1.1.5. aijack.attack.labelleakage package#

1.1.5.1. Submodules#

1.1.5.2. aijack.attack.labelleakage.normattack module#

class aijack.attack.labelleakage.normattack.NormAttackSplitNNManager(*args, **kwargs)[source]#

Bases: aijack.manager.base.BaseManager

attach(cls)[source]#
aijack.attack.labelleakage.normattack.attach_normattack_to_splitnn(cls, attack_criterion, target_client_index=0, device='cpu')[source]#

Attaches a normalization attack to a SplitNN model.

Parameters

  • cls – The SplitNN model class.

  • attack_criterion – The criterion for the attack.

  • target_client_index (int, optional) – Index of the target client. Defaults to 0.

  • device (str, optional) – Device for computation. Defaults to “cpu”.

Returns

A wrapper class with attached normalization attack.

Return type

class

1.1.5.3. Module contents#

Subpackage for label leakage attack, which infere the private label information of the training dataset.

class aijack.attack.labelleakage.NormAttackSplitNNManager(*args, **kwargs)[source]#

Bases: aijack.manager.base.BaseManager

attach(cls)[source]#
aijack.attack.labelleakage.attach_normattack_to_splitnn(cls, attack_criterion, target_client_index=0, device='cpu')[source]#

Attaches a normalization attack to a SplitNN model.

Parameters

  • cls – The SplitNN model class.

  • attack_criterion – The criterion for the attack.

  • target_client_index (int, optional) – Index of the target client. Defaults to 0.

  • device (str, optional) – Device for computation. Defaults to “cpu”.

Returns

A wrapper class with attached normalization attack.

Return type

class