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.

A neuron-weighted learning algorithm and its hardware implementation in associative memories

Weighted Learning of Bidirectional Associative Memories by Global Minimization

Optimal learning algorithm in bidirectional associative memories

A Spiking Neural Network Model For Associative Memory Using Temporal Codes

Efficient learning algorithm for associative memory neural network

Memory Circuit Design, Implementation and Analysis Based on Memristor Full-function Pavlov Associative

A neural network circuit with associative learning and forgetting process based on memristor neuromorphic device

A brain-inspired algorithm that mitigates catastrophic forgetting of artificial and spiking neural networks with low computational cost

Extending memory capacity of neural associative memory based on recursive synaptic bit reuse.

Memristive discrete chaotic neural network and its application in associative memory

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

Weighted Synapses Without Carry Operations for RRAM-Based Neuromorphic Systems

Neuromorphic Incremental on-chip Learning with Hebbian Weight Consolidation

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

A Digitalized RRAM-based Spiking Neuron Network System with 3-Bit Weight and Unsupervised Online Learning Scheme

Memristor-based neural networks with weight simultaneous perturbation training

A robust learning algorithm for the extended AM neural network

In-situ learning in multilayer locally-connected memristive spiking neural network

A Forward Learning Algorithm for Neural Memory Ordinary Differential Equations

High-Order Associative Learning Based on Memristive Circuits for Efficient Learning

Learning algorithms of bidirectional associative memory based on max-min criterion