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.
© 1988 Optical Society of America
R. Jackson Marks II, Les E. Atlas, Seho Oh, and Kwan F. Cheung, "Optical-processor architectures for alternating-projection neural networks," Opt. Lett. 13, 533-535 (1988)