Training and Operation of an Integrated Neuromorphic Network Based on Metal-Oxide Memristors

Abstract

Despite all the progress of semiconductor integrated circuit technology, the extreme complexity of the human cerebral cortex makes the hardware implementation of neuromorphic networks with a comparable number of devices exceptionally challenging. One of the most prospective candidates to provide comparable complexity, while operating much faster and with manageable power dissipation, are so-called CrossNets based on hybrid CMOS/memristor circuits. In these circuits, the usual CMOS stack is augmented with one or several crossbar layers, with adjustable two-terminal memristors at each crosspoint. Recently, there was a significant progress in improvement of technology of fabrication of such memristive crossbars and their integration with CMOS circuits, including first demonstrations of their vertical integration. Separately, there have been several demonstrations of discrete memristors as artificial synapses for neuromorphic networks. Very recently such experiments were extended to crossbar arrays of phase-change memristive devices. The adjustment of such devices, however, requires an additional transistor at each crosspoint, and hence the prospects of their scaling are less impressive than those of metal-oxide memristors, whose nonlinear I-V curves enable transistor-free operation. Here we report the first experimental implementation of a transistor-free metal-oxide memristor crossbar with device variability lowered sufficiently to demonstrate a successful operation of a simple integrated neural network, a single layer-perceptron. The network could be taught in situ using a coarse-grain variety of the delta-rule algorithm to perform the perfect classification of 3x3-pixel black/white images into 3 classes. We believe that this demonstration is an important step towards the implementation of much larger and more complex memristive neuromorphic networks.

arXiv comment21 pages, 12 figures

References in corpus (2)

• A ferroelectric memristor
• Brain-like associative learning using a nanoscale non-volatile phase change synaptic device array

Cited by in corpus (173)

