https://github.com/zw-zhang/eigengnn

This is a sample implementation of "Eigen-GNN: a Graph Structure Preserving Plug-in for GNNs, TKDE 2021".
https://github.com/zw-zhang/eigengnn

Last synced: about 1 month ago
JSON representation

This is a sample implementation of "Eigen-GNN: a Graph Structure Preserving Plug-in for GNNs, TKDE 2021".

• Host: GitHub
• URL: https://github.com/zw-zhang/eigengnn
• Owner: ZW-ZHANG
• License: mit
• Created: 2021-09-10T04:00:00.000Z (about 3 years ago)
• Default Branch: main
• Last Pushed: 2022-01-11T07:30:29.000Z (almost 3 years ago)
• Last Synced: 2023-09-18T09:39:37.859Z (about 1 year ago)
• Language: Python
• Size: 18.6 KB
• Stars: 4
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# EigenGNN

This is a sample implementation of "Eigen-GNN: a Graph Structure Preserving Plug-in for GNNs, TKDE 2021". [(Paper)](https://zw-zhang.github.io/files/2021_TKDE_EigenGNN.pdf)

### Requirements

* numpy

* scipy

* networkx

### Usage

As indicated by the title, EigenGNN is a plug-in instead of a stand-along model. Therefore, the code in the main file will only generate initial node representations (i.e., eigenvectors of a graph structure matrix), which can be used together with other GNNs.

The initial node representation can be generated as follows:

```bash

from EigenGNN import Eigen_multi

features = Eigen_multi(adj, d, adj_normalize, feature_abs)

```

An example to use EigenGNN with [GCN](https://github.com/tkipf/gcn) being the GNN backbone for node classification benchmark datasets (e.g., Cora, CiteSeer, PubMed) is provided in `example_gcn.py` and can be run as:

```bash

python example_gcn.py --method eigen --dataset cora

```

for only using the eigenspace, or 

```bash

python example_gcn.py --method feat_eigen --dataset cora

```

for concatenating the eigenspace with input node features.

Another example to use EigenGNN with [GIN](https://arxiv.org/pdf/1810.00826.pdf) being the GNN backbone for [graph isomorphism test](https://github.com/PurdueMINDS/RelationalPooling/) (modeled as graph classification) is provided in `example_gin.py` and can be run as:

```bash

python example_gin.py --cv-fold 0 --model-type EigenGin --num-mlp-hidden 1 --num-gnn-layers 1 --out-weight-dir dir1 --out-log-dir dir2

```

 

The last example is use EigenGNN with [PyTorch Geometric](https://github.com/pyg-team/pytorch_geometric) is provided in `example_pyg_gcn.py` and can be run as:

```bash

python example_pyg_gcn.py --use_eigengnn

```

### Cite

If you find this code useful, please cite our paper:

```

@article{zhang2021eigengnn,

  title={Eigen-GNN: a Graph Structure Preserving Plug-in for GNNs},

  author={Zhang, Ziwei and Cui, Peng and Pei, Jian and Wang, Xin and Zhu, Wenwu},

  journal={IEEE Transactions on Knowledge and Data Engineering},

  year={2021},

  publisher={IEEE}

}

```