BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Inference for binary Markov random fields without tears (or MCMC).
  - Nial Friel\, University of Glasgow
DTSTART:20070605T133000Z
DTEND:20070605T143000Z
UID:TALK6779@talks.cam.ac.uk
CONTACT:Nikolaos Demiris
DESCRIPTION:Binary Markov random fields have played an important role in t
 he development of MCMC methods. The Metropolis algorithm was designed to s
 ample from the Ising model\, for example. In this talk we will present an 
 non-MCMC based approach to carrying out inference for such models. Here I 
 will present an algorithm which allows\, for example\, direct sampling fro
 m the Markov random field\, computation of marginal distributions of latti
 ce points\, but crucially which also allows inference for the parameters o
 f the MRF. These algorithms can be carried out exactly if the dimension of
  the lattice is of moderate size. We present an approximate inference sche
 me for the case where the lattice is of large dimension. This is joint wor
 k with Havard Rue (NTNU\, Trondheim).
LOCATION:Large Seminar Room\, 1st Floor\, Institute of Public Health\, Uni
 versity Forvie Site\, Robinson Way\, Cambridge
END:VEVENT
END:VCALENDAR