BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:Isaac Newton Institute Seminar Series
SUMMARY:Privacy for Bayesian modelling - Anne-Sophie Chare
st (Université Laval)
DTSTART;TZID=Europe/London:20160728T153000
DTEND;TZID=Europe/London:20160728T163000
UID:TALK66916AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/66916
DESCRIPTION:The literature now contains a large set of methods
to privately estimate parameters from a classical
statistical model\, or to conduct a data mining o
r machine learning task. However\, little is known
about how to perform Bayesian statistics privatel
y. In this talk\, I will share my thoughts\, and a
few results\, about ways in which Bayesian modell
ing could be performed to offer some privacy guara
ntee. In particular\, I will discuss some attempts
at sampling from posterior predictive distributio
ns under the constraint of differential privacy (D
P). I will also discuss empirical differential pri
vacy\, a criterion designed to estimate the DP pri
vacy level offered by a certain Bayesian model\, a
nd present some recent results on the meaning and
limits of this privacy measure. A lot of what I wi
ll present is work in progress\, and I am hoping t
hat some of you may want to collaborate with me on
this research topic.
LOCATION:Seminar Room 2\, Newton Institute
CONTACT:INI IT
END:VEVENT
END:VCALENDAR