BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Bayesian Inference with Kernels - UCL
DTSTART:20101110T140000Z
DTEND:20101110T150000Z
UID:TALK27166@talks.cam.ac.uk
CONTACT:Microsoft Research Cambridge Talks Admins
DESCRIPTION:An embedding of probability distributions into a reproducing k
 ernel Hilbert space (RKHS) has been introduced: like the characteristic fu
 nction\, this provides a unique representation of a probability distributi
 on in a high dimensional feature space. This representation forms the basi
 s of an inference procedure on graphical models\, where the likelihoods ar
 e represented as RKHS functions. The resulting algorithm is completely non
 parametric: all aspects of the model are represented implicitly\, and lear
 ned from a training sample. Both exact inference on trees and loopy BP on 
 pairwise Markov random fields are demonstrated. Kernel message passing can
  be applied to general domains where kernels are defined\, handling challe
 nging cases such as discrete variables with huge domains\, or very complex
 \, non-Gaussian continuous distributions. In experiments\, the approach ou
 tperforms state-of-the-art techniques in a cross-lingual document retrieva
 l task and in the prediction of depth from 2-D images. Finally\, time perm
 itting\, a more general kernelized Bayes` law will be described\, in which
  a prior distribution embedding is updated to provide a posterior distribu
 tion embedding. This last approach makes weaker assumptions on the underly
 ing distributions\, but is somewhat more complex to implement. Joint work 
 with Danny Bickson\, Kenji Fukumizu\, Carlos Guestrin\, Yucheng Low\, Le S
 ong 
LOCATION:Small public lecture room\, Microsoft Research Ltd\, 7 J J Thomso
 n Avenue (Off Madingley Road)\, Cambridge
END:VEVENT
END:VCALENDAR