Bayesian Feature Weighting for Unsupervised Learning, with Application to Object Recognition

Paul Gustafson, Peter Carbonetto, Natalie Thompson, Nando de Freitas
Proceedings of the Ninth International Workshop on Artificial Intelligence and Statistics, PMLR R4:124-131, 2003.

Abstract

We present a method for variable selection/weighting in an unsupervised learning context using Bayesian shrinkage. The basis for the model is a finite mixture of multivariate Gaussian distributions. We demonstrate how the model parameters and cluster assignments can be computed simultaneously using an efficient EM algorithm. Applying our Bayesian shrinkage model to a complex problem in object recognition (Duygulu, Barnard, de Freitas and Forsyth 2002), our experiments yield good results.

Cite this Paper

BibTeX
@InProceedings{pmlr-vR4-gustafson03a, title = {Bayesian Feature Weighting for Unsupervised Learning, with Application to Object Recognition}, author = {Gustafson, Paul and Carbonetto, Peter and Thompson, Natalie and de Freitas, Nando}, booktitle = {Proceedings of the Ninth International Workshop on Artificial Intelligence and Statistics}, pages = {124--131}, year = {2003}, editor = {Bishop, Christopher M. and Frey, Brendan J.}, volume = {R4}, series = {Proceedings of Machine Learning Research}, month = {03--06 Jan}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/r4/gustafson03a/gustafson03a.pdf}, url = {https://proceedings.mlr.press/r4/gustafson03a.html}, abstract = {We present a method for variable selection/weighting in an unsupervised learning context using Bayesian shrinkage. The basis for the model is a finite mixture of multivariate Gaussian distributions. We demonstrate how the model parameters and cluster assignments can be computed simultaneously using an efficient EM algorithm. Applying our Bayesian shrinkage model to a complex problem in object recognition (Duygulu, Barnard, de Freitas and Forsyth 2002), our experiments yield good results.}, note = {Reissued by PMLR on 01 April 2021.} }
Endnote
%0 Conference Paper %T Bayesian Feature Weighting for Unsupervised Learning, with Application to Object Recognition %A Paul Gustafson %A Peter Carbonetto %A Natalie Thompson %A Nando de Freitas %B Proceedings of the Ninth International Workshop on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2003 %E Christopher M. Bishop %E Brendan J. Frey %F pmlr-vR4-gustafson03a %I PMLR %P 124--131 %U https://proceedings.mlr.press/r4/gustafson03a.html %V R4 %X We present a method for variable selection/weighting in an unsupervised learning context using Bayesian shrinkage. The basis for the model is a finite mixture of multivariate Gaussian distributions. We demonstrate how the model parameters and cluster assignments can be computed simultaneously using an efficient EM algorithm. Applying our Bayesian shrinkage model to a complex problem in object recognition (Duygulu, Barnard, de Freitas and Forsyth 2002), our experiments yield good results. %Z Reissued by PMLR on 01 April 2021.
APA
Gustafson, P., Carbonetto, P., Thompson, N. & de Freitas, N.. (2003). Bayesian Feature Weighting for Unsupervised Learning, with Application to Object Recognition. Proceedings of the Ninth International Workshop on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research R4:124-131 Available from https://proceedings.mlr.press/r4/gustafson03a.html. Reissued by PMLR on 01 April 2021.

Related Material