A new mechanical approach to handle generalized Hopfield neural networks (Q2182875)

One starts with a large presentation of results and points of view in the domain of neural networks from the statistical mechanics perspective. Some of the achievements of the first author are reviewed. In the second section of the article, machine learning and neural networks are introduced, one considers for learning the restricted Boltzmann machines (RBM) and for retrieval the generalized Boltzmann models (GHM). The third section is devoted to the mechanical formulation of the classical Hopfield networks and the fourth section to the mechanical formulation of the relativistic Hopfield networks. In the fifth section, Monte Carlo simulations to these network's performances are shown and in detail analyzed. The sixth section contains conclusions and trends for further developments.