Figure 11From: Adaptive neural-domain refinement for solving time-dependent differential equationsError comparison, (blue/solid) numerical error, (black/dotted) infinity norm, (orange/marked) training loss, (green/solid) verification lossBack to article page