Memristor-based Synaptic Networks and Logical Operations Using In-Situ Computing

TL;DR

This paper introduces memristor-based computational blocks capable of implementing learning rules like STDP and Hebbian learning, demonstrating their potential for in-situ neuromorphic computing with experimental validation.

Contribution

It presents novel memristor-based building blocks for supervised and unsupervised learning, including implementation of STDP and Hebbian learning in large-scale networks.

Findings

Memristor can implement STDP with a single device.

A 1x1000 memristor network demonstrates Hebbian learning.

Experimental validation using SPICE model confirms functionality.

Abstract

We present new computational building blocks based on memristive devices. These blocks, can be used to implement either supervised or unsupervised learning modules. This is achieved using a crosspoint architecture which is an efficient array implementation for nanoscale two-terminal memristive devices. Based on these blocks and an experimentally verified SPICE macromodel for the memristor, we demonstrate that firstly, the Spike-Timing-Dependent Plasticity (STDP) can be implemented by a single memristor device and secondly, a memristor-based competitive Hebbian learning through STDP using a $1\times 1000$ synaptic network. This is achieved by adjusting the memristor's conductance values (weights) as a function of the timing difference between presynaptic and postsynaptic spikes. These implementations have a number of shortcomings due to the memristor's characteristics such as memory decay, highly nonlinear switching behaviour as a function of applied voltage/current, and functional uniformity. These shortcomings can be addressed by utilising a mixed gates that can be used in conjunction with the analogue behaviour for biomimetic computation. The digital implementations in this paper use in-situ computational capability of the memristor.