BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Speeding-up Pseudo-marginal MCMC using a surrogate model - Thiery\
 , AH (National University of Singapore)
DTSTART:20140422T104000Z
DTEND:20140422T111500Z
UID:TALK52084@talks.cam.ac.uk
CONTACT:Mustapha Amrani
DESCRIPTION:The pseudo-marginal MCMC algorithm is a powerful tool for expl
 oring the posterior distribution when only unbiased stochastic estimates o
 f the target density are available. In many situations\, although there is
  no closed-form expression for this density\, computationally cheap determ
 inistic estimates are also available\; one can then use a delayed-acceptan
 ce strategy for exploiting these cheap approximations.\n\nAs powerful as t
 hey are\, the use of such algorithms are difficult in practice: it involve
 s tuning the MCMC proposals and choosing the computational budget that one
  is willing to invest in the creation of the unbiased estimates while taki
 ng into account the quality of the cheap deterministic approximations. In 
 this talk we discuss how high-dimensional asymptotic results can help in t
 he tuning of these delayed-acceptance pseudo-marginal MCMC algorithms.\n\n
 This is joint work with Chris Sherlock and Andrew Golightly.\n
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR