Shape‐Dependent Multi‐Weight Magnetic Artificial Synapses for Neuromorphic Computing
Abstract
Abstract In neuromorphic computing, artificial synapses provide a multi‐weight (MW) conductance state that is set based on inputs from neurons, analogous to the brain. Herein, artificial synapses based on magnetic materials that use a magnetic tunnel junction (MTJ) and a magnetic domain wall (DW) are explored. By fabricating lithographic notches in a DW track underneath a single MTJ, 3–5 stable resistance states that can be repeatably controlled electrically using spin‐orbit torque are achieved. The effect of geometry on the synapse behavior is explored, showing that a trapezoidal device has asymmetric weight updates with high controllability, while a rectangular device has higher stochasticity, but with stable resistance levels. The device data is input into neuromorphic computing simulators to show the usefulness of application‐specific synaptic functions. Implementing an artificial neural network (NN) applied to streamed Fashion‐MNIST data, the trapezoidal magnetic synapse can be used as a metaplastic function for efficient online learning. Implementing a convolutional NN for CIFAR‐100 image recognition, the rectangular magnetic synapse achieves near‐ideal inference accuracy, due to the stability of its resistance levels. This work shows MW magnetic synapses are a feasible technology for neuromorphic computing and provides design guidelines for emerging artificial synapse technologies.
- Authors:
-
- Electrical and Computer Engineering Department University of Texas at Austin Austin TX 78712 USA
- Applied Materials Santa Clara CA 95054 USA
- Sandia National Laboratories Albuquerque NM 87123 USA
- Electrical and Computer Engineering Department University of Texas at Dallas Richardson TX 75080 USA
- Publication Date:
- Sponsoring Org.:
- USDOE
- OSTI Identifier:
- 1995870
- Resource Type:
- Publisher's Accepted Manuscript
- Journal Name:
- Advanced Electronic Materials
- Additional Journal Information:
- Journal Name: Advanced Electronic Materials Journal Volume: 8 Journal Issue: 12; Journal ID: ISSN 2199-160X
- Publisher:
- Wiley Blackwell (John Wiley & Sons)
- Country of Publication:
- United States
- Language:
- English
Citation Formats
Leonard, Thomas, Liu, Samuel, Alamdar, Mahshid, Jin, Harrison, Cui, Can, Akinola, Otitoaleke G., Xue, Lin, Xiao, T. Patrick, Friedman, Joseph S., Marinella, Matthew J., Bennett, Christopher H., and Incorvia, Jean Anne C. Shape‐Dependent Multi‐Weight Magnetic Artificial Synapses for Neuromorphic Computing. United States: N. p., 2022.
Web. doi:10.1002/aelm.202200563.
Leonard, Thomas, Liu, Samuel, Alamdar, Mahshid, Jin, Harrison, Cui, Can, Akinola, Otitoaleke G., Xue, Lin, Xiao, T. Patrick, Friedman, Joseph S., Marinella, Matthew J., Bennett, Christopher H., & Incorvia, Jean Anne C. Shape‐Dependent Multi‐Weight Magnetic Artificial Synapses for Neuromorphic Computing. United States. https://doi.org/10.1002/aelm.202200563
Leonard, Thomas, Liu, Samuel, Alamdar, Mahshid, Jin, Harrison, Cui, Can, Akinola, Otitoaleke G., Xue, Lin, Xiao, T. Patrick, Friedman, Joseph S., Marinella, Matthew J., Bennett, Christopher H., and Incorvia, Jean Anne C. Sun .
"Shape‐Dependent Multi‐Weight Magnetic Artificial Synapses for Neuromorphic Computing". United States. https://doi.org/10.1002/aelm.202200563.
