Use of some sensitivity criteria for choosing networks with good generalization ability
In most applications of the multilayer perceptron (MLP) the main objective is to maximize the generalization ability of the network. We show that this ability is related to the sensitivity of the output of the MLP to small input changes. Several criteria have been proposed for the evaluation of the sensitivity. We propose a new index and present a way for improving these sensitivity criteria. Some numerical experiments allow a first comparison of the efficiencies of these criteria.
KeywordsNeural Network Artificial Intelligence Complex System Numerical Experiment Nonlinear Dynamics
Unable to display preview. Download preview PDF.
- J.H. Friedman. An overview of predictive learning and function approximation, in V. Cherkassky, J.H. Friedman, H. Wechsler eds.,From Statistics to Neural Networks, Theory and Pattern Recognition Applications, Springer-Verlag, pp. 1–61, 1993.Google Scholar
- A. S. Weigend, B. A. Huberman and D. E. Rumelhart. Predicting sunspots and exchange rates with connectionist networks, in M. Casdagli and S. Eubank eds.,Nonlinear Modeling and Forecasting, Addison-Wesley, pp. 395–432, 1992.Google Scholar
- L. Fu and T. Chen. Sensitivity analysis for input vector in multilayer feedforward neural networks,IEEE Internatioal Conférence on Neural Networks, San Francisco, California, pp. 215–218, 1993.Google Scholar
- K. Matsuoka. An approach to generalization problem in back-propagation learning,Proc. INNC'90, Paris, pp. 765–768, 1990.Google Scholar
- C. M. Bishop. Curvature-driven smoothing in backpropagation neural networks,Proc. INNC'90, Paris, pp. 749–752, 1990.Google Scholar
- A. R. Webb. Functional approximation by feedforward networks: a least-squares approach to generalization,IEEE Trans. on Neural Networks, vol. 5, no. 3, pp 363–371, 1994.Google Scholar