Other Graph Benchmarks
MalNet-Tiny
Structure
5,000Graphs
1,410Nodes
2,860Edges
5Classes
Coverage
26Results
1Variants
6Papers
- Message Passing on the Edge: Towards Scalable and Expressive GNNs Message Passing on the Edge: Towards Scalable and Expressive GNNs
- State Space Models over Directed Graphs State Space Models over Directed Graphs
- KDD 2024: Graph Mamba: Towards Learning on Graphs with State Space Models Graph Mamba: Towards Learning on Graphs with State Space Models KDD 2024
- NeurIPS 2022: Transformers over Directed Acyclic Graphs Transformers over Directed Acyclic Graphs NeurIPS 2022
- NeurIPS 2022: Recipe for a General, Powerful, Scalable Graph Transformer Recipe for a General, Powerful, Scalable Graph Transformer NeurIPS 2022
- NeurIPS 2020: A Large-Scale Database for Graph Representation Learning A Large-Scale Database for Graph Representation Learning NeurIPS 2020
5-class graph classification on Android function-call graphs, predicting one of five balanced malware families; nodes have no attributes. Evaluated by Accuracy.
Dataset details
Task
Malware detection graph classification on a tiny (balanced 5-class) subset of the MalNet-Large dataset. Graphs are function call graphs of Android apps. The primary catalog metric is Accuracy.
Data
Malware detection graph classification on a tiny (balanced 5-class) subset of the MalNet-Large dataset. Graphs are function call graphs of Android apps. The cataloged artifact reports 5,000 graphs, 5 target classes or tasks.
Features
Input construction is dataset-specific and follows the cited release: program call graphs, 3D point coordinates, text-derived vectors, product attributes, or movie metadata. Added edges, neighborhood graphs, and feature normalization are preprocessing choices.
Splits and evaluation
The catalog records these protocols or variants: Standard split. Evaluation uses Accuracy (higher is better). Exact masks or folds must come from the cited release.
Quirks and pitfalls
These entries come from unrelated upstream benchmarks and have no shared protocol. Report the exact graph construction, split, feature pipeline, and label granularity; similarly named Film, Actor, IMDB, or Amazon artifacts are not interchangeable.
Sources
SOTA Progression
new SOTA
matches SOTA at that time
other proposed result
other baseline result
Leaderboard
| # | Model | Accuracy ↑ | Paper |
|---|---|---|---|
| 1 | GMN | 0.9415
±
0.0020 | |
| 2 | DirGraphSSM | 0.9409
±
0.0023 | |
| 3 | DAG+SAT-plus | 0.9407
±
0.0020 | |
| 4 | Graph-Mamba | 0.9406
±
0.0018 | |
| 5 | DAG+Transformer-plus | 0.9404
±
0.0019 | |
| 6 | Exphormer | 0.9402
±
0.0020 | |
| 7 | Transformer | 0.9350
±
0.0060 | |
| 8 | DAG+GraphGPS | 0.9345
±
0.0041 | |
| 9 | EB-GNN | 0.9330
±
0.0066 | |
| 10 | GPS + Mamba | 0.9311
±
0.0042 | |
| 11 | HyMN | 0.9284
±
0.0052 | |
| 12 | GraphGPS | 0.9264
±
0.0078 | |
| 13 | Performer | 0.9264 | |
| 14 | SAT | 0.9256
±
0.0031 | |
| 15 | NC-GNN | 0.9250
±
0.0056 | |
| 16 | BigBird | 0.9234 | |
| 17 | GPS (Perf.) | 0.9214
±
0.0024 | |
| 18 | MPNN | 0.9110
±
0.0098 | |
| 19 | MagNet | 0.9095
±
0.0025 | |
| 20 | DiGCN | 0.9081
±
0.0017 | |
| 21 | DirGNN | 0.9073
±
0.0015 | |
| 22 | GIN | 0.8898
±
0.0055 | |
| 23 | Feather | 0.8600 | |
| 24 | LDP | 0.8600 | |
| 25 | GAT | 0.8509
±
0.0025 | |
| 26 | GCN | 0.8100
±
0.0001 |
SOTA performance
set a new SOTA at the time
matched the SOTA at the time