Abstract
A general neural minimization algorithm which can be applied to arbitrary types of polynomial energy functions is presented. The algorithm can be operated in a synchronous as well as asynchronous way. The synchronous algorithm can be implemented by highly parallel optical systems.
© 1989 Optical Society of America
PDF ArticleMore Like This
Convergence property of quadratic neural network
Shaoping Bian, Kebin Xu, and Jing Hong
THT26 OSA Annual Meeting (FIO) 1989
Minimizing errors in analog neural networks
H. John Caulfield
THJ7 OSA Annual Meeting (FIO) 1988
An Inter-Pattern Association Neural Network Model and its Optical Implementation
Taiwei Lu, Sudong Wu, Xin Xu, and Francis T. S. Yu
PD4 Optical Computing (IP) 1989