BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Sequential Monte Carlo methods for applications in Data Assimilati
 on - Beskos\, A (National University of Singapore)
DTSTART:20140423T124500Z
DTEND:20140423T132000Z
UID:TALK52129@talks.cam.ac.uk
CONTACT:Mustapha Amrani
DESCRIPTION:Sequential Monte Carlo (SMC) methods are nowadays routinely ap
 plied in a variety of complex applications: hidden Markov models\, dynamic
 al systems\, target tracking\, control problems\, just to name a few. Wher
 eas SMC have been greatly refined in the last decades and are now much bet
 ter understood\, they are still known to suffer from the curse of dimensio
 nality: algorithms can sometimes break down exponentially fast with the di
 mension of the state space. As a consequence\, practitioners in high-dimen
 sional Data Assimilation applications in atmospheric sciences\, oceanograp
 hy and elsewhere will typically use 3D-Var or Kalman-filter-type approxima
 tions that will provide biased estimates in the presence of non-linear mod
 el dynamics. \n\nThe talk will concentrate on a class of SMC algorithms an
 d will look at ways to reduce the cost of the algorithms as a function of 
 the dimension of the state space. Explicit asymptotic results will clarify
  the effect of the dimension at the properties of the algorithm and could 
 provide a platform for algorithmic optimisation in high dimensions. Applic
 ations will be shown in the context of Data Assimilation\, in a problem wh
 ere the objective is to target the posterior distribution of the initial c
 ondition of the Navier-Stokes equation given a Gaussian and noisy observat
 ions at different instances and locations of the spatial field. The dimens
 ion of the signal is in theory infinite-dimensional - in practice 64x64 or
  more depending on the resolution  thus posing great challenges for the de
 velopment and efficiency of SMC methods.\n
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR