Let D be a set with a probability measure +, +(D)=1, and let K be a compact subset of L q (D, +), where the infimum is taken over all g n of the form g n = n i=1 a i , i , with arbitrary , i # K and a i # R. It is shown that for f # conv(K \_ (&K )), under some mild restrictions, \ n ( f, K ) C q =
β¦ LIBER β¦
Random matrices and neural networks
β Scribed by Luca, A. ;Ricciardi, L. M. ;Vasudevan, R.
- Publisher
- Springer-Verlag
- Year
- 1970
- Weight
- 499 KB
- Volume
- 6
- Category
- Article
- ISSN
- 0023-5946
No coin nor oath required. For personal study only.
π SIMILAR VOLUMES
Random Approximants and Neural Networks
β
Y. Makovoz
π
Article
π
1996
π
Elsevier Science
π
English
β 492 KB
Brainwashing random asymmetric βneuralβ
β
P.C. McGuire; G.C. Littlewort; J. Rafelski
π
Article
π
1991
π
Elsevier Science
π
English
β 419 KB
Information storage matrices in neural n
β
N. Samardzija
π
Article
π
1990
π
Springer-Verlag
π
English
β 751 KB
M-matrices and global convergence of dis
β
Mauro Forti
π
Article
π
2007
π
John Wiley and Sons
π
English
β 334 KB
## Abstract The paper considers a general class of neural networks possessing discontinuous neuron activations and neuron interconnection matrices belonging to the class of __M__βmatrices or __H__βmatrices. A number of results are established on global __exponential__ convergence of the state and o
4996648 Neural network using random bina
π
Article
π
1991
π
Elsevier Science
π
English
β 69 KB
Univariate fuzzy-random neural network a
β
G.A. Anastassiou
π
Article
π
2004
π
Elsevier Science
π
English
β 954 KB