On-chip learning for domain wall synapse based Fully Connected Neural Network

Abstract

Spintronic devices are considered as promising candidates in implementing neuromorphic systems or hardware neural networks, which are expected to perform better than other existing computing systems for certain data classification and regression tasks. In this paper, we have designed a feedforward Fully Connected Neural Network (FCNN) with no hidden layer using spin orbit torque driven domain wall devices as synapses and transistor based analog circuits as neurons. A feedback circuit is also designed using transistors, which at every iteration computes the change in weights of the synapses needed to train the network using Stochastic Gradient Descent (SGD) method. Subsequently it sends write current pulses to the domain wall based synaptic devices which move the domain walls and updates the weights of the synapses. Through a combination of micromagnetic simulations, analog circuit simulations and numerically solving FCNN training equations, we demonstrate "on-chip" training of the designed FCNN on the MNIST database of handwritten digits in this paper. We report the training and test accuracies, energy consumed in the synaptic devices for the training and possible issues with hardware implementation of FCNN that can limit its test accuracy.

Submitted on November 5, 2018 for review in journal

References in corpus (16)

• Deep Learning in Neural Networks: An Overview
• Spin torque switching with the giant spin Hall effect of tantalum
• The design and verification of Mumax3
• Magnetic switching by spin torque from the spin Hall effect
• Current-driven dynamics of chiral ferromagnetic domain walls
• Spin Torque Ferromagnetic Resonance Induced by the Spin Hall Effect
• Neuromorphic computing with multi-memristive synapses
• Vowel recognition with four coupled spin-torque nano-oscillators
• Magnetic skyrmion-based synaptic devices
• Spin Hall torque magnetometry of Dzyaloshinskii domain walls
• Electrical signature of individual magnetic skyrmions in multilayered systems
• Roadmap for Emerging Materials for Spintronic Device Applications
• Proposal for an All-Spin Artificial Neural Network: Emulating Neural and Synaptic Functionalities Through Domain Wall Motion in Ferromagnets
• Hybrid Spintronic-CMOS Spiking Neural Network With On-Chip Learning: Devices, Circuits and Systems
• Determination of spin Hall effect and spin diffusion length of Pt from self-consistent fitting of damping enhancement and inverse spin-orbit torque measurements
• Resonate and Fire Neuron with Fixed Magnetic Skyrmions

Cited by in corpus (5)

• Energy Efficient Learning with Low Resolution Stochastic Domain Wall Synapse Based Deep Neural Networks
• Skyrmion-mediated Nonvolatile Ternary Memory
• Ultra-low Power Domain Wall Device for Spin-based Neuromorphic Computing
• Quantized Non-Volatile Nanomagnetic Synapse based Autoencoder for Efficient Unsupervised Network Anomaly Detection
• Artificial neurons based on antiferromagnetic auto-oscillators as a platform for neuromorphic computing