Abstract:Memristor-enabled neuromorphic computing systems provide a fast and energy-efficient approach to training neural networks<a href="#ref-CR1">1</a>,<a href="#ref-CR2">2</a>,<a href="#ref-CR3">3</a>,<a href="/articles/s41586-020-1942-4#ref-CR4">4</a>. However, convolutional neural networks (CNNs)—one of the most important models for image recognition<a href="/articles/s41586-020-1942-4#ref-CR5">5</a>—have not yet been fully hardware-implemented using memristor crossbars, which are cross-point arrays with a memristor device at each intersection. Moreover, achieving software-comparable results is highly challenging owing to the poor yield, large variation and other non-ideal characteristics of devices<a href="#ref-CR6">6</a>,<a href="#ref-CR7">7</a>,<a href="#ref-CR8">8</a>,<a href="/articles/s41586-020-1942-4#ref-CR9">9</a>. Here we report the fabrication of high-yield, high-performance and uniform memristor crossbar arrays for the implementation of CNNs, which integrate eight 2,048-cell memristor arrays to improve parallel-computing efficiency. In addition, we propose an effective hybrid-training method to adapt to device imperfections and improve the overall system performance. We built a five-layer memristor-based CNN to perform MNIST<a href="/articles/s41586-020-1942-4#ref-CR10">10</a> image recognition, and achieved a high accuracy of more than 96 per cent. In addition to parallel convolutions using different kernels with shared inputs, replication of multiple identical kernels in memristor arrays was demonstrated for processing different inputs in parallel. The memristor-based CNN neuromorphic system has an energy efficiency more than two orders of magnitude greater than that of state-of-the-art graphics-processing units, and is shown to be scalable to larger networks, such as residual neural networks. Our results are expected to enable a viable memristor-based non-von Neumann hardware solution for deep neural networks and edge computing.

A Brain-inspired Fully Hardware Hopfield Neural Network Based on Memristive Arrays.

A Brain-Inspired In-Memory Computing System for Neuronal Communication Via Memristive Circuits.

A Hardware Friendly Unsupervised Memristive Neural Network with Weight Sharing Mechanism

A Hardware Implementation of SNN-Based Spatio-Temporal Memory Model.

A Fully Hardware-Based Memristive Multilayer Neural Network

Memristive Devices and Networks for Brain‐Inspired Computing

Design and Implementation of a Flexible Neuromorphic Computing System for Affective Communication Via Memristive Circuits

Stochastic neuro-fuzzy system implemented in memristor crossbar arrays

Hardware implementation of memristor-based artificial neural networks

Memristive brain-like computing

A Brain-Inspired Hardware Architecture for Evolutionary Algorithms Based on Memristive Arrays

Harnessing Intrinsic Noise in Memristor Hopfield Neural Networks for Combinatorial Optimization

Hybrid memristor-CMOS neurons for in-situ learning in fully hardware memristive spiking neural networks

In-Memory Computing With Memristor Arrays

Dynamics and Implementation of FPGA for Memristor-Coupled Fractional-Order Hopfield Neural Networks

Fully Hardware Memristive Neuromorphic Computing Enabled by the Integration of Trainable Dendritic Neurons and High‐Density RRAM Chip

Design of Artificial Neurons of Memristive Neuromorphic Networks Based on Biological Neural Dynamics and Structures

Memristor synapse-coupled piecewise-linear simplified Hopfield neural network: Dynamics analysis and circuit implementation

A Novel Brain-Inspired Hierarchical Perception-Association Circuit Based on Memristor Arrays

Fully hardware-implemented memristor convolutional neural network

Dynamics in a Memristive Neural Network with Three Discrete Heterogeneous Neurons and Its Application