hal.structure.identifier dc.contributor.author Chopin, Nicolas * hal.structure.identifier dc.contributor.author Jacob, Pierre E. HAL ID: 16651 ORCID: 0000-0002-4662-1051 * dc.date.accessioned 2018-01-12T15:46:00Z dc.date.available 2018-01-12T15:46:00Z dc.date.issued 2011 dc.identifier.uri https://basepub.dauphine.fr/handle/123456789/17307 dc.language.iso en en dc.subject Free energy biasing en dc.subject Label switching en dc.subject Mixture en dc.subject Sequential Monte Carlo en dc.subject particle filter en dc.subject.ddc 519 en dc.title Free energy Sequential Monte Carlo, application to mixture modelling en dc.type Communication / Conférence dc.description.abstracten We introduce a new class of Sequential Monte Carlo (SMC) methods, which we call free energy SMC. This class is inspired by free energy methods, which originate from Physics, and where one samples from a biased distribution such that a given function $\xi(\theta)$ of the state $\theta$ is forced to be uniformly distributed over a given interval. From an initial sequence of distributions $(\pi_t)$ of interest, and a particular choice of $\xi(\theta)$, a free energy SMC sampler computes sequentially a sequence of biased distributions $(\tilde{\pi}_{t})$ with the following properties: (a) the marginal distribution of $\xi(\theta)$ with respect to $\tilde{\pi}_{t}$ is approximatively uniform over a specified interval, and (b) $\tilde{\pi}_{t}$ and $\pi_{t}$ have the same conditional distribution with respect to $\xi$. We apply our methodology to mixture posterior distributions, which are highly multimodal. In the mixture context, forcing certain hyper-parameters to higher values greatly faciliates mode swapping, and makes it possible to recover a symetric output. We illustrate our approach with univariate and bivariate Gaussian mixtures and two real-world datasets. en dc.relation.ispartoftitle Bayesian Statistics 9 en dc.relation.ispartofeditor Bernardo, José M. dc.relation.ispartofeditor Bayarri, M.J. dc.relation.ispartofeditor Berger, James O. dc.relation.ispartofeditor Dawid, A.P. dc.relation.ispartofeditor Heckermann, David dc.relation.ispartofeditor Smith, Adrian F. M. dc.relation.ispartofeditor West, Mike dc.relation.ispartofpublname Oxford University Press en dc.relation.ispartofpublcity Oxford en dc.relation.ispartofdate 2011 dc.subject.ddclabel Probabilités et mathématiques appliquées en dc.relation.ispartofisbn 9780199694587 en dc.relation.conftitle Bayesian Statistics 9 en dc.relation.confdate 2011-06 dc.relation.confcity Benidorm en dc.relation.confcountry Spain en dc.relation.forthcoming non en dc.identifier.doi 10.1093/acprof:oso/9780199694587.003.0003 en dc.description.ssrncandidate non en dc.description.halcandidate non en dc.description.readership recherche en dc.description.audience International en dc.relation.Isversionofjnlpeerreviewed non en dc.relation.Isversionofjnlpeerreviewed non en dc.date.updated 2017-12-22T17:16:34Z hal.author.function aut hal.author.function aut
﻿

## Files in this item

FilesSizeFormatView

There are no files associated with this item.