secure_learning.regularizers

Contains penalty functions.

Classes

BaseRegularizer	Base class for regularizations.
DifferentiableRegularizer	Differentiable regularizations can be included via their gradient.
L2Regularizer	Implementation for L2 regularization:
NonDifferentiableRegularizer	Non-differential regularization can be included via a proximal
L1Regularizer	Implementation for L1 regularization: $f(w) = ||w||_1$.

Module Contents

class secure_learning.regularizers.BaseRegularizer

Base class for regularizations.

class secure_learning.regularizers.DifferentiableRegularizer

Differentiable regularizations can be included via their gradient.

class secure_learning.regularizers.L2Regularizer(alpha: float)

Implementation for L2 regularization: $$f(w) = frac{alpha}{2} times ||w||^2_2`.$$

alpha

class secure_learning.regularizers.NonDifferentiableRegularizer

Non-differential regularization can be included via a proximal method.

class secure_learning.regularizers.L1Regularizer(alpha: float)

Implementation for L1 regularization: $f(w) = ||w||_1$.

alpha