Tom Hayward, a PI on the MARCH team recently gave a seminar at Imperial College London

 

From Stochasticity to Functionality: Magnetic Domain Walls for Probabilistic and Neuromorphic Computing

Tom Hayward

Functional Magnetic Materials Group, Department of Materials Science and Engineering, University of Sheffield

 

Domain walls (DWs) in ferromagnetic nanowires have been a topic of intense research interest due to proposals to use them as data carriers in energy efficient, non-volatile logic and memory devices. However, despite their apparent technological potential, these devices have been challenging to realise because DW pinning and propagation is highly stochastic, making digital devices unreliable [1]. While materials engineering approaches can be used to supress stochasticity [2], it is also interesting to consider whether alternative computational paradigms could prove more resistant to, or even benefit from, the DWs’ stochastic behaviours, thus converting them from technologically inhibitive phenomena into a functional properties. Exploring whether neuromorphic (brain-like) paradigms could benefit from such an approach is of particular significance due to the increasing adoption of artificial intelligence (AI) across society, and the huge energy costs of running AI algorithms on conventional CMOS computers.  

 

In this talk I will present key experimental results that illustrate how these domain wall devices can be used to realise two different neuromorphic computing paradigms in hardware: feed-forward neural networks and reservoir computers [3,4]. The work presented will include experimental demonstrations of how these devices can be used to performed benchmark machine learning tasks such as spoken/written digit recognition and time series prediction. I will discuss how DW devices can be further adapted to create technologies that both senses time evolving external stimuli and interprets their meaning – i.e. sensors that both “think” and “feel”.

 

[1] T.J. Hayward and K.A. Omari, Beyond the quasi-particle: stochastic domain wall dynamics in soft ferromagnetic nanowires, Journal of Physics D: Applied Physics 50, 8, 084006 (2017).

[2] T.J. Broomhall, A.W. Rushforth, M.C. Rosamond, E.H. Linfield, and T.J. Hayward, Physical Review Applied 13, 024039 (2020).

[3] R. W. Dawidek, T. J. Hayward, I. T. Vidamour, T. J. Broomhall, M. Al Mamoori, A. Mullen, S. J. Kyle, P. W. Fry, N. J. Steinke, J. F. K. Cooper, F. Maccherozzi, S. S.Dhesi, L. Aballe, M. Foerster, J. Prat, E. Vasilaki, M. O. A. Ellis, D. A. Allwood, Advanced Functional Materials 31, 2008389 (2021)

[4] I.T. Vidamour, C.Swindells, G. Venkat, P.W.Fry, A. Welbourne, R.Rowan-Robinson, D. Backes, F. Maccherozzi, S.S. Dhesi, E. Vasilaki, D.A. Allwood and T.J. Hayward, arXiv preprint arXiv:2206.04446 (2022).