BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Adaptive algorithms for Stratified Sampling Monte Carlo - Alexandr
 a Carpentier (Statistical Laboratory)
DTSTART:20130213T140000Z
DTEND:20130213T153000Z
UID:TALK41571@talks.cam.ac.uk
CONTACT:Robin Evans
DESCRIPTION:We consider the problem of estimating the integral of a functi
 on f over a domain. Although no analytic expression for f is available\, i
 t is possible to obtain n samples from f\, chosen anywhere in the domain. 
 A popular method for computing the integral of the function is to stratify
  the space in strata and sample points in the strata.  \n\nWe propose an a
 lgorithm for returning a stratified estimate of the integral. We prove tha
 t this algorithm adapts online the number of samples in each stratum to th
 e amount of variation of the function in the stratum. In particular\, this
  enables to allocate more samples where the function varies more\, and be 
 almost as efficient as an "oracle" strategy that has access to the variati
 ons of the functions in each stratum. More precisions on this aspect is\ni
 n paper (Carpentier and Munos\, 2011). We also provide some results on (i)
  how to choose the number of strata in an efficient way\nand (ii) how to a
 dapt the strata themselves to the specific shape of the function. We expre
 ss those results with finite-time bounds on a proxy of the variance of the
  estimate (returned by the algorithms we present).\n
LOCATION:MR9\, CMS
END:VEVENT
END:VCALENDAR