Spintronic Integrate-Fire-Reset Neuron with Stochasticity for Neuromorphic Computing

• DOI: 10.1021/acs.nanolett.2c02409
• Publication Date: 2022-10-19T17:11:26Z
• Authors: Qu Yang; Rahul Mishra; Yunuo Cen; Guoyi Shi; Raghav Sharma; Xuanyao Fong; Hyunsoo Yang
• Keywords: Neuromorphic engineering; Memristor

Spintronics has been recently extended to neuromorphic computing because of its energy efficiency and scalability. However, a biorealistic spintronic neuron with probabilistic "spiking" spontaneous reset functionality not demonstrated yet. Here, we propose device based on the heavy metal (HM)/ferromagnet (FM)/antiferromagnet (AFM) spin–orbit torque (SOT) heterostructure. The can autoreset itself after firing due exchange bias AFM. process is inherently stochastic competition between SOT AFM pinning effects. We also implement restricted Boltzmann machine (RBM) integration multilayer perceptron (SI-MLP) using our proposed neuron. Despite bit-width limitation, model achieve an accuracy 97.38% in pattern recognition, which even higher than baseline (96.47%). Our results offer solution emulate biologically realistic spiking neurons.