BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Scalable algorithms for Markov process parameter inference - Darre
 n Wilkinson (Newcastle University)
DTSTART:20160408T084500Z
DTEND:20160408T093000Z
UID:TALK65377@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:Inferring the parameters of continuous-time Markov process mod
 els using  partial discrete-time observations is an important practical pr
 oblem in many  fields of scientific research. Such models are very often "
 intractable"\, in the  sense that the transition kernel of the process can
 not be described in closed  form\, and is difficult to approximate well. N
 evertheless\, it is often possible  to forward simulate realisations of tr
 ajectories of the process using stochastic  simulation. There have been a 
 number of recent developments in the literature  relevant to the parameter
  estimation problem\, involving a mixture of  approximate\, sequential and
  Markov chain Monte Carlo methods. This talk will  compare some of the dif
 ferent "likelihood free" algorithms that have been  proposed\, including s
 equential ABC and particle marginal Metropolis Hastings\,  paying particul
 ar attention to how well they scale with model complexity.  Emphasis will 
 be placed on the problem of Bayesian pa rameter inference for the  rate co
 nstants of stochastic biochemical network models\, using noisy\, partial  
 high-resolution time course data.
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR