In this paper, we prove that any continuous mapping can be approximately realized by Rumelhart-Hinton-Williams' multilayer neural networks with at least one hidden layer whose output functions are sigmoid functions. The starting point of the proof for the one hidden layer case is an integral formula
β¦ LIBER β¦
Approximate realization of identity mappings by three-layer neural networks
β Scribed by Ken-Ichi Funahashi
- Book ID
- 112079302
- Publisher
- John Wiley and Sons
- Year
- 1990
- Tongue
- English
- Weight
- 555 KB
- Volume
- 73
- Category
- Article
- ISSN
- 1042-0967
No coin nor oath required. For personal study only.
π SIMILAR VOLUMES
On the approximate realization of contin
β
Ken-Ichi Funahashi
π
Article
π
1989
π
Elsevier Science
π
English
β 733 KB
Constructive function-approximation by t
β
Shin Suzuki
π
Article
π
1998
π
Elsevier Science
π
English
β 493 KB
Constructive theorems of three-layer artificial neural networks with (1) trigonometric, (2) piecewise linear, and (3) sigmoidal hidden-layer units are proved in this paper. These networks approximate 2p-periodic pth-order Lebesgue-integrable functions (L p 2p ) on R m to R n for p Υ 1 with L p 2p ΒΉ
On the approximation of stochastic proce
β
Turchetti, C.; Conti, M.; Crippa, P.; Orcioni, S.
π
Article
π
1998
π
IEEE
π
English
β 785 KB
On the approximation of a quadratic map
β
Henrik SaxΓ©n
π
Article
π
1996
π
Elsevier Science
π
English
β 725 KB
Degree of Approximation by Neural and Tr
β
H.N. Mhaskar; C.A. Micchelli
π
Article
π
1995
π
Elsevier Science
π
English
β 921 KB
Approximation bounds of three-layered ne
β
Noboru Murata
π
Article
π
1996
π
John Wiley and Sons
π
English
β 642 KB