Domain Wall Motion-Based Dual-Threshold Activation Unit for Low-Power Classification of Non-Linearly Separable Functions

Recently, a great deal of scientific endeavour has been devoted to developing spin-based neuromorphic platforms owing to the ultra-low-power benefits offered by spin devices and the inherent correspondence between spintronic phenomena and the desired neuronal, synaptic behavior. While domain wall motion-based threshold activation unit has previously been demonstrated for neuromorphic circuits, it remains well known that neurons with threshold activation cannot completely learn nonlinearly separable functions. This paper addresses this fundamental limitation by proposing a novel domain wall motion-based dual-threshold activation unit with additional nonlinearity in its function. Furthermore, a new learning algorithm is formulated for a neuron with this activation function. We perform 100 trials of tenfold training and testing of our neural networks on real-world datasets taken from the UCI machine learning repository. On an average, the proposed algorithm achieves <inline-formula><tex-math notation="LaTeX">$\text{1.04}\times - \text{6.54}\times$</tex-math></inline-formula> lower misclassification rate (MCR) than the traditional perceptron learning algorithm. In a circuit-level simulation, the neural networks with the proposed activation unit are observed to outperform the perceptron networks by as much as <inline-formula><tex-math notation="LaTeX">$\text{2.98}\times$</tex-math></inline-formula> MCR. The energy consumption of a neuron having the proposed domain wall motion-based activation unit averages to <inline-formula><tex-math notation="LaTeX">$\text{35 fJ}$</tex-math></inline-formula> approximately.

[1]  H. Ohno,et al.  Current-induced domain wall motion in perpendicularly magnetized CoFeB nanowire , 2011 .

[2]  G. Tatara,et al.  Theory of current-driven domain wall motion: spin transfer versus momentum transfer. , 2004, Physical review letters.

[3]  Kaushik Roy,et al.  AxNN: Energy-efficient neuromorphic systems using approximate computing , 2014, 2014 IEEE/ACM International Symposium on Low Power Electronics and Design (ISLPED).

[4]  L. Pileggi,et al.  Experimental demonstration of four-terminal magnetic logic device with separate read- and write-paths , 2014, 2014 IEEE International Electron Devices Meeting.

[5]  J. C. Sloncxewski Current-driven excitation of magnetic multilayers , 2003 .

[6]  Karin Strauss,et al.  Accelerating Deep Convolutional Neural Networks Using Specialized Hardware , 2015 .

[7]  Xuehai Zhou,et al.  PuDianNao: A Polyvalent Machine Learning Accelerator , 2015, ASPLOS.

[8]  Ron Kohavi,et al.  A Study of Cross-Validation and Bootstrap for Accuracy Estimation and Model Selection , 1995, IJCAI.

[9]  C. Rettner,et al.  Current-Controlled Magnetic Domain-Wall Nanowire Shift Register , 2008, Science.

[10]  Hiroyuki Awano,et al.  Direct Observation of Domain Wall Motion Induced by Low-Current Density in TbFeCo Wires , 2011 .

[11]  Ligang Gao,et al.  High precision tuning of state for memristive devices by adaptable variation-tolerant algorithm , 2011, Nanotechnology.

[12]  Johannes Schemmel,et al.  A wafer-scale neuromorphic hardware system for large-scale neural modeling , 2010, Proceedings of 2010 IEEE International Symposium on Circuits and Systems.

[13]  Hideo Ohno,et al.  Electric-field-induced magnetization switching in CoFeB/MgO magnetic tunnel junctions with high junction resistance , 2016 .

[14]  R. Williams,et al.  How We Found The Missing Memristor , 2008, IEEE Spectrum.

[15]  Jia Wang,et al.  DaDianNao: A Machine-Learning Supercomputer , 2014, 2014 47th Annual IEEE/ACM International Symposium on Microarchitecture.

[16]  Joos Vandewalle,et al.  Multi-Valued and Universal Binary Neurons: Theory, Learning and Applications , 2012 .

[17]  D. Strukov,et al.  Phenomenological modeling of memristive devices , 2014, 1406.4219.

