A linear optimization neural network for associative memory (Q2491006)

This short article investigates new alternative models for neural networks for associative memory. Traditionally, such models were based on optimizing a quadratic function, at the expense of a simple structure and implementation. The authors present an alternative to the quadratic function that is more efficient, based on a linear neural network for associative memory. The main part of the article concentrates in describing this model, which is then followed by a short numerical example and a useful list of references.