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:Isaac Newton Institute Seminar Series
SUMMARY:Bayesian Calibration of Computer Model Ensembles -
Pratola\, M (Los Alamos National Laboratory)
DTSTART;TZID=Europe/London:20110909T113000
DTEND;TZID=Europe/London:20110909T120000
UID:TALK32738AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/32738
DESCRIPTION:Using field observations to calibrate complex math
ematical models of a physical process allows one t
o obtain statistical estimates of model parameters
and construct predictions of the observed process
that ideally incorporate all sources of uncertain
ty. Many of the methods in the literature use resp
onse surface approaches\, and have demonstrated su
ccess in many applications. However there are nota
ble limitations\, such as when one has a small ens
emble of model runs where the model outputs are hi
gh dimensional. In such instances\, arriving at a
response surface model that reasonably describes t
he process can be dicult\, and computational issu
es may also render the approach impractical.\nIn t
his talk we present an approach that has numerous
beneifts compared to some popular methods. First\,
we avoid the problems associated with defining a
particular regression basis or covariance model by
making a Gaussian assumption on the ensemble. By
applying Bayes theorem\, the posterior distributio
n of unknown calibration parameters and prediction
s of the field process can be constructed. Second\
, as the approach relies on the empirical moments
of the distribution\, computational and stationari
ty issues are much reduced compared to some popula
r alternatives. Finally\, in the situation that ad
ditional observations are arriving over time\, our
method can be seen as a fully Bayesian generaliza
tion of the popular Ensemble Kalman Filter.\n\n
LOCATION:Seminar Room 1\, Newton Institute
CONTACT:Mustapha Amrani
END:VEVENT
END:VCALENDAR