TU Dortmund
IMDB-MULTI
Structure
1,500Graphs
13Nodes
66Edges
3Classes
Coverage
120Results
1Variants
19Papers
- ICEPool: Enhancing Graph Pooling Networks with Inter-cluster Connectivity ICEPool: Enhancing Graph Pooling Networks with Inter-cluster Connectivity
- NeurIPS 2024: Cluster-wise Graph Transformer with Dual-granularity Kernelized Attention Cluster-wise Graph Transformer with Dual-granularity Kernelized Attention NeurIPS 2024
- Disentangled Generative Graph Representation Learning Disentangled Generative Graph Representation Learning
- AAAI 2024: TopoGCL: Topological Graph Contrastive Learning TopoGCL: Topological Graph Contrastive Learning AAAI 2024
- Graph Learning with Distributional Edge Layouts Graph Learning with Distributional Edge Layouts
- ICML 2023: Graph Mixup with Soft Alignments Graph Mixup with Soft Alignments ICML 2023
- ICML 2023: Randomized Schur Complement Views for Graph Contrastive Learning Randomized Schur Complement Views for Graph Contrastive Learning ICML 2023
- NeurIPS 2023: Fine-grained Expressivity of Graph Neural Networks Fine-grained Expressivity of Graph Neural Networks NeurIPS 2023
- NeurIPS 2023: The expressive power of pooling in Graph Neural Networks The expressive power of pooling in Graph Neural Networks NeurIPS 2023
- ICML 2022: Structural Entropy Guided Graph Hierarchical Pooling Structural Entropy Guided Graph Hierarchical Pooling ICML 2022
- TMLR 2022: Empowering GNNs via Edge-Aware Weisfeiler-Leman Algorithm Empowering GNNs via Edge-Aware Weisfeiler-Leman Algorithm TMLR 2022
- NeurIPS 2021: AutoGEL: An Automated Graph Neural Network with Explicit Link Information AutoGEL: An Automated Graph Neural Network with Explicit Link Information NeurIPS 2021
- ICLR 2021: Graph Neural Networks with Learnable Structural and Positional Representations Graph Neural Networks with Learnable Structural and Positional Representations ICLR 2021
- Pooling Architecture Search for Graph Classification Pooling Architecture Search for Graph Classification
- ICLR 2021: Topological Graph Neural Networks Topological Graph Neural Networks ICLR 2021
- ICLR 2020: On Graph Neural Networks versus Graph-Augmented MLPs On Graph Neural Networks versus Graph-Augmented MLPs ICLR 2020
- ICML 2020: Contrastive Multi-View Representation Learning on Graphs Contrastive Multi-View Representation Learning on Graphs ICML 2020
- ICML 2020: Graph Homomorphism Convolution Graph Homomorphism Convolution ICML 2020
- ICML 2018: Graph Capsule Convolutional Neural Networks Graph Capsule Convolutional Neural Networks ICML 2018
3-class graph classification Multi-class graph classification of movie collaboration networks (3 genres). Uses the released node/edge labels or attributes; several social datasets are featureless. Evaluated by Accuracy.
Dataset details
Task
Multi-class graph classification of movie collaboration networks (3 genres). The primary catalog metric is Accuracy.
Data
Multi-class graph classification of movie collaboration networks (3 genres). The cataloged artifact reports 1,500 graphs, 3 target classes or tasks.
Features
The TU text format may contain discrete node or edge labels and optional continuous attributes. Social-network datasets commonly have no intrinsic node attributes, so papers often add constant or degree features; that choice changes the input.
Splits and evaluation
The catalog records these protocols or variants: 10-fold CV. Evaluation uses Accuracy (higher is better). Exact masks or folds must come from the cited release.
Quirks and pitfalls
TU datasets have no canonical train/test split. Report cross-validation folds, seeds, feature construction, and whether duplicate or isomorphic graphs were removed; published results can otherwise differ for protocol rather than model quality.
