ML Anthology

A Distributed Genetic Algorithm Improving the Generalization Behavior of Neural Networks

Jürgen Branke, Udo Kohlmorgen, Hartmut Schmeck
ECML-PKDD 1995 pp. 107-121
doi:10.1007/3-540-59286-5_52 /ecmlpkdd/1995/branke1995ecml-distributed/

Abstract

Artificial neural networks sometimes generalize poorly to unknown inputs, if they have been trained perfectly on relatively small training sets using standard learning algorithms like e.g. backpropagation. In this paper a distributed genetic algorithm is designed and used to improve the network's generalization capabilities by reducing the number of different weights in the neural network.

PDF ECML-PKDD Semantic Scholar

Cite

Text

Branke et al. "A Distributed Genetic Algorithm Improving the Generalization Behavior of Neural Networks." European Conference on Machine Learning, 1995. doi:10.1007/3-540-59286-5_52

Markdown

[Branke et al. "A Distributed Genetic Algorithm Improving the Generalization Behavior of Neural Networks." European Conference on Machine Learning, 1995.](https://mlanthology.org/ecmlpkdd/1995/branke1995ecml-distributed/) doi:10.1007/3-540-59286-5_52

BibTeX

@inproceedings{branke1995ecml-distributed,
  title     = {{A Distributed Genetic Algorithm Improving the Generalization Behavior of Neural Networks}},
  author    = {Branke, Jürgen and Kohlmorgen, Udo and Schmeck, Hartmut},
  booktitle = {European Conference on Machine Learning},
  year      = {1995},
  pages     = {107-121},
  doi       = {10.1007/3-540-59286-5_52},
  url       = {https://mlanthology.org/ecmlpkdd/1995/branke1995ecml-distributed/}
}