Multi-domain adaptive learning based on feasibility splitting and adaptive projected subgradient method

Masahiro Yukawa, Konstantinos Slavakis, Isao Yamada

Research output: Contribution to journalArticle

16 Citations (Scopus)

Abstract

We propose the multi-domain adaptive learning that enables us to find a point meeting possibly time-varying specifications simultaneously in multiple domains, e.g. space, time, frequency, etc. The novel concept is based on the idea of feasibility splitting - dealing with feasibility in each individual domain. We show that the adaptive projected subgradient method (Yamada, 2003) realizes the multi-domain adaptive learning by employing (i) a projected gradient operator with respect to a 'fixed' proximity function reflecting the time-invariant specifications and (ii) a subgradient projection with respect to 'time-varying' objective functions reflecting the time-varying specifications. The resulting algorithm is suitable for real-time implementation, because it requires no more than metric projections onto closed convex sets each of which accommodates the specification in each domain. A convergence analysis and numerical examples are presented.

Original languageEnglish
Pages (from-to)456-466
Number of pages11
JournalIEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences
VolumeE93-A
Issue number2
DOIs
Publication statusPublished - 2010 Feb
Externally publishedYes

Fingerprint

Subgradient Method
Adaptive Learning
Specification
Specifications
Time-varying
Projected Gradient
Metric Projection
Subgradient
Closed set
Convergence Analysis
Convex Sets
Proximity
Mathematical operators
Objective function
Projection
Real-time
Numerical Examples
Invariant
Operator

Keywords

  • Adaptive algorithm
  • Convex feasibility problem
  • Convex projection
  • Projected gradient method

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Computer Graphics and Computer-Aided Design
  • Applied Mathematics
  • Signal Processing

Cite this

Multi-domain adaptive learning based on feasibility splitting and adaptive projected subgradient method. / Yukawa, Masahiro; Slavakis, Konstantinos; Yamada, Isao.

In: IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences, Vol. E93-A, No. 2, 02.2010, p. 456-466.

Research output: Contribution to journalArticle

@article{483e575470364926a5b400c54bf81eb7,
title = "Multi-domain adaptive learning based on feasibility splitting and adaptive projected subgradient method",
abstract = "We propose the multi-domain adaptive learning that enables us to find a point meeting possibly time-varying specifications simultaneously in multiple domains, e.g. space, time, frequency, etc. The novel concept is based on the idea of feasibility splitting - dealing with feasibility in each individual domain. We show that the adaptive projected subgradient method (Yamada, 2003) realizes the multi-domain adaptive learning by employing (i) a projected gradient operator with respect to a 'fixed' proximity function reflecting the time-invariant specifications and (ii) a subgradient projection with respect to 'time-varying' objective functions reflecting the time-varying specifications. The resulting algorithm is suitable for real-time implementation, because it requires no more than metric projections onto closed convex sets each of which accommodates the specification in each domain. A convergence analysis and numerical examples are presented.",
keywords = "Adaptive algorithm, Convex feasibility problem, Convex projection, Projected gradient method",
author = "Masahiro Yukawa and Konstantinos Slavakis and Isao Yamada",
year = "2010",
month = "2",
doi = "10.1587/transfun.E93.A.456",
language = "English",
volume = "E93-A",
pages = "456--466",
journal = "IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences",
issn = "0916-8508",
publisher = "Maruzen Co., Ltd/Maruzen Kabushikikaisha",
number = "2",

}

TY - JOUR

T1 - Multi-domain adaptive learning based on feasibility splitting and adaptive projected subgradient method

AU - Yukawa, Masahiro

AU - Slavakis, Konstantinos

AU - Yamada, Isao

PY - 2010/2

Y1 - 2010/2

N2 - We propose the multi-domain adaptive learning that enables us to find a point meeting possibly time-varying specifications simultaneously in multiple domains, e.g. space, time, frequency, etc. The novel concept is based on the idea of feasibility splitting - dealing with feasibility in each individual domain. We show that the adaptive projected subgradient method (Yamada, 2003) realizes the multi-domain adaptive learning by employing (i) a projected gradient operator with respect to a 'fixed' proximity function reflecting the time-invariant specifications and (ii) a subgradient projection with respect to 'time-varying' objective functions reflecting the time-varying specifications. The resulting algorithm is suitable for real-time implementation, because it requires no more than metric projections onto closed convex sets each of which accommodates the specification in each domain. A convergence analysis and numerical examples are presented.

AB - We propose the multi-domain adaptive learning that enables us to find a point meeting possibly time-varying specifications simultaneously in multiple domains, e.g. space, time, frequency, etc. The novel concept is based on the idea of feasibility splitting - dealing with feasibility in each individual domain. We show that the adaptive projected subgradient method (Yamada, 2003) realizes the multi-domain adaptive learning by employing (i) a projected gradient operator with respect to a 'fixed' proximity function reflecting the time-invariant specifications and (ii) a subgradient projection with respect to 'time-varying' objective functions reflecting the time-varying specifications. The resulting algorithm is suitable for real-time implementation, because it requires no more than metric projections onto closed convex sets each of which accommodates the specification in each domain. A convergence analysis and numerical examples are presented.

KW - Adaptive algorithm

KW - Convex feasibility problem

KW - Convex projection

KW - Projected gradient method

UR - http://www.scopus.com/inward/record.url?scp=77950789961&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77950789961&partnerID=8YFLogxK

U2 - 10.1587/transfun.E93.A.456

DO - 10.1587/transfun.E93.A.456

M3 - Article

AN - SCOPUS:77950789961

VL - E93-A

SP - 456

EP - 466

JO - IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences

JF - IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences

SN - 0916-8508

IS - 2

ER -