IMDB-BINARY
Description
IMDB-BINARY
Dataset Overview
Lifting Methods
Structural-based Liftings
- Cellular: Cycle-based lifting
- Simplicial: Clique complex lifting
- Hypergraph: k-hop lifting
Feature Lifting
- Projected sum
Model Performance
| Model | Accuracy (%) | Std Dev (±) |
|---|---|---|
| GCN | 72.00 | 2.48 |
| GIN | 70.96 | 1.93 |
| GAT | 69.76 | 2.65 |
| AST | 70.32 | 3.27 |
| EDGNN | 69.12 | 2.92 |
| UniGNN2 | 71.04 | 1.31 |
| CWN | 70.40 | 2.02 |
| CCCN | 69.12 | 2.82 |
| SCCNN | 70.88 | 2.25 |
| SCN | 70.80 | 2.38 |
Key Insights
- GCN achieves the best performance with 72.00% accuracy
- Most models perform in the 69-71% range, with UniGNN2 and SCCNN also above 71%
- GCN shows the highest accuracy, while UniGNN2 is the most stable (±1.31)
- AST shows the highest variability (±3.27)
Reproducibility
python -m topobench model=graph/gcn dataset=graph/IMDB-BINARY optimizer.parameters.lr=0.01 model.feature_encoder.out_channels=64 model.backbone.num_layers=2 model.feature_encoder.proj_dropout=0.5 dataset.dataloader_params.batch_size=256 dataset.split_params.data_seed=0,3,5,7,9 trainer.max_epochs=500 trainer.min_epochs=50 trainer.check_val_every_n_epoch=5 callbacks.early_stopping.patience=10 logger.wandb.project=TopoBench --multirun @imdb-binary.md