Oscillatory Neural Networks Using VO2 Based Phase Encoded Logic.

Juan Núñez,María J. Avedillo,Bernabé Linares-Barranco,Manuel Jiménez,José M. Quintana,Elisabetta Corti,Aida Todri-Sanial,Siegfried Karg

doi:10.3389/fnins.2021.655823

Abstract

Nano-oscillators based on phase-transition materials are being explored for the implementation of different non-conventional computing paradigms. In particular, vanadium dioxide (VO2) devices are used to design autonomous non-linear oscillators from which oscillatory neural networks (ONNs) can be developed. In this work, we propose a new architecture for ONNs in which sub-harmonic injection locking (SHIL) is exploited to ensure that the phase information encoded in each neuron can only take two values. In this sense, the implementation of ONNs from neurons that inherently encode information with two-phase values has advantages in terms of robustness and tolerance to variability present in VO2 devices. Unlike conventional interconnection schemes, in which the sign of the weights is coded in the value of the resistances, in our proposal the negative (positive) weights are coded using static inverting (non-inverting) logic at the output of the oscillator. The operation of the proposed architecture is shown for pattern recognition applications.

Highlights

Phase-transition materials (PTMs) like vanadium dioxide (VO2), with their abrupt switching between states with very different resistivity, are being explored for implementing non-boolean computational paradigms such as neuromorphic architectures
Corti et al (2021) have shown that this approach using VO2 oscillators can be exploited for the implementation of commercial high-accuracy image processing architectures based on convolutional neural networks (CNN). Motivated by the latter type of application, in this paper we describe the implementation of an oscillatory neural networks (ONNs) using VO2 based phase encoded logic (PeL)
ONNs With VO2 Oscillators Figure 1 shows the ONN proposed in Corti et al (2018, 2020), and the VO2 oscillator used as neuron

Summary

Introduction

Phase-transition materials (PTMs) like vanadium dioxide (VO2), with their abrupt switching between states with very different resistivity, are being explored for implementing non-boolean computational paradigms such as neuromorphic architectures. The field of OBC is not a new idea, with outstanding contributions in the field of logic in the 1950s (von Neumann, 1957; Goto, 1959) In recent years, this idea has received considerable interest and has become an active research area due to the appearance of devices, operating based on very different physical phenomena, with the ability to implement very compact oscillators and with very low energy consumption. They are reported to reduce energy per cycle by more than an order of magnitude when compared to CMOS ring oscillators (ROs). They rank second in terms of energy efficiency, behind only superconducting oscillators

Methods

Results

Conclusion