Predicting magnetic anisotropy energies using site-specific spin-orbit coupling energies and machine learning: Application to iron-cobalt nitrides

Accelerating materials discovery using integrated deep machine learning approaches

“…While the g-CGCNN model has been shown to be very useful for accelerating the discovery of novel complex ternary compounds, [25][26][27] the accuracy and efficiency of the CGCNN model can be signicantly improved further if the model is retrained using the data specically targeting the materials being studied. 25,36 For La-Si-P ternary system, we retrained the CGCNN using 228 284 structures and formation energies from the dataset that was prepared to train an ANN-ML interatomic potential for the La-Si-P system (see below and ESI †). We refer to this later CGCNN model as a specic CGCMM (s-CGCNN) model.…”