Title :
Hidden control neural network
Author_Institution :
INFOCOM Dept., Rome Univ., Italy
fDate :
3/1/1994 12:00:00 AM
Abstract :
The hidden control neural network is a multilayer perceptron operating as a nonlinear predictor, driven simultaneously by the available signal and by a further signal to be determined during training, denoted as a hidden control. The possibility of using HCNNs in more general contexts has not been made evident. The authors wonder if the peculiar scheme of the HCNN is applicable in the general case. Moreover, in the affirmative case they are interested to know the relation between the predictor, the control signal and the underlying system
Keywords :
feedforward neural nets; learning (artificial intelligence); HCNNs; control signal; hidden control neural network; multilayer perceptron; nonlinear predictor; training; Design methodology; Filter bank; Finite impulse response filter; Mirrors; Multilayer perceptrons; Neural networks; Nonlinear control systems; Power system modeling; Robustness; Speech synthesis;
Journal_Title :
Circuits and Systems II: Analog and Digital Signal Processing, IEEE Transactions on
