pygip.models.attack.DataFreeMEA¶
Classes
|
|
|
Type I: Uses victim outputs + gradients for surrogate training. |
|
Type II: Uses victim outputs only (hard labels). |
|
Type III: Two surrogates with victim supervision + consistency. |
|
- class pygip.models.attack.DataFreeMEA.DFEAAttack(dataset, attack_node_fraction, model_path=None)[source]¶
Bases:
BaseAttack
- _abc_impl = <_abc_data object>¶
- supported_api_types = {'dgl'}¶
- class pygip.models.attack.DataFreeMEA.DFEATypeI(dataset, attack_node_fraction, model_path=None)[source]¶
Bases:
DFEAAttack
Type I: Uses victim outputs + gradients for surrogate training.
- _abc_impl = <_abc_data object>¶
- class pygip.models.attack.DataFreeMEA.DFEATypeII(dataset, attack_node_fraction, model_path=None)[source]¶
Bases:
DFEAAttack
Type II: Uses victim outputs only (hard labels).
- _abc_impl = <_abc_data object>¶
- class pygip.models.attack.DataFreeMEA.DFEATypeIII(dataset, attack_node_fraction, model_path=None)[source]¶
Bases:
DFEAAttack
Type III: Two surrogates with victim supervision + consistency.
- _abc_impl = <_abc_data object>¶