Accurate property prediction for structures without knowledge of exact atomic positions is achieved with a machine-learning model. The model is formulated as neural-message passing on Voronoi graphs using a new… Click to show full abstract
Accurate property prediction for structures without knowledge of exact atomic positions is achieved with a machine-learning model. The model is formulated as neural-message passing on Voronoi graphs using a new method for the encoding of local symmetry information.
               
Click one of the above tabs to view related content.