Figure 2.

The PINN-evaluated solution for $|h(x,t)|$ from the Schrödinger equation at different timeslices, $t=0,0.39,0.78,1.37$, from left to right for error-free boundary data (top) and corrupted boundary data with (bottom). The additive errors on the boundary data are taken independently from samples of zero mean Gaussian distribution with σ = 0.1. The points marked with the blue cross (x) pointer in the leftmost set of plots indicate the samples on the initial timeslice used to train the PINN.