On Some Factors Influencing MLP Error Surface.


Miroslaw Kordos,
Faculty of Automatic Control, Electronics and Computer Science,
The Silesian University of Technology, Gliwice, Poland.

Wlodzislaw Duch,
School of Computer Engineering, Nanyang Technological University, Singapore,
and Department of Informatics, Nicholas Copernicus University,
Grudziadzka 5, 87-100 Torun, Poland.

Abstract.

Visualization of MLP error surfaces helps to understand the influence of network structure and training data on neural learning dynamics. PCA is used to determine two orthogonal directions that capture almost all variance in the weight space. 3-dimensional plots show many aspects of the original error surfaces.

The Seventh International Conference on Artificial Intelligence and Soft Computing (ICAISC), Zakopane, 7-11.06.2004 (submitted 11/03)

Preprint for comments in PDF, 519 KB.

BACK to the publications of W. Duch.
BACK to the on-line publications of the Department of Informatics, NCU.