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.

21 pages, 12 figures

References in corpus (1)

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

Cited by in corpus (87)

• A Survey of Neuromorphic Computing and Neural Networks in Hardware
• Magnetic skyrmion-based synaptic devices
• Spiking Neural Networks Hardware Implementations and Challenges: a Survey
• Variability in Resistive Memories
• Stochastic resonance in a metal-oxide memristive device
• Two-dimensional layered materials for memristive and neuromorphic applications
• 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
• Roadmap for Unconventional Computing with Nanotechnology
• Stochastic Spiking Neural Networks Enabled by Magnetic Tunnel Junctions: From Nontelegraphic to Telegraphic Switching Regimes
• Parallel in-memory wireless computing
• Memristor variability and stochastic physical properties modeling from a multivariate time series approach
• Side-channel attack analysis on in-memory computing architectures
• Delay dynamics of neuromorphic optoelectronic nanoscale resonators: Perspectives and applications
• A differential memristive synapse circuit for on-line learning in neuromorphic computing systems
• Synaptic dynamics in complex self-assembled nanoparticle networks
• Training a Probabilistic Graphical Model with Resistive Switching Electronic Synapses
• In-memory eigenvector computation in time O(1)
• Energy Efficient Learning with Low Resolution Stochastic Domain Wall Synapse Based Deep Neural Networks
• A new family of bioSFQ logic/memory cells
• PUMA: A Programmable Ultra-efficient Memristor-based Accelerator for Machine Learning Inference
• Harnessing Intrinsic Noise in Memristor Hopfield Neural Networks for Combinatorial Optimization
• 3D Interconnected Magnetic Nanowire Networks as Potential Integrated Multistate Memristors
• Quantum-limited stochastic optical neural networks operating at a few quanta per activation
• X-CHANGR: Changing Memristive Crossbar Mapping for Mitigating Line-Resistance Induced Accuracy Degradation in Deep Neural Networks
• Implementation of Multilayer Perceptron Network with Highly Uniform Passive Memristive Crossbar Circuits
• Temperature overshoot as the cause of physical changes in resistive switching devices during electro-formation
• End-to-End Memristive HTM System for Pattern Recognition and Sequence Prediction
• Training Coupled Phase Oscillators as a Neuromorphic Platform using Equilibrium Propagation
• 4K-Memristor Analog-Grade Passive Crossbar Circuit
• Mimicking synaptic plasticity with wedged Pt/Co/Pt spin-orbit torque device
• Efficient Training of the Memristive Deep Belief Net Immune to Non-Idealities of the Synaptic Devices
• Online Training of Spiking Recurrent Neural Networks with Phase-Change Memory Synapses
• Edge AI without Compromise: Efficient, Versatile and Accurate Neurocomputing in Resistive Random-Access Memory
• Non-volatile Hierarchical Temporal Memory: Hardware for Spatial Pooling
• 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
• Advancing Memristive Analog Neuromorphic Networks: Increasing Complexity, and Coping with Imperfect Hardware Components
• Oxygen vacancy dynamics in Pt/TiOx/TaOy/Pt memristors: exchange with the environment and internal electromigration
• A four-state magnetic tunnel junction switchable with spin-orbit torques
• 5 Parallel Prism: A topology for pipelined implementations of convolutional neural networks using computational memory
• Gate-tunable, normally-on to normally-off memristance transition in patterned LaAlO3/SrTiO3 interfaces
• Reminding Forgetful Organic Neuromorphic Device Networks
• High-speed metamagnetic resistive switching of FeRh through Joule heating
• Transient dynamics of pulse-driven memristors in the presence of a stable fixed point
• 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
• Oxide layer thickness effects on the resistance switching characteristics of Ti/TiO2-NT/Au structure
• Storage Class Memory: Principles, Problems, and Possibilities
• Bridging the Gap Between Neural Networks and Neuromorphic Hardware with A Neural Network Compiler
• Contemporary implementations of spiking bio-inspired neural networks
• Two-junction model in different percolation regimes of silver nanowires networks
• Quantum transport properties of tantalum-oxide resistive switching filaments
• Hybrid SLC-MLC RRAM Mixed-Signal Processing-in-Memory Architecture for Transformer Acceleration via Gradient Redistribution
• Charge-based computing with analogue reconfigurable gates
• Leveraging Voltage-Controlled Magnetic Anisotropy to Solve Sneak Path Issues in Crossbar Arrays
• Highly-Secure Physically Unclonable Cryptographic Primitives Using Nonlinear Conductance and Analog State Tuning in Memristive Crossbar Arrays
• Compact Device Models for FinFET and Beyond
• Phase Change Memtransistive Synapse
• Spintronics for neuromorphic computing
• Coexistance of volatile and non-volatile memristive effects in phase-separated La${}_{0.5}$Ca${}_{0.5}$MnO${}_3$-based devices
• MaD: Mapping and debugging framework for implementing deep neural network onto a neuromorphic chip with crossbar array of synapses
• Ferroelectric tunnel junctions
• Emulating homoeostatic effects with metal-oxide memristors T-dependence
• Hardware-friendly Neural Network Architecture for Neuromorphic Computing
• Memory-efficient training with streaming dimensionality reduction
• 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${}_2$
• Memristor Hardware-Friendly Reinforcement Learning
• A Physical Unclonable Function with Redox-based Nanoionic Resistive Memory
• Effect of memristorś potentiation-depression curves peculiarities in the convergence of physical perceptrons
• A Configurable Memristor-based Finite Impulse Response Filter
• Markov chain Hebbian learning algorithm with ternary synaptic units
• A light-stimulated neuromorphic device based on graphene hybrid phototransistor
• A Unified Platform to Evaluate STDP Learning Rule and Synapse Model using Pattern Recognition in a Spiking Neural Network
• Capacity, Fidelity, and Noise Tolerance of Associative Spatial-Temporal Memories Based on Memristive Neuromorphic Network
• In-situ learning harnessing intrinsic resistive memory variability through Markov Chain Monte Carlo Sampling
• Practical Demonstration of a Memristive Fuse
• A Survey of Brain Inspired Technologies for Engineering
• SOT-MRAM based Sigmoidal Neuron for Neuromorphic Architectures
• Volatile resistive-switched state in a bulk organic conductor with a sharp metal-insulator transition
• AutoGMap: Learning to Map Large-scale Sparse Graphs on Memristive Crossbars
• Fast learning synapses with molecular spin valves via selective magnetic potentiation
• Deep Neural Network Optimized to Resistive Memory with Nonlinear Current-Voltage Characteristics
• Deterministic Role of Concentration Surplus of Cation Vacancy over Anion Vacancy in Bipolar Memristive NiO
• Glass-like Electronics in Vanadium Dioxide
• Realization of A Non-Markov Chain in A Single 2D Crystal RRAM