Sample sizes for multiple-output threshold networks
- 1 February 1991
- journal article
- Published by Taylor & Francis in Network: Computation in Neural Systems
- Vol. 2 (1) , 107-117
- https://doi.org/10.1088/0954-898x/2/1/006
Abstract
This paper applies the theory of probably approximately correct (PAC) le-g to multiplcmtput feedfomard tbhold networks. It is &om that the sample size for reliable learning can be bounded above by a quantity independent of the number of outputs of the network.Keywords
This publication has 8 references indexed in Scilit:
- Learnability and the Vapnik-Chervonenkis dimensionJournal of the ACM, 1989
- A general lower bound on the number of examples needed for learningInformation and Computation, 1989
- What Size Net Gives Valid Generalization?Neural Computation, 1989
- Bounds on the learning capacity of some multi-layer networksBiological Cybernetics, 1989
- On the capabilities of multilayer perceptronsJournal of Complexity, 1988
- A theory of the learnableCommunications of the ACM, 1984
- On the density of families of setsJournal of Combinatorial Theory, Series A, 1972
- Geometrical and Statistical Properties of Systems of Linear Inequalities with Applications in Pattern RecognitionIEEE Transactions on Electronic Computers, 1965