用二元光学元件实现分布式存储的神经网络计算

Implementing Dispersed Memory of the Interconnect Matrix of Neural Network by Using a Binary Optical Element

本文介绍了用二元光学元件和分布式存储方式存储Hofield模型神经网络互连矩阵的方法及光学系统,进行了神经网络的计算的实验。

In this paper, a dispersed memory approach is proposed for recording the interconnect matrix of a Hopfield model neural network, in which a binary optical element is employed. A such optical system is established in laboratory for demonstrating the experiment of associate memory.