BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:&quot\;Some aspects in high-dimensional Bayesian model choice&quot
 \; - Dr David Rossell\, University of Warwick
DTSTART:20160426T133000Z
DTEND:20160426T143000Z
UID:TALK65253@talks.cam.ac.uk
CONTACT:Alison Quenault
DESCRIPTION:Given a collection of candidate probability models for an obse
 rved data y\, a fundamental statistical task is to evaluate which models a
 re more likely to have generated y. Tackling this problem within a Bayesia
 n framework requires one to complement the probability model for y (likeli
 hood) with a prior probability model on the parameters describing each of 
 the candidate models\, as well as to specify model prior probabilities and
  possibly a utility function. We shall review some recent strategies for h
 igh-dimensional model choice\, and then discuss what we denominate the "mo
 del separation principle". This principle states that the models under con
 sideration should be minimally different from each other\, else it becomes
  hard to distinguish them on the basis of the observed y. In the common se
 tting where some of the models are nested this principle is violated\, as 
 say Model 1 is a particular case of Model 2 and thus these models are not 
 well separated. We shall review a class of prior distributions called non-
 local priors (NLPs) as a way to enforce the model separation principle and
  some of the NLP properties\, focusing on parsimony and accelerated conver
 gence rates in high-dimensional inference. We shall illustrate their use i
 n ongoing work related to regression and robust regression.
LOCATION:Large  Seminar Room\, 1st Floor\, Institute of Public Health\, Un
 iversity Forvie Site\, Robinson Way\, Cambridge
END:VEVENT
END:VCALENDAR