GNNBoundary: Towards Explaining Graph Neural Networks through the Lens of Decision Boundaries

Xiaoqi Wang¹,   Han-Wei Shen¹,  

¹The Ohio State University,  

ICLR 2024

🚀 Overview

📖 Introduction

While Graph Neural Networks (GNNs) have achieved remarkable performance on various machine learning tasks on graph data, they also raised questions regarding their transparency and interpretability. Recently, there have been extensive research efforts to explain the decision-making process of GNNs. These efforts often focus on explaining why a certain prediction is made for a particular instance, or what discriminative features the GNNs try to detect for each class. However, to the best of our knowledge, there is no existing study on understanding the decision boundaries of GNNs, even though the decision-making process of GNNs is directly determined by the decision boundaries. To bridge this research gap, we propose a model-level explainability method called GNNBoundary, which attempts to gain deeper insights into the decision boundaries of graph classifiers. Specifically, we first develop an algorithm to identify the pairs of classes whose decision regions are adjacent. For an adjacent class pair, the near-boundary graphs between them are effectively generated by optimizing a novel objective function specifically designed for boundary graph generation. Thus, by analyzing the near-boundary graphs, the important characteristics of decision boundaries can be uncovered. To evaluate the efficacy of GNNBoundary, we conduct experiments on both synthetic and public real-world datasets. The results demonstrate that, via the analysis of faithful near-boundary graphs generated by GNNBoundary, we can thoroughly assess the robustness and generalizability of the explained GNNs.

Paper: https://openreview.net/pdf?id=WIzzXCVYiH

🔗 Prior Work - GNNInterpreter

GNNBoundary is inspired by our prior work, GNNInterpreter, which aims to explain the high-level decision-making process of GNNs. Please check this repository for more details.

Paper: https://openreview.net/forum?id=rqq6Dh8t4d

🔥 How to use

Notebooks

• gnnboundary_collab.ipynb contains the demo for the COLLAB dataset experiment in the paper.
• gnnboundary_enzymes.ipynb contains the demo for the ENZYME dataset experiment in the paper.
• gnnboundary_motif.ipynb contains the demo for the Motif dataset experiment in the paper.
• model_training.ipynb contains the demo for GNN classifier training.

Model Checkpoints

• You can find the GNN classifier checkpoints in the ckpts folder.
• See model_training.ipynb for how to load the model checkpoints.

Datasets

• Here's the link for downloading the processed datasets.
• After downloading the datasets zip, please unzip it in the root folder.

Environment

Codes in this repo have been tested on python3.10 + pytorch2.1 + pyg2.5.

To reproduce the exact python environment, please run:

conda create -n gnnboundary poetry jupyter
conda activate gnnboundary
poetry install
ipython kernel install --user --name=gnnboundary --display-name="GNNBoundary"

Note: In case poetry fails to install the dependencies, you can manually install them using pip:

pip install -r requirements.txt

🖼️ Demo

🖊️ Citation

If you used our code or find our work useful in your research, please consider citing:

@inproceedings{wang2024gnnboundary,
title={{GNNB}oundary: Towards Explaining Graph Neural Networks through the Lens of Decision Boundaries},
author={Xiaoqi Wang and Han Wei Shen},
booktitle={The Twelfth International Conference on Learning Representations},
year={2024},
url={https://openreview.net/forum?id=WIzzXCVYiH}
}

🙏 Acknowledgement

The work was supported in part by the US Department of Energy SciDAC program DE-SC0021360, National Science Foundation Division of Information and Intelligent Systems IIS-1955764, and National Science Foundation Office of Advanced Cyberinfrastructure OAC-2112606.