BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Rejection Sampling Variational Inference - Francisco J. R. Ruiz (C
 olumbia University &amp\; University of Cambridge)
DTSTART:20161122T113000Z
DTEND:20161122T123000Z
UID:TALK69280@talks.cam.ac.uk
CONTACT:39846
DESCRIPTION:Talk based on https://arxiv.org/abs/1610.05683\, for which the
  abstract is:\n\n"Variational inference using the reparameterization trick
  has enabled large-scale approximate Bayesian inference in complex probabi
 listic models\, leveraging stochastic optimization to sidestep intractable
  expectations. The reparameterization trick is applicable when we can simu
 late a random variable by applying a (differentiable) deterministic functi
 on on an auxiliary random variable whose distribution is fixed. For many d
 istributions of interest (such as the gamma or Dirichlet)\, simulation of 
 random variables relies on rejection sampling. The discontinuity introduce
 d by the accept--reject step means that standard reparameterization tricks
  are not applicable. We propose a new method that lets us leverage reparam
 eterization gradients even when variables are outputs of a rejection sampl
 ing algorithm. Our approach enables reparameterization on a larger class o
 f variational distributions. In several studies of real and synthetic data
 \, we show that the variance of the estimator of the gradient is significa
 ntly lower than other state-of-the-art methods. This leads to faster conve
 rgence of stochastic optimization variational inference."
LOCATION:CBL Room BE-438
END:VEVENT
END:VCALENDAR