发明名称 |
Chaotic recurrent neural net and learning procedure |
摘要 |
The chaotic recurrent neural net consists of N chaotic neural nets, which each receive an external input ai(t) and the output of N-1 chaotic neural nets among the N chaotic neural nets, to perform an operation given by a dynamic mathematical equation. This equation is tau 1 Yi(t+1) = EPSILON Wij Zi (t) + kYi (t) - alpha Xi(t) where Xi(t) = fi Yi(t). fi is the output function of a neuron. Wij is a synapse connection coefficient of the feedback input from the jth neuron to the ith neuron and Xi(t) is the output of the ith neuron at time t. tau i, a and k are a time delay constant, a non-negative parameter and a refractory time attenuation constant. Zi(t) represents Xi(t) when i belongs to the neuron group I, or ai(t) when i belongs to the external input group E.
|
申请公布号 |
DE19504664(A1) |
申请公布日期 |
1995.08.17 |
申请号 |
DE19951004664 |
申请日期 |
1995.02.13 |
申请人 |
GOLD STAR ELECTRON CO., LTD., CHUNGCHEONGBUK, KR |
发明人 |
CHUNG, HO-SUN, TAEGU, KR;TAK, HYE-YOUNG, TAEGU, KR |
分类号 |
G06G7/60;G06F15/18;G06N3/04;(IPC1-7):G06F15/18 |
主分类号 |
G06G7/60 |
代理机构 |
|
代理人 |
|
主权项 |
|
地址 |
|