Welcome to ai4materials’s documentation!¶
The current documentation is not actively mantained and thus might not be up-to-date. For the most recent documentation, please visit ai4materials github repository https://github.com/angeloziletti/ai4materials.
ai4materials allows to perform complex analysis of materials science data using machine learning. It also provide functions to pre-process (on parallel processors), save and subsequently load materials science datasets, thus easing the traceability, reproducibility, and prototyping of new models.
ai4materials allows perform crystal-structure classification and analysis, as introduced in:
| [1] | A. Leitherer, A. Ziletti, and L. M. Ghiringhelli, “Robust recognition and exploratory analysis of crystal structures via Bayesian deep learning”, https://arxiv.org/abs/2103.09777 (2021) |
Installation instructions can be found in the ai4materials github repository: https://github.com/angeloziletti/ai4materials.
On the left panel, you can find a few examples that showcase what ai4materials can do.
Moreover, ai4materials can also reproduce results from the following publications:
| [2] | A. Ziletti, D. Kumar, M. Scheffler, and L. M. Ghiringhelli, “Insightful classification of crystal structures using deep learning,” Nature Communications, vol. 9, pp. 2775, 2018. [Link to article] |
| [3] | L. M. Ghiringhelli, J. Vybiral, S. V. Levchenko, C. Draxl, and M. Scheffler, “Big Data of Materials Science: Critical Role of the Descriptor,” Physical Review Letters, vol. 114, no. 10, p. 105503 . [Link to article] |
Code author: Angelo Ziletti <angelo.ziletti@gmail.com>
Contents:
- Installation
- Data preprocessing
- Representing crystal structures: descriptors
- Example: atomic features
- Example: two-dimensional diffraction fingerprint
- Example: three-dimensional diffraction fingerprint
- Submodules
- ai4materials.descriptors.atomic_features module
- ai4materials.descriptors.base_descriptor module
- ai4materials.descriptors.diffraction1d module
- ai4materials.descriptors.diffraction2d module
- ai4materials.descriptors.diffraction3d module
- ai4materials.descriptors.ft_soap_descriptor module
- ai4materials.descriptors.prdf module
- ai4materials.descriptors.quippy_soap_descriptor module
- ai4materials.descriptors.soap_model module
- Module contents
- Creating and loading materials science datasets
- Regression and classification models
- Example regression: LASSO+l0 method
- Example classification: convolutional neural network for crystal-structure classification
- Submodules
- ai4materials.models.clustering module
- ai4materials.models.cnn_architectures module
- ai4materials.models.cnn_nature_comm_ziletti2018 module
- ai4materials.models.cnn_polycrystals module
- ai4materials.models.embedding module
- ai4materials.models.l1_l0 module
- ai4materials.models.sis module
- ai4materials.models.strided_pattern_matching module
- Module contents
- Neural network interpretation
- Visualization
- Utils
- Utils crystals
- Pristine and defective supercell generation
- Example: pristine supercell creation
- Example: defective supercell creation
- Submodules
- ai4materials.utils.unit_conversion module
- ai4materials.utils.utils_binaries module
- ai4materials.utils.utils_config module
- ai4materials.utils.utils_crystals module
- ai4materials.utils.utils_data_retrieval module
- ai4materials.utils.utils_mp module
- ai4materials.utils.utils_parsing module
- ai4materials.utils.utils_plotting module
- ai4materials.utils.utils_vol_data module
- Module contents
- Utils crystals
- Authors
- History