Incremental ELMVIS for Unsupervised Learning

Anton Akusok; Emil Eirola; Yoan Miche; Ian Oliver; Kaj-Mikael Björk; Andrey Gritsenko; Stephen Baek; Amaury Lendasse

doi:10.1007/978-3-319-57421-9_15

Incremental ELMVIS for Unsupervised Learning

Anton Akusok, Emil Eirola, Yoan Miche, Ian Oliver, Kaj-Mikael Björk, Andrey Gritsenko, Stephen Baek, Amaury Lendasse

Department of Computer Science

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceedings › Scientific › peer-review

Abstract

An incremental version of the ELMVIS+ method is proposed in this paper. It iteratively selects a few best fitting data samples from a large pool, and adds them to the model. The method keeps high speed of ELMVIS+ while allowing for much larger possible sample pools due to lower memory requirements. The extension is useful for reaching a better local optimum with greedy optimization of ELMVIS, and the data structure can be specified in semi-supervised optimization. The major new application of incremental ELMVIS is not to visualization, but to a general dataset processing. The method is capable of learning dependencies from non-organized unsupervised data—either reconstructing a shuffled dataset, or learning dependencies in complex high-dimensional space. The results are interesting and promising, although there is space for improvements.

Original language	English
Title of host publication	Proceedings of ELM-2016
Editors	Jiuwen Cao, Erik Cambria, Amaury Lendasse, Yoan Miche, Chi Man Vong
Place of Publication	Cham
Publisher	Springer
Pages	183-193
Number of pages	11
ISBN (Print)	978-3-319-57421-9
DOIs	https://doi.org/10.1007/978-3-319-57421-9_15
Publication status	Published - 2018
MoE publication type	A4 Conference publication
Event	International Conference on Extreme Learning Machines - Singapore, Singapore Duration: 13 Dec 2016 → 15 Dec 2016

Publication series

Name	Proceedings in Adaptation, Learning and Optimization
Publisher	Springer
Volume	9
ISSN (Print)	2363-6084

Conference

Conference	International Conference on Extreme Learning Machines
Abbreviated title	ELM
Country/Territory	Singapore
City	Singapore
Period	13/12/2016 → 15/12/2016

Keywords

engineering
artificial intelligence
computational intelligence

Access to Document

10.1007/978-3-319-57421-9_15

Cite this

Akusok, A., Eirola, E., Miche, Y., Oliver, I., Björk, K.-M., Gritsenko, A., Baek, S., & Lendasse, A. (2018). Incremental ELMVIS for Unsupervised Learning. In J. Cao, E. Cambria, A. Lendasse, Y. Miche, & C. M. Vong (Eds.), Proceedings of ELM-2016 (pp. 183-193). (Proceedings in Adaptation, Learning and Optimization; Vol. 9). Springer. https://doi.org/10.1007/978-3-319-57421-9_15

@inproceedings{f6a9fd3ea9f748adb447d7415bfdfff9,

title = "Incremental ELMVIS for Unsupervised Learning",

abstract = "An incremental version of the ELMVIS+ method is proposed in this paper. It iteratively selects a few best fitting data samples from a large pool, and adds them to the model. The method keeps high speed of ELMVIS+ while allowing for much larger possible sample pools due to lower memory requirements. The extension is useful for reaching a better local optimum with greedy optimization of ELMVIS, and the data structure can be specified in semi-supervised optimization. The major new application of incremental ELMVIS is not to visualization, but to a general dataset processing. The method is capable of learning dependencies from non-organized unsupervised data—either reconstructing a shuffled dataset, or learning dependencies in complex high-dimensional space. The results are interesting and promising, although there is space for improvements.",

keywords = "engineering, artificial intelligence, computational intelligence",

author = "Anton Akusok and Emil Eirola and Yoan Miche and Ian Oliver and Kaj-Mikael Bj{\"o}rk and Andrey Gritsenko and Stephen Baek and Amaury Lendasse",

year = "2018",

doi = "10.1007/978-3-319-57421-9_15",

language = "English",

isbn = "978-3-319-57421-9",

series = "Proceedings in Adaptation, Learning and Optimization",

publisher = "Springer",

pages = "183--193",

editor = "Jiuwen Cao and Erik Cambria and Amaury Lendasse and Yoan Miche and Vong, {Chi Man}",

booktitle = "Proceedings of ELM-2016",

address = "Germany",

note = "International Conference on Extreme Learning Machines, ELM ; Conference date: 13-12-2016 Through 15-12-2016",

