BEGIN:VCALENDAR
VERSION:2.0
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:
X-WR-TIMEZONE:Europe/Paris
BEGIN:VTIMEZONE
TZID:Europe/Paris
X-LIC-LOCATION:Europe/Paris
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700329T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701025T030000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
SUMMARY:
Proving Differential Privacy via Probabilistic Couplings
STATUS:CONFIRMED
ATTENDEE;CN="Pierre-Yves Strub
":
MAILTO:no@spam.com
DESCRIPTION:
Differential privacy is a promising formal approach to data
privacy\, which provides a quantitative bound on the privacy
cost of an algorithm that operates on sensitive information
. Several tools have been developed for the formal verificat
ion of differentially private algorithms\, including program
logics and type systems. However\, these tools do not captu
re fundamental techniques that have emerged in recent years\
, and cannot be used for reasoning about cutting-edge differ
entially private algorithms. Existing techniques fail to han
dle three broad classes of algorithms: 1) algorithms where p
rivacy de- pends on accuracy guarantees\, 2) algorithms that
are analyzed with the advanced composition theorem\, which
shows slower growth in the privacy cost\, 3) algorithms that
interactively accept adaptive inputs. In this presentation\
, I will develop compositional methods for formally verifyin
g differential privacy for algorithms whose analysis goes be
yond the composition theorem. The methods are based on the o
bservation that differential privacy has deep connections wi
th a generalization of probabilistic couplings\, an establis
hed mathematical tool for reasoning about stochastic process
es. Even when the composition theorem is not helpful\, we ca
n often prove privacy by a coupling argument. I will illustr
ate this approach through a single running example\, which e
xemplifies the three classes of algorithms and explores new
variants of the Sparse Vector technique\, a well-studied alg
orithm from the privacy literature.
DTSTART;TZID=Europe/Paris:20161122T110000
DURATION:PT1H
URL;VALUE=URI:http://www.lsv.ens-cachan.fr/Seminaires/?sem=201611221
100
UID:LSVsemLSV.201611221100@lsv.ens-cachan.fr
LOCATION:Salle de ConfĂ©rence (Pavillon des Jardins)
END:VEVENT
END:VCALENDAR