Magnetocaloric properties predicted by deep learning with compositional features for bulk metallic glasses

A convolutional neural network (CNN) is a widely recognized deep-learning model that performs regression or classification from domain knowledge using a kernel function that extracts specific features of the input data. This study demonstrates the prediction performance of the magnetic entropy changes (−Δ S_mag) and transition temperatures ( T _r) of bulk metallic glasses (BMGs) with magnetocaloric effects using a CNN regression model. Compositional descriptors were generated using the “Magpie” method and were used as input data. The prediction performance was evaluated using the determination coefficient ( R ² score) and root mean square error (RMSE), resulting in R ² score of 0.920 and 0.917 and RMSE of 1.285 J/kgK and 45.619 K for −Δ S_mag and T _r, respectively. In the case study, we further applied the CNN model to unseen datasets for extrapolated −Δ S_mag and T _r prediction and compared the results with the experimental values, resulting in excellent agreement with the reported experimental results.