BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Pre-Processing for Approximate Bayesian Computation in Image Analy
 sis - Mengersen\, K (Queensland University of Technology)
DTSTART:20140424T124500Z
DTEND:20140424T132000Z
UID:TALK52166@talks.cam.ac.uk
CONTACT:Mustapha Amrani
DESCRIPTION:Co-authors: Matthew Moores (QUT\, Australia)\, Christian Rober
 t (U. Paris Dauphine\, France) \n\nExisting algorithms for approximate Bay
 esian computation (ABC) assume that it is feasible to simulate pseudo-data
  from the model at each iteration. However\, the computational cost of the
 se simulations can be prohibitive for high dimensional data. An important 
 example is the Potts model\, which is commonly used in image analysis. The
  dimension of the state vector in this model is equal to the size of the d
 ata\, which can be millions of pixels. We introduce a preparatory computat
 ion step before model fitting to improve the scalability of ABC. The outpu
 t of this precomputation can be reused across multiple datasets. We illust
 rate this method by estimating the smoothing parameter for satellite image
 s\, demonstrating that the pre-computation step can sufficiently reduce th
 e average runtime required for model fitting to enable analysis in realist
 ic\, if not yet real\, time. \n
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR