Title :
Neural network design through embedded representations and evolutionary computation
Author :
Lee, Cin-Young ; Antonsson, Erik K.
fDate :
6/24/1905 12:00:00 AM
Abstract :
An evolutionary algorithm is developed to address simultaneous learning of weights and connection topologies for a feedforward neural network. The algorithm is dependent on an embedded representation of the network, in which architecture specification is determined from the interactions of the embedded nodes. Preliminary results are presented and discussed
Keywords :
evolutionary computation; feedforward neural nets; learning (artificial intelligence); neural net architecture; architecture specification; connection topologies learning; embedded representations; evolutionary algorithm; evolutionary computation; feedforward neural network; neural network design; weights learning; Artificial neural networks; Backpropagation algorithms; Biological cells; Computer architecture; Evolutionary computation; Feedforward neural networks; Genetic mutations; Network topology; Neural networks; Stochastic processes;
Conference_Titel :
Neural Networks, 2002. IJCNN '02. Proceedings of the 2002 International Joint Conference on
Conference_Location :
Honolulu, HI
Print_ISBN :
0-7803-7278-6
DOI :
10.1109/IJCNN.2002.1005602
