Autoencoder-based causal discovery based on multivariate post-nonlinear model

This is an official implementation of the multivariate nonlinear causal discovery method using post-nonlinear causal model in the following papers.

• A Multivariate Causal Discovery based on Post-Nonlinear Model, CLeaR 2022
• Estimation Of Post-Nonlinear Causal Models Using Autoencoding Structure, ICASSP 2020

Requirements

• python 3.10 (may work on >=3.8 but tested only on 3.10)
  • numpy
  • scipy
  • torch

Example

./sample.sh

See example.py for more details.

Note: If you use parallelization with max_workers parameter, it is recommended to disable Numpy's multithreading by, for example, export OMP_NUM_THREADS=1.