BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Bayesian inference for Markov processes with application to bioche
 mical network dynamics - Darren Wilkinson\, University of Newcastle
DTSTART:20111014T150000Z
DTEND:20111014T160000Z
UID:TALK32897@talks.cam.ac.uk
CONTACT:Richard Samworth
DESCRIPTION:A number of interesting statistical applications require the\n
 estimation of parameters underlying a nonlinear multivariate\ncontinuous t
 ime Markov process model\, using partial and noisy discrete\ntime observat
 ions of the system state. Bayesian inference for this\nproblem is difficul
 t due to the fact that the discrete time transition\ndensity of the Markov
  process is typically intractable and\ncomputationally intensive to approx
 imate. It turns out to be possible\nto develop MCMC algorithms which are e
 xact\, provided that one can\nsimulate exact realisations of the process f
 orwards in time. Such\nalgorithms\, often termed "likelihood free" or "plu
 g-and-play" are very\nattractive\, as they allow separation of the problem
  of model\ndevelopment and simulation implementation from the development 
 of\ninferential algorithms. Such techniques break down in the case of\nper
 fect observation or high-dimensional data\, but more efficient\nalgorithms
  can be developed if one is prepared to deviate from the\nlikelihood free 
 paradigm\, at least in the case of diffusion processes.\nThe methods will 
 be illustrated using examples from population\ndynamics and stochastic bio
 chemical network dynamics.
LOCATION:MR12\, CMS\, Wilberforce Road\, Cambridge\, CB3 0WB
END:VEVENT
END:VCALENDAR