Premium
Two‐dimensional inner‐product model for auto‐hetero‐associative memory and its new optoelectronic implementation
Author(s) -
Zhang Jiajun,
Zhang Li,
He Anzhi,
Yan Dapeng
Publication year - 1992
Publication title -
microwave and optical technology letters
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.304
H-Index - 76
eISSN - 1098-2760
pISSN - 0895-2477
DOI - 10.1002/mop.4650050710
Subject(s) - content addressable memory , computer science , hopfield network , bidirectional associative memory , encode , coding (social sciences) , electronic engineering , associative property , algorithm , artificial neural network , artificial intelligence , mathematics , engineering , chemistry , biochemistry , statistics , pure mathematics , gene
The associative memory of the two‐dimensional (2‐D) Hopfield model described in terms of inner product is realized by using a multiimaging hybrid system and an incoherent correlation hybrid system, respectively. A new method, in which polarization coding and half wave plate are employed, is proposed to encode the bipolar data in memory masks. Four 8 × 8 images are used in our experiment, and the experimental results agree well with the numerical simulations. On the basis of this, we also present the inner‐product expression of the modified Hopfield model for the heteroassociative memory which can be implemented in the same system described for autoassociation, as long as the first memory mask is changed and the thresholds in the feedback are adjusted accordingly. © 1992 John Wiley & Sons, Inc.