• Memory and information processing in neuromorphic systems
• Neuromorphic computing with multi-memristive synapses
• Robust memristors based on layered two-dimensional materials
• Deep physical neural networks enabled by a backpropagation algorithm for arbitrary physical systems
• Biological plausibility and stochasticity in scalable VO2 active memristor neurons
• Magnetic skyrmion artificial synapse for neuromorphic computing
• Memristor Crossbars with 4.5 Terabits-per-Inch-Square Density and Two Nanometer Dimension
• Training of photonic neural networks through in situ backpropagation
• Acceleration of Deep Neural Network Training with Resistive Cross-Point Devices
• All-Optically Controlled Memristor
• Long short-term memory networks in memristor crossbars
• A Survey of Neuromorphic Computing and Neural Networks in Hardware
• Neuro-memristive Circuits for Edge Computing: A review
• Photoferroelectric oxides
• Magnetic skyrmion-based synaptic devices
• Spiking Neural Networks Hardware Implementations and Challenges: a Survey
• Challenges in materials and devices for Resistive-Switching-based Neuromorphic Computing
• Proposal for an All-Spin Artificial Neural Network: Emulating Neural and Synaptic Functionalities Through Domain Wall Motion in Ferromagnets
• Variability in Resistive Memories
• Hardware for Machine Learning: Challenges and Opportunities
• Tuning a binary ferromagnet into a multi-state synapse with spin-orbit torque induced plasticity
• Electrochemical Ion Insertion: From Atoms to Devices
• Physical reservoir computing built by spintronic devices for temporal information processing
• Stochastic resonance in a metal-oxide memristive device
• Truly concomitant and independently expressed short- and long-term plasticity in Bi2O2Se-based three-terminal memristor
• Mixed-precision deep learning based on computational memory
• Two-dimensional layered materials for memristive and neuromorphic applications
• Polaritonic neuromorphic computing outperforms linear classifiers
• Mitigating Asymmetric Nonlinear Weight Update Effects in Hardware Neural Network based on Analog Resistive Synapse
• Scalable massively parallel computing using continuous-time data representation in nanoscale crossbar array
• Energy Efficient Computing Systems: Architectures, Abstractions and Modeling to Techniques and Standards
• A neuromorphic systems approach to in-memory computing with non-ideal memristive devices: From mitigation to exploitation
• Roadmap for Unconventional Computing with Nanotechnology
• Stochastic Spiking Neural Networks Enabled by Magnetic Tunnel Junctions: From Nontelegraphic to Telegraphic Switching Regimes
• Committee machines -- a universal method to deal with non-idealities in memristor-based neural networks
• Device and System Level Design Considerations for Analog-Non-Volatile-Memory Based Neuromorphic Architectures
• Technology Aware Training in Memristive Neuromorphic Systems based on non-ideal Synaptic Crossbars
• A sub-1-volt analog metal oxide memristive-based synaptic device for energy-efficient spike-based computing systems
• Multifilamentary character of anticorrelated capacitive and resistive switching in memristive structures based on (CoFeB)x(LiNbO3)100-x nanocomposite
• Adaptive Learning Rule for Hardware-based Deep Neural Networks Using Electronic Synapse Devices
• Tunneling anomalous Hall effect in the nanogranular CoFe-B-Al-O films near the metal-insulator transition
• Parallel in-memory wireless computing
• Fabry-Perot Lasers as Enablers for Parallel Reservoir Computing
• Picosecond Multilevel Resistive Switching in Tantalum Oxide Thin Films
• A spiking neuron constructed by the skyrmion-based spin torque nano-oscillator
• Training LSTM Networks with Resistive Cross-Point Devices
• Memristor variability and stochastic physical properties modeling from a multivariate time series approach
• An atomic Boltzmann machine capable of on-chip learning
• A superconducting nanowire spiking element for neural networks
• Side-channel attack analysis on in-memory computing architectures
• An exploration of parameter redundancy in deep networks with circulant projections
• Perspective: Organic electronic materials and devices for neuromorphic engineering
• Efficient Processing of Deep Neural Networks: A Tutorial and Survey
• Stateful characterization of resistive switching TiO2 with electron beam induced currents
• Nanoscale-femtosecond dielectric response of Mott insulators captured by two-colour near-field ultrafast electron microscopy
• Delay dynamics of neuromorphic optoelectronic nanoscale resonators: Perspectives and applications
• A differential memristive synapse circuit for on-line learning in neuromorphic computing systems
• Analog Weights in ReRAM DNN Accelerators
