Normalizing flows transform a latent distribution through an invertible
neural network for a flexible and pleasingly simple approach to generative
modelling, while preserving an exact likelihood. We propose FlowGMM, an
end-to-end approach to generative semi supervised learning with normalizing
flows, using a latent Gaussian mixture model. FlowGMM is distinct in its
simplicity, unified treatment of labelled and unlabelled data with an exact
likelihood, interpretability, and broad applicability beyond image data. We
show promising results on a wide range of applications, including AG-News and
Yahoo Answers text data, tabular data, and semi-supervised image
classification. We also show that FlowGMM can discover interpretable structure,
provide real-time optimization-free feature visualizations, and specify well
calibrated predictive distributions.
Описание
[1912.13025] Semi-Supervised Learning with Normalizing Flows
%0 Journal Article
%1 izmailov2019semisupervised
%A Izmailov, Pavel
%A Kirichenko, Polina
%A Finzi, Marc
%A Wilson, Andrew Gordon
%D 2019
%K flows generative-models semi-supervised
%T Semi-Supervised Learning with Normalizing Flows
%U http://arxiv.org/abs/1912.13025
%X Normalizing flows transform a latent distribution through an invertible
neural network for a flexible and pleasingly simple approach to generative
modelling, while preserving an exact likelihood. We propose FlowGMM, an
end-to-end approach to generative semi supervised learning with normalizing
flows, using a latent Gaussian mixture model. FlowGMM is distinct in its
simplicity, unified treatment of labelled and unlabelled data with an exact
likelihood, interpretability, and broad applicability beyond image data. We
show promising results on a wide range of applications, including AG-News and
Yahoo Answers text data, tabular data, and semi-supervised image
classification. We also show that FlowGMM can discover interpretable structure,
provide real-time optimization-free feature visualizations, and specify well
calibrated predictive distributions.
@article{izmailov2019semisupervised,
abstract = {Normalizing flows transform a latent distribution through an invertible
neural network for a flexible and pleasingly simple approach to generative
modelling, while preserving an exact likelihood. We propose FlowGMM, an
end-to-end approach to generative semi supervised learning with normalizing
flows, using a latent Gaussian mixture model. FlowGMM is distinct in its
simplicity, unified treatment of labelled and unlabelled data with an exact
likelihood, interpretability, and broad applicability beyond image data. We
show promising results on a wide range of applications, including AG-News and
Yahoo Answers text data, tabular data, and semi-supervised image
classification. We also show that FlowGMM can discover interpretable structure,
provide real-time optimization-free feature visualizations, and specify well
calibrated predictive distributions.},
added-at = {2020-01-05T22:02:27.000+0100},
author = {Izmailov, Pavel and Kirichenko, Polina and Finzi, Marc and Wilson, Andrew Gordon},
biburl = {https://www.bibsonomy.org/bibtex/25cdd4e74642fb0696d827fa590614720/kirk86},
description = {[1912.13025] Semi-Supervised Learning with Normalizing Flows},
interhash = {9b5fc11b89d12c8bbfa02b17176636b4},
intrahash = {5cdd4e74642fb0696d827fa590614720},
keywords = {flows generative-models semi-supervised},
note = {cite arxiv:1912.13025},
timestamp = {2020-01-05T22:05:18.000+0100},
title = {Semi-Supervised Learning with Normalizing Flows},
url = {http://arxiv.org/abs/1912.13025},
year = 2019
}