BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Advanced Gaussian Process approximation methods - Dr. Richard Turn
 er\, Thang Bui
DTSTART:20140522T140000Z
DTEND:20140522T153000Z
UID:TALK52792@talks.cam.ac.uk
CONTACT:16537
DESCRIPTION:Inference and learning in regression or classification tasks u
 sing Gaussian Processes is expensive ($O(n^3)$) and practitioners often ne
 ed to resort to approximation techniques. There is now a large family of s
 o-called sparse approximation methods which summarise the observed data us
 ing a small number of pseudo-datapoints. These sparse methods can be categ
 orised into two non-exclusive classes: indirect posterior approximations w
 hich employ a modified prior designed to approximately match the original 
 (e.g. FITC\, PITC\, sparse spectrum GPs) and direct posterior approximatio
 ns which explicitly optimise an approximation to the posterior. In this ta
 lk we will discuss two sparse methods in the latter scheme. The first appr
 oximates the posterior using the variational free energy approach (Titsias
 \, 2009) and the second uses expectation propagation (Qi et al\, 2010). 
LOCATION:Engineering Department\, CBL Room 438
END:VEVENT
END:VCALENDAR