This abstract class represents the concept of loss index composed of an error term and a regularization term. More...

#include <loss_index.h>

Inheritance diagram for LossIndex:

Public Types
enum class	RegularizationMethod { L1 , L2 , NoRegularization }
	Enumeration of available regularization methods. More...

Public Member Functions
	LossIndex ()

	LossIndex (NeuralNetwork , DataSet )

virtual	~LossIndex ()
	Destructor. More...

NeuralNetwork *	get_neural_network_pointer () const
	Returns a pointer to the neural network object associated to the error term. More...

DataSet *	get_data_set_pointer () const
	Returns a pointer to the data set object associated to the error term. More...

const type &	get_regularization_weight () const
	Returns regularization weight. More...

const bool &	get_display () const

bool	has_neural_network () const

bool	has_data_set () const

RegularizationMethod	get_regularization_method () const
	Returns the regularization method. More...

void	set ()

void	set (NeuralNetwork *)

void	set (DataSet *)

void	set (NeuralNetwork , DataSet )

void	set (const LossIndex &)

void	set_threads_number (const int &)

void	set_neural_network_pointer (NeuralNetwork *)

virtual void	set_data_set_pointer (DataSet *)
	Sets a new data set on which the error term is to be measured. More...

void	set_default ()
	Sets the members of the error term to their default values: More...

void	set_regularization_method (const RegularizationMethod &)

void	set_regularization_method (const string &)

void	set_regularization_weight (const type &)

void	set_display (const bool &)

virtual void	set_normalization_coefficient ()

bool	has_selection () const
	Returns true if there are selection samples and false otherwise. More...

type	calculate_eta () const

type	calculate_h (const type &) const

Tensor< type, 1 >	calculate_gradient_numerical_differentiation ()

Tensor< type, 2 >	calculate_jacobian_numerical_differentiation ()

void	calculate_errors (const DataSetBatch &, const NeuralNetworkForwardPropagation &, LossIndexBackPropagation &) const

virtual void	calculate_error (const DataSetBatch &, const NeuralNetworkForwardPropagation &, LossIndexBackPropagation &) const =0

virtual void	calculate_output_delta (const DataSetBatch &, NeuralNetworkForwardPropagation &, LossIndexBackPropagation &) const =0

void	calculate_layers_delta (const DataSetBatch &, NeuralNetworkForwardPropagation &, LossIndexBackPropagation &) const

void	calculate_error_gradient (const DataSetBatch &, const NeuralNetworkForwardPropagation &, LossIndexBackPropagation &) const

void	back_propagate (const DataSetBatch &, NeuralNetworkForwardPropagation &, LossIndexBackPropagation &) const

void	calculate_errors_lm (const DataSetBatch &, const NeuralNetworkForwardPropagation &, LossIndexBackPropagationLM &) const

virtual void	calculate_squared_errors_lm (const DataSetBatch &, const NeuralNetworkForwardPropagation &, LossIndexBackPropagationLM &) const

virtual void	calculate_error_lm (const DataSetBatch &, const NeuralNetworkForwardPropagation &, LossIndexBackPropagationLM &) const

virtual void	calculate_output_delta_lm (const DataSetBatch &, NeuralNetworkForwardPropagation &, LossIndexBackPropagationLM &) const

void	calculate_layers_delta_lm (const DataSetBatch &, NeuralNetworkForwardPropagation &, LossIndexBackPropagationLM &) const

virtual void	calculate_error_gradient_lm (const DataSetBatch &, LossIndexBackPropagationLM &) const

void	calculate_squared_errors_jacobian_lm (const DataSetBatch &, NeuralNetworkForwardPropagation &, LossIndexBackPropagationLM &) const

virtual void	calculate_error_hessian_lm (const DataSetBatch &, LossIndexBackPropagationLM &) const

void	back_propagate_lm (const DataSetBatch &, NeuralNetworkForwardPropagation &, LossIndexBackPropagationLM &) const

type	calculate_regularization (const Tensor< type, 1 > &) const

void	calculate_regularization_gradient (const Tensor< type, 1 > &, Tensor< type, 1 > &) const

void	calculate_regularization_hessian (const Tensor< type, 1 > &, Tensor< type, 2 > &) const

void	from_XML (const tinyxml2::XMLDocument &)

virtual void	write_XML (tinyxml2::XMLPrinter &) const

void	regularization_from_XML (const tinyxml2::XMLDocument &)

void	write_regularization_XML (tinyxml2::XMLPrinter &) const

virtual string	get_error_type () const
	Returns a string with the default type of error term, "USER_PERFORMANCE_TERM". More...

virtual string	get_error_type_text () const
	Returns a string with the default type of error term in text format, "USER_PERFORMANCE_TERM". More...

string	write_regularization_method () const

void	check () const

Protected Attributes
NonBlockingThreadPool *	non_blocking_thread_pool = nullptr

ThreadPoolDevice *	thread_pool_device = nullptr

NeuralNetwork *	neural_network_pointer = nullptr
	Pointer to a neural network object. More...

DataSet *	data_set_pointer = nullptr
	Pointer to a data set object. More...

