Abstract
Graph neural networks (GNNs) have been used previously for identifying new crystalline materials. However, geometric structure is not usually taken into consideration, or only partially. Here, we develop a geometric-information-enhanced crystal graph neural network (GeoCGNN) to predict the properties of crystalline materials. By considering the distance vector between each node and its neighbors, our model can learn full topological and spatial geometric structure information. Furthermore, we incorporate an effective method based on the mixed basis functions to encode the geometric information into our model, which outperforms other GNN methods in a variety of databases. For example, for predicting formation energy our model is 25.6%, 14.3% and 35.7% more accurate than CGCNN, MEGNet and iCGCNN models, respectively. For band gap, our model outperforms CGCNN by 27.6% and MEGNet by 12.4%.
Highlights
Graph neural networks (GNNs) have been used previously for identifying new crystalline materials
The Crystal Graph Convolutional Neural Network (CGCNN)[19] chose the distance between atoms to represent the edges in the crystal graph
We propose a GNN model to accurately predict properties for any crystalline materials, which is invariant to global 3D rotations, translations, and node permutations
Summary
Graph neural networks (GNNs) have been used previously for identifying new crystalline materials. We develop a geometric-information-enhanced crystal graph neural network (GeoCGNN) to predict the properties of crystalline materials. The utilization of attention masks is an excellent way to encode the geometrical structure among atoms Since it is just beginning of the use in the field of crystal material prediction, there is still room for development. Our main contributions can be summarized as below: (1) We propose a message passing neural network (MPNN)31-based GNN architecture with high prediction accuracy for the formation energy and band gap; and (2) we provide an effective way to encode the local geometrical information in the process of aggregation, that is, an attention mask composed by Gaussian radial basis and plane waves
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
