OGB-LSC: A Large-Scale Challenge for Machine Learning on Graphs

Benchmark Model Rank Results
graph-regression-on-pcqm4m-lscGIN-virtual#2Test MAE: 14.87Validation MAE: 0.1396
graph-regression-on-pcqm4m-lscGCN-Virtual#3Test MAE: 15.79Validation MAE: 0.1536
graph-regression-on-pcqm4m-lscGIN#4Test MAE: 16.78
graph-regression-on-pcqm4m-lscGCN#5Test MAE: 18.38Validation MAE: 0.1684
graph-regression-on-pcqm4m-lscMLP-fingerprint#6Test MAE: 20.68Validation MAE: 0.2044
graph-regression-on-pcqm4mv2-lscMLP-Fingerprint#20Validation MAE: 0.1753Test MAE: 0.1760
knowledge-graphs-on-wikikg90m-lscTransE-Concat#1Test MRR: 85.48Validation MRR: 0.8494
knowledge-graphs-on-wikikg90m-lscComplEx-Concat#2Test MRR: 0.8637Validation MRR: 0.8425
knowledge-graphs-on-wikikg90m-lscComplEx-RoBERTa#3Test MRR: 0.7186Validation MRR: 0.7052
knowledge-graphs-on-wikikg90m-lscTransE-RoBERTa#4Test MRR: 0.6288Validation MRR: 0.6039
node-classification-on-mag240m-lscR-GraphSAGE (NS)#1Test Accuracy: 68.94
node-classification-on-mag240m-lscGAT (NS)#2Test Accuracy: 66.63
node-classification-on-mag240m-lscGraphSAGE (NS)#3Test Accuracy: 66.25
node-classification-on-mag240m-lscSIGN#4Test Accuracy: 66.09Validation Accuracy: 66.64