BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Bayesian optimization: A framework for optimal computational effor
 t for experimental design - Winterfors\, E (Ruprecht-Karls-Universitt Heid
 elberg)
DTSTART:20110719T160000Z
DTEND:20110719T163000Z
UID:TALK32092@talks.cam.ac.uk
CONTACT:Mustapha Amrani
DESCRIPTION:DOE on models involving time or space dynamics is often very c
 omputationally demanding. Predicting a single experimental outcome may req
 uire significant computation\, let alone evaluating a design criterion  an
 d optimizing it with respect to design parameters. To find the exact optim
 um of the design criterion would typically take infinite computation\, and
  any finite computation will yield a result possessing some uncertainty (d
 ue to approximation of the design criterion as well as stopping the optimi
 zation procedure). Ideally\, one would like to optimize not only the desig
 n criterion\, but also the way it is approximated and optimized in order t
 o get the largest likely improvement in the design criterion relative to t
 he computational effort spent. Using a Bayesian method for the optimizatio
 n of the design criterion (not only for calculating the design criterion) 
 can accomplish such an optimal trade-off between (computational) resources
  spent planning the experiment and expected gain from carrying it out.  Th
 is talk will lay out the concepts and theory necessary to perform a fully 
 Bayesian optimization that maximizes the expected improvement of the desig
 n criterion in relation the computational effort spent.\n
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR