Materials and devices for spin-based neuromorphic computing – application of spintronics to fzzy logic

Abstract

Research in the area of Spintronics is gradually gaining popularity in recent times due to the stagnation of Moore’s Law and the electronics industry. The research conducted in this paper aims to adopt Spin-based Domain Wall devices for application in Neuromorphic Computing where methods for low-power consumption Artificial Intelligence and Machine Learning is explored. The mechanism of Spin Orbit Torque will be explored in this paper, exploring a novel design of a dual-tungsten layer which is able to achieve low power requirement for movement of domain walls. Also, a new logic system called Fuzzy Logic is in the process of being tested for implementation through spin-based domain wall devices. The device focuses on the idea of geometrical domain wall pinning and also multi-resistive states that is necessary for this fuzzy logic system over the classical Boolean logic. As a proof of concept, the movement of a domain wall has shown to be controlled and domain wall devices for Fuzzy Logic implementation has been shown to be possible through said devices. This will assist in the agenda of promoting lower power consumption Artificial Intelligence by mimicking a synaptic learning as stated in Neuromorphic Computing.