UnlockNN
UnlockNN: Uncertainty quantification for neural network models of chemical systems - Published in JOSS (2022)
Science Score: 93.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
○CITATION.cff file
-
✓codemeta.json file
Found codemeta.json file -
✓.zenodo.json file
Found .zenodo.json file -
✓DOI references
Found 1 DOI reference(s) in JOSS metadata -
✓Academic publication links
Links to: joss.theoj.org -
○Committers with academic emails
-
○Institutional organization owner
-
✓JOSS paper metadata
Published in Journal of Open Source Software
Keywords
Scientific Fields
Repository
A Python package for adding uncertainties to neural network models of chemical systems.
Basic Info
Statistics
- Stars: 26
- Watchers: 5
- Forks: 7
- Open Issues: 7
- Releases: 5
Topics
Metadata Files
README.md
unlockNN
A Python package for interpreting and extracting uncertainties in neural network models of chemical systems based upon Gaussian processes.
Statement of need
Neural networks (NNs) are powerful tools for materials property prediciton (MPP) based on structural information. After training, they offer a cheaper alternative to density function theory (DFT) and are therefore promising for high throughput screening of materials. However, most current implementations of NNs for MPP lack uncertainty quantifiers. Knowledge of the certainty in an estimate is particularly important for machine learning models, as the reliability of a prediction depends on the existence of functionally similar structures in the training dataset, which cannot be readily determined.
UnlockNN contains utilities for adding uncertainty quantification to Keras-based models. This is achieved by replacing the last layer of the model with a variational Gaussian process (VGP), a modification of a Gaussian process that improves scalability to larger data sets. The caveat is that the modified model must undergo further training in order to calibrate the uncertainty quantifier; however, this typically only requires a small number of training iterations.
UnlockNN also contains a specific configuration for adding uncertainty quantification to MEGNet: a powerful graph NN model for predicting properties of molecules and crystals.
Installation
The package can be installed by cloning this repository and building it using either anaconda or pip, or it can be downloaded directly from PyPi.
To install from PyPi, run pip install unlockNN.
To install from source:
bash
git clone https://github.com/a-ws-m/unlockNN.git
cd unlockNN
conda env create -f environment.yml # Optional: create a virtual environment with conda
pip install .
The dev_environment.yml contains additional dependencies for development, testing and building documentation.
It can be installed using conda env create -f dev_environment.yml.
Documentation
Full documentation is available for the project here.
Benchmarks
Benchmark details and results can be found in the benchmarking subdirectory.
License and attribution
Code licensed under the MIT License.
Development notes
Reporting issues
Please use the Issue tracker to report bugs in the software, suggest feature improvements, or seek support.
Contributing to unlockNN
Contributions are very welcome as we look to make unlockNN more flexible and efficient. Please use the Fork and Pull workflow to make contributions and follow the contribution guidelines:
- Use the environment defined in
dev_environment.yml. This installsblack, the formatter used for this project, as well as utilities for building documentation (sphinxand theinsegeltheme), enabling the testing suite (pytestandpytest-cov) and publishing to PyPi (build, but this will be handled by the package maintainer). - Use
blackto format all Python files that you edit:black {edited_file.py}orpython -m black {edited_file.py}. - Write tests for new features in the appropriate directory. Run tests using
pytest tests/, or optionally withpytest --cov=unlocknn tests/to generate coverage on the fly. - After testing that
pytestworks for your current environment, runtoxin the root directory of the project to check that all versions of Python are compatible. - Use Google-style Docstrings. Check docstrings with
pydocstyle:pydocstyle {edited_file.py}. - Check docstrings are valid Sphinx RST and that the documentation compiles without errors: in the
docsdirectory, runmake html. - Feel free to clean up others' code as you go along.
List of developers
Contributors to unlockNN:
Huge thanks to Keith Butler, Aron Walsh and Kazuki Morita for supervising the project at its inception and for their immense support.
Owner
- Name: Alex Moriarty
- Login: a-ws-m
- Kind: user
- Repositories: 8
- Profile: https://github.com/a-ws-m
JOSS Publication
UnlockNN: Uncertainty quantification for neural network models of chemical systems
Authors
Department of Materials, Imperial College London, London, UK
Department of Materials, Imperial College London, London, UK
SciML, STFC Scientific Computing Division, Rutherford Appleton Laboratories, UK
Department of Materials, Imperial College London, London, UK, Department of Materials Science and Engineering, Yonsei University, Seoul, Korea
Editor
Øystein Sørensen
Tags
graph neural networks uncertainty quantification machine learning material science chemistryGitHub Events
Total
- Watch event: 4
- Fork event: 1
Last Year
- Watch event: 4
- Fork event: 1
Committers
Last synced: 7 months ago
Top Committers
| Name | Commits | |
|---|---|---|
| a-ws-m | a****4@g****m | 612 |
| Øystein Sørensen | o****n@h****m | 2 |
Issues and Pull Requests
Last synced: 6 months ago
All Time
- Total issues: 19
- Total pull requests: 22
- Average time to close issues: 2 months
- Average time to close pull requests: 2 days
- Total issue authors: 4
- Total pull request authors: 2
- Average comments per issue: 2.05
- Average comments per pull request: 0.0
- Merged pull requests: 22
- Bot issues: 0
- Bot pull requests: 0
Past Year
- Issues: 0
- Pull requests: 0
- Average time to close issues: N/A
- Average time to close pull requests: N/A
- Issue authors: 0
- Pull request authors: 0
- Average comments per issue: 0
- Average comments per pull request: 0
- Merged pull requests: 0
- Bot issues: 0
- Bot pull requests: 0
Top Authors
Issue Authors
- pmeier (14)
- sgbaird (3)
- Het-Shah (1)
- matthewmelissa049 (1)
Pull Request Authors
- a-ws-m (20)
- osorensen (2)
Top Labels
Issue Labels
Pull Request Labels
Packages
- Total packages: 1
-
Total downloads:
- pypi 21 last-month
- Total dependent packages: 0
- Total dependent repositories: 1
- Total versions: 3
- Total maintainers: 1
pypi.org: unlocknn
Uncertainty quantification for neural network models of chemical systems.
- Homepage: https://github.com/a-ws-m/unlockNN
- Documentation: https://unlocknn.readthedocs.io/
- License: MIT
-
Latest release: 2.0.3
published over 3 years ago
Rankings
Maintainers (1)
Dependencies
- megnet *
- pyarrow *
- pymatgen ==2021.2.8
- requests *
- tensorflow *
- tensorflow-probability *
- megnet >=1.1.4
- numpy <=1.19.5
- pyarrow >=1.0.1
- pymatgen <=2021.2.8
- requests *
- tensorflow >=2.5
- tensorflow-probability >=0.10.1
- typish *
- megnet >=1.1.4
- numpy <=1.19.5
- pyarrow >=1.0.1
- pymatgen <=2021.2.8
- requests *
- tensorflow >=2.2
- tensorflow-probability >=0.10.1
- typish *