Predicting quantum materials properties using novel faithful machine learning

Machine learning (ML) is accelerating the progress of materials prediction and classification, with particular success in crystallographic graph neural network designs. While classical ML methods remain accessible, advanced deep networks are still challenging to build and train. We introduce two new adaptations, and refine two existing ML networks, for generic crystalline quantum materials property prediction and optimization. These new models achieve strong performance in topological classification, predicting band gaps, magnetic classifications, formation energies, and symmetry groups, while the crystallographic graph neural network acheives state-of-the-art predictions for topological materials. All networks extend to provide general quantum crystalline materials property predictions. To support this, full implementations and automated methods for data handling and materials predictions are provided, facilitating the use of deep ML methods in quantum materials science. Finally, dataset error rates are analyzed using an ensemble model to identify highly atypical materials for further investigation.