BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Particle islands and archipelagos: some large sample theory - Olss
 on\, JR (KTH - Royal Institute of Technology)
DTSTART:20140424T093000Z
DTEND:20140424T100500Z
UID:TALK52164@talks.cam.ac.uk
CONTACT:Mustapha Amrani
DESCRIPTION:Co-authors: Christelle Vergé\, Pierre Del Moral\, and Eric Mo
 ulines \n\nThis talk discusses parallelisation of sequential Monte Carlo m
 ethods via the particle island framework (Vergé et al.\, 2013) and presen
 ts some novel convergence results for methods of this sort. More specifica
 lly\, we introduce the concept of weighted archipelagos (i.e. sets of weig
 hted particle islands\, where each island is itself a weighted sample of p
 articles) and define three different operations on such archipelagos\, nam
 ely: selection on the island level\, selection on the particle level\, and
  mutation. We then establish that these operations preserve a set of conve
 rgence properties\, including asymptotic normality\, of the archipelago as
  the number of islands as well as the number of particles of each island t
 end jointly to infinity. Moreover\, we provide recursive formulas for the 
 asymptotic variance associated with each operation. As our results allow a
 rbitrary compositions of the mentioned operations to be analysed\, we may 
 use the same for establishing the convergence properties of not only the d
 ouble bootstrap algorithm but also generalisations of this algorithm. 
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR