BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY: Dirichlet Process Mixture Models and Bayesian Nonparametric Densi
 ty Estimation - Andrew Gordon Wilson ()
DTSTART:20120510T130000Z
DTEND:20120510T143000Z
UID:TALK38213@talks.cam.ac.uk
CONTACT:Konstantina Palla
DESCRIPTION:Often we are unsure about what probability density functions t
 o use in our models. Ideally we would like the flexibility to infer any tr
 ue underlying density but without overfitting.  Surprisingly\, this is pos
 sible using Bayesian nonparametric approaches like the Dirichlet process i
 nfinite mixture model.  I will give a tutorial on Dirichlet process mixtur
 e models\, and discuss alternative Bayesian Nonparametric approaches to de
 nsity estimation\, including the Gaussian process density sampler (Adams e
 t. al\, 2009) and Pitman Yor diffusion trees (Knowles and Ghahramani\, 201
 1).
LOCATION:Engineering Department\, CBL Room 438
END:VEVENT
END:VCALENDAR