Nano-oscillator-based classification with a machine learning-compatible architecture

TL;DR

This paper introduces a nano-oscillator classification architecture with a gradient-based offline learning algorithm, demonstrating improved accuracy, parameter efficiency, and compatibility with existing nano-technologies and noise tolerance.

Contribution

It presents a novel nano-oscillator classification architecture with a gradient-based offline learning algorithm, enhancing performance and practicality over previous online learning methods.

Findings

Significant classification improvements over online learning approaches.

Comparable performance to standard neural networks with fewer parameters.

Architecture is compatible with existing nano-technologies and tolerant to phase noise.

Abstract

Pattern classification architectures leveraging the physics of coupled nano-oscillators have been demonstrated as promising alternative computing approaches, but lack effective learning algorithms. In this work, we propose a nano-oscillator based classification architecture where the natural frequencies of the oscillators are learned linear combinations of the inputs, and define an offline learning algorithm based on gradient back-propagation. Our results show significant classification improvements over a related approach with online learning. We also compare our architecture with a standard neural network on a simple machine learning case, which suggests that our approach is economical in terms of numbers of adjustable parameters. The introduced architecture is also compatible with existing nano-technologies: the architecture does not require changes in the coupling between nano-oscillators, and it is tolerant to oscillator phase noise.