Research map / Graph learning / Graph self-supervised learning
Open Graph Benchmark: Datasets for Machine Learning on Graphs
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 |
|---|---|---|---|
| realistic sales-rank split instead of random split | GraphSAGE · ogbn-products | test accuracy: 88.20 -> 78.50 | Hurt |
| time-based split instead of random split | GCN · ogbn-arxiv | test accuracy: 73.54 -> 71.74 | Hurt |
| use graph structure (GCN) instead of MLP | node classifier · ogbn-arxiv | test accuracy: 55.50 -> 71.74 | Helped |
| use SGC graph preprocessing instead of MLP | node classifier · ogbn-papers100M | test accuracy: 47.24 -> 63.29 | Helped |
| GCN instead of MLP | node classifier · ogbn-proteins | test ROC-AUC: 72.04 -> 72.51 | No clear effect |
| mini-batch GraphSAINT training instead of full-batch R-GCN | R-GCN · ogbn-mag | test accuracy: 39.77 -> 47.51 | Helped |
| ClusterGCN mini-batch training instead of full-batch | R-GCN · ogbn-mag | test accuracy: 39.77 -> 37.32 | Hurt |
| use heterogeneous relations (R-GCN) instead of homogeneous GraphSAGE | node classifier · ogbn-mag | test accuracy: 31.53 -> 39.77 | Helped |
| increase embedding dimension 5x | RotatE · ogbl-wikikg2 | test MRR: 0.2530 -> 0.4332 | Helped |
| add additional molecular features | GIN + virtual node · ogbg-mol* (10 datasets) | ROC-AUC/RMSE | Helped |
| add virtual nodes | GIN · ogbg-mol* (10 datasets) | ROC-AUC/RMSE | Helped |
Closest papers by meaning
- RDB2G-Bench: A Comprehensive Benchmark for Automatic Graph Modeling of Relational Databases (NeurIPS 2025 · similarity 0.90)
- No Metric to Rule Them All: Toward Principled Evaluations of Graph-Learning Datasets (ICML 2025 · similarity 0.90)
- OgBench: A Framework for Evaluating Graph Neural Networks on Omics Data (NeurIPS 2026 · similarity 0.90)
- 4DBInfer: A 4D Benchmarking Toolbox for Graph-Centric Predictive Modeling on RDBs (NeurIPS 2024 · similarity 0.89)
- GLBench: A Comprehensive Benchmark for Graph with Large Language Models (NeurIPS 2024 · similarity 0.88)
- IGL-Bench: Establishing the Comprehensive Benchmark for Imbalanced Graph Learning (ICLR 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.