Tree approximate message passing (Tree-AMP)
Implements gaussian expectation propagation for any tree-like probabilistic graphical model.
Documentation website: https://sphinxteam.github.io/tramp.docs
Requirements
- python>=3.6
- numpy/pandas/scipy/matplotlib
- networkx==1.11
- daft
Warning Currently the package does not support networkx 2.xx and will throw errors. We plan to upgrade to networkx 2.xx at some point.
Install
To install the package, go to the folder where setup.py is located and run:
pip install .
or if you want to install in development mode (changes to the repository will immediately affect the installed package without needing to re-install):
pip install --editable .
To install the package on a remote machine directly from the github repo:
pip install git+https://github.com/sphinxteam/tramp.git
See installing from sources for more details. In both cases, the necessary requirements should be automatically installed.
Citation
The package is presented in arXiv:2004.01571. To cite this work, please use:
@misc{Baker2021TreeAMP,
title={Tree-AMP: Compositional Inference with Tree Approximate Message Passing},
author={Antoine Baker and Benjamin Aubin and Florent Krzakala and Lenka Zdeborová},
year={2021},
eprint={2004.01571},
archivePrefix={arXiv},
primaryClass={stat.ML}
}
Examples
See the corresponding gallery in the documentation website.
Acknowledgments
Both the SPHINX team and the SMILE team acknowledge funding from:
