Fig. 3: Floquet extreme learning for highly nonlinear maps.
From: Electromagnetic wave-based extreme deep learning with nonlinear time-Floquet entanglement
a–c Comparison between ground truth and predicted values for three different nonlinear functions: \({y}_{1}=\alpha \sin (4\pi {\zeta }^{in})(| {\zeta }^{in}| /\pi )\), and y2 = rect(ζin) and \({y}_{3}=\sin (\pi {\zeta }^{in})/(\pi {\zeta }^{in})\), respectively. d–f The corresponding values of root-mean-square error (RMSE) upon increasing the numbers of involved Floquet harmonics at the readout nodes.
