Code and Data for the ICAPS 2022 paper "Learning General Optimal Policies with Graph Neural Networks: Expressive Power, Transparency, and Limits"

This archive contains three files:

• file 'code.zip' contains the code used to train the models;
• file 'data.zip' contains the datasets used to train and validate models, and formulas used to test if features are learned by trained models;
• file 'models.zip' contains the final trained models together with logs from both training and evaluation, and the linear transformation between learned and hand-crafted features.