Neuro-Inspired Computing with Resistive Switching Devices [Guest Editorial

Abstract

The articles in this special section examine the current progress in neuro-inspired computing with resistive switching devices, and they share their views on future research directions in this area. Inspired by the biological neural networks giving rise to human intelligence, artificial neural networks have achieved revolutionary milestones in computer vision and speech recognition, extending their use for numerous applications ranging from autonomous driving to disease detection. However, computation with large-scale artificial neural networks requires substantial power and time due to iterative updates of a massive number of network parameters and the constant transfer of data between memory and a processor, which is a major drawback arising from von Neumann architecture. Recent progress in nanoscale resistive-switching devices has opened up new opportunities toward the development of low-power neuro-inspired hardware, which can empower high-density on-chip data storage and in-memory computing.