Graphtester documentation#
Graphtester is a Python package for comprehensive analysis of the theoretical capabilities of GNNs for various datasets, tasks, and scores, using and extension of the Weisfeiler-Lehman framework.
You can read more about the theoretical background in our paper: https://arxiv.org/abs/2306.17482
If you use Graphtester in your research, please cite:
@inproceedings{akbiyik2023graphtester,
title = {{Graphtester: Exploring Theoretical Boundaries of GNNs on Graph Datasets}},
author = {Eren Akbiyik and Florian Grötschla and Béni Egressy and Roger Wattenhofer},
year = 2023,
month = {July},
booktitle = {{Data-centric Machine Learning Research (DMLR) Workshop at ICML 2023, Honolulu, Hawaii}}
}