Online Program Home
My Program

Abstract Details

Activity Number: 215 - Non- and Semiparametric Methods to Accommodate Dependency and Heterogeneity in Complex Data
Type: Invited
Date/Time: Monday, July 30, 2018 : 2:00 PM to 3:50 PM
Sponsor: Section on Nonparametric Statistics
Abstract #326473
Title: Nonparametric Bayesian Priors for Hidden Markov Random Fields
Author(s): Florence Forbes* and Hongliang Lu and Julyan Arbel
Companies: INRIA and INRIA and INRIA
Keywords: Bayesian non parametrics; Markov random fields; image segmentation

Hidden Markov random field models are widely used for clustering data under spatial constraints. They can be seen as spatial extensions of independent mixture models. As for standard mixtures, one concern is the automatic selection of the proper number of components in the mixture. A number of criteria exist to select this number automatically based on penalized likelihood (eg. AIC, BIC, ICL etc.) but they usually require to run several models for different number of classes to choose the best one. Other techniques (eg. reversible jump) use a fully Bayesian setting including a prior on the class number but at the cost of prohibitive computational times. In this work, we propose to investigate alternatives based on the more recent field of Bayesian nonparametrics. In particular, Dirichlet process mixture models  (DPMM) have emerged as promising candidates for clustering applications where the number of clusters is unknown.  Most applications of DPMM involve observations which are supposed to be independent. For more complex tasks such as unsupervised image segmentation with spatial relationships or dependencies between the observations, DPMM are not satisfying.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2018 program