Sparse system identification by exponentially weighted adaptive parallel projection and generalized soft-thresholding

Masao Yamagishi, Masahiro Yukawa, Isao Yamada

Research output: Chapter in Book/Report/Conference proceedingConference contribution

10 Citations (Scopus)

Abstract

In this paper, we propose a novel online scheme named the adaptive proximal forward-backward splitting method to suppress the sum of'smooth' and'nonsmooth' convex functions, both of which are assumed time-varying. We derive a powerful algorithm for sparse system identification by defining each function as a certain average squared distance ('smooth') and a weighted ℓ1-norm ('nonsmooth'). The smooth term brings an exponentially weighted average of the metric projections of the current estimate onto linear varieties, of which the number grows as new measurements arrive. The presented recursive formula realizes an efficient computation of the average (the exponentially weighted adaptive parallel projection), which contributes the fast and stable convergence. The nonsmooth term, on the other hand, brings the weighted soft-thresholding, contributing the enhancement of the filter sparsity. The weights are adaptively controlled according to the filter coefficients so that the softthresholding gives significant impacts solely to inactive coefficients (coefficients close to zero). The numerical example demonstrates the efficacy of the proposed algorithm.

Original languageEnglish
Title of host publicationAPSIPA ASC 2010 - Asia-Pacific Signal and Information Processing Association Annual Summit and Conference
Pages367-370
Number of pages4
Publication statusPublished - 2010
Externally publishedYes
Event2nd Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2010 - Biopolis, Singapore
Duration: 2010 Dec 142010 Dec 17

Other

Other2nd Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2010
CountrySingapore
CityBiopolis
Period10/12/1410/12/17

Fingerprint

Identification (control systems)

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Information Systems

Cite this

Yamagishi, M., Yukawa, M., & Yamada, I. (2010). Sparse system identification by exponentially weighted adaptive parallel projection and generalized soft-thresholding. In APSIPA ASC 2010 - Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (pp. 367-370)

Sparse system identification by exponentially weighted adaptive parallel projection and generalized soft-thresholding. / Yamagishi, Masao; Yukawa, Masahiro; Yamada, Isao.

APSIPA ASC 2010 - Asia-Pacific Signal and Information Processing Association Annual Summit and Conference. 2010. p. 367-370.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Yamagishi, M, Yukawa, M & Yamada, I 2010, Sparse system identification by exponentially weighted adaptive parallel projection and generalized soft-thresholding. in APSIPA ASC 2010 - Asia-Pacific Signal and Information Processing Association Annual Summit and Conference. pp. 367-370, 2nd Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2010, Biopolis, Singapore, 10/12/14.
Yamagishi M, Yukawa M, Yamada I. Sparse system identification by exponentially weighted adaptive parallel projection and generalized soft-thresholding. In APSIPA ASC 2010 - Asia-Pacific Signal and Information Processing Association Annual Summit and Conference. 2010. p. 367-370
Yamagishi, Masao ; Yukawa, Masahiro ; Yamada, Isao. / Sparse system identification by exponentially weighted adaptive parallel projection and generalized soft-thresholding. APSIPA ASC 2010 - Asia-Pacific Signal and Information Processing Association Annual Summit and Conference. 2010. pp. 367-370
@inproceedings{879058a85802456f8388063a1d22e37a,
title = "Sparse system identification by exponentially weighted adaptive parallel projection and generalized soft-thresholding",
abstract = "In this paper, we propose a novel online scheme named the adaptive proximal forward-backward splitting method to suppress the sum of'smooth' and'nonsmooth' convex functions, both of which are assumed time-varying. We derive a powerful algorithm for sparse system identification by defining each function as a certain average squared distance ('smooth') and a weighted ℓ1-norm ('nonsmooth'). The smooth term brings an exponentially weighted average of the metric projections of the current estimate onto linear varieties, of which the number grows as new measurements arrive. The presented recursive formula realizes an efficient computation of the average (the exponentially weighted adaptive parallel projection), which contributes the fast and stable convergence. The nonsmooth term, on the other hand, brings the weighted soft-thresholding, contributing the enhancement of the filter sparsity. The weights are adaptively controlled according to the filter coefficients so that the softthresholding gives significant impacts solely to inactive coefficients (coefficients close to zero). The numerical example demonstrates the efficacy of the proposed algorithm.",
author = "Masao Yamagishi and Masahiro Yukawa and Isao Yamada",
year = "2010",
language = "English",
pages = "367--370",
booktitle = "APSIPA ASC 2010 - Asia-Pacific Signal and Information Processing Association Annual Summit and Conference",

}

TY - GEN

T1 - Sparse system identification by exponentially weighted adaptive parallel projection and generalized soft-thresholding

AU - Yamagishi, Masao

AU - Yukawa, Masahiro

AU - Yamada, Isao

PY - 2010

Y1 - 2010

N2 - In this paper, we propose a novel online scheme named the adaptive proximal forward-backward splitting method to suppress the sum of'smooth' and'nonsmooth' convex functions, both of which are assumed time-varying. We derive a powerful algorithm for sparse system identification by defining each function as a certain average squared distance ('smooth') and a weighted ℓ1-norm ('nonsmooth'). The smooth term brings an exponentially weighted average of the metric projections of the current estimate onto linear varieties, of which the number grows as new measurements arrive. The presented recursive formula realizes an efficient computation of the average (the exponentially weighted adaptive parallel projection), which contributes the fast and stable convergence. The nonsmooth term, on the other hand, brings the weighted soft-thresholding, contributing the enhancement of the filter sparsity. The weights are adaptively controlled according to the filter coefficients so that the softthresholding gives significant impacts solely to inactive coefficients (coefficients close to zero). The numerical example demonstrates the efficacy of the proposed algorithm.

AB - In this paper, we propose a novel online scheme named the adaptive proximal forward-backward splitting method to suppress the sum of'smooth' and'nonsmooth' convex functions, both of which are assumed time-varying. We derive a powerful algorithm for sparse system identification by defining each function as a certain average squared distance ('smooth') and a weighted ℓ1-norm ('nonsmooth'). The smooth term brings an exponentially weighted average of the metric projections of the current estimate onto linear varieties, of which the number grows as new measurements arrive. The presented recursive formula realizes an efficient computation of the average (the exponentially weighted adaptive parallel projection), which contributes the fast and stable convergence. The nonsmooth term, on the other hand, brings the weighted soft-thresholding, contributing the enhancement of the filter sparsity. The weights are adaptively controlled according to the filter coefficients so that the softthresholding gives significant impacts solely to inactive coefficients (coefficients close to zero). The numerical example demonstrates the efficacy of the proposed algorithm.

UR - http://www.scopus.com/inward/record.url?scp=79958127505&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79958127505&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:79958127505

SP - 367

EP - 370

BT - APSIPA ASC 2010 - Asia-Pacific Signal and Information Processing Association Annual Summit and Conference

ER -