Modified Decomposition Framework and Algorithm for Many-objective Topology and Weight Evolution of Neural Networks

Adham Salih; Amiram Moshaiov

doi:10.1109/CEC45853.2021.9504953

Modified Decomposition Framework and Algorithm for Many-objective Topology and Weight Evolution of Neural Networks

Adham Salih, Amiram Moshaiov

School of Mechanical Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

6 Scopus citations

Abstract

This paper presents a modified decomposition framework to support the Many-Objective Topology and Weight Evolution of Artificial Neural Networks (MaO-TWEANNs). Next, an algorithm, which is termed NEWS/D, is devised using the proposed framework. To validate its optimization capabilities, a numerical study is carried out. The performed numerical study includes demonstration problems ranging from three to seven objectives for which the ideal points are known. Finally, an additional numerical study is performed with respect to a possible real-life application. The latter study suggests that evolving class experts for multi-class classification problems could be enhanced using NEWS/D in a non-intuitive approach.

Original language	English
Title of host publication	2021 IEEE Congress on Evolutionary Computation, CEC 2021 - Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	1478-1485
Number of pages	8
ISBN (Electronic)	9781728183923
DOIs	https://doi.org/10.1109/CEC45853.2021.9504953
State	Published - 2021
Event	2021 IEEE Congress on Evolutionary Computation, CEC 2021 - Virtual, Krakow, Poland Duration: 28 Jun 2021 → 1 Jul 2021

Publication series

Name	2021 IEEE Congress on Evolutionary Computation, CEC 2021 - Proceedings

Conference

Conference	2021 IEEE Congress on Evolutionary Computation, CEC 2021
Country/Territory	Poland
City	Virtual, Krakow
Period	28/06/21 → 1/07/21

Keywords

Artificial neural networks
Decomposition approach
Many-objective optimization
Neural architecture search
Neuro-evolution
Topology and weight evolution

Access to Document

10.1109/CEC45853.2021.9504953

Cite this

Salih, A., & Moshaiov, A. (2021). Modified Decomposition Framework and Algorithm for Many-objective Topology and Weight Evolution of Neural Networks. In 2021 IEEE Congress on Evolutionary Computation, CEC 2021 - Proceedings (pp. 1478-1485). (2021 IEEE Congress on Evolutionary Computation, CEC 2021 - Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CEC45853.2021.9504953

Salih, Adham ; Moshaiov, Amiram. / Modified Decomposition Framework and Algorithm for Many-objective Topology and Weight Evolution of Neural Networks. 2021 IEEE Congress on Evolutionary Computation, CEC 2021 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2021. pp. 1478-1485 (2021 IEEE Congress on Evolutionary Computation, CEC 2021 - Proceedings).

@inproceedings{0e8180347b9249758f0fa2827e91e70e,

title = "Modified Decomposition Framework and Algorithm for Many-objective Topology and Weight Evolution of Neural Networks",

abstract = "This paper presents a modified decomposition framework to support the Many-Objective Topology and Weight Evolution of Artificial Neural Networks (MaO-TWEANNs). Next, an algorithm, which is termed NEWS/D, is devised using the proposed framework. To validate its optimization capabilities, a numerical study is carried out. The performed numerical study includes demonstration problems ranging from three to seven objectives for which the ideal points are known. Finally, an additional numerical study is performed with respect to a possible real-life application. The latter study suggests that evolving class experts for multi-class classification problems could be enhanced using NEWS/D in a non-intuitive approach.",

keywords = "Artificial neural networks, Decomposition approach, Many-objective optimization, Neural architecture search, Neuro-evolution, Topology and weight evolution",

author = "Adham Salih and Amiram Moshaiov",

note = "Publisher Copyright: {\textcopyright} 2021 IEEE; 2021 IEEE Congress on Evolutionary Computation, CEC 2021 ; Conference date: 28-06-2021 Through 01-07-2021",

