BEGIN:VCALENDAR
PRODID:-//Columba Systems Ltd//NONSGML CPNG/SpringViewer/ICal Output/3.3-
 M3//EN
VERSION:2.0
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VEVENT
DTSTAMP:20260528T092840Z
DTSTART:20260602T110000Z
DTEND:20260602T120000Z
SUMMARY:SQUIDS-NASC Joint Seminar: Metropolis-Hastings Acceptance Behavio
 r for Bayesian Inversion with Random Forward Solvers
UID:{http://www.columbasystems.com/customers/uom/gpp/eventid/}z28q-mppajb
 ib-8l2vkc
DESCRIPTION:Speaker: Emil Løvbak  (Karlsruhe Institute of Technology\, Ge
 rmany)\n\nAbstract: In various application domains\, one wishes to deter
 mine which parameter values should be used for a model to match its simu
 lation output with measurement data. In practice however\, measurement e
 rror on the data means that\, at best\, one can produce a so-called post
 erior probability distribution of these parameter values\, given an assu
 med noise model. The Metropolis-Hastings is a straightforward approach t
 hat constructs a Markov chain with this posterior distribution as its in
 variant distribution. The parameter samples in the chain are selected th
 rough an accept-reject strategy\, that accepts proposal samples\, based 
 on their likelihood\, relative to that of the previous accepted sample.\
 n\nEvaluating this likelihood requires the solution of the given model. 
 Therefore\, any errors in the discrete solver will result in errors in t
 he likelihood evaluation. In this presentation\, we discuss the case whe
 re Metropolis-Hastings is run on top of a stochastic solver\, such as a 
 Monte Carlo particle solver. In this case\, the likelihood --- and thus 
 the acceptance probability --- becomes a random variable who’s varianc
 e scales with the number of random trajectories simulated by the solver.
  We discuss the mismatch between theory and practice in this setting. To
  this end\, we combine classical error analysis and simulation results t
 o understand the behavior of the pseudomarginal Markov chains in this se
 tting. We then present practical approaches for efficient estimation in 
 such settings.
STATUS:TENTATIVE
TRANSP:TRANSPARENT
CLASS:PUBLIC
LOCATION:TBA\, Alan Turing Building\, Manchester
END:VEVENT
END:VCALENDAR