BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Probabilistic numerics: treating numerical computation as learning
 \, or\; it's Bayes all the way down - Michael Osborne (Oxford University)
DTSTART:20150512T100000Z
DTEND:20150512T110000Z
UID:TALK59442@talks.cam.ac.uk
CONTACT:Dr Jes Frellsen
DESCRIPTION:This talk will introduce the probabilistic numerics framework.
  Probabilistic numerics interprets numerical procedures (e.g. optimisation
 \, linear algebra\, integration) as demanding Bayesian inference. This int
 erpretation allows: uncertainty management at all levels of an algorithm\;
  for the benefits of structure in numerical tasks to be realised\, and\; f
 or no more costly computation to be allocated to any constituent numerical
  algorithm than is necessary to achieve our overall goals. The talk will p
 articularly focus on recent work in probabilistic approaches to numerical 
 integration: Bayesian quadrature\, a robust alternative to MCMC methods. A
 pplications of the techniques will be demonstrated to domains including as
 trometry and sensor networks\, illustrating the superior wall-clock perfor
 mance of probabilistic numeric techniques.
LOCATION:Engineering Department\, LR5 (Baker Building\, Trumpington Street
 )
END:VEVENT
END:VCALENDAR