Communication-efficient distributed online learning with kernels

Michael Kamp; Sebastian Bothe; Mario Boley; Michael Mock

doi:10.1007/978-3-319-46227-1_50

Communication-efficient distributed online learning with kernels

Michael Kamp, Sebastian Bothe, Mario Boley, Michael Mock

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

Abstract

We propose an efficient distributed online learning protocol for low-latency real-time services. It extends a previously presented protocol to kernelized online learners that represent their models by a support vector expansion. While such learners often achieve higher predictive performance than their linear counterparts, communicating the support vector expansions becomes inefficient for large numbers of support vectors. The proposed extension allows for a larger class of online learning algorithms—including those alleviating the problem above through model compression. In addition, we characterize the quality of the proposed protocol by introducing a novel criterion that requires the communication to be bounded by the loss suffered.

Original language	English
Title of host publication	Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2016, Proceedings
Editors	Niels Landwehr, Jilles Giuseppe, Giuseppe Manco, Paolo Frasconi
Publisher	Springer Verlag
Pages	805-819
Number of pages	15
ISBN (Print)	9783319462264
DOIs	https://doi.org/10.1007/978-3-319-46227-1_50
State	Published - 2016
Externally published	Yes
Event	15th European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2016 - Riva del Garda, Italy Duration: 19 Sep 2016 → 23 Sep 2016

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	9852 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	15th European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2016
Country/Territory	Italy
City	Riva del Garda
Period	19/09/16 → 23/09/16

Bibliographical note

Publisher Copyright:
© Springer International Publishing AG 2016.

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-319-46227-1_50

Cite this

Kamp, M., Bothe, S., Boley, M., & Mock, M. (2016). Communication-efficient distributed online learning with kernels. In N. Landwehr, J. Giuseppe, G. Manco, & P. Frasconi (Eds.), Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2016, Proceedings (pp. 805-819). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9852 LNAI). Springer Verlag. https://doi.org/10.1007/978-3-319-46227-1_50

Kamp, Michael ; Bothe, Sebastian ; Boley, Mario et al. / Communication-efficient distributed online learning with kernels. Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2016, Proceedings. editor / Niels Landwehr ; Jilles Giuseppe ; Giuseppe Manco ; Paolo Frasconi. Springer Verlag, 2016. pp. 805-819 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{f38befc46e0947ae9bb06ac261106dce,

title = "Communication-efficient distributed online learning with kernels",

abstract = "We propose an efficient distributed online learning protocol for low-latency real-time services. It extends a previously presented protocol to kernelized online learners that represent their models by a support vector expansion. While such learners often achieve higher predictive performance than their linear counterparts, communicating the support vector expansions becomes inefficient for large numbers of support vectors. The proposed extension allows for a larger class of online learning algorithms—including those alleviating the problem above through model compression. In addition, we characterize the quality of the proposed protocol by introducing a novel criterion that requires the communication to be bounded by the loss suffered.",

author = "Michael Kamp and Sebastian Bothe and Mario Boley and Michael Mock",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing AG 2016.; 15th European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2016 ; Conference date: 19-09-2016 Through 23-09-2016",

year = "2016",

doi = "10.1007/978-3-319-46227-1_50",

language = "English",

isbn = "9783319462264",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "805--819",

editor = "Niels Landwehr and Jilles Giuseppe and Giuseppe Manco and Paolo Frasconi",

booktitle = "Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2016, Proceedings",

address = "Germany",

}

Kamp, M, Bothe, S, Boley, M & Mock, M 2016, Communication-efficient distributed online learning with kernels. in N Landwehr, J Giuseppe, G Manco & P Frasconi (eds), Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2016, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9852 LNAI, Springer Verlag, pp. 805-819, 15th European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2016, Riva del Garda, Italy, 19/09/16. https://doi.org/10.1007/978-3-319-46227-1_50

Communication-efficient distributed online learning with kernels. / Kamp, Michael; Bothe, Sebastian; Boley, Mario et al.
Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2016, Proceedings. ed. / Niels Landwehr; Jilles Giuseppe; Giuseppe Manco; Paolo Frasconi. Springer Verlag, 2016. p. 805-819 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9852 LNAI).

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

TY - GEN

T1 - Communication-efficient distributed online learning with kernels

AU - Kamp, Michael

AU - Bothe, Sebastian

AU - Boley, Mario

AU - Mock, Michael

N1 - Publisher Copyright: © Springer International Publishing AG 2016.

PY - 2016

Y1 - 2016

N2 - We propose an efficient distributed online learning protocol for low-latency real-time services. It extends a previously presented protocol to kernelized online learners that represent their models by a support vector expansion. While such learners often achieve higher predictive performance than their linear counterparts, communicating the support vector expansions becomes inefficient for large numbers of support vectors. The proposed extension allows for a larger class of online learning algorithms—including those alleviating the problem above through model compression. In addition, we characterize the quality of the proposed protocol by introducing a novel criterion that requires the communication to be bounded by the loss suffered.

AB - We propose an efficient distributed online learning protocol for low-latency real-time services. It extends a previously presented protocol to kernelized online learners that represent their models by a support vector expansion. While such learners often achieve higher predictive performance than their linear counterparts, communicating the support vector expansions becomes inefficient for large numbers of support vectors. The proposed extension allows for a larger class of online learning algorithms—including those alleviating the problem above through model compression. In addition, we characterize the quality of the proposed protocol by introducing a novel criterion that requires the communication to be bounded by the loss suffered.

UR - http://www.scopus.com/inward/record.url?scp=84988653784&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-46227-1_50

DO - 10.1007/978-3-319-46227-1_50

M3 - Conference contribution

AN - SCOPUS:84988653784

SN - 9783319462264

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 805

EP - 819

BT - Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2016, Proceedings

A2 - Landwehr, Niels

A2 - Giuseppe, Jilles

A2 - Manco, Giuseppe

A2 - Frasconi, Paolo

PB - Springer Verlag

T2 - 15th European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2016

Y2 - 19 September 2016 through 23 September 2016

ER -

Kamp M, Bothe S, Boley M, Mock M. Communication-efficient distributed online learning with kernels. In Landwehr N, Giuseppe J, Manco G, Frasconi P, editors, Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2016, Proceedings. Springer Verlag. 2016. p. 805-819. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-46227-1_50

Communication-efficient distributed online learning with kernels

Abstract

Publication series

Conference

Bibliographical note

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this