Associative Memory and Pattern Recognition of Fractal and Hopfield Neural Networks
返回论文页
|更新时间:2023-12-11
|
Associative Memory and Pattern Recognition of Fractal and Hopfield Neural Networks
Acta Scientiarum Naturalium Universitatis SunYatseniVol. 35, Issue S2, Pages: 150-154(1996)
作者机构:
中山大学无线电电子学系
作者简介:
基金信息:
DOI:
CLC:
Published:1996,
Published Online:25 December 1996,
扫 描 看 全 文
Li Xuechun, Ma Zhengming. Associative Memory and Pattern Recognition of Fractal and Hopfield Neural Networks. [J]. Acta Scientiarum Naturalium Universitatis SunYatseni 35(S2):150-154(1996)
DOI:
Li Xuechun, Ma Zhengming. Associative Memory and Pattern Recognition of Fractal and Hopfield Neural Networks. [J]. Acta Scientiarum Naturalium Universitatis SunYatseni 35(S2):150-154(1996)DOI:
Associative Memory and Pattern Recognition of Fractal and Hopfield Neural Networks
For the application of neural networks on information proceeding and model recognition
research and comparition were carried out on HNN and FNN. This paper proposed the construction method of subpatterns of FNN. New samples will be the stable states if they are composed of subpatterns of stored samples. (Make sure that there are no conflict between the value of internet neurons) Sufficient condition were proposed for samples safe storage
stable points and associated memory. The results of research and experiment show that compared with HNN
FNN has more potentialities and superiority on information storage and fast converging to stable point.