Nano device fabrication for in-memory and in-sensor reservoir computing

Abstract

Highlights Reservoir computing (RC), with its smaller network size and straightforward training process, has become a popular machine learning algorithm in the current landscape. Nano-memristors, characterized by their high integration density and the ability to achieve storage and computation in a unified manner, are regarded as promising devices to accelerate machine learning. The aim of current study is to present the current applications of in-memory and in-sensor RC based on nano-memristors and to outline the potential developments for the next steps.