Title :
A quick learning rule to expand stable attraction basins in autoassociative neural networks
Author :
Qingshan, Zhou ; Guoxiang, Zhou ; JianDong, Hu
Author_Institution :
Dept. of Telecommun. Eng., Beijing Univ. of Posts & Telecommun., China
Abstract :
In this paper, a quick repeated learning rule, which is based on the Hebb rule and the Hamming distance distribution of the pattern set to be learned, is studied. With the help of the proposed learning rule, not only can the learned patterns be addressable, but an attraction basin with a predetermined radius is established for each attractor
Keywords :
Hebbian learning; content-addressable storage; neural nets; Hamming distance distribution; Hebb rule; autoassociative neural networks; quick learning rule; radius of attraction; stable attraction basins; Electronic mail; Equations; Hamming distance; Hidden Markov models; Intelligent networks; Magnesium compounds; Neural networks; Stability;
Conference_Titel :
Neural Networks, 1995. Proceedings., IEEE International Conference on
Conference_Location :
Perth, WA
Print_ISBN :
0-7803-2768-3
DOI :
10.1109/ICNN.1995.488974