Year: 1991
Type: dissertation
Abstract: An optical quadratic neural network (OQNN) utilizing \nfour-wave mixing in barium titanate (BaTi03 ) has been investigated \nand developed. This network implements a feedback \nloop using a charged-coupled device (CCD) camera, a Macintosh \nII microcomputer, two monochrome liquid crystal televisions \n(LCTVs), and various optical elements. For training, the \nnetwork employs the supervised quadratic perceptron algorithm \nto associate binary-valued input vectors with specified \ntarget vectors. Using a spatial multiplexing scheme for two \nbipolar neurons, the quadratic network was able to associate \nan input vector with various target vectors. In addition, \nthe network was reconfigured to discriminate two different \ninput patterns simultaneously. Both analytical and experimental \nresults are presented.
Author: Alex V. Huynh
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2788750731