Table 4 Summary of the average RMSE values for the isotropic and the anisotropic Rabi models. The number of input nodes is 96 and the noiseless case for the JC model is shown for comparison.
From: Supervised learning of the Jaynes–Cummings Hamiltonian
Model | Method | \(\epsilon\) | \(\omega _c\) | \(\omega _a\) | \(g\) | \(g'\) | Avg. | Energy |
|---|---|---|---|---|---|---|---|---|
JC | vDNN | 0 | 2.14e-05 | 5.57e-05 | 4.12e-05 | - | 5.01e-05 | 6.13e-04 |
Isotropic Rabi | vDNN | 0 | 6.44e-05 | 2.02e-04 | 1.32e-04 | - | 1.65e-04 | 1.03e-03 |
0.4 | 3.11e-04 | 7.23e-03 | 3.77e-03 | - | 5.27e-03 | 1.63e-02 | ||
0.8 | 6.28e-04 | 1.46e-02 | 7.55e-03 | - | 1.06e-02 | 3.24e-02 | ||
dDNN | 0 | 7.72e-05 | 1.01e-03 | 5.06e-04 | - | 7.43e-04 | 2.60e-03 | |
0.4 | 1.42e-04 | 1.94e-03 | 6.70e-04 | - | 1.29e-03 | 3.87e-03 | ||
0.8 | 2.44e-04 | 3.45e-03 | 1.03e-03 | - | 2.23e-03 | 6.32e-03 | ||
Anisotropic Rabi | vDNN | 0 | 1.11e-03 | 3.70e-03 | 3.94e-03 | 3.58e-03 | 3.84e-03 | 1.86e-02 |
0.4 | 5.10e-03 | 6.14e-02 | 5.40e-02 | 5.31e-02 | 5.62e-02 | 1.13e-01 | ||
0.8 | 1.95e-02 | 1.09e-01 | 9.70e-02 | 9.13e-02 | 1.02e-01 | 3.47e-01 | ||
dDNN | 0 | 1.29e-03 | 2.46e-02 | 2.33e-02 | 2.27e-02 | 2.34e-02 | 3.62e-02 | |
0.4 | 1.41e-03 | 3.08e-02 | 2.82e-02 | 2.76e-02 | 2.88e-02 | 4.01e-02 | ||
0.8 | 1.68e-03 | 4.29e-02 | 3.68e-02 | 3.64e-02 | 3.89e-02 | 4.86e-02 |
