mlpack::ann::OrthogonalRegularizer Class Reference

Implementation of the OrthogonalRegularizer. More...

#include <orthogonal_regularizer.hpp>

Public Member Functions
	OrthogonalRegularizer (double factor=1.0)
	Create the regularizer object. More...
template<typename MatType >
void	Evaluate (const MatType &weight, MatType &gradient)
	Calculate the gradient for regularization. More...
template<typename Archive >
void	serialize (Archive &ar, const uint32_t)
	Serialize the regularizer (nothing to do).

Public Attributes
double	factor
	The constant for the regularization.

Detailed Description

Implementation of the OrthogonalRegularizer.

Orthogonality of weights is a desirable property because multiplication by an orthogonal matrix leaves the norm of the matrix unchanged. The orthogonal regularization technique encourages weights to be orthogonal.

For more information, see the following.

@inproceedings{WanICML2013,
  title={Neural Photo Editing with Introspective Adversarial Networks},
  booktitle = {5th International Conference on Learning Representations
               (ICLR - 17)},
  author = {Andrew Brock and Theodore Lim and J.M. Ritchie and Nick Weston},
  year = {2017}
}

Constructor & Destructor Documentation

OrthogonalRegularizer()

mlpack::ann::OrthogonalRegularizer::OrthogonalRegularizer

(

double

factor = 1.0

)

Create the regularizer object.

Parameters

factor

The factor for regularization.

Member Function Documentation

Evaluate()

template<typename MatType >

void mlpack::ann::OrthogonalRegularizer::Evaluate	(	const MatType &	weight,
		MatType &	gradient
	)

Calculate the gradient for regularization.

Template Parameters

MatType

Type of weight matrix.

Parameters

weight	The weight matrix to be regularized.
gradient	The calculated gradient.

---

The documentation for this class was generated from the following files:

• src/mlpack/methods/ann/regularizer/orthogonal_regularizer.hpp
• src/mlpack/methods/ann/regularizer/orthogonal_regularizer_impl.hpp