Formation of Weighting Coefficients in an Artificial Neural Network Based on the Memristive Effect in Metal–Oxide–Metal Nanostructures

Abstract

An approach to formation and training of an artificial neural network (ANN) based on thin-film memristive metal–oxide–metal nanostructures, which exhibit the effect of bipolar resistive switching, has been proposed. An experimental electric circuit of a small-sized ANN (a two-layer perceptron with 32 memristive elements) has been constructed. An algorithm for formation of weighting coefficients (ANN training), which takes into account probable spread of technological parameters of memristive structures has been developed.