DPGNN: Differential Privacy Preservation in Graph Neural Networks

This is the repository implement the DPGNN

Reqirement:

Numpy
Pandas
tensorflow > 2.4.1 (for face detection)
tensorflow 1.14 (for SHAP extraction)
Pytorch
Dgl
Scikit-learn

Instruction:

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