This is a repository for using semi-supervised learning to classify neutron diffraction data. The work is described in a paper currently under review

Files included:

• Data
  • BravaisLattice_Data.pt: A binary file that contains 10 random simulated powder diffraction patterns from the testing dataset. Each piece of data is labeled with the correct Bravais Lattice.
  • SpaceGroup_Data.pt: A binary file that contains 10 random simulated powder diffraction patterns from the testing dataset. Each piece of data is labeled with the correct Space Group.
  • mapping.pt: A binary file that is used to convert categorical labels to text.
  • Data.py: Defines the DiffractionDataset class, a Pytorch Dataset that can use the BravaisLatticeData.pt and the SpaceGroupData.pt files.
  • ICSD_IDs.txt: A text file containing the ICSD IDs used in this work.
• Models
  • Generator.py: Defines the Generator model used in the Semi-Supervised Generative Adversarial Network (SGAN).
  • ResNet.py: Defines the ResNet classifier that is used as a supervised classifier and as the Discriminator of the SGAN.
  • BravaisModels.pth: Supervised and Semi-supervised Bravais Lattice classifiers
  • SpaceGroupModels.pth: Supervised and Semi-supervised space group classifiers
• Notebooks
  • PlotData.ipynb: Loads data from the BravaisLattice_Data.pt and SpaceGroup_Data.pt files as a 2θ vs. normalized intensity graph.
  • LoadBravaisLatticeModels.ipynb: Evaluates the accuracy of the supervised and semi-supervised Bravais Lattice models.
  • LoadSpaceGroupModels.ipynb: Evaluates the accuracy of the supervised and semi-supervised space group models. This notebook also includes a comparison of the top-5 accuracies of each model.