Fig. 6: Performance of signal transformation and prediction tasks. | Nature Communications

Fig. 6: Performance of signal transformation and prediction tasks.

From: Electrodeposition of magnetic nanonetworks featuring triangular motifs and parallel ridges on a macroscopic scale

Fig. 6: Performance of signal transformation and prediction tasks.The alternative text for this image may have been generated using AI.

a Schematic of the magnetic network reservoir computing scheme. Input values from −1 to 1 are mapped to the field range of −100 to 100 mT as the input to the reservoir. The ferromagnetic resonance (FMR) spectra corresponding to the different magnetic states of the magnetic nanonetwork are used as the output of the physical reservoir. The network pattern of the simulation is retrieved from the SEM micrograph of a self-organized network. b Distribution of the magnetization of the magnetic network at different moments of the magnetic field signal input. Transformation of sine wave input datasets to target waveforms of sawtooth wave (c, d), square wave (e, f), sin2(t) wave (g, h). Future prediction of t  +  15(i, j) and t + 20 (k, l) for a Mackey-Glass chaotic differential time series. In both tasks, the training results of applying the magnetic network reservoir exhibited superior alignment with the target waveform and had a smaller MSE than the training results of bypassing the reservoir. Source data are provided as a Source Data file.

Back to article page