Optical implementation of a multi-layer neural network by SLR and MSLM

IEEE 1991 International Conference on Systems Engineering Pub Date : 1900-01-01 DOI:10.1109/ICSYSE.1991.161112

A. McAulay, X. Xu, J. Wang

引用次数: 1

Abstract

An optical implementation of a multilayer neural network by spatial light rebroadcaster (SLR) and microchannel spatial light modulation (MSLM) is presented. In the system, the SLR is used as a buffer, which is an optical storage material that can be written with an Ar laser and read with an IR laser. The MSLM has a very sharp low-level boundary, and is used for thresholding. Since SLR and MSLM are used to handle data and thresholding instead of using a computer, a high-speed processing system can be achieved. The speed limitation of the system is the response times of the SLR and MLSM.<>

查看原文

微信好友朋友圈 QQ好友复制链接

本刊更多论文

一种多层神经网络的光学实现

提出了一种利用空间光中继器(SLR)和微通道空间光调制(MSLM)实现多层神经网络的方法。在该系统中，单反被用作缓冲器，它是一种光存储材料，可以用Ar激光器写入，用IR激光器读取。MSLM具有非常清晰的低水平边界，并用于阈值设定。由于单反和MSLM是用来处理数据和阈值，而不是使用计算机，因此可以实现高速处理系统。系统的速度限制是单反和MLSM的响应时间。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文去求助

来源期刊

IEEE 1991 International Conference on Systems Engineering

自引率

0.00%

发文量

期刊最新文献

Modelling and simulation of a heat exchanger Multidisciplinary modeling and design of a space system Robust stabilization of an aero-elastic system Inductive character learning and classification with genetic algorithms Determining bus arbitration policies and data transfer techniques for multiprocessor systems