VGAE-based_Model_Poisoning_Attack_FL

These codes are about variational graph autoencoder-based model poisoning attack against federated learning.

Requirements

• Install requirements via pip install -r requirements.txt

How to run 👇

python FL_VGAE_Attack_main.py 

Note that the variable num_clients_index in the FL_VGAE_Attack_main.py must be kept consistent with the variable input_dim in args.py

References

1. https://github.com/aswarth123/Federated_Learning_MNIST
2. https://github.com/DaehanKim/vgae_pytorch
3. https://arxiv.org/abs/1611.07308