ignavierng / notears-admm

Towards Federated Bayesian Network Structure Learning with Continuous Optimization

Federated Structure Learning with Continuous Optimization

This repository contains an implementation of the structure learning methods described in "Towards Federated Bayesian Network Structure Learning with Continuous Optimization".

If you find it useful, please consider citing:

@inproceedings{Ng2022federated,
  author = {Ng, Ignavier and Zhang, Kun},
  title = {Towards Federated Bayesian Network Structure Learning with Continuous Optimization},
  booktitle = {International Conference on Artificial Intelligence and Statistics},
  year = {2022},
}

Requirements

• Python 3.6+
• numpy
• scipy
• python-igraph
• torch

Running NOTEARS(-MLP) with ADMM

• See examples/linear.ipynb and examples/nonlinear.ipynb for a demo in the linear and nonlinear cases, respectively.

Acknowledgments

• A large part of the code, including some helper functions, is obtained and modified from the implementation of NOTEARS, and we are grateful to the authors of NOTEARS for releasing their code.
• The code to post-process the output is modified and obtained from the implementation of GOLEM.