2.1.7. aijack.defense.paillier package

2.1.7.1. Submodules

2.1.7.2. aijack.defense.paillier.fed_wrapper module

class aijack.defense.paillier.fed_wrapper.PaillierGradientClientManager(*args, **kwargs)

Bases: aijack.manager.base.BaseManager

Client Manager for secure aggregation with Paillier Encryption

attach(cls)

aijack.defense.paillier.fed_wrapper.attach_paillier_to_client_for_encrypted_grad(cls, pk, sk)

Makes the client class communicate the encrypted gradients with paillier encryption scheme.

Parameters

• cls – client class

• pk – public key

• sk – secret key

2.1.7.3. aijack.defense.paillier.torch_wrapper module

class aijack.defense.paillier.torch_wrapper.PaillierTensor(paillier_array)

Bases: object

torch.Tensor-like object for Paillier Encryption

add(other)

cpu()

decrypt(sk, device='cpu')

detach()

linear(w, bias)

matmul(other)

mul(other)

numpy()

sub(other)

tensor(sk=None)

aijack.defense.paillier.torch_wrapper.implements(torch_function)

Registers a torch function override for PaillierTensor

2.1.7.4. Module contents