Graph Classification on CSL
80Accuracy (Max)sGNN5
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| sGNN52019.05 | 80 | 80 | 0 | — | |
| Ring-GNNInput node features=identical features2019.05 | 80 | 10 | 15.7 | — | |
| RP-GIN2019.05 | 53.3 | 10 | 12.9 | — | |
| sGNN22019.05 | 30 | 30 | 0 | — | |
| LGNN2019.05 | 30 | 30 | 0 | — | |
| GIN2019.05 | 10 | 10 | 0 | — | |
| 2-IGN2019.05 | 10 | 10 | 0 | — | |
| sGNN12019.05 | 10 | 10 | 0 | — | |
| GATTraining/Evaluation Protocol=fully supervised2026.06 | — | — | — | 22.33 | |
| GAUGETraining/Evaluation Protocol=pretrain-finetune2026.06 | — | — | — | 92.56 | |
| GCNTraining/Evaluation Protocol=fully supervised2026.06 | — | — | — | 26 | |
| GINTraining/Evaluation Protocol=fully supervised2026.06 | — | — | — | 29.67 | |
| GraphSAGETraining/Evaluation Protocol=fully supervised2026.06 | — | — | — | 21.67 | |
| SAMGPTTraining/Evaluation Protocol=pretrain-finetune2026.06 | — | — | — | 64.17 |