BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:MRC Biostatistics Unit Seminars
SUMMARY:&quot\;Design for ABC and history matching with Ga
 ussian processes&quot\; - Dr Richard Wilkinson\, S
 heffield University
DTSTART;TZID=Europe/London:20160301T143000
DTEND;TZID=Europe/London:20160301T153000
UID:TALK63429AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/63429
DESCRIPTION:Abstract: History matching and approximate Bayesia
 n computation (ABC) are techniques for parameter e
 stimation for complex models. In both cases\, Gaus
 sian processes can be used to accelerate inference
  by building an emulator/surrogate model either of
  the simulator itself\, or of a summary such as th
 e likelihood function.  Design approaches for buil
 ding emulators usually focus on how to build accur
 ate global approximations. However\, if our aim is
  calibration\, then this can be wasteful. In this 
 talk I will discuss the use of Gaussian processes\
 , and describe design strategies that can be used 
 for history matching and calibration\, both in a d
 eterministic and stochastic setting. Entropy based
  designs that minimize the uncertainty in the clas
 sification surface are particularly useful\, and I
  will show how these can be efficiently computed u
 sing Bayesian optimisation.
LOCATION:Large  Seminar Room\, 1st Floor\, Institute of Pub
 lic Health\, University Forvie Site\, Robinson Way
 \, Cambridge
CONTACT:Alison Quenault
END:VEVENT
END:VCALENDAR