论文信息 - On Some Factors Influencing MLP Error Surface

On Some Factors Influencing MLP Error Surface

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.

Wlodzislaw Duch | Miroslaw Kordos | Wlodzislaw Duch | M. Kordos

[1] Héctor J. Sussmann,et al. Uniqueness of the weights for minimal feedforward nets with a given input-output map , 1992, Neural Networks.

[2] Catherine Blake,et al. UCI Repository of machine learning databases , 1998 .

[3] Simon Haykin,et al. Neural Networks: A Comprehensive Foundation , 1998 .

[4] Włodzisław Duch,et al. Multilayer Perceptron Trained with Numerical Gradient , 2003 .

[5] Christopher J. Merz,et al. UCI Repository of Machine Learning Databases , 1996 .

[6] Lawrence D. Jackel,et al. Large Automatic Learning, Rule Extraction, and Generalization , 1987, Complex Syst..

[7] Marcus Gallagher,et al. Multi-layer Perceptron Error Surfaces: Visualization, Structure and Modelling , 2000 .

[8] Włodzisław Duch,et al. Search-based Training for Logical Rule Extraction by Multilayer Perceptron , 2003 .