Fingerprinting Magnetic States in Interconnected Nanoring Arrays via Spin Wave Spectra

Arrays of interconnected magnetic nanorings show emergent dynamics arising from stochastic interactions at junctions between rings. These global emergent properties have already shown potential as a platform for reservoir computing. However, their experimental computational performance has been limited by the use of a single-dimensional global measurement, their magnetoresistance, as an output, masking much of the underlying state information. Here, we show in both simulation and experiment the potential of using the spin wave spectra of such arrays of nanorings to produce a fingerprint of the number and orientation of the internal magnetic states, providing a route towards a multi-dimensional output with which to perform computation. We show that the spectra fulfill the two criteria for reservoir computing: a nonlinear response to an input stimulus, and a fading memory.