Advances in Neuromorphic Computing Devices: Insights on Both Conventional and Unconventional Architectures.

Abstract

Neuromorphic circuits and devices have been introduced in the last decades as elements of a key strategy for developing of new paradigms of computation, inspired by the intent to mimic elementary neuron structure and biological mechanisms, for the overcoming of energy and timeconsuming bottlenecks achieved by digital computing (DC) technologies. Although the term "neuromorphic" is in common use, its meaning is often misunderstood and indistinctly associated with many different technologies, based on both conventional and unconventional electronic components and architectures. Here an overview of the different technological strategies used for developing neuromorphic computing systems is proposed, with an insight on the neuromorphic features they implement and a special focus on the technological strategies and patents that exploit unconventional computing paradigms.