Fig. 8. Box plot representation of the MLPs' learning performance across the compositional space of MD17 database. All ML models were trained using energies and forces information. Each box plot represents the distribution of the root-mean squared error (RMSE) for the total energies of all molecules in the MD17 database calculated with respect to the true labels of a test set consisting of 20 000 geometries. The mean value of RMSE calculated for all different molecules is represented as white dots, while black markers represent outliers.