Skip to Main content Skip to Navigation
Journal articles

Online but Accurate Inference for Latent Variable Models with Local Gibbs Sampling

Christophe Dupuy 1, 2 Francis Bach 3, 2 
2 SIERRA - Statistical Machine Learning and Parsimony
DI-ENS - Département d'informatique - ENS Paris, CNRS - Centre National de la Recherche Scientifique, Inria de Paris
Abstract : We study parameter inference in large-scale latent variable models. We first propose an unified treatment of online inference for latent variable models from a non-canonical exponential family, and draw explicit links between several previously proposed frequentist or Bayesian methods. We then propose a novel inference method for the frequentist estimation of parameters, that adapts MCMC methods to online inference of latent variable models with the proper use of local Gibbs sampling. Then, for latent Dirich-let allocation,we provide an extensive set of experiments and comparisons with existing work, where our new approach outperforms all previously proposed methods. In particular, using Gibbs sampling for latent variable inference is superior to variational inference in terms of test log-likelihoods. Moreover, Bayesian inference through variational methods perform poorly, sometimes leading to worse fits with latent variables of higher dimensionality.
Complete list of metadata
Contributor : Christophe Dupuy Connect in order to contact the contributor
Submitted on : Tuesday, January 30, 2018 - 1:20:54 AM
Last modification on : Wednesday, June 8, 2022 - 12:50:05 PM
Long-term archiving on: : Friday, May 25, 2018 - 8:58:19 AM


Files produced by the author(s)


  • HAL Id : hal-01284900, version 4
  • ARXIV : 1603.02644



Christophe Dupuy, Francis Bach. Online but Accurate Inference for Latent Variable Models with Local Gibbs Sampling. Journal of Machine Learning Research, Microtome Publishing, 2017, 18, pp.1-45. ⟨hal-01284900v4⟩



Record views


Files downloads