Share Email Print

Proceedings Paper

Integrated Optical Implementation Of The Hopfield Neural Network Model
Author(s): C. DeCusatis; P. Das; D . M. Litynski
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

An integrated optics implementation of content addressable associative memory based on the Hopfield Neural Network Model is proposed. The device uses frequency multiplexed surface acoustic waves (FM-SAW) to input the synaptic matrix, which may be stored as a volume hologram using photorefractive memory. Stored information may then be recovered in real time by applying a partial or distorted input to the system. This input is presented in the form of coherent optical pulses, which may be modulated by an external CCD camera. Phase conjugate mirrors are used to obtain the nonlinear thresholding inherent in the Hopfield model, and to provide feedback and gain.

Paper Details

Date Published: 8 February 1988
PDF: 7 pages
Proc. SPIE 0963, Optical Computing '88, (8 February 1988); doi: 10.1117/12.947929
Show Author Affiliations
C. DeCusatis, Rensselaer Polytechnic Institute (United States)
P. Das, Rensselaer Polytechnic Institute (United States)
D . M. Litynski, United States Military Academy (United States)

Published in SPIE Proceedings Vol. 0963:
Optical Computing '88
Pierre H. Chavel; Joseph W. Goodman; Gerard Roblin, Editor(s)

© SPIE. Terms of Use
Back to Top