Title :
The refracted bifurcating neural network
Author :
Hu, X.L. ; Zhang, Y.T.
Author_Institution :
Dept. of Electron. Eng., Chinese Univ. of Hong Kong, Shatin, China
Abstract :
In this paper, a refracted bifurcating neural network (RBNN) is proposed for associative memory study, based on the understanding of refractory mechanism in neuron firing and a bifurcating neural network (BNN). The absolute refractory period is introduced into the individual neuron model of the BNN to construct RBNN with an attempt to improve the convergent speed of the original network. The results of this study suggest that introduction of refractory period into BNN can achieve comparable pattern recalling rates as those for the original network. By adjusting suitably the value of the absolute refractory period and the amplitude of relaxation level, RBNN can achieve faster convergence than that of BNN for similar pattern recalling rates.
Keywords :
brain models; cellular biophysics; neural nets; neurophysiology; absolute refractory period; associative memory study; bifurcating neural network; comparable pattern recalling rates; convergent speed; individual neuron model; neuron firing; refracted bifurcating neural network; refractory mechanism; relaxation level amplitude; Associative memory; Bifurcation; Biological neural networks; Biological system modeling; Chaos; Convergence; Hopfield neural networks; Neural networks; Neurons; Oscillators;
Conference_Titel :
Engineering in Medicine and Biology, 2002. 24th Annual Conference and the Annual Fall Meeting of the Biomedical Engineering Society EMBS/BMES Conference, 2002. Proceedings of the Second Joint
Print_ISBN :
0-7803-7612-9
DOI :
10.1109/IEMBS.2002.1053140