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.
CITATION STYLE
Branke, J., Kohlmorgen, U., & Schmeck, H. (1995). A distributed genetic algorithm improving the generalization behavior of neural networks. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 912, pp. 107–121). Springer Verlag. https://doi.org/10.1007/3-540-59286-5_52
Mendeley helps you to discover research relevant for your work.