DPGNN: Differential Privacy Preservation in Graph Neural Networks

This is the repository implement the DPGNN

Reqirement:

BRand algorithm is implemented in the feature_level.py.
ERand is implemented in folder ERand. The instruction for ERand is attached that folder.
GCN training is implemented in the main folder

Name		Name	Last commit message	Last commit date
Latest commit History 20 Commits
Attack		Attack
Data		Data
Datasets		Datasets
ERand		ERand
Models		Models
SavedModel		SavedModel
Trainer		Trainer
Utils		Utils
main		main
.gitignore		.gitignore
LICENSE		LICENSE
README.md		README.md
SupplementalResults.pdf		SupplementalResults.pdf
config.py		config.py
extract_shap.py		extract_shap.py
feature_baseline_nus.py		feature_baseline_nus.py
feature_baseline_ppi.py		feature_baseline_ppi.py
feature_baseline_reddit.py		feature_baseline_reddit.py
feature_level.py		feature_level.py
feature_level_mir.py		feature_level_mir.py
feature_level_nus.py		feature_level_nus.py
feature_level_ppi.py		feature_level_ppi.py
feature_level_reddit.py		feature_level_reddit.py
flickr_feat_level.py		flickr_feat_level.py
resnet50_feat_extraction.py		resnet50_feat_extraction.py
test.py		test.py
train_edge_level.py		train_edge_level.py
zoom_plot.py		zoom_plot.py