• Synaptic dynamics in complex self-assembled nanoparticle networks
• Training a Probabilistic Graphical Model with Resistive Switching Electronic Synapses
• Perceptrons from Memristors
• Resistive switching in reverse: voltage driven formation of a transverse insulating barrier
• TraNNsformer: Neural network transformation for memristive crossbar based neuromorphic system design
• In-memory eigenvector computation in time O(1)
• Emerging spintronics phenomena and applications
• Energy Efficient Learning with Low Resolution Stochastic Domain Wall Synapse Based Deep Neural Networks
• Device and Circuit Interaction Analysis of Stochastic Behaviors in Cross-Point RRAM Arrays
• A new family of bioSFQ logic/memory cells
• Harnessing Intrinsic Noise in Memristor Hopfield Neural Networks for Combinatorial Optimization
• PUMA: A Programmable Ultra-efficient Memristor-based Accelerator for Machine Learning Inference
• 3D Interconnected Magnetic Nanowire Networks as Potential Integrated Multistate Memristors
• Quantum-limited stochastic optical neural networks operating at a few quanta per activation
• Voltage control of domain walls in magnetic nanowires for energy efficient neuromorphic devices
• Supervised learning of an opto-magnetic neural network with ultrashort laser pulses
• X-CHANGR: Changing Memristive Crossbar Mapping for Mitigating Line-Resistance Induced Accuracy Degradation in Deep Neural Networks
• Temperature overshoot as the cause of physical changes in resistive switching devices during electro-formation
• Locality of interactions for planar memristive circuits
• Implementation of Multilayer Perceptron Network with Highly Uniform Passive Memristive Crossbar Circuits
• End-to-End Memristive HTM System for Pattern Recognition and Sequence Prediction
• Entangled Quantum Memristors
• Training Coupled Phase Oscillators as a Neuromorphic Platform using Equilibrium Propagation
• Streaming Batch Eigenupdates for Hardware Neuromorphic Networks
• 4K-Memristor Analog-Grade Passive Crossbar Circuit
• Tunable stochasticity in an artificial spin network
• Heterogeneous Memristive Devices Enabled by Magnetic Tunnel Junction Nanopillars Surrounded by Resistive Silicon Switches
• Mimicking synaptic plasticity with wedged Pt/Co/Pt spin-orbit torque device
• In-situ Stochastic Training of MTJ Crossbar based Neural Networks
• Efficient Training of the Memristive Deep Belief Net Immune to Non-Idealities of the Synaptic Devices
• Photonic kernel machine learning for ultrafast spectral analysis
• Online Training of Spiking Recurrent Neural Networks with Phase-Change Memory Synapses
• Memcapacitive Devices in Logic and Crossbar Applications
• Design and Characterization of Superconducting Nanowire-Based Processors for Acceleration of Deep Neural Network Training
• Zero-shifting Technique for Deep Neural Network Training on Resistive Cross-point Arrays
• Non-volatile Hierarchical Temporal Memory: Hardware for Spatial Pooling
• Edge AI without Compromise: Efficient, Versatile and Accurate Neurocomputing in Resistive Random-Access Memory
• Band gap reconstruction at the interface between black phosphorus and a gold electrode
• Neural Networks as Geometric Chaotic Maps
• Breaking the quantum PIN code of atomic synapses
• Precise neural network computation with imprecise analog devices
• RESPARC: A Reconfigurable and Energy-Efficient Architecture with Memristive Crossbars for Deep Spiking Neural Networks
• Interfacial resistive switching by multiphase polarization in ion-intercalation nanofilms
• Physical Neural Networks with Self-Learning Capabilities
• Self-Adaptive Spike-Time-Dependent Plasticity of Metal-Oxide Memristors
• The electronic structures of TiO2/Ti4O7, Ta2O5/TaO2 interfaces and the interfacial effects of dopants
• Oxygen vacancy dynamics in Pt/TiOx/TaOy/Pt memristors: exchange with the environment and internal electromigration
• Gate-tunable, normally-on to normally-off memristance transition in patterned LaAlO3/SrTiO3 interfaces
• 5 Parallel Prism: A topology for pipelined implementations of convolutional neural networks using computational memory
• Advancing Memristive Analog Neuromorphic Networks: Increasing Complexity, and Coping with Imperfect Hardware Components
• A four-state magnetic tunnel junction switchable with spin-orbit torques
• Transient dynamics of pulse-driven memristors in the presence of a stable fixed point
• Reminding Forgetful Organic Neuromorphic Device Networks
• High-speed metamagnetic resistive switching of FeRh through Joule heating
• Low Power Artificial Neural Network Architecture
• Metastable Memristive Lines for Signal Transmission and Information Processing Applications
• Memristive integrative sensors for neuronal activity
• GXNOR-Net: Training deep neural networks with ternary weights and activations without full-precision memory under a unified discretization framework
