Abstract
Artificial neural networks are becoming more common because they can do tasks that are difficult, or even impossible, for a common numerical processor. Their strength, compared with normal numerical processors, resides in the use of a very large number of simple processing elements that are highly interconnected. A typical artificial neural network mimics some behavioral characteristics of a brain, for example it is trained not programmed.
© 1996 Optical Society of America
PDF ArticleMore Like This
George J. M. Aitken
AThA.4 Adaptive Optics (AO) 1996
Yuri Owechko
MD4 Optical Computing (IP) 1989
Shuo Zheng, Min Yang, Guofeng Yan, Yanjun Zhu, Hua Zhang, Chaonan Yao, Yuchen Shao, and Jian Wang
SpTu1J.6 Signal Processing in Photonic Communications (SPPCom) 2022