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:Deterministic RBF Surrogate Methods for Uncertaint
y Quantification\, Global Optimization and Paralle
l HPC Applications - Christine Shoemaker (National
University of Singapore)
DTSTART;TZID=Europe/London:20180208T100000
DTEND;TZID=Europe/London:20180208T110000
UID:TALK100204AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/100204
DESCRIPTION:Co-author: Antoine Espinet (Cornell Univers
ity)
This talk will describe
general-purpose algorithms for global optimizatio
n These algorithms can be used to estimate model p
arameters to fit complex simulation models to data
\, to select among alternative options for design
or management\, or to quantify model uncertainty.
In general the numerical results indicate these al
gorithms do very well in comparison to alternative
s\, including Gaussian Process based approaches..
Prof. Shoemaker&rsquo\;s group has developed open
source (free) PySOT optimization software that is
available online (18\,000 downloads) . The algori
thms can be run in serial or parallel. The focus o
f the talk will be on SOARS\, an Uncertainty Quant
ification method for using optimization-based sa
mpling to build a surrogate likelihood function fo
llowed by additional sampling The algorithms build
s a surrogate approximation of the likelihood func
tion based on simulations done during the optimiza
tion search. Then MCMC is performed by evaluating
the surrogate likelihood function rather than the
original expensive-to-evaluate function. Numerica
l results indicate the SOARS algorithm is very acc
urate when compared to the posterior densities com
puted when using the expensive exact likelihood f
unction. I also discuss an application to a mode
l of the underground movement of a plume of geolog
ically sequestered carbon dioxide. The uncertain
ty in the parameter values obtained from the MCMC
analysis on the surrogate likelihood function can
be used to assess alternative strategies for ident
ifying a cost-effective plan that will most effici
ently give a reliable forecast of a carbon dioxide
underground plume. This includes joint work with
David Ruppert\, Antoine Espinet\, Nikolay Bliznyuk
\, and Yilun Wang.
Related Links
LOCATION:Seminar Room 1\, Newton Institute
CONTACT:INI IT
END:VEVENT
END:VCALENDAR