Boundary Embedding Shaping with Adaptive Contrastive Learning for Graph Structural Disentanglement
For graph neural network practitioners, BES offers a plug-in module that improves classification by focusing on boundary-region entanglement, a previously underexplored bottleneck.
The paper addresses graph structural entanglement in GNNs, where spurious correlations from irrelevant neighbors blur decision boundaries, especially for nodes near class boundaries. The proposed Boundary Embedding Shaping (BES) module improves node classification by an average of 3.3% (up to 5.0% on WikiCS) and achieves superior link prediction accuracy.
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.