Multilayer Perceptron Trained with Numerical Gradient (2003)
| Venue: | Proc. of Int. Conf. on Artificial Neural Networks (ICANN |
| Citations: | 6 - 5 self |
BibTeX
@INPROCEEDINGS{Duch03multilayerperceptron,
author = {Włodzisław Duch and Mirosław Kordos},
title = {Multilayer Perceptron Trained with Numerical Gradient},
booktitle = {Proc. of Int. Conf. on Artificial Neural Networks (ICANN},
year = {2003},
pages = {106--109}
}
Bookmark
 |
 |
 |
 |
 |
 |
OpenURL
Abstract
Abstract—An application of numerical gradient (NG) to training of MLP networks is presented. Several versions of the algorithm and the influence of various parameters on the training process are discussed. Optimization of network parameters based on global search with numerical gradient is presented. Examples of two-dimensional projection of the error surface are shown and the influence of various numerical gradient parameters on the error surface is presented. The speed and accuracy of this method is compared with the search-based MLP training algorithm.
Citations
| 3912 | CM: Neural Networks for Pattern Recognition - Bishop - 1995 |
| 122 | An empirical comparison of pattern recognition, neural nets, and machine learning classi®cation methods - Weiss, Kapouleas - 1989 |
| 10 | Optimization by Simulated Annealing, Science 220 - Vecchi - 1983 |
| 7 | Optimization and global minimization methods suitable for neural networks - Duch, Korczak - 1998 |
| 6 | Search-based Training for Logical Rule Extraction by Multilayer Perceptron. ICONIP/ICANN 2003 (this volume - Kordos, Duch |
| 2 | Analog computation and learning in VLSI - Koosh - 2001 |
