BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Anglican\; Particle MCMC inference for Probabilistic Programs - Ja
 n-Willem van de Meent (Columbia University)
DTSTART:20140115T110000Z
DTEND:20140115T120000Z
UID:TALK49946@talks.cam.ac.uk
CONTACT:Zoubin Ghahramani
DESCRIPTION:Probabilistic programming languages hold the promise of dramat
 ically accelerating the development of both new statistical models and inf
 erence strategies. In probabilistic programs\, variables can take on rando
 m values at run time and inference is performed by calculating expectation
  values over all execution traces that are in agreement with a set of obse
 rved data. This allows statistical models to be represented in a concise a
 nd intuitive manner\, enabling more rapid iteration over model variants. I
 nference schemes\, when implemented as a backend to a programming framewor
 k\, can easily be tested on a large collection of models\, enabling a much
  more systematic comparison of the efficacy of inference strategies. \n\nW
 e introduce Anglican\, a probabilistic language that uses particle Markov 
 chain Monte Carlo to perform inference. Our approach is simple to implemen
 t\, easy to parallelize\, and supports accurate inference in models that m
 ake use of complex control flow\, including stochastic recursion. It also 
 includes primitives from Bayesian nonparametric statistics. Our experiment
 s show that this approach can be more efficient than previously introduced
  single-site Metropolis-Hastings methods.\n
LOCATION:Engineering Department\, CBL Room BE-438
END:VEVENT
END:VCALENDAR