Non-Gaussian discriminative factor models via the max-margin rank-likelihood

dc.contributor.author

Yuan, X

dc.contributor.author

Henao, R

dc.contributor.author

Tsalik, EL

dc.contributor.author

Langley, RJ

dc.contributor.author

Carin, L

dc.contributor.editor

Bach, Francis R

dc.contributor.editor

Blei, David M

dc.date.accessioned

2016-12-02T16:02:53Z

dc.date.available

2016-12-02T16:02:53Z

dc.date.issued

2015-01-01

dc.description.abstract

Copyright © 2015 by the author(s).We consider the problem of discriminative factor analysis for data that are in general non-Gaussian. A Bayesian model based on the ranks of the data is proposed. We first introduce a new max-margin version of the rank-likelihood. A discriminative factor model is then developed, integrating the max-margin rank-likelihood and (linear) Bayesian support vector machines, which are also built on the max-margin principle. The discriminative factor model is further extended to the nonlinear case through mixtures of local linear classifiers, via Dirichlet processes. Fully local conjugacy of the model yields efficient inference with both Markov Chain Monte Carlo and variational Bayes approaches. Extensive experiments on benchmark and real data demonstrate superior performance of the proposed model and its potential for applications in computational biology.

dc.identifier.uri

https://hdl.handle.net/10161/13121

dc.publisher

JMLR.org

dc.relation.ispartof

32nd International Conference on Machine Learning, ICML 2015

dc.title

Non-Gaussian discriminative factor models via the max-margin rank-likelihood

dc.type

Journal article

duke.contributor.orcid

Henao, R|0000-0003-4980-845X

duke.contributor.orcid

Tsalik, EL|0000-0002-6417-2042

pubs.begin-page

1254

pubs.end-page

1263

pubs.organisational-group

Clinical Science Departments

pubs.organisational-group

Duke

pubs.organisational-group

Electrical and Computer Engineering

pubs.organisational-group

Medicine

pubs.organisational-group

Medicine, Infectious Diseases

pubs.organisational-group

Pratt School of Engineering

pubs.organisational-group

School of Medicine

pubs.publication-status

Published

pubs.volume

2

Files

Original bundle

Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
1504.07468v3.pdf
Size:
774.07 KB
Format:
Adobe Portable Document Format