Computational prediction of crystal materials properties can help to do large-scale in-silicon screening. Recent studies of material informatics have focused on expert design of multi-dimensional interpretable material descriptors/features. However, successes of deep learning such as Convolutional Neural Networks (CNN) in image recognition and speech recognition have demonstrated their automated feature extraction capability to effectively capture the characteristics of the data and achieve superior prediction performance. Here, we propose CNN-OFM-Magpie, a CNN model with OFM (Orbital-field Matrix) and Magpie descriptors to predict the formation energy of 4030 crystal material by exploiting the complementarity of two-dimensional OFM features and Magpie features. Experiments showed that our method achieves better performance than conventional regression algorithms such as support vector machines and Random Forest. It is also better than CNN models using only the OFM features, the Magpie features, or the basic one-hot encodings. This demonstrates the advantages of CNN and feature fusion for materials property prediction. Finally, we visualized the two-dimensional OFM descriptors and analyzed the features extracted by the CNN to obtain greater understanding of the CNN-OFM model.
Digital Object Identifier (DOI)
Published in Crystals, Volume 9, Issue 4, 2019, pages 191-.
© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Cao, Z., Dan, Y., Xiong, Z., Niu, C., Li, X., Qian, S., & Hu, J. (2019). Convolutional Neural Networks for Crystal Material Property Prediction Using Hybrid Orbital-Field Matrix and Magpie Descriptors. Crystals, 9(4), 191.