[18]  Berger Emission of spin waves by a magnetic multilayer traversed by a current. , 1996, Physical review. B, Condensed matter.

[19]  S. Fukami,et al.  20-nm magnetic domain wall motion memory with ultralow-power operation , 2013, 2013 IEEE International Electron Devices Meeting.

[20]  G. Beach,et al.  Current-driven dynamics of chiral ferromagnetic domain walls. , 2013, Nature materials.

[21]  Z. Diao,et al.  Spin transfer switching and spin polarization in magnetic tunnel junctions with MgO and AlOx barriers , 2005, cond-mat/0510204.

[22]  Miad Faezipour,et al.  Hyperbolic Modeling of Subthalamic Nucleus Cells to Investigate the Effect of Dopamine Depletion , 2017, Comput. Intell. Neurosci..

[23]  Kaushik Roy,et al.  DWM-TAPESTRI - An energy efficient all-spin cache using domain wall shift based writes , 2013, 2013 Design, Automation & Test in Europe Conference & Exhibition (DATE).

[24]  Ninghui Sun,et al.  DianNao: a small-footprint high-throughput accelerator for ubiquitous machine-learning , 2014, ASPLOS.

[25]  Weisheng Zhao,et al.  High Speed, High Stability and Low Power Sensing Amplifier for MTJ/CMOS Hybrid Logic Circuits , 2009, IEEE Transactions on Magnetics.

[26]  Peter L. Bartlett,et al.  Using random weights to train multilayer networks of hard-limiting units , 1992, IEEE Trans. Neural Networks.

[27]  Jim D. Garside,et al.  Overview of the SpiNNaker System Architecture , 2013, IEEE Transactions on Computers.

[28]  Lawrence T. Pileggi,et al.  mLogic: Ultra-low voltage non-volatile logic circuits using STT-MTJ devices , 2012, DAC Design Automation Conference 2012.

[29]  Andrew S. Cassidy,et al.  A million spiking-neuron integrated circuit with a scalable communication network and interface , 2014, Science.

[30]  Tianshi Chen,et al.  ShiDianNao: Shifting vision processing closer to the sensor , 2015, 2015 ACM/IEEE 42nd Annual International Symposium on Computer Architecture (ISCA).

[31]  S. Parkin,et al.  Chiral spin torque at magnetic domain walls. , 2013, Nature nanotechnology.

[32]  Saburo Muroga,et al.  Enumeration of Threshold Functions of Eight Variables , 1970, IEEE Transactions on Computers.

[33]  Kaushik Roy,et al.  Ultra low power associative computing with spin neurons and resistive crossbar memory , 2013, 2013 50th ACM/EDAC/IEEE Design Automation Conference (DAC).

[34]  C. Rettner,et al.  Dynamics of Magnetic Domain Walls Under Their Own Inertia , 2010, Science.

[35]  Farnood Merrikh-Bayat,et al.  Efficient training algorithms for neural networks based on memristive crossbar circuits , 2015, 2015 International Joint Conference on Neural Networks (IJCNN).

[36]  Seul Jung,et al.  Hardware Implementation of a Real-Time Neural Network Controller With a DSP and an FPGA for Nonlinear Systems , 2007, IEEE Transactions on Industrial Electronics.

[37]  Jeong-Heon Park,et al.  Dependence of Voltage and Size on Write Error Rates in Spin-Transfer Torque Magnetic Random-Access Memory , 2016, IEEE Magnetics Letters.

[38]  S. Fukami,et al.  Low-current perpendicular domain wall motion cell for scalable high-speed MRAM , 2006, 2009 Symposium on VLSI Technology.

[39]  Kuei-Hung Shen,et al.  Racetrack Memory: A high-performance, low-cost, non-volatile memory based on magnetic domain walls , 2011, 2011 International Electron Devices Meeting.

[40]  Deliang Fan,et al.  Energy Efficient Reconfigurable Threshold Logic Circuit with Spintronic Devices , 2017, IEEE Transactions on Emerging Topics in Computing.

[41]  Jeffrey Bokor,et al.  Deterministic Domain Wall Motion Orthogonal To Current Flow Due To Spin Orbit Torque , 2014, Scientific Reports.