Functional Equivalence and Genetic Learning of RBF Networks
In this paper a functional equivalence property of feedforward networks is introduced and studied for the case of radial basis function networks with Gaussian activation function and metrics induced by an inner product. The description of functional equivalent parameterizations is used for proposition of new genetic learning rules that operate only on a small part of the whole weight space.
KeywordsCrossover Operator Weight Space Radial Basis Function Network Hide Unit Feedforward Network
Unable to display preview. Download preview PDF.
- 2.Chen, A.M., Hecht-Nielsen, R.: On the geometry of feedforward neural network weight spaces. In Proceedings of the 2nd IEE Conerence on Artificial Neural Networks 1–4, London: IEE Press, 1991.Google Scholar
- 5.Hecht-Nielsen, R.: On the algebraic structure of feedforward network weight spaces. In Advanced Neural Computers, 129–135, Elsevier, 1990.Google Scholar
- 6.Hlaváčková, K., Neruda, R.: Radial Basis Function Networks, Neural Network World, 3, 1, 93–101 (1993).Google Scholar
- 7.Kainen, P. et al: A New Criterion for Choosing an Activation Function: Uniqueness Leads to Faster Learning, Neural Parallel and Scientific Computations (in press)Google Scholar
- 9.Kůrková, V., Neruda, R. Uniqueness of Functional Representation by Gaussian Basis Function Networks. In Proceedings of ICANN′94, 417–474, Springer: London, 1994.Google Scholar