Localized Multiple Kernel Learning—A Convex Approach

Yunwen Lei; Alexander Binder; Urun Dogan; Marius Kloft

Localized Multiple Kernel Learning—A Convex Approach

Yunwen Lei, Alexander Binder, Urun Dogan, Marius Kloft

Department of Mathematics

Research output: Chapter in book/report/conference proceeding › Conference proceeding › peer-review

5 Citations (Scopus)

Abstract

We propose a localized approach to multiple kernel learning that can be formulated as a convex optimization problem over a given cluster structure. For which we obtain generalization error guarantees and derive an optimization algorithm based on the Fenchel dual representation. Experiments on real-world datasets from the application domains of computational biology and computer vision show that convex localized multiple kernel learning can achieve higher prediction accuracies than its global and non-convex local counterparts.

Original language	English
Title of host publication	JMLR
Subtitle of host publication	Workshop and Conference Proceedings
Editors	Robert J. Durrant, Kee Eung Kim
Pages	81-96
Number of pages	16
Volume	63
Publication status	Published - Nov 2016
Event	8th Asian Conference on Machine Learning, ACML 2016 - Hamilton, New Zealand Duration: 16 Nov 2016 → 18 Nov 2016

Conference

Conference	8th Asian Conference on Machine Learning, ACML 2016
Country/Territory	New Zealand
City	Hamilton
Period	16/11/16 → 18/11/16

Scopus Subject Areas

Software
Control and Systems Engineering
Statistics and Probability
Artificial Intelligence

User-Defined Keywords

Generalization analysis
Localized algorithms
Multiple kernel learning

Access to Document

http://proceedings.mlr.press/v63/lei63.pdf

Cite this

@inproceedings{46963e8d295f4023a5570ed5514be169,

title = "Localized Multiple Kernel Learning—A Convex Approach",

abstract = "We propose a localized approach to multiple kernel learning that can be formulated as a convex optimization problem over a given cluster structure. For which we obtain generalization error guarantees and derive an optimization algorithm based on the Fenchel dual representation. Experiments on real-world datasets from the application domains of computational biology and computer vision show that convex localized multiple kernel learning can achieve higher prediction accuracies than its global and non-convex local counterparts.",

keywords = "Generalization analysis, Localized algorithms, Multiple kernel learning",

author = "Yunwen Lei and Alexander Binder and Urun Dogan and Marius Kloft",

note = "Funding Information: YL acknowledges support from the NSFC/RGC Joint Research Scheme [RGC Project No. N CityU120/14 and NSFC Project No. 11461161006]. AB acknowledges support from Singapore University of Technology and Design Startup Grant SRIS15105. MK acknowledges support from the German Research Foundation (DFG) award KL 2698/2-1 and from the Federal Ministry of Science and Education (BMBF) awards 031L0023A and 031B0187B. Publisher Copyright: {\textcopyright} 2016 Y. Lei, A. Binder, U. Dogan & M. Kloft. ; 8th Asian Conference on Machine Learning, ACML 2016 ; Conference date: 16-11-2016 Through 18-11-2016",

year = "2016",

month = nov,

language = "English",

volume = "63",

pages = "81--96",

editor = "Durrant, {Robert J.} and Kim, {Kee Eung}",

booktitle = "JMLR",

}

TY - GEN

T1 - Localized Multiple Kernel Learning—A Convex Approach

AU - Lei, Yunwen

AU - Binder, Alexander

AU - Dogan, Urun

AU - Kloft, Marius

N1 - Funding Information: YL acknowledges support from the NSFC/RGC Joint Research Scheme [RGC Project No. N CityU120/14 and NSFC Project No. 11461161006]. AB acknowledges support from Singapore University of Technology and Design Startup Grant SRIS15105. MK acknowledges support from the German Research Foundation (DFG) award KL 2698/2-1 and from the Federal Ministry of Science and Education (BMBF) awards 031L0023A and 031B0187B. Publisher Copyright: © 2016 Y. Lei, A. Binder, U. Dogan & M. Kloft.

PY - 2016/11

Y1 - 2016/11

N2 - We propose a localized approach to multiple kernel learning that can be formulated as a convex optimization problem over a given cluster structure. For which we obtain generalization error guarantees and derive an optimization algorithm based on the Fenchel dual representation. Experiments on real-world datasets from the application domains of computational biology and computer vision show that convex localized multiple kernel learning can achieve higher prediction accuracies than its global and non-convex local counterparts.

AB - We propose a localized approach to multiple kernel learning that can be formulated as a convex optimization problem over a given cluster structure. For which we obtain generalization error guarantees and derive an optimization algorithm based on the Fenchel dual representation. Experiments on real-world datasets from the application domains of computational biology and computer vision show that convex localized multiple kernel learning can achieve higher prediction accuracies than its global and non-convex local counterparts.

KW - Generalization analysis

KW - Localized algorithms

KW - Multiple kernel learning

UR - http://www.scopus.com/inward/record.url?scp=85059295887&partnerID=8YFLogxK

M3 - Conference proceeding

AN - SCOPUS:85059295887

VL - 63

SP - 81

EP - 96

BT - JMLR

A2 - Durrant, Robert J.

A2 - Kim, Kee Eung

T2 - 8th Asian Conference on Machine Learning, ACML 2016

Y2 - 16 November 2016 through 18 November 2016

ER -

Localized Multiple Kernel Learning—A Convex Approach

Abstract

Conference

Scopus Subject Areas

User-Defined Keywords

Access to Document

Other files and links

Fingerprint

Cite this