Fig. 4: Resilience inference performance on synthetic networked systems.

We evaluate resilience inference performance of ResInf along with other baseline methods on synthetic network datasets. These datasets include mixed types of perturbed networks of mutualistic (a), gene regulatory (b), and neuronal (c) dynamics, respectively (Supplementary Table 2). The parameter of each dynamics is the setting outlined in index 1 of Supplementary Table 4. The dataset of each dynamics contains 2000 samples, which are partitioned into training, validation, and test sets at a ratio of 8:1:1. For the mutualistic, gene regulatory, and neuronal dynamics, we consider 10, 5, and 11 different initial conditions, respectively. Box plots depict the median (central line) of F1-scores (n = 35 with different random seeds), the first and third quartiles (box), whiskers extending to 1.5 times the interquartile range from the first and third quartiles, respectively, and outliers are represented as individual points.