@article{osti_1995870,
title = {Shape‐Dependent Multi‐Weight Magnetic Artificial Synapses for Neuromorphic Computing},
author = {Leonard, Thomas and Liu, Samuel and Alamdar, Mahshid and Jin, Harrison and Cui, Can and Akinola, Otitoaleke G. and Xue, Lin and Xiao, T. Patrick and Friedman, Joseph S. and Marinella, Matthew J. and Bennett, Christopher H. and Incorvia, Jean Anne C.},
abstractNote = {Abstract In neuromorphic computing, artificial synapses provide a multi‐weight (MW) conductance state that is set based on inputs from neurons, analogous to the brain. Herein, artificial synapses based on magnetic materials that use a magnetic tunnel junction (MTJ) and a magnetic domain wall (DW) are explored. By fabricating lithographic notches in a DW track underneath a single MTJ, 3–5 stable resistance states that can be repeatably controlled electrically using spin‐orbit torque are achieved. The effect of geometry on the synapse behavior is explored, showing that a trapezoidal device has asymmetric weight updates with high controllability, while a rectangular device has higher stochasticity, but with stable resistance levels. The device data is input into neuromorphic computing simulators to show the usefulness of application‐specific synaptic functions. Implementing an artificial neural network (NN) applied to streamed Fashion‐MNIST data, the trapezoidal magnetic synapse can be used as a metaplastic function for efficient online learning. Implementing a convolutional NN for CIFAR‐100 image recognition, the rectangular magnetic synapse achieves near‐ideal inference accuracy, due to the stability of its resistance levels. This work shows MW magnetic synapses are a feasible technology for neuromorphic computing and provides design guidelines for emerging artificial synapse technologies.},
doi = {10.1002/aelm.202200563},
journal = {Advanced Electronic Materials},
number = 12,
volume = 8,
place = {United States},
year = {Sun Sep 11 00:00:00 EDT 2022},
month = {Sun Sep 11 00:00:00 EDT 2022}
}
https://doi.org/10.1002/aelm.202200563
Works referenced in this record:
Spin-dependent tunneling conductance of sandwiches
journal, January 2001
- Butler, W. H.; Zhang, X. -G.; Schulthess, T. C.
- Physical Review B, Vol. 63, Issue 5
The multifarious hippocampal mossy fiber pathway: a review
journal, June 2000
- Henze, D. A.; Urban, N. N.; Barrionuevo, G.
- Neuroscience, Vol. 98, Issue 3
Nanoscale domain wall devices with magnetic tunnel junction read and write
journal, June 2021
- Raymenants, E.; Bultynck, O.; Wan, D.
- Nature Electronics, Vol. 4, Issue 6
Magnetic Domain Wall Based Synaptic and Activation Function Generator for Neuromorphic Accelerators
journal, December 2019
- Siddiqui, Saima A.; Dutta, Sumit; Tang, Astera
- Nano Letters, Vol. 20, Issue 2
Domain wall-magnetic tunnel junction spin–orbit torque devices and circuits for in-memory computing
journal, March 2021
- Alamdar, Mahshid; Leonard, Thomas; Cui, Can
- Applied Physics Letters, Vol. 118, Issue 11
Spin-torque-driven MTJs with extended free layer for logic applications
journal, June 2018
- Raymenants, Eline; Vaysset, Adrien; Wan, Danny
- Journal of Physics D: Applied Physics, Vol. 51, Issue 27
Probabilistic Inference in General Graphical Models through Sampling in Stochastic Networks of Spiking Neurons
journal, December 2011
- Pecevski, Dejan; Buesing, Lars; Maass, Wolfgang
- PLoS Computational Biology, Vol. 7, Issue 12
Area and Energy Efficient Joint 2T SOT-MRAM-Based on Diffusion Region Sharing With Adjacent Cells
journal, March 2022
- Jang, Yunho; Park, Jongsun
- IEEE Transactions on Circuits and Systems II: Express Briefs, Vol. 69, Issue 3
Magnetic Tunnel Junction Based Long-Term Short-Term Stochastic Synapse for a Spiking Neural Network with On-Chip STDP Learning
journal, July 2016
- Srinivasan, Gopalakrishnan; Sengupta, Abhronil; Roy, Kaushik
- Scientific Reports, Vol. 6, Issue 1
Active particle feedback control with a single-shot detection convolutional neural network
journal, July 2020
- Fränzl, Martin; Cichos, Frank
- Scientific Reports, Vol. 10, Issue 1
Impact of Non-Ideal Characteristics of Resistive Synaptic Devices on Implementing Convolutional Neural Networks
journal, September 2019
- Sun, Xiaoyu; Yu, Shimeng
- IEEE Journal on Emerging and Selected Topics in Circuits and Systems, Vol. 9, Issue 3
Hippocampal Reactivation of Random Trajectories Resembling Brownian Diffusion
journal, April 2019
- Stella, Federico; Baracskay, Peter; O’Neill, Joseph
- Neuron, Vol. 102, Issue 2
Three-terminal magnetic tunnel junction synapse circuits showing spike-timing-dependent plasticity
journal, September 2019
- Akinola, Otitoaleke; Hu, Xuan; Bennett, Christopher H.