year = "2021",

doi = "10.1109/CEC45853.2021.9504953",

language = "אנגלית",

series = "2021 IEEE Congress on Evolutionary Computation, CEC 2021 - Proceedings",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "1478--1485",

booktitle = "2021 IEEE Congress on Evolutionary Computation, CEC 2021 - Proceedings",

address = "ארצות הברית",

}

Salih, A & Moshaiov, A 2021, Modified Decomposition Framework and Algorithm for Many-objective Topology and Weight Evolution of Neural Networks. in 2021 IEEE Congress on Evolutionary Computation, CEC 2021 - Proceedings. 2021 IEEE Congress on Evolutionary Computation, CEC 2021 - Proceedings, Institute of Electrical and Electronics Engineers Inc., pp. 1478-1485, 2021 IEEE Congress on Evolutionary Computation, CEC 2021, Virtual, Krakow, Poland, 28/06/21. https://doi.org/10.1109/CEC45853.2021.9504953

Modified Decomposition Framework and Algorithm for Many-objective Topology and Weight Evolution of Neural Networks. / Salih, Adham; Moshaiov, Amiram.
2021 IEEE Congress on Evolutionary Computation, CEC 2021 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2021. p. 1478-1485 (2021 IEEE Congress on Evolutionary Computation, CEC 2021 - Proceedings).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Modified Decomposition Framework and Algorithm for Many-objective Topology and Weight Evolution of Neural Networks

AU - Salih, Adham

AU - Moshaiov, Amiram

PY - 2021

Y1 - 2021

N2 - This paper presents a modified decomposition framework to support the Many-Objective Topology and Weight Evolution of Artificial Neural Networks (MaO-TWEANNs). Next, an algorithm, which is termed NEWS/D, is devised using the proposed framework. To validate its optimization capabilities, a numerical study is carried out. The performed numerical study includes demonstration problems ranging from three to seven objectives for which the ideal points are known. Finally, an additional numerical study is performed with respect to a possible real-life application. The latter study suggests that evolving class experts for multi-class classification problems could be enhanced using NEWS/D in a non-intuitive approach.

AB - This paper presents a modified decomposition framework to support the Many-Objective Topology and Weight Evolution of Artificial Neural Networks (MaO-TWEANNs). Next, an algorithm, which is termed NEWS/D, is devised using the proposed framework. To validate its optimization capabilities, a numerical study is carried out. The performed numerical study includes demonstration problems ranging from three to seven objectives for which the ideal points are known. Finally, an additional numerical study is performed with respect to a possible real-life application. The latter study suggests that evolving class experts for multi-class classification problems could be enhanced using NEWS/D in a non-intuitive approach.

KW - Artificial neural networks

KW - Decomposition approach

KW - Many-objective optimization

KW - Neural architecture search

KW - Neuro-evolution

KW - Topology and weight evolution

UR - http://www.scopus.com/inward/record.url?scp=85124584091&partnerID=8YFLogxK

U2 - 10.1109/CEC45853.2021.9504953

DO - 10.1109/CEC45853.2021.9504953

M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???

AN - SCOPUS:85124584091

T3 - 2021 IEEE Congress on Evolutionary Computation, CEC 2021 - Proceedings

SP - 1478

EP - 1485

BT - 2021 IEEE Congress on Evolutionary Computation, CEC 2021 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2021 IEEE Congress on Evolutionary Computation, CEC 2021

Y2 - 28 June 2021 through 1 July 2021

ER -

Salih A, Moshaiov A. Modified Decomposition Framework and Algorithm for Many-objective Topology and Weight Evolution of Neural Networks. In 2021 IEEE Congress on Evolutionary Computation, CEC 2021 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2021. p. 1478-1485. (2021 IEEE Congress on Evolutionary Computation, CEC 2021 - Proceedings). doi: 10.1109/CEC45853.2021.9504953

Modified Decomposition Framework and Algorithm for Many-objective Topology and Weight Evolution of Neural Networks

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this