Fig. 5: Accuracy of the reservoir computer on noisy time series.
From: Robust neural networks using stochastic resonance neurons
Accuracy of the reservoir computer when training on (a) noisy Mackey-Glass (MG) and (b noisy Rössler sequences with various signal-to-noise ratios (SNR) in the training sequences. Solid lines represent performance using stochastic resonance (SR), while dashed lines correspond to the standard sigmoid function.
