Introduction

Hopfield networks were introduced in the early 1980's by John Hopfield, a well-known physicist. They originated as neurophysiologically-motivated models for content-addressable memories (CAM).

Content-addressable memories differ from conventional computer memories in that information is retrieved from a CAM by specifying either a part of the desired information or some properties of it, instead of specifying the location in the memory where the information is stored. This is reminiscent of the way in which human memory works; for example, we can recall a long word beginning with the letter A meaning capable of using both hands with equal facility.