AUTHOR=Viatkin Dmitry , Garcia-Zapirain Begonya , Méndez-Zorrilla Amaia , Zakharov Maxim 

TITLE=Deep Learning Approach for Prediction of Critical Temperature of Superconductor Materials Described by Chemical Formulas

JOURNAL=Frontiers in Materials

VOLUME=Volume 8 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/materials/articles/10.3389/fmats.2021.714752

DOI=10.3389/fmats.2021.714752

ISSN=2296-8016

ABSTRACT=This paper proposes a novel neural network architecture and their ensembles to predict critical superconductivity temperature of materials based on their chemical formula. The research de-scribes the methods and process of extracting data from the chemical formula and preparing these extracted data for use in neural network training using TensorFlow. In our approach, re-current neural networks are used including long short-term memory layers and neural networks based on one-dimensional convolution layers for data analysis. The proposed model is an en-semble of pre-trained neural network architectures for prediction of the critical temperature of superconductors based on their chemical formula. The architecture of seven pre-trained neural networks is based on the long short-term memory layers and convolution layers. In the final ensemble, six neural networks are used: one network based on LSTM and 4 based on convolu-tional neural networks and one embedding ensemble of convolution neural networks. LSTM neural network and convolution neural network were trained in 300 epochs. Ensemble of mod-els were trained in 20 epochs. All neural networks are trained in two stages. At both stages, the optimizer Adam was used. In the first stage, training was carried out by the function of losses Mean Absolute Error (MAE) with the value of optimizer learning rate equal to 0.001. In the sec-ond stage, the previously trained model was trained by the function of losses Mean Squared Er-ror (MSE) with a learning rate equal to 0.0001. Final ensemble is trained with a learning rate equal to 0.00001. The final ensemble model has the following accuracy values: Mean Absolute Error (MAE) is 4.068, Mean Squared Error (MSE) is 67.272, and Coefficient of determination (R2) is 0.923. The final model can predict the critical temperature for the chemistry formula with an accuracy of 4.068 degrees.