XRDidentifier

Pytorch implementation of XRD spectral identification from COD database.
Details is written in the NeurIPS ML4PS workshop paper (https://ml4physicalsciences.github.io/2021/files/NeurIPS_ML4PS_2021_9.pdf).
Blog explains this codes in a more intuitive way. (https://medium.com/towards-data-science/automatic-spectral-identification-using-deep-metric-learning-with-1d-regnet-and-adacos-8b7fb36f2d5f)
Please consider citing my paper if this is helpful for your paper (see the bottom).

Features

expert model

1D-CNN (1D-RegNet) + Hierarchical Deep metric learning (AdaCos + Angular Penalty Softmax Loss)

mixture of experts

73 expert models tailered to general chemical elements with sparsely-gated layer

data augmentation

Physics-informed data augmentation

Requirements

• Python 3.6
• PyTorch 1.4
• pymatgen
• scikit-learn

Dataset Construction

In the paper, I used ICSD dataset, but it is forbidden to redistribute the CIFs followed by their license. I will write the CIF dataset construction method using COD instead.

1. download cif files from COD

Go to the COD homepage, search and download the cif URL list.
http://www.crystallography.net/cod/search.html

python3 download_cif_from_cod.py --input ./COD-selection.txt --output ./cif

2. convert cif into XRD spectra

First, check the cif files. (some files are broken or physically meaningless)

python3 read_cif.py --input ./cif --output ./lithium_datasets.pkl

lithium_datasets.pkl will be created.

Second, convert the checked results into XRD spectra database.

python3 convertXRDspectra.py --input ./lithium_datasets.pkl --batch 8 --n_aug 5

XRD_epoch5.pkl will be created.

Train expert models

python3 train_expert.py --input ./XRD_epoch5.pkl --output learning_curve.csv --batch 16 --n_epoch 100

Output data

• Trained model -> regnet1d_adacos_epoch100.pt
• Learning curve -> learning_curve.csv
• Correspondence between numerical int label and crystal names -> material_labels.csv

(Under construction) Train Mixture-of-Experts model

The below is not ready currently. But if the dataset is not such large, the expert model should work.

Cite as

Please cite this work as

@article{adachimixture,
  title={Mixture-of-Experts Ensemble with Hierarchical Deep Metric Learning for Spectroscopic Identification},
  journal={Advances in Neural Information Processing Systems 34 (NeurIPS 2021) Workshop: Machine Learning and the Physical Science},
  year={2021}
}

Citation

Papers

• AdaCos: https://arxiv.org/abs/1905.00292
• 1D-RegNet: https://arxiv.org/abs/2008.04063
• Physics-informed data augmentation: https://arxiv.org/abs/1811.08425v2
• Sparsely-gated layer: https://arxiv.org/abs/1701.06538

Implementation

• AdaCos: https://github.com/4uiiurz1/pytorch-adacos/blob/master/metrics.py
• 1D-RegNet: https://github.com/hsd1503/resnet1d
• Physics-informed data augmentation: https://github.com/PV-Lab/autoXRD
• Top k accuracy: https://gist.github.com/weiaicunzai/2a5ae6eac6712c70bde0630f3e76b77b
• Angular Penalty Softmax Loss: https://github.com/cvqluu/Angular-Penalty-Softmax-Losses-Pytorch
• Sparsely-gated layer: https://github.com/davidmrau/mixture-of-experts