Share Email Print

Proceedings Paper

Photorefractive optical neural network
Author(s): Bernard H. Soffer; Yuri Owechko; Gilmore J. Dunning
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

This paper discusses holographic neumi netwrk architectures in which the connection weights between neurons are implemented as gratings in a photorefrtive crystal. In particular, it will discuss the self-pumped optical neural network (SPONN), which is a fineed optical architecture which features massive parallelism and a much greater degree of interconnectivity than bus-oriented or hypercube electronic architectures.Connecüons between neurons are implemented as sets of angularly and spatially multiplexed volume phase gratings. SPONN is potentially capable of implementing neural networks consisting of 1O1O6 neurons with 109.1010 interconnections. The mapping of neural network models onto the architecture occurs naturally without the need for multiplexing neurons or dealing with the contention, muting, and communication bottleneck problems of electronic parallel computers. This simplifies the programming of the optical system.

Paper Details

Date Published: 1 July 1990
PDF: 2 pages
Proc. SPIE 1319, Optics in Complex Systems, (1 July 1990); doi: 10.1117/12.22274
Show Author Affiliations
Bernard H. Soffer, Hughes Research Labs. (United States)
Yuri Owechko, Hughes Aircraft Co. (United States)
Gilmore J. Dunning, Hughes Research Labs. (United States)

Published in SPIE Proceedings Vol. 1319:
Optics in Complex Systems
F. Lanzl; H.-J. Preuss; G. Weigelt, Editor(s)

© SPIE. Terms of Use
Back to Top