Quantifying the Computational Capability of a Nanomagnetic Reservoir Computing Platform with Emergent Magnetization Dynamics

Vidamour, Ian T.; Ellis, Matthew O. A.; Griffin, David Ray; Venkat, Guru; Swindells, C.; Dawidek, Richard W. S.; Broomhall, T. J.; Steinke, Nina‐Juliane; Cooper, Joshaniel F. K.; Francisco, Maccherozzi,; Dhesi, S. S.; Stepney, Susan; Vasilaki, Eleni; Allwood, D. A.; Hayward, Thomas J.

doi:10.48550/arxiv.2111.14603

Cited by 1 publication

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To demonstrate the suitability of the ring array reservoirs to more complex computational tasks on real-world data, a standard benchmark test of classifying spoken digits was undertaken, which has previously been performed with other proposed RC platforms 16,22 . Using input data pre-processed using Mel-frequency cepstral filters 32 , the data is linearly inseparable, with classification accuracy limited to around 75% 17,22,33 . The role of the reservoir is to provide a mapping of input data into higher dimensional reservoir space via its inherent nonlinearity, allowing the linear readout layer to establish hyperplanes which can classify the data more accurately.…”

Section: Spoken Digit Recognition With Ring Array Reservoirsmentioning

confidence: 99%

Reservoir Computing with Emergent Dynamics in a Magnetic Metamaterial

Vidamour

Swindells

Venkat

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

In Materio reservoir computing (RC) leverages the response of physical systems to perform computation. Dynamic systems with emergent behaviours (where local interactions lead to complex global behaviours) are especially promising for RC, as computational capability is determined by the complexity of the transformation provided. However, it is often difficult to extract these complex behaviours via device tractable measurements that can be interfaced with standard electronics. In this paper, we measure the emergent response of interconnected magnetic nanoring arrays using simple electric transport measurements, observing distinct computationally promising dynamic behaviours in device response. Then, we employ three distinct reservoir architectures that exploit each of the behaviours to perform benchmark tasks with contrasting computational requirements on a single device with state-of-the-art accuracies for spintronic computing platforms.

show abstract