Complex dynamics and applications of memristive devices and systems

Abstract

Due to a complex dynamic behaviour and ability to imitate the important functions of biological synapses and neurons, memristive devices and systems allow not only the hardware implementation of artificial neural networks, but also a qualitative breakthrough towards the symbiosis of artificial electronic circuits and living biological systems in order to solve the urgent problems of robotics, artificial intelligence and medicine. In this report, the essence of complex dynamics of memristive devices is explained through a number of valuable effects hidden in the generalized definition of memristor providing various applications based on synaptic plasticity models and rules as well as neuronal functionality. Two different approaches for interfacing memristive systems and biological neural networks in vitro and in vivo are discussed based on a perceptron with an array of programmable memristive weights or based on memristive stochastic plasticity and neural synchrony as part of the brain-like spiking architectures integrated with CMOS technology.