BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Turing: Rejuvenating Probabilistic Programming in Julia - Hong Ge 
 (University of Cambridge)
DTSTART:20160602T083000Z
DTEND:20160602T093000Z
UID:TALK66453@talks.cam.ac.uk
CONTACT:Louise Segar
DESCRIPTION:In this talk\, I will present a probabilistic programming syst
 em (PPL) called Turing. The implementation of Turing is based on Julia\, a
  fast and modern programming language for technical computing. Novel aspec
 ts of this system include 1) the use of coroutines for scalable implementa
 tion of existing general-purpose inference algorithms\, and 2) new syntax 
 features that improve statistical efficiency of general-purpose inference 
 engines. I will discuss some lessons we have learnt so far\, on both PPL d
 esign and novel Monte Carlo methods.\n\nMore concretely\, I will first rev
 iew a (classical) result that some widely used importance sampling methods
  (e.g. IS\, PMC\, SMC) are still valid when exact importance weights are r
 eplaced with non-negative unbiased Monte Carlo estimates or pseudo margina
 ls. This result allow us to export the celebrated pseudo-marginal method f
 rom the MCMC framework to importance sampling methods. Then\, following th
 is generic result\, we develop a pseudo-marginal particle filtering (PM-PF
 ) method and apply it to PPL inference. Some experiments show the PM-PF me
 thod is consistently more accurate than similar algorithms: particle filte
 ring\, particle Gibbs.
LOCATION:Engineering Department\, CBL Room BE-438
END:VEVENT
END:VCALENDAR