Abstract
In spite of the general wisdom that much more examples need to test than to train a neural network, vice-versa we show that testing the approximation capability of a neural network generally demands smaller sample size than training it.
We move in an extended PAC learning framework and use some recent results in terms of sentry functions of a concept class to statistically proof our claims.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Apolloni B. (1992) Design of Algorithms for Neural Networks. Supervised Learning. Computers and artificial intelligence, vol. 5, 457–480, 1992.
Apolloni B., Chiaravalli S. PAC Learning of concept classes through the boundaries of their items. J. Theoretical Computer Science, vol. 172, 91–120, 1997.
Apolloni B., C. Ferretti, G. Mauri, Approximation of optimization problems and learnability. Proc. of GAA 92,Roma, 261–268, 1992.
Barron A., Neural networks approximation, Procs. of 7th Yale Workshop on Adaprive and Learning Systems, pp.69–72, 1992.
Baum E.,B., and D. Haussler, What size net gives valid generalizations?, Neural Comp. 1, 151–160, 1989
Blumer A., Ehrenfeucht A., Haussler D., Warmuth M., Learnability and the Vapnik- Chervonenkis dimension, J. of ACM 36, 929 - 965, 1989.
D. Haussler, Generalizing the PAC model for neural net and other learning applications. Res. Rep. UCSC-CRL-89-30, University of California, Santa Cruz, 1989.
Lawrence S, A.C. Tsoi and C.Lee. Giles, Local minima and generalization, Procs. ICNN, vol. 1, pp. 371–376, 1996.
Valiant L.G. A theory of the learnable. Communication of ACM 27, 11 (1984), pp. 1134–1142.
Vapnik V.V., Estimation of dependences based on empirical data, Springer-Verlag N.Y.,1982.
Vapnik V.V., The nature of Statistical Learning Theory, Springer-Verlag N.Y.,1982.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1998 Springer-Verlag London Limited
About this paper
Cite this paper
Apolloni, B. (1998). What Size Needs Testing?. In: Marinaro, M., Tagliaferri, R. (eds) Neural Nets WIRN VIETRI-97. Perspectives in Neural Computing. Springer, London. https://doi.org/10.1007/978-1-4471-1520-5_5
Download citation
DOI: https://doi.org/10.1007/978-1-4471-1520-5_5
Publisher Name: Springer, London
Print ISBN: 978-1-4471-1522-9
Online ISBN: 978-1-4471-1520-5
eBook Packages: Springer Book Archive