Research map / Graph learning / GNN architectures
SelfSAGCN: Self-Supervised Semantic Alignment for Graph Convolution Network
Xu Yang, Cheng Deng, Zhiyuan Dang, Kun Wei, Junchi Yan
Check 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 |
|---|---|---|---|
| Use SelfSAGCN instead of Shoestring | semi-supervised node classification · Cora, 1 label per class | accuracy: 60.2(0.9) -> 63.5(0.8) | Improved |
| Use SelfSAGCN instead of Shoestring | semi-supervised node classification · Citeseer, 1 label per class | accuracy: 52.2(1.3) -> 57.5(3.7) | Improved |
| Use SelfSAGCN instead of Shoestring | semi-supervised node classification · Pubmed, 1 label per class | accuracy: 60.3(6.1) -> 65.0(3.3) | Improved |
| Use SelfSAGCN instead of Shoestring | semi-supervised node classification · Cora, 2 labels per class | accuracy: 68.3(0.9) -> 71.7(0.6) | Improved |
| Use SelfSAGCN instead of Shoestring | semi-supervised node classification · Citeseer, 2 labels per class | accuracy: 60.1(1.3) -> 68.3(1.2) | Improved |
| Use SelfSAGCN instead of DAGNN | semi-supervised node classification · Pubmed, 2 labels per class | accuracy: 65.3(2.7) -> 70.9(1.9) | Improved |
| Use SelfSAGCN instead of Shoestring | semi-supervised node classification · Cora, 5 labels per class | accuracy: 73.0(1.2) -> 75.1(0.7) | Improved |
| Use SelfSAGCN instead of Shoestring | semi-supervised node classification · Citeseer, 5 labels per class | accuracy: 64.2(1.5) -> 71.9(1.3) | Improved |
| Use SelfSAGCN instead of SGC | semi-supervised node classification · Pubmed, 5 labels per class | accuracy: 71.4(2.9) -> 73.0(2.1) | Improved |
| Use SelfSAGCN instead of DAGNN | image graph classification · STL-10, 1 label per class | accuracy: 87.5(0.5) -> 92.1(0.2) | Improved |
| Use SelfSAGCN instead of DAGNN | image graph classification · CIFAR-10, 1 label per class | accuracy: 61.2(0.8) -> 65.8(0.7) | Improved |
| Use SelfSAGCN instead of DAGNN | image graph classification · STL-10, 2 labels per class | accuracy: 89.0(0.9) -> 94.3(0.3) | Improved |
Closest papers by meaning
- Graph Alignment for Benchmarking Graph Neural Networks and Learning Positional Encodings (ICML 2026 · similarity 0.89)
- Deeper Insights Into Graph Convolutional Networks for Semi-Supervised Learning (AAAI 2018 · similarity 0.89)
- Beyond Graph Convolutional Network: An Interpretable Regularizer-Centered Optimization Framework (AAAI 2023 · similarity 0.89)
- Simple and Deep Graph Convolutional Networks (ICML 2020 · similarity 0.89)
- Graph Your Own Prompt (NeurIPS 2025 · similarity 0.87)
- On the Stability of Graph Convolutional Neural Networks: A Probabilistic Perspective (NeurIPS 2025 · similarity 0.87)
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.
