Saved in:
Bibliographic Details
Main Authors: Plouet, Erwan, Sanz-Hernández, Dédalo, Vecchiola, Aymeric, Grollier, Julie, Mizrahi, Frank
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2408.02835
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917944247713792
author Plouet, Erwan
Sanz-Hernández, Dédalo
Vecchiola, Aymeric
Grollier, Julie
Mizrahi, Frank
author_facet Plouet, Erwan
Sanz-Hernández, Dédalo
Vecchiola, Aymeric
Grollier, Julie
Mizrahi, Frank
contents The ability to process time-series at low energy cost is critical for many applications. Recurrent neural network, which can perform such tasks, are computationally expensive when implementing in software on conventional computers. Here we propose to implement a recurrent neural network in hardware using spintronic oscillators as dynamical neurons. Using numerical simulations, we build a multi-layer network and demonstrate that we can use backpropagation through time (BPTT) and standard machine learning tools to train this network. Leveraging the transient dynamics of the spintronic oscillators, we solve the sequential digits classification task with $89.83\pm2.91~\%$ accuracy, as good as the equivalent software network. We devise guidelines on how to choose the time constant of the oscillators as well as hyper-parameters of the network to adapt to different input time scales.
format Preprint
id arxiv_https___arxiv_org_abs_2408_02835
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Training a multilayer dynamical spintronic network with standard machine learning tools to perform time series classification
Plouet, Erwan
Sanz-Hernández, Dédalo
Vecchiola, Aymeric
Grollier, Julie
Mizrahi, Frank
Disordered Systems and Neural Networks
Mesoscale and Nanoscale Physics
Artificial Intelligence
Machine Learning
The ability to process time-series at low energy cost is critical for many applications. Recurrent neural network, which can perform such tasks, are computationally expensive when implementing in software on conventional computers. Here we propose to implement a recurrent neural network in hardware using spintronic oscillators as dynamical neurons. Using numerical simulations, we build a multi-layer network and demonstrate that we can use backpropagation through time (BPTT) and standard machine learning tools to train this network. Leveraging the transient dynamics of the spintronic oscillators, we solve the sequential digits classification task with $89.83\pm2.91~\%$ accuracy, as good as the equivalent software network. We devise guidelines on how to choose the time constant of the oscillators as well as hyper-parameters of the network to adapt to different input time scales.
title Training a multilayer dynamical spintronic network with standard machine learning tools to perform time series classification
topic Disordered Systems and Neural Networks
Mesoscale and Nanoscale Physics
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2408.02835