- Journal of Physics D: Applied Physics, Vol. 52, Issue 49
Bioinspired Programming of Memory Devices for Implementing an Inference Engine
journal, August 2015
- Querlioz, Damien; Bichler, Olivier; Vincent, Adrien Francis
- Proceedings of the IEEE, Vol. 103, Issue 8
Synaptic metaplasticity in binarized neural networks
journal, May 2021
- Laborieux, Axel; Ernoult, Maxence; Hirtzlin, Tifenn
- Nature Communications, Vol. 12, Issue 1
Neural algorithms and computing beyond Moore's law
journal, March 2019
- Aimone, James B.
- Communications of the ACM, Vol. 62, Issue 4
Stochastic Spin-Orbit Torque Devices as Elements for Bayesian Inference
journal, October 2017
- Shim, Yong; Chen, Shuhan; Sengupta, Abhronil
- Scientific Reports, Vol. 7, Issue 1
Elementary depinning processes of magnetic domain walls under fields and currents
journal, October 2014
- Nguyen, V. D.; Torres, W. Savero; Laczkowski, P.
- Scientific Reports, Vol. 4, Issue 1
Magnetic domain wall neuron with lateral inhibition
journal, October 2018
- Hassan, Naimul; Hu, Xuan; Jiang-Wei, Lucian
- Journal of Applied Physics, Vol. 124, Issue 15
Multiscale Co-Design Analysis of Energy, Latency, Area, and Accuracy of a ReRAM Analog Neural Training Accelerator
journal, March 2018
- Marinella, Matthew J.; Agarwal, Sapan; Hsia, Alexander
- IEEE Journal on Emerging and Selected Topics in Circuits and Systems, Vol. 8, Issue 1
Area-Efficient SOT-MRAM With a Schottky Diode
journal, August 2016
- Seo, Yeongkyo; Kwon, Kon-Woo; Roy, Kaushik
- IEEE Electron Device Letters, Vol. 37, Issue 8
Efficacy and Stability of Quantal GABA Release at a Hippocampal Interneuron–Principal Neuron Synapse
journal, August 2000
- Kraushaar, Udo; Jonas, Peter
- The Journal of Neuroscience, Vol. 20, Issue 15
Maximized lateral inhibition in paired magnetic domain wall racetracks for neuromorphic computing
journal, May 2020
- Cui, Can; Akinola, Otitoaleke G.; Hassan, Naimul
- Nanotechnology, Vol. 31, Issue 29
Spin-Transfer Torque Magnetic Memory as a Stochastic Memristive Synapse for Neuromorphic Systems
journal, April 2015
- Vincent, Adrien F.; Larroque, Jerome; Locatelli, Nicolas
- IEEE Transactions on Biomedical Circuits and Systems, Vol. 9, Issue 2
Proposal for an All-Spin Artificial Neural Network: Emulating Neural and Synaptic Functionalities Through Domain Wall Motion in Ferromagnets
journal, December 2016
- Sengupta, Abhronil; Shim, Yong; Roy, Kaushik
- IEEE Transactions on Biomedical Circuits and Systems, Vol. 10, Issue 6
Graded-Anisotropy-Induced Magnetic Domain Wall Drift for an Artificial Spintronic Leaky Integrate-and-Fire Neuron
journal, June 2019
- Brigner, Wesley H.; Hu, Xuan; Hassan, Naimul
- IEEE Journal on Exploratory Solid-State Computational Devices and Circuits, Vol. 5, Issue 1
Magnetic Tunnel Junction Mimics Stochastic Cortical Spiking Neurons
journal, July 2016
- Sengupta, Abhronil; Panda, Priyadarshini; Wijesinghe, Parami
- Scientific Reports, Vol. 6, Issue 1
A domain wall-magnetic tunnel junction artificial synapse with notched geometry for accurate and efficient training of deep neural networks
journal, May 2021
- Liu, Samuel; Xiao, T. Patrick; Cui, Can
- Applied Physics Letters, Vol. 118, Issue 20
Intrinsic optimization using stochastic nanomagnets
journal, March 2017
- Sutton, Brian; Camsari, Kerem Yunus; Behin-Aein, Behtash
- Scientific Reports, Vol. 7, Issue 1
Asymmetric ground state spin configuration of transverse domain wall on symmetrically notched ferromagnetic nanowires
journal, July 2010
- Djuhana, Dede; Piao, Hong-Guang; Lee, Sang-Hyuk
- Applied Physics Letters, Vol. 97, Issue 2
Research progress on solutions to the sneak path issue in memristor crossbar arrays
journal, January 2020
- Shi, Lingyun; Zheng, Guohao; Tian, Bobo
- Nanoscale Advances, Vol. 2, Issue 5