Research map / Graph learning / Graph self-supervised learning
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
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add BES plug-in on top of GCN encoders | GCN · WikiCS node classification | accuracy: 80.48 -> 85.52 | Improved |
| Add BES plug-in | GCN · Pubmed node classification | accuracy: 87.36 -> 91.03 | Improved |
| Add BES plug-in | GCN · Cora link prediction | accuracy: 92.30 -> 94.45 | Improved |
| Use graph-topology neighbors instead of embedding-space neighbors for shift score (ablation) | BES · Cora, CiteSeer, Pubmed, WikiCS node classification | accuracy: WikiCS 85.28 vs 85.30 | No clear effect |
| Add attention layers of BES to the pre-trained encoder embedding (ablation) | BES · WikiCS | accuracy | Improved |
| Increase number of BES attention layers (ablation) | BES · WikiCS | accuracy | No clear effect |
| Use larger beta (more boundary nodes) (ablation) | BES · Hyperparameter sensitivity | accuracy | Improved |
| Use smaller alpha (gentler updates) (ablation) | BES · Hyperparameter sensitivity | accuracy | Improved |
Closest papers by meaning
- Contrastive Auxiliary Learning with Structure Transformation for Heterogeneous Graphs (AAAI 2025 · similarity 0.87)
- Towards characterizing the value of edge embeddings in Graph Neural Networks (ICML 2025 · similarity 0.87)
- ReGCL: Rethinking Message Passing in Graph Contrastive Learning (AAAI 2024 · similarity 0.87)
- Graph Structure Refinement with Energy-based Contrastive Learning (AAAI 2025 · similarity 0.86)
- Enhancing Size Generalization in Graph Neural Networks through Disentangled Representation Learning (ICML 2024 · similarity 0.86)
- GPEN: Global Position Encoding Network for Enhanced Subgraph Representation Learning (ICML 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
