Table 2 Parameter baselines and ranges for the parameter error task. The parameter sweep modifies individual parameters whilst holding the rest at the baseline setting.
From: Modeling nonlinear oscillator networks using physics-informed hybrid reservoir computing
Parameter | Baseline | Range |
|---|---|---|
\(D_r\) | 300 | [50, 1000] |
Spectral radius | 0.4 | [0.1, 2.0] |
Input scaling | 0.15 | [0.05, 2.00] |
\(\sigma _K\) | 0.05 | [0.004, 0.480] |
\(\sigma _\omega\) | 0.05 | [0.004, 0.480] |
Knowledge ratio | 0.5 | [0.05, 1.00] |
Regularization strength | \(10^{-6}\) | Baseline |
