Incremental growth in modular neural networks

Christopher MacLeod, Grant Maxwell, Sethuraman Muthuraman

Research output: Contribution to journalArticlepeer-review

7 Citations (Scopus)

Abstract

This paper outlines an algorithm for incrementally growing Artificial Neural Networks. The algorithm allows the network to expand by adding new sub-networks or modules to an existing structure; the modules are trained using an Evolutionary Algorithm. Only the latest module added to the network is trained, the previous structure remains fixed. The algorithm allows information from different data domains to be integrated into the network and because the search space in each iteration is small, large and complex networks with a modular structure can emerge naturally. The paper describes an application of the algorithm to a legged robot and discusses its biological inspiration.

Original languageEnglish
Pages (from-to)660-666
Number of pages7
JournalEngineering Applications of Artificial Intelligence
Volume22
Issue number4-5
DOIs
Publication statusPublished - Jun 2009

Keywords

  • Evolutionary algorithm
  • Genetic algorithm
  • Incremental evolution
  • Incremental growth
  • Modular artificial neural network
  • Robotics

Fingerprint

Dive into the research topics of 'Incremental growth in modular neural networks'. Together they form a unique fingerprint.

Cite this