Recent progress on logic and algorithms for optical neural networks (ONN)

Abstract

Summary form only given. The need for optical implementations of neural networks is rooted in a number of obvious shortcomings of digital/electronic techniques. The author aims at giving at broad outline, in quite technical terms, of the state of the art in optical implementations. Special attention is paid to various types of logics, naturally attached to parallel data transportation by light. Algorithms built on these logics are presented through a number of examples the advantages, as well as disadvantages, of optical implementations are elucidated.