https://github.com/zw-zhang/robustgcn
This is a sample implementation of "Robust Graph Convolutional Networks Against Adversarial Attacks", KDD 2019.
https://github.com/zw-zhang/robustgcn
Last synced: 2 months ago
JSON representation
This is a sample implementation of "Robust Graph Convolutional Networks Against Adversarial Attacks", KDD 2019.
- Host: GitHub
- URL: https://github.com/zw-zhang/robustgcn
- Owner: ZW-ZHANG
- Created: 2020-08-28T08:34:38.000Z (almost 5 years ago)
- Default Branch: master
- Last Pushed: 2020-12-08T15:05:15.000Z (over 4 years ago)
- Last Synced: 2025-04-14T02:06:08.300Z (2 months ago)
- Language: Python
- Homepage:
- Size: 133 KB
- Stars: 10
- Watchers: 2
- Forks: 3
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# RobustGCN
This is a sample implementation of "[Robust Graph Convolutional Networks Against Adversarial Attacks](https://zw-zhang.github.io/files/2019_KDD_RGCN.pdf)", KDD 2019.### Requirements
```
tensorflow >= 1.12
numpy >= 1.14.2
scipy >= 1.1.0
networkx >= 2.0.0
gcn (note that you need to follow https://github.com/tkipf/gcn to correctly install gcn instead of using pip)
```
### Example Usage
```
python src/train.py --dataset cora
```
### Full Command List
```
optional arguments:
--dataset Dataset string.
--learning_rate Initial learning rate.
--epochs Number of epochs to train.
--hidden Number of units in hidden layer.
--dropout Dropout rate (1 - keep probability).
--para_var Parameter of variance-based attention.
--para_kl Parameter of kl regularization.
--para_l2 Parameter for l2 loss.
--early_stopping Tolerance for early stopping (# of epochs).
```
### Cite
If you find this code useful, please cite our paper:
```
@inproceedings{zhu2019robust,
title={Robust graph convolutional networks against adversarial attacks},
author={Zhu, Dingyuan and Zhang, Ziwei and Cui, Peng and Zhu, Wenwu},
booktitle={Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery \& Data Mining},
pages={1399--1407},
year={2019}
}
```
### Acknowledgement
Our code is adapted from the Tensorflow implementation of GCN by Thomas Kipf (https://github.com/tkipf/gcn).