Classification of multi-frequency RF signals by extreme learning, using magnetic tunnel junctions as neurons and synapses

TL;DR

This paper demonstrates that magnetic tunnel junctions can process multi-frequency RF signals using extreme learning, enabling low-energy neural network classification with accuracy comparable to software implementations, advancing embedded RF AI.

Contribution

It introduces a hardware neural network using magnetic tunnel junctions for parallel RF signal processing and classification with a backpropagation-free learning method.

Findings

Achieved RF signal classification accuracy comparable to software neural networks.

Demonstrated parallel processing of multi-frequency RF inputs with magnetic tunnel junctions.

Validated the approach with experimental data, showing potential for low-energy embedded RF AI.

Abstract

Extracting information from radiofrequency (RF) signals using artificial neural networks at low energy cost is a critical need for a wide range of applications from radars to health. These RF inputs are composed of multiples frequencies. Here we show that magnetic tunnel junctions can process analogue RF inputs with multiple frequencies in parallel and perform synaptic operations. Using a backpropagation-free method called extreme learning, we classify noisy images encoded by RF signals, using experimental data from magnetic tunnel junctions functioning as both synapses and neurons. We achieve the same accuracy as an equivalent software neural network. These results are a key step for embedded radiofrequency artificial intelligence.