Evolutionary algorithms for training neural networks

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

Abstract

This paper surveys the various approaches used to apply evolutionary algorithms to develop artificial neural networks that solve pattern recognition, classification, and other tasks. These approaches are classified into four groups, each addressing one aspect of an artificial neural network: (a) evolving connection weights; (b) evolving neural architectures; (c) evolving an ensemble of networks; and (d) evolving node functions. Hybrid approaches are also discussed.

Original languageEnglish (US)
Title of host publicationModeling and Simulation for Military Applications
DOIs
StatePublished - Sep 14 2006
EventModeling and Simulation for Military Applications - Kissimmee, FL, United States
Duration: Apr 18 2006Apr 21 2006

Publication series

NameProceedings of SPIE - The International Society for Optical Engineering
Volume6228
ISSN (Print)0277-786X

Other

OtherModeling and Simulation for Military Applications
CountryUnited States
CityKissimmee, FL
Period4/18/064/21/06

Keywords

  • Ensemble models
  • Evolutionary algorithms
  • Model learning
  • Neural networks
  • Optimization
  • Parameter learning

ASJC Scopus subject areas

  • Electronic, Optical and Magnetic Materials
  • Condensed Matter Physics
  • Computer Science Applications
  • Applied Mathematics
  • Electrical and Electronic Engineering

Fingerprint Dive into the research topics of 'Evolutionary algorithms for training neural networks'. Together they form a unique fingerprint.

  • Cite this

    Mohan, C. K. (2006). Evolutionary algorithms for training neural networks. In Modeling and Simulation for Military Applications [62280Q] (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 6228). https://doi.org/10.1117/12.670263