Federated Self-Explaining GNNs with Anti-shortcut Augmentations

Linan Yue, Qi Liu, Weibo Gao, Ye Liu, Kai Zhang, Yichao Du, Li Wang, Fangzhou Yao
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:58019-58036, 2024.

Abstract

Graph Neural Networks (GNNs) have demonstrated remarkable performance in graph classification tasks. However, ensuring the explainability of their predictions remains a challenge. To address this, graph rationalization methods have been introduced to generate concise subsets of the original graph, known as rationales, which serve to explain the predictions made by GNNs. Existing rationalizations often rely on shortcuts in data for prediction and rationale composition. In response, de-shortcut rationalization methods have been proposed, which commonly leverage counterfactual augmentation to enhance data diversity for mitigating the shortcut problem. Nevertheless, these methods have predominantly focused on centralized datasets and have not been extensively explored in the Federated Learning (FL) scenarios. To this end, in this paper, we propose a Federated Graph Rationalization (FedGR) with anti-shortcut augmentations to achieve self-explaining GNNs, which involves two data augmenters. These augmenters are employed to produce client-specific shortcut conflicted samples at each client, which contributes to mitigating the shortcut problem under the FL scenarios. Experiments on real-world benchmarks and synthetic datasets validate the effectiveness of FedGR under the FL scenarios.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-yue24b, title = {Federated Self-Explaining {GNN}s with Anti-shortcut Augmentations}, author = {Yue, Linan and Liu, Qi and Gao, Weibo and Liu, Ye and Zhang, Kai and Du, Yichao and Wang, Li and Yao, Fangzhou}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {58019--58036}, 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/yue24b/yue24b.pdf}, url = {https://proceedings.mlr.press/v235/yue24b.html}, abstract = {Graph Neural Networks (GNNs) have demonstrated remarkable performance in graph classification tasks. However, ensuring the explainability of their predictions remains a challenge. To address this, graph rationalization methods have been introduced to generate concise subsets of the original graph, known as rationales, which serve to explain the predictions made by GNNs. Existing rationalizations often rely on shortcuts in data for prediction and rationale composition. In response, de-shortcut rationalization methods have been proposed, which commonly leverage counterfactual augmentation to enhance data diversity for mitigating the shortcut problem. Nevertheless, these methods have predominantly focused on centralized datasets and have not been extensively explored in the Federated Learning (FL) scenarios. To this end, in this paper, we propose a Federated Graph Rationalization (FedGR) with anti-shortcut augmentations to achieve self-explaining GNNs, which involves two data augmenters. These augmenters are employed to produce client-specific shortcut conflicted samples at each client, which contributes to mitigating the shortcut problem under the FL scenarios. Experiments on real-world benchmarks and synthetic datasets validate the effectiveness of FedGR under the FL scenarios.} }
Endnote
%0 Conference Paper %T Federated Self-Explaining GNNs with Anti-shortcut Augmentations %A Linan Yue %A Qi Liu %A Weibo Gao %A Ye Liu %A Kai Zhang %A Yichao Du %A Li Wang %A Fangzhou Yao %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-yue24b %I PMLR %P 58019--58036 %U https://proceedings.mlr.press/v235/yue24b.html %V 235 %X Graph Neural Networks (GNNs) have demonstrated remarkable performance in graph classification tasks. However, ensuring the explainability of their predictions remains a challenge. To address this, graph rationalization methods have been introduced to generate concise subsets of the original graph, known as rationales, which serve to explain the predictions made by GNNs. Existing rationalizations often rely on shortcuts in data for prediction and rationale composition. In response, de-shortcut rationalization methods have been proposed, which commonly leverage counterfactual augmentation to enhance data diversity for mitigating the shortcut problem. Nevertheless, these methods have predominantly focused on centralized datasets and have not been extensively explored in the Federated Learning (FL) scenarios. To this end, in this paper, we propose a Federated Graph Rationalization (FedGR) with anti-shortcut augmentations to achieve self-explaining GNNs, which involves two data augmenters. These augmenters are employed to produce client-specific shortcut conflicted samples at each client, which contributes to mitigating the shortcut problem under the FL scenarios. Experiments on real-world benchmarks and synthetic datasets validate the effectiveness of FedGR under the FL scenarios.
APA
Yue, L., Liu, Q., Gao, W., Liu, Y., Zhang, K., Du, Y., Wang, L. & Yao, F.. (2024). Federated Self-Explaining GNNs with Anti-shortcut Augmentations. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:58019-58036 Available from https://proceedings.mlr.press/v235/yue24b.html.

Related Material