Sparse representation for machine learning the properties of defects in 2D materials

A non-transitory computer-readable medium comprising:

• instructions for performing a method for predicting at least one property of a crystal of a material, wherein the crystal exhibits at least one point defect, the instructions including:
• sampling structures of materials and point defects;
• providing a neural network, wherein with the neural network:
• receiving as inputs:
• a structure of the material as one of the sampled structures of materials and which is represented by a cloud of defect points, and
• an ideal crystal unit cell structure represented by a global state vector, and
• outputting at least one predicted property of the crystal using a generated set of data,
• wherein within the neural network, the at least one point defect is defined according to:
• a set of coordinates, a type of point defect, and a crystal unit cell structure.

A system for predicting at least one property of a crystal of a material exhibiting at least one point defect, the system comprising:

• a module for sampling structures of materials and the at least one point defect,
• a neural network for predicting the at least one property, and a processor for executing instructions including:
• sampling structures of materials and point defects;
• providing a neural network, wherein with the neural network:
• receiving as inputs:
• a structure of the material as one of the sampled structures of materials and which is represented by a cloud of defect points, and
• an ideal crystal unit cell structure represented by a global state vector, and outputting at least one predicted property of the crystal using a generated set of data;
• wherein within the neural network, the at least one point defect is defined according to:
• a set of coordinates and a type of point defect, and a crystal unit cell structure.