tapas.report.attack_summary

Classes to summarise the output of an attack in terms of a range of metrics.

AttackSummary are produced by ThreatModel.test calls, and typically contain all relevant outputs of the test. They provide an interface to access and interpret these outputs.

Classes

AIAttackSummary(labels, predictions[, ...])

Class summarising the outputs of an attribute inference attack.

AttackSummary()

Summarise the results of an attack in a specific threat model.

BinaryAIAttackSummary(labels, predictions[, ...])

Class summarising the outputs of an attribute inference attack with binary values.

BinaryLabelInferenceAttackSummary(labels, ...)

LabelInferenceAttackSummary, where the label is binary.

LabelInferenceAttackSummary(labels, predictions)

Class summarising main performance metrics of a label-inference attack.

MIAttackSummary(labels, predictions[, ...])

Class summarising main performance metrics of a membership inference attack.