BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Sequential Monte Carlo methods for graphical models - Schn\, TB (U
 ppsala Universitet)
DTSTART:20140425T094000Z
DTEND:20140425T101500Z
UID:TALK52187@talks.cam.ac.uk
CONTACT:Mustapha Amrani
DESCRIPTION:Co-authors: Christian A. Naesseth (Linkoping University)\, Fre
 drik Lindsten (University of Cambridge) \n\nWe develop a sequential Monte 
 Carlo (SMC) algorithm for inference in general probabilistic graphical mod
 el. Via a sequential decomposition of the PGM we find a sequence of auxili
 ary distributions defined on a monotonically increasing sequence of probab
 ility spaces. By targeting these auxiliary distributions using purpose bui
 lt SMC samplers we are able to approximate the full joint distribution def
 ined by the graphical model. Our SMC sampler also provides an unbiased est
 imate of the partition function (normalization constant) and we show how i
 t can be used within a particle Markov chain Monte Carlo framework. This a
 llows for better approximations of the marginals and for unknown parameter
 s to be estimated. The proposed inference algorithms can deal with an arbi
 trary graph structure and the domain of the random variables in the graph 
 can be discrete or continuous.\n\nRelated Links: http://arxiv.org/pdf/1402
 .0330v1.pdf - Associated paper \nhttp://user.it.uu.se/~thosc112/index.html
  - Speaker (Thomas Schn) home page \n\n
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR