|View source on GitHub|
Maximum Mean Discrepency between predictions on two groups of examples.
model_remediation.min_diff.losses.MMDLoss( kernel='gaussian', predictions_transform=None, name: Optional[Text] = None )
String (name of kernel) or
Optional transform function to be applied to the
predictions. This can be used to smooth out the distributions or limit the
range of predictions.
The choice of whether to apply a transform to the predictions is task and
data dependent. For example, for classifiers, it might make sense to apply
Name used for logging and tracking. Defaults to
The Maximum Mean Discrepancy (MMD) is a measure of the distance between the distributions of prediction scores on two groups of examples. The metric guarantees that the result is 0 if and only if the two distributions it is comparing are exactly the same.
membership input indicates with a numerical value whether
each example is part of the sensitive group with a numerical value. This
currently only supports hard membership of
For more details, see the paper.