BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Particle filters and curse of dimensionality - Rebeschini\, P (Pri
 nceton University)
DTSTART:20140424T104000Z
DTEND:20140424T111500Z
UID:TALK52163@talks.cam.ac.uk
CONTACT:Mustapha Amrani
DESCRIPTION:Co-author: Ramon van Handel (Princeton University) \n\nA probl
 em that arises in many applications is to compute the conditional distribu
 tions of stochastic models given observed data. While exact computations a
 re rarely possible\, particle filtering algorithms have proved to be very 
 useful for approximating such conditional distributions. Unfortunately\, t
 he approximation error of particle filters grows exponentially with dimens
 ion\, a phenomenon known as curse of dimensionality. This fact has rendere
 d particle filters of limited use in complex data assimilation problems th
 at arise\, for example\, in weather forecasting. In this talk I will argue
  that it is possible to develop local particle filtering algorithms whose 
 approximation error is dimension-free. By exploiting conditional decay of 
 correlations properties of high-dimensional models\, we prove for the simp
 lest possible algorithm of this type an error bound that is uniform both i
 n time and in the model dimension. (Joint work with R. van Handel)\n\nRela
 ted Links: http://arxiv.org/abs/1301.6585 - Preprint \n\n
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR