Generating In-Distribution Proxy Graphs for Explaining Graph Neural Networks

Zhuomin Chen, Jiaxing Zhang, Jingchao Ni, Xiaoting Li, Yuchen Bian, Md Mezbahul Islam, Ananda Mondal, Hua Wei, Dongsheng Luo
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:7712-7730, 2024.

Abstract

Graph Neural Networks (GNNs) have become a building block in graph data processing, with wide applications in critical domains. The growing needs to deploy GNNs in high-stakes applications necessitate explainability for users in the decision-making processes. A popular paradigm for the explainability of GNNs is to identify explainable subgraphs by comparing their labels with the ones of original graphs. This task is challenging due to the substantial distributional shift from the original graphs in the training set to the set of explainable subgraphs, which prevents accurate prediction of labels with the subgraphs. To address it, in this paper, we propose a novel method that generates proxy graphs for explainable subgraphs that are in the distribution of training data. We introduce a parametric method that employs graph generators to produce proxy graphs. A new training objective based on information theory is designed to ensure that proxy graphs not only adhere to the distribution of training data but also preserve explanatory factors. Such generated proxy graphs can be reliably used to approximate the predictions of the labels of explainable subgraphs. Empirical evaluations across various datasets demonstrate our method achieves more accurate explanations for GNNs.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-chen24bd, title = {Generating In-Distribution Proxy Graphs for Explaining Graph Neural Networks}, author = {Chen, Zhuomin and Zhang, Jiaxing and Ni, Jingchao and Li, Xiaoting and Bian, Yuchen and Islam, Md Mezbahul and Mondal, Ananda and Wei, Hua and Luo, Dongsheng}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {7712--7730}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/chen24bd/chen24bd.pdf}, url = {https://proceedings.mlr.press/v235/chen24bd.html}, abstract = {Graph Neural Networks (GNNs) have become a building block in graph data processing, with wide applications in critical domains. The growing needs to deploy GNNs in high-stakes applications necessitate explainability for users in the decision-making processes. A popular paradigm for the explainability of GNNs is to identify explainable subgraphs by comparing their labels with the ones of original graphs. This task is challenging due to the substantial distributional shift from the original graphs in the training set to the set of explainable subgraphs, which prevents accurate prediction of labels with the subgraphs. To address it, in this paper, we propose a novel method that generates proxy graphs for explainable subgraphs that are in the distribution of training data. We introduce a parametric method that employs graph generators to produce proxy graphs. A new training objective based on information theory is designed to ensure that proxy graphs not only adhere to the distribution of training data but also preserve explanatory factors. Such generated proxy graphs can be reliably used to approximate the predictions of the labels of explainable subgraphs. Empirical evaluations across various datasets demonstrate our method achieves more accurate explanations for GNNs.} }
Endnote
%0 Conference Paper %T Generating In-Distribution Proxy Graphs for Explaining Graph Neural Networks %A Zhuomin Chen %A Jiaxing Zhang %A Jingchao Ni %A Xiaoting Li %A Yuchen Bian %A Md Mezbahul Islam %A Ananda Mondal %A Hua Wei %A Dongsheng Luo %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-chen24bd %I PMLR %P 7712--7730 %U https://proceedings.mlr.press/v235/chen24bd.html %V 235 %X Graph Neural Networks (GNNs) have become a building block in graph data processing, with wide applications in critical domains. The growing needs to deploy GNNs in high-stakes applications necessitate explainability for users in the decision-making processes. A popular paradigm for the explainability of GNNs is to identify explainable subgraphs by comparing their labels with the ones of original graphs. This task is challenging due to the substantial distributional shift from the original graphs in the training set to the set of explainable subgraphs, which prevents accurate prediction of labels with the subgraphs. To address it, in this paper, we propose a novel method that generates proxy graphs for explainable subgraphs that are in the distribution of training data. We introduce a parametric method that employs graph generators to produce proxy graphs. A new training objective based on information theory is designed to ensure that proxy graphs not only adhere to the distribution of training data but also preserve explanatory factors. Such generated proxy graphs can be reliably used to approximate the predictions of the labels of explainable subgraphs. Empirical evaluations across various datasets demonstrate our method achieves more accurate explanations for GNNs.
APA
Chen, Z., Zhang, J., Ni, J., Li, X., Bian, Y., Islam, M.M., Mondal, A., Wei, H. & Luo, D.. (2024). Generating In-Distribution Proxy Graphs for Explaining Graph Neural Networks. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:7712-7730 Available from https://proceedings.mlr.press/v235/chen24bd.html.

Related Material