High capacity content addressable memory with mixed order hyper networks

Abstract

A mixed order hyper network (MOHN) is a neural network in which weights can connect any number of neurons, rather than the usual two. MOHNs can be used as content addressable memories (CAMs) with higher capacity than standard Hopfield networks. They can also be used for regression learning of functions in f: {−1, 1}n→ ℝ in which the turning points are equivalent to memories in a CAM. This paper presents a number of methods for learning an energy function from data that can act as either a CAM or a regression model and presents the advantages of using such an approach.