Infinite variational autoencoder for semi-supervised learning

M. Ehsan Abbasnejad; Anthony Dick; Anton Van Den Hengel

doi:10.1109/CVPR.2017.90

Infinite variational autoencoder for semi-supervised learning

M. Ehsan Abbasnejad
, Anthony Dick
, Anton Van Den Hengel

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research › peer-review

77 Citations (Scopus)

Abstract

This paper presents an infinite variational autoencoder (VAE) whose capacity adapts to suit the input data. This is achieved using a mixture model where the mixing coefficients are modeled by a Dirichlet process, allowing us to integrate over the coefficients when performing inference. Critically, this then allows us to automatically vary the number of autoencoders in the mixture based on the data. Experiments show the flexibility of our method, particularly for semi-supervised learning, where only a small number of training samples are available.

Original language	English
Title of host publication	Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017
Publisher	IEEE, Institute of Electrical and Electronics Engineers
Pages	781-790
Number of pages	10
ISBN (Electronic)	9781538604571
DOIs	https://doi.org/10.1109/CVPR.2017.90
Publication status	Published - 2017
Externally published	Yes
Event	IEEE Conference on Computer Vision and Pattern Recognition 2017 - Honolulu, United States of America Duration: 21 Jul 2017 → 26 Jul 2017 http://cvpr2017.thecvf.com/ https://ieeexplore.ieee.org/xpl/conhome/8097368/proceeding (Proceedings)

Conference

Conference	IEEE Conference on Computer Vision and Pattern Recognition 2017
Abbreviated title	CVPR 2017
Country/Territory	United States of America
City	Honolulu
Period	21/07/17 → 26/07/17
Internet address	http://cvpr2017.thecvf.com/ https://ieeexplore.ieee.org/xpl/conhome/8097368/proceeding (Proceedings)

Access to Document

10.1109/CVPR.2017.90

Cite this

@inproceedings{e7c8bf4ac8834181bd7305d24264a1de,

title = "Infinite variational autoencoder for semi-supervised learning",

abstract = "This paper presents an infinite variational autoencoder (VAE) whose capacity adapts to suit the input data. This is achieved using a mixture model where the mixing coefficients are modeled by a Dirichlet process, allowing us to integrate over the coefficients when performing inference. Critically, this then allows us to automatically vary the number of autoencoders in the mixture based on the data. Experiments show the flexibility of our method, particularly for semi-supervised learning, where only a small number of training samples are available.",

author = "Abbasnejad, \{M. Ehsan\} and Anthony Dick and \{Van Den Hengel\}, Anton",

note = "Publisher Copyright: {\textcopyright} 2017 IEEE.; IEEE Conference on Computer Vision and Pattern Recognition 2017, CVPR 2017 ; Conference date: 21-07-2017 Through 26-07-2017",

year = "2017",

doi = "10.1109/CVPR.2017.90",

language = "English",

pages = "781--790",

booktitle = "Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017",

publisher = "IEEE, Institute of Electrical and Electronics Engineers",

address = "United States of America",

url = "http://cvpr2017.thecvf.com/, https://ieeexplore.ieee.org/xpl/conhome/8097368/proceeding",

}

Abbasnejad, ME, Dick, A & Van Den Hengel, A 2017, Infinite variational autoencoder for semi-supervised learning. in Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017. IEEE, Institute of Electrical and Electronics Engineers, pp. 781-790, IEEE Conference on Computer Vision and Pattern Recognition 2017, Honolulu, Hawaii, United States of America, 21/07/17. https://doi.org/10.1109/CVPR.2017.90

Infinite variational autoencoder for semi-supervised learning. / Abbasnejad, M. Ehsan; Dick, Anthony; Van Den Hengel, Anton.
Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017. IEEE, Institute of Electrical and Electronics Engineers, 2017. p. 781-790.

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research › peer-review

TY - GEN

T1 - Infinite variational autoencoder for semi-supervised learning

AU - Abbasnejad, M. Ehsan

AU - Dick, Anthony

AU - Van Den Hengel, Anton

PY - 2017

Y1 - 2017

N2 - This paper presents an infinite variational autoencoder (VAE) whose capacity adapts to suit the input data. This is achieved using a mixture model where the mixing coefficients are modeled by a Dirichlet process, allowing us to integrate over the coefficients when performing inference. Critically, this then allows us to automatically vary the number of autoencoders in the mixture based on the data. Experiments show the flexibility of our method, particularly for semi-supervised learning, where only a small number of training samples are available.

AB - This paper presents an infinite variational autoencoder (VAE) whose capacity adapts to suit the input data. This is achieved using a mixture model where the mixing coefficients are modeled by a Dirichlet process, allowing us to integrate over the coefficients when performing inference. Critically, this then allows us to automatically vary the number of autoencoders in the mixture based on the data. Experiments show the flexibility of our method, particularly for semi-supervised learning, where only a small number of training samples are available.

UR - https://www.scopus.com/pages/publications/85044444902

U2 - 10.1109/CVPR.2017.90

DO - 10.1109/CVPR.2017.90

M3 - Conference Paper

AN - SCOPUS:85044444902

SP - 781

EP - 790

BT - Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017

PB - IEEE, Institute of Electrical and Electronics Engineers

T2 - IEEE Conference on Computer Vision and Pattern Recognition 2017

Y2 - 21 July 2017 through 26 July 2017

ER -

Infinite variational autoencoder for semi-supervised learning

Abstract

Conference

Access to Document

Other files and links

Cite this