• Optimization of the multi-mem response of topotactic redox La${}_{1/2}$Sr${}_{1/2}$Mn${}_{1/2}$Co${}_{1/2}$O${}_{3-x}$
• A system of different layers of abstraction for artificial intelligence
• Quantum transport properties of tantalum-oxide resistive switching filaments
• Two-junction model in different percolation regimes of silver nanowires networks
• Contemporary implementations of spiking bio-inspired neural networks
• Storage Class Memory: Principles, Problems, and Possibilities
• Oxide layer thickness effects on the resistance switching characteristics of Ti/TiO2-NT/Au structure
• Training and pattern recognition by an opto-magnetic neural network
• Bridging the Gap Between Neural Networks and Neuromorphic Hardware with A Neural Network Compiler
• Hybrid SLC-MLC RRAM Mixed-Signal Processing-in-Memory Architecture for Transformer Acceleration via Gradient Redistribution
• Towards Trainable Media: Using Waves for Neural Network-Style Training
• Charge-based computing with analogue reconfigurable gates
• Highly-Secure Physically Unclonable Cryptographic Primitives Using Nonlinear Conductance and Analog State Tuning in Memristive Crossbar Arrays
• Leveraging Voltage-Controlled Magnetic Anisotropy to Solve Sneak Path Issues in Crossbar Arrays
• The role of single oxygen or metal induced defect and correlated multiple defects in the formation of conducting filaments
• Field-Programmable Crossbar Array (FPCA) for Reconfigurable Computing
• Ferroelectric tunnel junctions
• Recent Advances in Convolutional Neural Network Acceleration
• MaD: Mapping and debugging framework for implementing deep neural network onto a neuromorphic chip with crossbar array of synapses
• Effect of memristorś potentiation-depression curves peculiarities in the convergence of physical perceptrons
• Memristor Hardware-Friendly Reinforcement Learning
• Design space exploration of Ferroelectric FET based Processing-in-Memory DNN Accelerator
• Memory-efficient training with streaming dimensionality reduction
• Compact Device Models for FinFET and Beyond
• Hardware-friendly Neural Network Architecture for Neuromorphic Computing
• Emulating homoeostatic effects with metal-oxide memristors T-dependence
• Multi-mem behavior at reduced voltages in La${}_{1/2}$Sr${}_{1/2}$Mn${}_{1/2}$Co${}_{1/2}$O${}_{3-x}$ perovskite modified with Sm:CeO${}_2$
• Spintronics for neuromorphic computing
• Phase Change Memtransistive Synapse
• A Physical Unclonable Function with Redox-based Nanoionic Resistive Memory
• Coexistance of volatile and non-volatile memristive effects in phase-separated La${}_{0.5}$Ca${}_{0.5}$MnO${}_3$-based devices
• A Unified Platform to Evaluate STDP Learning Rule and Synapse Model using Pattern Recognition in a Spiking Neural Network
• In-situ learning harnessing intrinsic resistive memory variability through Markov Chain Monte Carlo Sampling
• Comparison of Update and Genetic Training Algorithms in a Memristor Crossbar Perceptron
• Capacity, Fidelity, and Noise Tolerance of Associative Spatial-Temporal Memories Based on Memristive Neuromorphic Network
• Markov chain Hebbian learning algorithm with ternary synaptic units
• A light-stimulated neuromorphic device based on graphene hybrid phototransistor
• A Configurable Memristor-based Finite Impulse Response Filter
• Real-time image processing with a 2D semiconductor neural network vision sensor
• A Survey of Brain Inspired Technologies for Engineering
• Practical Demonstration of a Memristive Fuse
• SOT-MRAM based Sigmoidal Neuron for Neuromorphic Architectures
• Glass-like Electronics in Vanadium Dioxide
• AutoGMap: Learning to Map Large-scale Sparse Graphs on Memristive Crossbars
• 2022 Roadmap on Neuromorphic Computing and Engineering
• A technology agnostic RRAM characterisation methodology protocol
• Deep Neural Network Optimized to Resistive Memory with Nonlinear Current-Voltage Characteristics
• A Reconfigurable Low Power High Throughput Architecture for Deep Network Training
• Neuromorphic MoS2 memtransistors fabricated by localised helium ion beam irradiation
• Deterministic Role of Concentration Surplus of Cation Vacancy over Anion Vacancy in Bipolar Memristive NiO
• Artificial Retina Using A Hybrid Neural Network With Spatial Transform Capability
• Realization of A Non-Markov Chain in A Single 2D Crystal RRAM
• Memory effect in the charge transport in strongly disordered antimony films
• Fast learning synapses with molecular spin valves via selective magnetic potentiation
• Single-electron tunneling PbS/InP neuromorphic computing building blocks
• Volatile resistive-switched state in a bulk organic conductor with a sharp metal-insulator transition
• Multibit memory operation of metal-oxide bi-layer memristors