Optical-processor architectures for alternating-projection neural networks

R. Jackson Marks, Les E. Atlas, Seho Oh, Kwan F. Cheung
1988 Optics Letters  
Optical-processor architectures for various forms of the alternating-projection neural network are considered. Required iteration is performed by passive optical feedback. No electronics or slow optics (e.g., phase conjugators) are used in the feedback path. The processor can be taught a new training vector by viewing it only once. If the desired outputs are trained to be either ±1, then the network can be configured to converge in one iteration. Optical neural network architectures have been
more » ... ectures have been proposed by a number of researchers.--5 Using the alternating-projection neural network (APNN) algorithm developed elsewhere, 6 -9 we propose two architectures for a corresponding optical implementation. (A contrast of APNN performance to other neural network architectures is in Ref. 9.) Required iteration is ac-
doi:10.1364/ol.13.000533 pmid:19745956 fatcat:hl6qvxusejciljugeuk4kawfce