BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Foundations of Nonparametric Bayesian Methods (Part I) - Peter Orb
 anz (University of Cambridge)
DTSTART:20081009T150000Z
DTEND:20081009T170000Z
UID:TALK14246@talks.cam.ac.uk
CONTACT:Peter Orbanz
DESCRIPTION:This 3-part tutorial will address a machine learning audience\
 , not assumed to be familiar with measure theory or the theory stochastic 
 processes. The course is intended to provide (1) an overview of what nonpa
 rametric Bayesian models exist beyond those already used in\nmachine learn
 ing\, and (2) a basic understanding of the mathematical construction of ''
 process'' models\, both existing ones and new models on a variety of possi
 ble domains.\n\nPart I: Basics\n\nAt least half of the first part will pro
 bably be spent reviewing concepts from measure-theoretic probability (and 
 motivating why we need them for Bayesian nonparametrics). We will then def
 ine Bayesian estimation in these terms\, introduce the basic construction 
 tools for stochastic processes\, and see how the Gaussian and Dirichlet pr
 ocesses are constructed from finite-dimensional Gaussian and Dirichlet dis
 tributions. \n\nWebpage:\nhttp://mlg.eng.cam.ac.uk/porbanz/npb-tutorial.ht
 ml
LOCATION:Engineering Department\, CBL Room 438
END:VEVENT
END:VCALENDAR