Boundary Embedding Shaping with Adaptive Contrastive Learning for Graph Structural Disentanglement

Jiaqing Chen, Zidu Yin, Yichao Cai, Yuhang Liu, Zhen Zhang, Dong Gong, Javen Qinfeng Shi
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:17913-17939, 2026.

Abstract

Graph neural networks (GNNs) excel at aggregating neighbor information for classification, yet their performance is hindered by graph structural entanglement, where spurious correlations from semantically irrelevant neighbors contaminate node embeddings. This challenge is most acute for nodes near class boundaries in the embedding space, where amplified structural noise blurs decision boundaries and destabilizes predictions. Existing robust GNN methods largely treat all nodes uniformly, ignoring boundary vulnerabilities. In this paper, to improve classification performance, we tackle graph structural disentanglement by identifying boundary-region entanglement as the primary bottleneck and propose Boundary Embedding Shaping (BES), an adaptive contrastive learning GNN plug-in module that selectively suppresses spurious structural noise at decision boundaries with minimal model parameter perturbation. Extensive experiments demonstrate that BES consistently improves boundary discrimination and outperforms existing leading methods. Notably, BES boosts GCN performance by an average of 3.3% in node classification (up to 5.0% on WikiCS) and achieves superior accuracy in link prediction.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-chen26fu, title = {Boundary Embedding Shaping with Adaptive Contrastive Learning for Graph Structural Disentanglement}, author = {Chen, Jiaqing and Yin, Zidu and Cai, Yichao and Liu, Yuhang and Zhang, Zhen and Gong, Dong and Shi, Javen Qinfeng}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {17913--17939}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/chen26fu/chen26fu.pdf}, url = {https://proceedings.mlr.press/v306/chen26fu.html}, abstract = {Graph neural networks (GNNs) excel at aggregating neighbor information for classification, yet their performance is hindered by graph structural entanglement, where spurious correlations from semantically irrelevant neighbors contaminate node embeddings. This challenge is most acute for nodes near class boundaries in the embedding space, where amplified structural noise blurs decision boundaries and destabilizes predictions. Existing robust GNN methods largely treat all nodes uniformly, ignoring boundary vulnerabilities. In this paper, to improve classification performance, we tackle graph structural disentanglement by identifying boundary-region entanglement as the primary bottleneck and propose Boundary Embedding Shaping (BES), an adaptive contrastive learning GNN plug-in module that selectively suppresses spurious structural noise at decision boundaries with minimal model parameter perturbation. Extensive experiments demonstrate that BES consistently improves boundary discrimination and outperforms existing leading methods. Notably, BES boosts GCN performance by an average of 3.3% in node classification (up to 5.0% on WikiCS) and achieves superior accuracy in link prediction.} }
Endnote
%0 Conference Paper %T Boundary Embedding Shaping with Adaptive Contrastive Learning for Graph Structural Disentanglement %A Jiaqing Chen %A Zidu Yin %A Yichao Cai %A Yuhang Liu %A Zhen Zhang %A Dong Gong %A Javen Qinfeng Shi %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-chen26fu %I PMLR %P 17913--17939 %U https://proceedings.mlr.press/v306/chen26fu.html %V 306 %X Graph neural networks (GNNs) excel at aggregating neighbor information for classification, yet their performance is hindered by graph structural entanglement, where spurious correlations from semantically irrelevant neighbors contaminate node embeddings. This challenge is most acute for nodes near class boundaries in the embedding space, where amplified structural noise blurs decision boundaries and destabilizes predictions. Existing robust GNN methods largely treat all nodes uniformly, ignoring boundary vulnerabilities. In this paper, to improve classification performance, we tackle graph structural disentanglement by identifying boundary-region entanglement as the primary bottleneck and propose Boundary Embedding Shaping (BES), an adaptive contrastive learning GNN plug-in module that selectively suppresses spurious structural noise at decision boundaries with minimal model parameter perturbation. Extensive experiments demonstrate that BES consistently improves boundary discrimination and outperforms existing leading methods. Notably, BES boosts GCN performance by an average of 3.3% in node classification (up to 5.0% on WikiCS) and achieves superior accuracy in link prediction.
APA
Chen, J., Yin, Z., Cai, Y., Liu, Y., Zhang, Z., Gong, D. & Shi, J.Q.. (2026). Boundary Embedding Shaping with Adaptive Contrastive Learning for Graph Structural Disentanglement. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:17913-17939 Available from https://proceedings.mlr.press/v306/chen26fu.html.

Related Material