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

Adham Salih, Amiram Moshaiov

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

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 languageEnglish
Title of host publication2021 IEEE Congress on Evolutionary Computation, CEC 2021 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1478-1485
Number of pages8
ISBN (Electronic)9781728183923
DOIs
StatePublished - 2021
Event2021 IEEE Congress on Evolutionary Computation, CEC 2021 - Virtual, Krakow, Poland
Duration: 28 Jun 20211 Jul 2021

Publication series

Name2021 IEEE Congress on Evolutionary Computation, CEC 2021 - Proceedings

Conference

Conference2021 IEEE Congress on Evolutionary Computation, CEC 2021
Country/TerritoryPoland
CityVirtual, Krakow
Period28/06/211/07/21

Funding

FundersFunder number
Ministry of Science and Technology, Israel

    Keywords

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

    Fingerprint

    Dive into the research topics of 'Modified Decomposition Framework and Algorithm for Many-objective Topology and Weight Evolution of Neural Networks'. Together they form a unique fingerprint.

    Cite this