This repository contains the implementation of MEG (IJCNN 2021).
We assume miniconda (or anaconda) to be installed.
Run the following commands:
source setup/install.sh [cpu | cu92 | cu101 | cu102]
conda activate meg
Train the DGN to be explained by running:
python train_dgn.py [tox21 | esol] <experiment_name>
To generate counterfactual explanations for a specific sample, run:
python train_meg.py [tox21 | esol] <experiment_name> --sample <INTEGER>
Results will be saved at runs/<dataset_name>/<experiment_name>/meg_output.
@conference{numeroso2021,
title={MEG: Generating Molecular Counterfactual Explanations for Deep Graph Networks},
author={Danilo Numeroso and Davide Bacciu},
year={2021},
date={2021-07-18},
booktitle={Proceedings of the International Joint Conference on Neural Networks (IJCNN 2021)},
organization={IEEE},
keywords={deep learning for graphs, explainable AI, graph data, structured data processing},
pubstate={forthcoming},
tppubtype={conference}
}