Abstract:In recent years, the scaling down that Moore’s Law relies on has been gradually slowing down, and the traditional von Neumann architecture has been limiting the improvement of computing power. Thus, neuromorphic in-memory computing hardware has been proposed and is becoming a promising alternative. However, there is still a long way to make it possible, and one of the problems is to provide an efficient, reliable, and achievable neural network for hardware implementation. In this paper, we proposed a two-layer fully connected spiking neural network based on binary MRAM (Magneto-resistive Random Access Memory) synapses with low hardware cost. First, the network used an array of multiple binary MRAM cells to store multi-bit fixed-point weight values. This helps to simplify the read/write circuit. Second, we used different kinds of spike encoders that ensure the sparsity of input spikes, to reduce the complexity of peripheral circuits, such as sense amplifiers. Third, we designed a single-step learning rule, which fit well with the fixed-point binary weights. Fourth, we replaced the traditional exponential Leak-Integrate-Fire (LIF) neuron model to avoid the massive cost of exponential circuits. The simulation results showed that, compared to other similar works, our SNN with 1184 neurons and 313,600 synapses achieved an accuracy of up to 90.6% in the MNIST recognition task with full-resolution (28 × 28) and full-bit-depth (8-bit) images. In the case of low-resolution (16 × 16) and black-white (1-bit) images, the smaller version of our network with 384 neurons and 32,768 synapses still maintained an accuracy of about 77%, extending its application to ultra-low-cost situations. Both versions need less than 30,000 samples to reach convergence, which is a >50% reduction compared to other similar networks. As for robustness, it is immune to the fluctuation of MRAM cell resistance.

Highly Uniform Two-Terminal Artificial Synapses Based on Polycrystalline Hf(0.5)Zr(0.5)O(2)for Sparsified Back Propagation Networks

HfZrOx-based Capacitive Synapses with Highly Linear and Symmetric Multilevel Characteristics for Neuromorphic Computing

Artificial HfO2/TiOx Synapses with Controllable Memory Window and High Uniformity for Brain-Inspired Computing

Artificial HfO<sub>2</sub>/TiO<sub>x</sub> Synapses with Controllable Memory Window and High Uniformity for Brain-Inspired Computing

Artificial HfO 2 /TiO x Synapses with Controllable Memory Window and High Uniformity for Brain-Inspired Computing.

High‐Performance Neuromorphic Computing Based on Ferroelectric Synapses with Excellent Conductance Linearity and Symmetry

Analog Synapses Based on Nonvolatile FETs With Amorphous ZrO 2 Dielectric for Spiking Neural Network Applications

An Efficient and Accurate Memristive Memory for Array-based Spiking Neural Networks

Novel Two-Terminal Synapse/Neuron Based on an Antiferroelectric Hafnium Zirconium Oxide Device for Neuromorphic Computing.

Highly Stable Artificial Synapses Based on Ferroelectric Tunnel Junctions for Neuromorphic Computing Applications

In-situ Learning in Hardware Compatible Multi-layer Memristive Spiking Neural Network

Reconfigurable Ag/HfO2/NiO/Pt Memristors with Stable Synchronous Synaptic and Neuronal Functions for Renewable Homogeneous Neuromorphic Computing System

A Low-Cost Hardware-Friendly Spiking Neural Network Based on Binary MRAM Synapses, Accelerated Using In-Memory Computing

Binary Resistive-Switching-Device-Based Electronic Synapse with Spike-Rate-Dependent Plasticity for Online Learning

Low‐Voltage Oscillatory Neurons for Memristor‐Based Neuromorphic Systems

Artificial synapse arrays based on SiO x /TiO x memristive crossbar with high uniformity for neuromorphic computing

Compact Artificial Neuron Based on Anti-Ferroelectric Transistor

Back-Propagation Neural Network Based on Analog Memristive Synapse

Configurable NbOx Memristors as Artificial Synapses or Neurons Achieved by Regulating the Forming Compliance Current for the Spiking Neural Network

Artificial synapse arrays based on SiOx/TiOx memristive crossbar with high uniformity for neuromorphic computing