}

Akusok, A, Eirola, E, Miche, Y, Oliver, I, Björk, K-M, Gritsenko, A, Baek, S & Lendasse, A 2018, Incremental ELMVIS for Unsupervised Learning. in J Cao, E Cambria, A Lendasse, Y Miche & CM Vong (eds), Proceedings of ELM-2016. Proceedings in Adaptation, Learning and Optimization, vol. 9, Springer, Cham, pp. 183-193, International Conference on Extreme Learning Machines, Singapore, Singapore, 13/12/2016. https://doi.org/10.1007/978-3-319-57421-9_15

Incremental ELMVIS for Unsupervised Learning. / Akusok, Anton; Eirola, Emil; Miche, Yoan et al.
Proceedings of ELM-2016. ed. / Jiuwen Cao; Erik Cambria; Amaury Lendasse; Yoan Miche; Chi Man Vong. Cham: Springer, 2018. p. 183-193 (Proceedings in Adaptation, Learning and Optimization; Vol. 9).

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceedings › Scientific › peer-review

TY - GEN

T1 - Incremental ELMVIS for Unsupervised Learning

AU - Akusok, Anton

AU - Eirola, Emil

AU - Miche, Yoan

AU - Oliver, Ian

AU - Björk, Kaj-Mikael

AU - Gritsenko, Andrey

AU - Baek, Stephen

AU - Lendasse, Amaury

PY - 2018

Y1 - 2018

N2 - An incremental version of the ELMVIS+ method is proposed in this paper. It iteratively selects a few best fitting data samples from a large pool, and adds them to the model. The method keeps high speed of ELMVIS+ while allowing for much larger possible sample pools due to lower memory requirements. The extension is useful for reaching a better local optimum with greedy optimization of ELMVIS, and the data structure can be specified in semi-supervised optimization. The major new application of incremental ELMVIS is not to visualization, but to a general dataset processing. The method is capable of learning dependencies from non-organized unsupervised data—either reconstructing a shuffled dataset, or learning dependencies in complex high-dimensional space. The results are interesting and promising, although there is space for improvements.

AB - An incremental version of the ELMVIS+ method is proposed in this paper. It iteratively selects a few best fitting data samples from a large pool, and adds them to the model. The method keeps high speed of ELMVIS+ while allowing for much larger possible sample pools due to lower memory requirements. The extension is useful for reaching a better local optimum with greedy optimization of ELMVIS, and the data structure can be specified in semi-supervised optimization. The major new application of incremental ELMVIS is not to visualization, but to a general dataset processing. The method is capable of learning dependencies from non-organized unsupervised data—either reconstructing a shuffled dataset, or learning dependencies in complex high-dimensional space. The results are interesting and promising, although there is space for improvements.

KW - engineering

KW - artificial intelligence

KW - computational intelligence

U2 - 10.1007/978-3-319-57421-9_15

DO - 10.1007/978-3-319-57421-9_15

M3 - Conference article in proceedings

SN - 978-3-319-57421-9

T3 - Proceedings in Adaptation, Learning and Optimization

SP - 183

EP - 193

BT - Proceedings of ELM-2016

A2 - Cao, Jiuwen

A2 - Cambria, Erik

A2 - Lendasse, Amaury

A2 - Miche, Yoan

A2 - Vong, Chi Man

PB - Springer

CY - Cham

T2 - International Conference on Extreme Learning Machines

Y2 - 13 December 2016 through 15 December 2016

ER -

Incremental ELMVIS for Unsupervised Learning

Abstract

Publication series

Conference

Keywords

Access to Document

Fingerprint

Cite this