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:Statistics
SUMMARY:Alpha-Stable Poisson-Kingman Processes: Some Appli
cations and Methodologies - Yee Whye Teh\, Univers
ity of Oxford
DTSTART;TZID=Europe/London:20140314T160000
DTEND;TZID=Europe/London:20140314T170000
UID:TALK51095AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/51095
DESCRIPTION:Poisson-Kingman partitions reposed on alpha-stable
processes have been well-studied\, e.g. by Pitman
(2003) who first introduced them\, and Gnedin and
Pitman (2006) who characterized them as a particu
lar class of Gibbs-type exchangeable random partit
ions. They have recently\nenjoyed increasing inte
rest in the Bayesian nonparametrics community for
being the largest known class of random probabilit
y measures that have significant advantages in ter
ms of mathematical and computational tractability.
\n\nIn this talk I will give an overview to alpha-
stable Poisson-Kingman processes from the perspect
ive of using them in statistical modelling applica
tions. Specifically I will describe their use in
Bayesian\nnonparametric mixture models for cluster
ing and density estimation problems\, and in estim
ating discovery probabilities in special sampling
problems. I will also describe computational techn
iques which allow us to perform efficient Markov c
hain Monte Carlo inference\nfor these processes.\n
\nJoint work with Stefano Favaro and Maria Lomeli.
LOCATION:MR12\, Centre for Mathematical Sciences\, Wilberf
orce Road\, Cambridge
CONTACT:
END:VEVENT
END:VCALENDAR