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:Signal Processing and Communications Lab Seminars
SUMMARY:Multilevel sequential Monte Carlo Samplers. - Dr A
jay Jasra\, National University of Singapore
DTSTART;TZID=Europe/London:20150528T150000
DTEND;TZID=Europe/London:20150528T160000
UID:TALK58691AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/58691
DESCRIPTION:In this talk we consider the approximation of expe
ctations w.r.t. probability distributions associat
ed to the solution of partial differential equatio
ns (PDEs)\; this scenario appears routinely in Bay
esian inverse problems. In practice\, one often ha
s to solve the associated PDE numerically\, using\
, for instance finite element methods and leading
to a discretisation bias\, with the step-size leve
l h_L. In addition\, the expectation cannot be com
puted analytically and one often resorts to Monte
Carlo methods. In the context of this problem\, it
is known that the introduction of the multilevel
Monte Carlo (MLMC) method can reduce the amount of
computational effort to estimate expectations\, f
or a given level of error. This is achieved via a
telescoping identity associated to a Monte Carlo a
pproximation of a sequence of probability distribu
tions. In many practical problems of interest\, on
e cannot achieve an i.i.d. sampling of the associa
ted sequence of probability distributions. A seque
ntial Monte Carlo (SMC) version of the MLMC method
is introduced to deal with this problem. It is sh
own that under appropriate assumptions\, the attra
ctive property of a reduction of the amount of com
putational effort to estimate expectations\, for a
given level of error\, can be maintained within t
he SMC context. This is a joint work with Alex Be
skos (UCL)\, Kody Law (KAUST)\, Raul Tempone (KAUS
T) and Yan Zhou (NUS).\n\n*BIO*: Ajay Jasra receiv
ed his PhD degree in statistics from Imperial Coll
ege London in 2005.\nSince 2011 he has been tenure
d associate professor at the Department of Statist
ics and Applied Probability at the National Univer
sity of Singapore. Between 2005-2008 he has held v
arious post-doctoral positions at the University o
f Oxford\, University of Cambridge and the Institu
te of Statistical Mathematics in Tokyo. He was als
o Chapman Fellow of Mathematics at Imperial Colleg
e London in that period. Between 2008-2011 he was
assistant professor at Imperial College London. He
is currently associate editor at Statistics and C
omputing\, American Journal of Algorithms and Comp
uting and Stat and has over 60 publications.\n\n
LOCATION:LR5\, Department of Engineering
CONTACT:Dr Ramji Venkataramanan
END:VEVENT
END:VCALENDAR