Faithful Explanations for Graph Classification using Logic

Official Repository of the ECML-PKDD 2025 Paper "Faithful Explanations for Graph Classification using Logic"

Model Training

To train a single model: python train_baseline.py

To perform a grid search: python optimize_baseline.py

Explanations

To replicate experiments and explain the model with all the methods in the paper use the notebook PostHoc.ipynb to save explanations and Results.ipynb to gather the results.