Accelerated dual-averaging primal–dual method for composite convex minimization

Conghui Tan; Yuqiu Qian; Shiqian Ma; Tong Zhang

doi:10.1080/10556788.2020.1713779

Accelerated dual-averaging primal–dual method for composite convex minimization

Conghui Tan, Yuqiu Qian, Shiqian Ma^*, Tong Zhang

^*Corresponding author for this work

Department of Computer Science and Engineering

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

6 Citations (Scopus)

Abstract

Dual averaging-type methods are widely used in industrial machine learning applications due to their ability to promoting solution structure (e.g. sparsity) efficiently. In this paper, we propose a novel accelerated dual-averaging primal–dual algorithm for minimizing a composite convex function. We also derive a stochastic version of the proposed method that solves empirical risk minimization, and its advantages on handling sparse data are demonstrated both theoretically and empirically.

Original language	English
Pages (from-to)	741-766
Number of pages	26
Journal	Optimization Methods and Software
Volume	35
Issue number	4
DOIs	https://doi.org/10.1080/10556788.2020.1713779
Publication status	Published - 3 Jul 2020

Bibliographical note

Publisher Copyright:
© 2020, © 2020 Informa UK Limited, trading as Taylor & Francis Group.

Keywords

Dual averaging algorithm
acceleration
empirical risk minimization
primal–dual
sparse data

Access to Document

10.1080/10556788.2020.1713779

Cite this

@article{ff22a2ce6af4451490984cedf513d3bf,

title = "Accelerated dual-averaging primal–dual method for composite convex minimization",

abstract = "Dual averaging-type methods are widely used in industrial machine learning applications due to their ability to promoting solution structure (e.g. sparsity) efficiently. In this paper, we propose a novel accelerated dual-averaging primal–dual algorithm for minimizing a composite convex function. We also derive a stochastic version of the proposed method that solves empirical risk minimization, and its advantages on handling sparse data are demonstrated both theoretically and empirically.",

keywords = "Dual averaging algorithm, acceleration, empirical risk minimization, primal–dual, sparse data",

author = "Conghui Tan and Yuqiu Qian and Shiqian Ma and Tong Zhang",

note = "Publisher Copyright: {\textcopyright} 2020, {\textcopyright} 2020 Informa UK Limited, trading as Taylor \& Francis Group.",

year = "2020",

month = jul,

day = "3",

doi = "10.1080/10556788.2020.1713779",

language = "English",

volume = "35",

pages = "741--766",

journal = "Optimization Methods and Software",

issn = "1055-6788",

publisher = "Taylor and Francis Ltd.",

number = "4",

}

TY - JOUR

T1 - Accelerated dual-averaging primal–dual method for composite convex minimization

AU - Tan, Conghui

AU - Qian, Yuqiu

AU - Ma, Shiqian

AU - Zhang, Tong

PY - 2020/7/3

Y1 - 2020/7/3

N2 - Dual averaging-type methods are widely used in industrial machine learning applications due to their ability to promoting solution structure (e.g. sparsity) efficiently. In this paper, we propose a novel accelerated dual-averaging primal–dual algorithm for minimizing a composite convex function. We also derive a stochastic version of the proposed method that solves empirical risk minimization, and its advantages on handling sparse data are demonstrated both theoretically and empirically.

AB - Dual averaging-type methods are widely used in industrial machine learning applications due to their ability to promoting solution structure (e.g. sparsity) efficiently. In this paper, we propose a novel accelerated dual-averaging primal–dual algorithm for minimizing a composite convex function. We also derive a stochastic version of the proposed method that solves empirical risk minimization, and its advantages on handling sparse data are demonstrated both theoretically and empirically.

KW - Dual averaging algorithm

KW - acceleration

KW - empirical risk minimization

KW - primal–dual

KW - sparse data

UR - https://www.webofscience.com/wos/woscc/full-record/WOS:000508103600001

UR - https://openalex.org/W2999988540

UR - https://www.scopus.com/pages/publications/85078478666

U2 - 10.1080/10556788.2020.1713779

DO - 10.1080/10556788.2020.1713779

M3 - Journal Article

SN - 1055-6788

VL - 35

SP - 741

EP - 766

JO - Optimization Methods and Software

JF - Optimization Methods and Software

IS - 4

ER -

Accelerated dual-averaging primal–dual method for composite convex minimization

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this