Sources
SOTA Progression
new SOTA
matches SOTA at that time
other proposed result
other baseline result
Leaderboard
| # | Model | Accuracy ↑ | equivalence classes | Paper |
|---|---|---|---|---|
| 1 | GraphSNN | 0.7790
±
0.0360 | — | |
| 2 | SIN | 0.7560
±
0.0320 | — | |
| 3 | CIN | 0.7560
±
0.0370 | — | |
| 4 | Empowering GNNs | 0.7520
±
0.0450 | — | |
| 5 | GIN | 0.7510
±
0.0510 | — | |
| 6 | GNN-AK+ | 0.7500
±
0.0420 | — | |
| 7 | PPGN | 0.7300
±
0.0580 | — | |
| 8 | rLap | 0.5947
±
0.0742 | — | |
| 9 | EdgeDroppingDegree | 0.5907
±
0.0500 | — | |
| 10 | EdgeAddition | 0.5900
±
0.0958 | — | |
| 11 | EdgeDroppingPR | 0.5780
±
0.0996 | — | |
| 12 | RandomWalkSubgraph | 0.5780
±
0.0947 | — | |
| 13 | NodeDropping | 0.5753
±
0.0850 | — | |
| 14 | PPRDiffusion | 0.5713
±
0.0772 | — | |
| 15 | MarkovDiffusion | 0.5680
±
0.0771 | — | |
| 16 | EdgeDropping | 0.5613
±
0.0831 | — | |
| 17 | EdgeDroppingEVC | 0.5500
±
0.0701 | — | |
| 18 | H_d | 0.5477
±
0.0263 | — | |
| 19 | H_d + H_g | 0.5469
±
0.0206 | — | |
| 20 | Graph Transformer + DEL-K | 0.5354
±
0.0027 | — | |
| 21 | Graph Transformer +DEL-F | 0.5339
±
0.0033 | — | |
| 22 | GAT +DEL-F | 0.5327
±
0.0099 | — | |
| 23 | GPS(DEL-F) | 0.5295
±
0.0064 | — | |
| 24 | GAT + DEL-K | 0.5286
±
0.0033 | — | |
| 25 | GAT + MPNN | 0.5286
±
0.0029 | — | |
| 26 | TopoGCL | 0.5281
±
0.0031 | — | |
| 27 | Graph Transformer + MPNN | 0.5281
±
0.0025 | — | |
| 28 | GNTK | 0.5280
±
0.0460 | — | |
| 29 | GAT + Random distance | 0.5260
±
0.0058 | — | |
| 30 | SEP-G ICE | 0.5247
±
0.0100 | — | |
| 31 | Graph Transformer + Random distance | 0.5231
±
0.0045 | — | |
| 32 | GIN-0 | 0.5230
±
0.0280 | — | |
| 33 | GIN | 0.5230
±
0.0280 | — | |
| 34 | PAS | 0.5220
±
0.0373 | — | |
| 35 | GCL-TAGS | 0.5216
±
0.0072 | — | |
| 36 | Cluster-GT | 0.5213
±
0.0078 | — | |
| 37 | GIN-ε | 0.5210
±
0.0360 | — | |
| 38 | GCN-1-TOGL-1 | 0.5200
±
0.0400 | — | |
| 39 | GCN | 0.5190
±
0.0380 | — | |
| 40 | GraphMAE | 0.5163
±
0.0052 | — | |
| 41 | H_g | 0.5162
±
0.0061 | — | |
| 42 | SEP-G | 0.5153
±
0.0065 | — | |
| 43 | AWL kernel | 0.5150
±
0.0360 | — | |
| 44 | MI | 0.5120
±
0.0050 | — | |
| 45 | GraphSAGE | 0.5090
±
0.0220 | — | |
| 46 | GIN-JK | 0.5080
±
0.0302 | — | |
| 47 | S-Mixup | 0.5073
±
0.0366 | — | |
| 48 | GMT | 0.5066
±
0.0082 | — | |
| 49 | AD-GCL | 0.5060
±
0.0070 | — | |
| 50 | AutoGCL | 0.5060
±
0.0080 | — |
SOTA performance
set a new SOTA at the time
matched the SOTA at the time