Imperial College London
Portrait Picture

ProfessorWillBranford

Faculty of Natural Sciences, Department of Physics

Professor of Solid State Physics
 
 
 
//

Contact

 

+44 (0)20 7594 6674w.branford Website

 
 
//

Location

 

912Blackett LaboratorySouth Kensington Campus

//

Summary

 

Publications

Citation

BibTex format

@article{Gartside:2021,
author = {Gartside, JC and Stenning, KD and Vanstone, A and Dion, T and Holder, HH and Arroo, DM and Caravelli, F and Kurebayashi, H and Branford, WR},
journal = {Nature Nanotechnology},
pages = {460--469},
title = {Reconfigurable Training and Reservoir Computing in an Artificial  Spin-Vortex Ice via Spin-Wave Fingerprinting},
url = {http://arxiv.org/abs/2107.08941v3},
volume = {17},
year = {2021}
}
Download

RIS format (EndNote, RefMan)

TY  - JOUR
AB  - Strongly-interacting artificial spin systems are moving beyond mimickingnaturally-occurring materials to emerge as versatile functional platforms, fromreconfigurable magnonics to neuromorphic computing. Typically artificial spinsystems comprise nanomagnets with a single magnetisation texture: collinearmacrospins or chiral vortices. By tuning nanoarray dimensions we achievemacrospin/vortex bistability and demonstrate a four-state metamaterialspin-system 'Artificial Spin-Vortex Ice' (ASVI). ASVI can host Ising-likemacrospins with strong ice-like vertex interactions, and weakly-coupledvortices with low stray dipolar-field. Vortices and macrospins exhibitstarkly-differing spin-wave spectra with analogue-style mode-amplitude controland mode-frequency shifts of df = 3.8 GHz.  The enhanced bi-textural microstate space gives rise to emergent physicalmemory phenomena, with ratchet-like vortex training and history-dependentnonlinear fading memory when driven through global field cycles. We employspin-wave microstate fingerprinting for rapid, scaleable readout of vortex andmacrospin populations and leverage this for spin-wave reservoir computation.ASVI performs linear and non-linear mapping transformations of diverse inputsignals as well as chaotic time-series forecasting. Energy costs of machinelearning are spiralling unsustainably, developing low-energy neuromorphiccomputation hardware such as ASVI is crucial to achieving a zero-carboncomputational future.
AU  - Gartside,JC
AU  - Stenning,KD
AU  - Vanstone,A
AU  - Dion,T
AU  - Holder,HH
AU  - Arroo,DM
AU  - Caravelli,F
AU  - Kurebayashi,H
AU  - Branford,WR
EP  - 469
PY  - 2021///
SP  - 460
TI  - Reconfigurable Training and Reservoir Computing in an Artificial  Spin-Vortex Ice via Spin-Wave Fingerprinting
T2  - Nature Nanotechnology
UR  - http://arxiv.org/abs/2107.08941v3
UR  - http://dx.doi.org/10.1038/s41565-022-01091-7
UR  - http://hdl.handle.net/10044/1/107817
VL  - 17
ER  -
Download