Fig. 6.

Data applied for training (red circles) with doses $d$ = 1 and $d$ = 10 and simulation (blue triangles) with $d$ = 5 were generated with a two-compartment IV ODE model, see panel A. Furthermore, fits of the trained NODE Eq. (6) for the two training doses (solid red line) and simulation (dashed blue line) are presented. In panel B, the corresponding derivative versus state plot indicates that the NODE learned the two-compartmental behavior from the training data