Fig. 2: Formation energies predicted by the ML with CE features (DNN-CE) and DFT using various datasets. | npj Computational Materials

Fig. 2: Formation energies predicted by the ML with CE features (DNN-CE) and DFT using various datasets.

From: Center-environment deep transfer machine learning across crystal structures: from spinel oxides to perovskite oxides

Fig. 2

a Spinel oxides (source domain Modelspinel), b perovskite oxides (target domain Modelperovskite), c direct prediction of perovskite oxides using the source domain model (ModelSP), and d transfer learning model (ModelTL). The results were averaged over 10 runs of the 10-fold cross-validation in the training process.

Back to article page