributions to data privacy.\n
CATEGORIES:FDS Events,Colloquium,Seminar Series
END:VEVENT
BEGIN:VTIMEZONE
TZID:America/New_York
X-LIC-LOCATION:America/New_York
BEGIN:DAYLIGHT
DTSTART:20240310T030000
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
END:DAYLIGHT
END:VTIMEZONE
END:VCALENDAR