OPTICAL SELF-ROUTING COMPLEX NEURAL-NETWORK USING MIXED NEGABINARY NUMBERS

被引:0
|
作者
LIU, L
LI, GQ
SHAO, L
YIN, YZ
机构
来源
OPTICAL COMPUTING | 1995年 / 139卷
关键词
D O I
暂无
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
A general neural unit with only a complex-valued matrix depicting the integrated synaptic connection and lateral interaction is proposed. Various connections can be self-routed by the routing code contained in the input. Based on the mixed negabinary system, an inner-product algorithm is developed to calculate the complex matrix-vector multiplication without carries, signs, and decimal points. An incoherent optical correlator with spatial coding is suggested to execute the algorithm in parallel.
引用
收藏
页码:459 / 462
页数:4
相关论文
共 50 条