Research map / Graph learning / Graph self-supervised learning
SGFormer: Simplifying and Empowering Transformers for Large-Graph Representations
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 SGFormer instead of the strongest reported baseline | Node classification models · Cora | accuracy: 83.3 ± 0.5 -> 84.5 ± 0.8 | Helped |
| Use SGFormer instead of the strongest reported baseline | Node classification models · CiteSeer | accuracy: 72.5 ± 1.1 -> 72.6 ± 0.2 | No clear effect |
| Use SGFormer instead of the strongest reported baseline | Node classification models · PubMed | accuracy: 80.1 ± 0.2 -> 80.3 ± 0.6 | No clear effect |
| Use SGFormer instead of the strongest reported baseline | Node classification models · Actor | accuracy: 36.9 ± 1.0 -> 37.9 ± 1.1 | Helped |
| Use SGFormer instead of the strongest reported baseline | Node classification models · Squirrel | accuracy: 41.0 ± 2.8 -> 41.8 ± 2.2 | No clear effect |
| Use SGFormer instead of the strongest reported baseline | Node classification models · Chameleon | accuracy: 42.8 ± 3.3 -> 44.9 ± 3.9 | Helped |
| Use SGFormer instead of the strongest reported baseline | Node classification models · Deezer | accuracy: 66.4 ± 0.7 -> 67.1 ± 1.1 | No clear effect |
| Use SGFormer instead of the strongest reported baseline | Large-graph node prediction models · ogbn-proteins | ROC-AUC: 77.45 ± 1.15 -> 79.53 ± 0.38 | Helped |
| Use SGFormer instead of the strongest reported baseline | Large-graph node prediction models · Amazon2M | accuracy: 87.85 ± 0.24 -> 89.09 ± 0.10 | Helped |
| Use SGFormer instead of the strongest reported baseline | Large-graph node prediction models · pokec | accuracy: 70.32 ± 0.45 -> 73.76 ± 0.24 | Helped |
| Use SGFormer instead of the strongest reported baseline | Large-graph node prediction models · ogbn-arxiv | accuracy: 71.74 ± 0.29 -> 72.63 ± 0.13 | Helped |
| Use SGFormer instead of the strongest reported baseline | Large-graph node prediction models · ogbn-papers100M | accuracy: 65.11 ± 0.14 -> 66.01 ± 0.37 | Helped |
Closest papers by meaning
- Generalizable Insights for Graph Transformers in Theory and Practice (NeurIPS 2025 · similarity 0.90)
- What Improves the Generalization of Graph Transformers? A Theoretical Dive into the Self-attention and Positional Encoding (ICML 2024 · similarity 0.90)
- A Scalable and Effective Alternative to Graph Transformers (AAAI 2025 · similarity 0.90)
- Towards Principled Graph Transformers (NeurIPS 2024 · similarity 0.90)
- VCR-Graphormer: A Mini-batch Graph Transformer via Virtual Connections (ICLR 2024 · similarity 0.89)
- Less is More: on the Over-Globalizing Problem in Graph Transformers (ICML 2024 · similarity 0.89)
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.