Accelerated proximal stochastic dual coordinate ascent for regularized loss minimization

Shai Shalev-Shwartz; Tong Zhang

doi:10.1007/s10107-014-0839-0

Accelerated proximal stochastic dual coordinate ascent for regularized loss minimization

Shai Shalev-Shwartz
, Tong Zhang^*

^*Corresponding author for this work

The Rachel and Selim Benin School of Engineering and Computer Science

Research output: Contribution to journal › Article › peer-review

131 Scopus citations

Abstract

We introduce a proximal version of the stochastic dual coordinate ascent method and show how to accelerate the method using an inner-outer iteration procedure. We analyze the runtime of the framework and obtain rates that improve state-of-the-art results for various key machine learning optimization problems including SVM, logistic regression, ridge regression, Lasso, and multiclass SVM. Experiments validate our theoretical findings.

Original language	English
Pages (from-to)	105-145
Number of pages	41
Journal	Mathematical Programming
Volume	155
Issue number	1-2
DOIs	https://doi.org/10.1007/s10107-014-0839-0
State	Published - 1 Jan 2016

Bibliographical note

Publisher Copyright:
© 2014, Springer-Verlag Berlin Heidelberg and Mathematical Optimization Society.

Keywords

90C06
90C15
90C25

Access to Document

10.1007/s10107-014-0839-0

Cite this

@article{6eabbfe50c164336991a3b194eed5027,

title = "Accelerated proximal stochastic dual coordinate ascent for regularized loss minimization",

abstract = "We introduce a proximal version of the stochastic dual coordinate ascent method and show how to accelerate the method using an inner-outer iteration procedure. We analyze the runtime of the framework and obtain rates that improve state-of-the-art results for various key machine learning optimization problems including SVM, logistic regression, ridge regression, Lasso, and multiclass SVM. Experiments validate our theoretical findings.",

keywords = "90C06, 90C15, 90C25",

author = "Shai Shalev-Shwartz and Tong Zhang",

note = "Publisher Copyright: {\textcopyright} 2014, Springer-Verlag Berlin Heidelberg and Mathematical Optimization Society.",

year = "2016",

month = jan,

day = "1",

doi = "10.1007/s10107-014-0839-0",

language = "אנגלית",

volume = "155",

pages = "105--145",

journal = "Mathematical Programming",

issn = "0025-5610",

publisher = "Springer-Verlag GmbH and Co. KG",

number = "1-2",

}

TY - JOUR

T1 - Accelerated proximal stochastic dual coordinate ascent for regularized loss minimization

AU - Shalev-Shwartz, Shai

AU - Zhang, Tong

PY - 2016/1/1

Y1 - 2016/1/1

N2 - We introduce a proximal version of the stochastic dual coordinate ascent method and show how to accelerate the method using an inner-outer iteration procedure. We analyze the runtime of the framework and obtain rates that improve state-of-the-art results for various key machine learning optimization problems including SVM, logistic regression, ridge regression, Lasso, and multiclass SVM. Experiments validate our theoretical findings.

AB - We introduce a proximal version of the stochastic dual coordinate ascent method and show how to accelerate the method using an inner-outer iteration procedure. We analyze the runtime of the framework and obtain rates that improve state-of-the-art results for various key machine learning optimization problems including SVM, logistic regression, ridge regression, Lasso, and multiclass SVM. Experiments validate our theoretical findings.

KW - 90C06

KW - 90C15

KW - 90C25

UR - https://www.scopus.com/pages/publications/84953283129

U2 - 10.1007/s10107-014-0839-0

DO - 10.1007/s10107-014-0839-0

M3 - ???researchoutput.researchoutputtypes.contributiontojournal.article???

AN - SCOPUS:84953283129

SN - 0025-5610

VL - 155

SP - 105

EP - 145

JO - Mathematical Programming

JF - Mathematical Programming

IS - 1-2

ER -

Accelerated proximal stochastic dual coordinate ascent for regularized loss minimization

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this