14898 search results - page 162 / 2980 » Design of Neural Networks |
71
Voted
SBRN
2000
IEEE
15 years 2 months ago
2000
IEEE
One of the main obstacles to the widespread use of artijcial neural networks is the difJiculty of adequately define valuesfor their free parameters. This article discusses how Rad...
78
Voted
IJON
2008
14 years 10 months ago
2008
The term neural network evolution usually refers to network topology evolution leaving the network's parameters to be trained using conventional algorithms. In this paper we ...
IPPS
2006
IEEE
15 years 4 months ago
2006
IEEE
This work presents a new approach to solve the location management problem by using the location areas approach. A combination of a genetic algorithm and the Hopfield neural netwo...
85
Voted
NN
2000
Springer
14 years 10 months ago
2000
Springer
In Lp-spaces with p [1, ) there exists a best approximation mapping to the set of functions computable by Heaviside perceptron networks with n hidden units; however for p (1, ) ...
71
Voted
GECCO
2005
Springer
15 years 3 months ago
2005
Springer
Recurrent neural networks are theoretically capable of learning complex temporal sequences, but training them through gradient-descent is too slow and unstable for practical use i...