Data set for Research on Single Soliton Transmission Characteristics of One-Dimensional Nonlinear Schrödinger Equation and Nonlinear Coefficient Calibration Based on Physics-Informed Neural Networks

Contents:Figure 2 (Fig_obs_scatter_x): distribution of observation points used for training.Figure 3 (Fig_beta_reconstruction): comparison of true β(x) and PINN-reconstructed β(x).Figure 4 (Fig_beta_abs_error_curve): spatial distribution of the absolute error |β_PINN(x) – β_true(x)|.Figure 5 (Fig_beta_prime): spatial derivative β'(x) obtained via automatic differentiation.Figure 6 (Fig_cloud_intensity_true): SSFM benchmark intensity |ψ_true(x,t)|² in the (x,t) plane.Figure 7 (Fig_cloud_intensity_pred): PINN-predicted intensity |ψ_pred(x,t)|².Figure 8 (Fig_cloud_intensity_abs_error): absolute error of intensity in space-time.Figure 9 (Fig_cloud_phase_abs_error): absolute error of the phase arg(ψ) in space-time.Figure 10 (Fig_phase_slices): phase profiles at selected time slices (true vs. PINN).Figure 11 (Fig_soliton_center_trajectory): soliton center-of-mass trajectory x_c(t) (true vs. PINN).Figure 12 (Fig_soliton_peak_amplitude): temporal evolution of the peak amplitude.Figure 13 (Fig_soliton_fwhm): temporal evolution of the intensity full-width at half-maximum (FWHM).Figure 14 (Fig_energy_conservation): conserved mass/power M(t) = ∫|ψ|² dx over time.Figure 15 (Fig_energy_relative_deviation): relative drift δ_M(t) = |M(t)–M(0)|/M(0).Figure 16 (Fig_momentum_conservation): momentum P(t) computed from PINN and SSFM.Figure 17 (Fig_training_loss): total loss curve (log scale) during PINN training.Figure 18 (Fig_beta_param_trace): evolution of β parameters (A, c, σ) during training.Figure 19 (Fig_beta_grad_norm): gradient norm with respect to β network parameters.All CSV files are formatted with clear headers and can be used directly to reproduce the corresponding plots in the manuscript.