ISING MODEL AND NEURAL NETWORKS

Volodymyr Tkachuk (Personal webpage)

The talk is the short review of the relation of Ising model with neutral networks. The first model which describes the behavior of neural networks was introduced by J.J. Hopfield. This model was proposed for explanation of the mechanism of associative memory. Formally the Hopfield model is identical to the Ising model of spin glass. Therefore, the methods of statistical mechanics can be used for studies of the properties of the Hopfield model.