Open Graph Benchmark (OGB) - Biology

← Back to all benchmarks

Date: 2020-05-02

Name: Open Graph Benchmark OGB - Biology

Domain: Biology & Medicine

Focus: Biological graph property prediction

Task Types: Node property prediction, Link property prediction, Graph property prediction

Metrics: Accuracy, ROC-AUC

Models: GCN, GraphSAGE, GAT

AI/ML Motif: Sequence Prediction/Forecasting

Resources

Benchmark: Visit

Datasets: OGB Webpage

Keywords

node prediction link prediction graph classification

Citation

• Weihua Hu, Matthias Fey, Marinka Zitnik, Yuxiao Dong, Hongyu Ren, Bowen Liu, Michele Catasta, and Jure Leskovec. Open graph benchmark: datasets for machine learning on graphs. 2021. URL: https://arxiv.org/abs/2005.00687, arXiv:2005.00687.

@misc{hu2021opengraphbenchmarkdatasets,
    archiveprefix = {arXiv},
    author        = {Weihua Hu and Matthias Fey and Marinka Zitnik and Yuxiao Dong and Hongyu Ren and Bowen Liu and Michele Catasta and Jure Leskovec},
    eprint        = {2005.00687},
    primaryclass  = {cs.LG},
    title         = {Open Graph Benchmark: Datasets for Machine Learning on Graphs},
    url           = {https://arxiv.org/abs/2005.00687},
    year          = {2021}
}

Ratings

CategoryRating

Software

5.00

All necessary information is provided on the Github

Specification

4.00

Tasks (node/link/graph property prediction) are clearly specified with input/output formats and standardized protocols; splits are well-defined.

Dataset

5.00

Fully FAIR- datasets are versioned, split, and accessible via a standardized API; extensive metadata and documentation are included.

Metrics

5.00

Reproducible, quantitative metrics (e.g., ROC-AUC, accuracy) that are tightly aligned with the tasks.

Reference Solution

5.00

Multiple baselines implemented and documented (GCN, GAT, GraphSAGE).

Documentation

5.00

All necessary information is included in a paper.

Average rating: 4.83/5

Radar plot

Edit: edit this entry