Two k-winners-take-all networks with discontinuous activation functions

Qingshan Liu; Jun Wang

doi:10.1016/j.neunet.2007.12.044

Two k-winners-take-all networks with discontinuous activation functions

Qingshan Liu^*, Jun Wang

^*Corresponding author for this work

Research output: Journal Publications and Reviews › RGC 21 - Publication in refereed journal › peer-review

52 Citations (Scopus)

Abstract

This paper presents two k-winners-take-all (k-WTA) networks with discontinuous activation functions. The k-WTA operation is first converted equivalently into linear and quadratic programming problems. Then two k-winners-take-all networks are designed based on the linear and quadratic programming formulations. The networks are theoretically guaranteed to be capable of performing the k-WTA operation in real time. Simulation results show the effectiveness and performance of the networks. © 2007 Elsevier Ltd. All rights reserved.

Original language	English
Pages (from-to)	406-413
Journal	Neural Networks
Volume	21
Issue number	2-3
DOIs	https://doi.org/10.1016/j.neunet.2007.12.044
Publication status	Published - Mar 2008
Externally published	Yes

Research Keywords

Differential inclusion
Discontinuous activation function
Global convergence
Linear programming
Lyapunov stability
Quadratic programming
Recurrent neural network

Access Output

10.1016/j.neunet.2007.12.044

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

@article{7c695106b4ff47c9abd73e65202bc53d,

title = "Two k-winners-take-all networks with discontinuous activation functions",

abstract = "This paper presents two k-winners-take-all (k-WTA) networks with discontinuous activation functions. The k-WTA operation is first converted equivalently into linear and quadratic programming problems. Then two k-winners-take-all networks are designed based on the linear and quadratic programming formulations. The networks are theoretically guaranteed to be capable of performing the k-WTA operation in real time. Simulation results show the effectiveness and performance of the networks. {\textcopyright} 2007 Elsevier Ltd. All rights reserved.",

keywords = "Differential inclusion, Discontinuous activation function, Global convergence, Linear programming, Lyapunov stability, Quadratic programming, Recurrent neural network",

author = "Qingshan Liu and Jun Wang",

year = "2008",

month = mar,

doi = "10.1016/j.neunet.2007.12.044",

language = "English",

volume = "21",

pages = "406--413",

journal = "Neural Networks",

issn = "0893-6080",

publisher = "Elsevier Ltd.",

number = "2-3",

}

TY - JOUR

T1 - Two k-winners-take-all networks with discontinuous activation functions

AU - Liu, Qingshan

AU - Wang, Jun

PY - 2008/3

Y1 - 2008/3

N2 - This paper presents two k-winners-take-all (k-WTA) networks with discontinuous activation functions. The k-WTA operation is first converted equivalently into linear and quadratic programming problems. Then two k-winners-take-all networks are designed based on the linear and quadratic programming formulations. The networks are theoretically guaranteed to be capable of performing the k-WTA operation in real time. Simulation results show the effectiveness and performance of the networks. © 2007 Elsevier Ltd. All rights reserved.

AB - This paper presents two k-winners-take-all (k-WTA) networks with discontinuous activation functions. The k-WTA operation is first converted equivalently into linear and quadratic programming problems. Then two k-winners-take-all networks are designed based on the linear and quadratic programming formulations. The networks are theoretically guaranteed to be capable of performing the k-WTA operation in real time. Simulation results show the effectiveness and performance of the networks. © 2007 Elsevier Ltd. All rights reserved.

KW - Differential inclusion

KW - Discontinuous activation function

KW - Global convergence

KW - Linear programming

KW - Lyapunov stability

KW - Quadratic programming

KW - Recurrent neural network

UR - http://www.scopus.com/inward/record.url?scp=40649091560&partnerID=8YFLogxK

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-40649091560&origin=recordpage

U2 - 10.1016/j.neunet.2007.12.044

DO - 10.1016/j.neunet.2007.12.044

M3 - RGC 21 - Publication in refereed journal

C2 - 18243655

SN - 0893-6080

VL - 21

SP - 406

EP - 413

JO - Neural Networks

JF - Neural Networks

IS - 2-3

ER -

Two k-winners-take-all networks with discontinuous activation functions

Abstract

Research Keywords

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Cite this