# DynGFN: Bayesian Dynamic Causal Discovery using Generative Flow Networks [](https://arxiv.org/abs/2302.04178) [](https://github.com/pre-commit/pre-commit) [](https://pytorch.org/get-started/locally/) [](https://pytorchlightning.ai/) [](https://hydra.cc/) [](https://black.readthedocs.io/en/stable/) [](https://github.com/pre-commit/pre-commit) [](https://github.com/lazaratan/dyn-gfn/blob/main/LICENSE)

<img src="assets/dyngfn.jpg" title="DynGFN Oerview" style="display: inline-block; margin: 0 auto; max-width: 300px">

Description

Generative Flow Networks (GFlowNets) are an approach for learning generative models over discrete spaces. GFlowNets learn a stochastic policy $P_F(\tau)$ to sequentially sample an object $\mathbf{x}$ (e.g. a graph) from a discrete space $\mathcal{X}$. We propose a general framework for cyclic causal discovery of dynamic systems and use GFlowNets to learn dynamic causal structure. See this http link for the paper.

DynGFN is method for learning cyclic causal structure of dynamic systems from observational dynamic data. DynGFN consists of three components: (1) A graph sampler that samples graphical structure encoding the causal dependencies between variables, (2) A class of structural equation modules that models the functional relationships between the observed variables, indexed by parameters $\theta$, and (3) A hyper-network architecture that outputs the parameters $\theta$ of the structural equations as a function of the sampled graph, i.e. $\theta = h_{\phi}(G)$.

If you find this code useful in your research, please cite the following paper (expand for BibTeX):

How to run

Install dependencies

# clone project
git clone https://github.com/lazaratan/dyn-gfn.git
cd dyn-gfn

# [OPTIONAL] create conda environment
conda create -n myenv python=3.8
conda activate myenv

# install pytorch according to instructions
# https://pytorch.org/get-started/

# install requirements
pip install -r requirements.txt

Train model with default configuration

# train on CPU
python train.py

# train on GPU
python train.py trainer=gpu

Train model with chosen experiment configuration from configs/experiment/

python train.py experiment=experiment_name.yaml

You can override any parameter from command line like this

python train.py trainer.max_epochs=20 datamodule.batch_size=64

Project Structure

The directory structure of new project looks like this:

├── configs                   <- Hydra configuration files
│   ├── callbacks                <- Callbacks configs
│   ├── datamodule               <- Datamodule configs
│   ├── debug                    <- Debugging configs
│   ├── experiment               <- Experiment configs
│   ├── hparams_search           <- Hyperparameter search configs
│   ├── local                    <- Local configs
│   ├── log_dir                  <- Logging directory configs
│   ├── logger                   <- Logger configs
│   ├── model                    <- Model configs
│   ├── trainer                  <- Trainer configs
│   │
│   ├── test.yaml             <- Main config for testing
│   └── train.yaml            <- Main config for training
│
├── data                   <- Project data
│
├── logs                   <- Logs generated by Hydra and PyTorch Lightning loggers
│
├── notebooks              <- Jupyter notebooks. Naming convention is a number (for ordering),
│                             the creator's initials, and a short `-` delimited description,
│                             e.g. `1.0-jqp-initial-data-exploration.ipynb`.
│
├── scripts                <- Shell scripts
│
├── src                    <- Source code
│   ├── datamodules              <- Lightning datamodules
│   ├── models                   <- Lightning models
│   ├── utils                    <- Utility scripts
│   ├── vendor                   <- Third party code that cannot be installed using PIP/Conda
│   │
│   ├── testing_pipeline.py
│   └── training_pipeline.py
│
├── tests                  <- Tests of any kind
│   ├── helpers                  <- A couple of testing utilities
│   ├── shell                    <- Shell/command based tests
│   └── unit                     <- Unit tests
│
├── test.py               <- Run testing
├── train.py              <- Run training
│
├── .env.example              <- Template of the file for storing private environment variables
├── .gitignore                <- List of files/folders ignored by git
├── .pre-commit-config.yaml   <- Configuration of pre-commit hooks for code formatting
├── requirements.txt          <- File for installing python dependencies
├── setup.cfg                 <- Configuration of linters and pytest
└── README.md