Aditya239233 / GNNExplainer

Code for running experiments and benchmarking on GNNExplainer: Generating Explanations for Graph Neural Networks

GNNExplainer

CZ4071 - Network Science

This repository contains the modified code for the paper GNNExplainer: Generating Explanations for Graph Neural Networks by Rex Ying, Dylan Bourgeois, Jiaxuan You, Marinka Zitnik & Jure Leskovec, presented at NeurIPS 2019.

[Arxiv]

@misc{ying2019gnnexplainer,
    title={GNNExplainer: Generating Explanations for Graph Neural Networks},
    author={Rex Ying and Dylan Bourgeois and Jiaxuan You and Marinka Zitnik and Jure Leskovec},
    year={2019},
    eprint={1903.03894},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}

The code involves expirements to run benchmarks on GNN explainability on synthetic graphs generated using Barabási–Albert model

Results

	BA-shapes (Base)	Ba-shapes (motif)
Accuracy (from Paper)	0.925	0.836
Accuracy (from Replication)	0.971	0.911

Number of Nodes	30	300	3000
Accuracy	0.965	0.971	0.99

Contributors

1. Aditya Chandrasekhar
2. Vincent Yong Wei Jie