RegularizationMethod	regularization_method = RegularizationMethod::L2
	Pointer to a regularization method object. More...

type	regularization_weight = static_cast<type>(0.01)
	Regularization weight value. More...

bool	display = true
	Display messages to screen. More...

const Eigen::array< IndexPair< Index >, 1 >	AT_B = {IndexPair<Index>(0, 0)}

const Eigen::array< IndexPair< Index >, 1 >	A_B = {IndexPair<Index>(1, 0)}

const Eigen::array< IndexPair< Index >, 2 >	SSE = {IndexPair<Index>(0, 0), IndexPair<Index>(1, 1)}

const Eigen::array< int, 1 >	rows_sum = {Eigen::array<int, 1>({1})}

Detailed Description

This abstract class represents the concept of loss index composed of an error term and a regularization term.

The error term could be:

Cross Entropy Error.
Mean Squared Error.
Minkowski Error.
Normalized Squared Error.
Sum Squared Error.
Weighted Squared Error.

Definition at line 47 of file loss_index.h.

Member Enumeration Documentation

◆ RegularizationMethod

enum class RegularizationMethod

strong

Enumeration of available regularization methods.

Definition at line 66 of file loss_index.h.

Constructor & Destructor Documentation

◆ LossIndex() [1/2]

LossIndex ( )

explicit

Default constructor. It creates a default error term object, with all pointers initialized to nullptr. It also initializes all the rest of class members to their default values.

Definition at line 18 of file loss_index.cpp.

◆ LossIndex() [2/2]

LossIndex	(	NeuralNetwork *	new_neural_network_pointer,
		DataSet *	new_data_set_pointer
	)

explicit

Neural network and data set constructor. It creates a error term object associated to a neural network and to be measured on a data set. The rest of pointers are initialized to nullptr. It also initializes all the rest of class members to their default values.

Parameters

new_neural_network_pointer	Pointer to a neural network object.
new_data_set_pointer	Pointer to a data set object.

Definition at line 33 of file loss_index.cpp.

◆ ~LossIndex()

~LossIndex ( )

virtual

Destructor.

Definition at line 43 of file loss_index.cpp.

Member Function Documentation

◆ back_propagate()

void back_propagate	(	const DataSetBatch &	batch,
		NeuralNetworkForwardPropagation &	forward_propagation,
		LossIndexBackPropagation &	back_propagation
	)		const

Definition at line 419 of file loss_index.cpp.

◆ back_propagate_lm()

void back_propagate_lm	(	const DataSetBatch &	batch,
		NeuralNetworkForwardPropagation &	forward_propagation,
		LossIndexBackPropagationLM &	loss_index_back_propagation_lm
	)		const

This method calculates the second order loss. It is used for optimization of parameters during training. Returns a second order terms loss structure, which contains the values and the Hessian of the error terms function.

Todo:: Update method.

Definition at line 457 of file loss_index.cpp.

◆ calculate_error()

virtual void calculate_error	(	const DataSetBatch &	,
		const NeuralNetworkForwardPropagation &	,
		LossIndexBackPropagation &
	)		const

pure virtual

Implemented in MeanSquaredError, NormalizedSquaredError, and MinkowskiError.

◆ calculate_error_gradient()

void calculate_error_gradient	(	const DataSetBatch &	batch,
		const NeuralNetworkForwardPropagation &	forward_propagation,
		LossIndexBackPropagation &	back_propagation
	)		const

Definition at line 757 of file loss_index.cpp.

◆ calculate_error_gradient_lm()

void calculate_error_gradient_lm	(	const DataSetBatch &	batch,
		LossIndexBackPropagationLM &	loss_index_back_propagation_lm
	)		const

virtual

Reimplemented in WeightedSquaredError.

Definition at line 598 of file loss_index.cpp.

◆ calculate_error_hessian_lm()

virtual void calculate_error_hessian_lm	(	const DataSetBatch &	,
		LossIndexBackPropagationLM &
	)		const

inlinevirtual

Definition at line 215 of file loss_index.h.

◆ calculate_error_lm()

virtual void calculate_error_lm	(	const DataSetBatch &	,
		const NeuralNetworkForwardPropagation &	,
		LossIndexBackPropagationLM &
	)		const

inlinevirtual

Definition at line 196 of file loss_index.h.

◆ calculate_errors()

void calculate_errors	(	const DataSetBatch &	batch,
		const NeuralNetworkForwardPropagation &	forward_propagation,
		LossIndexBackPropagation &	back_propagation
	)		const

Definition at line 321 of file loss_index.cpp.

◆ calculate_errors_lm()

void calculate_errors_lm	(	const DataSetBatch &	batch,
		const NeuralNetworkForwardPropagation &	neural_network_forward_propagation,
		LossIndexBackPropagationLM &	loss_index_back_propagation
	)		const

Definition at line 366 of file loss_index.cpp.

◆ calculate_eta()

type calculate_eta ( ) const

Definition at line 1143 of file loss_index.cpp.

◆ calculate_gradient_numerical_differentiation()

Tensor< type, 1 > calculate_gradient_numerical_differentiation ( )

Definition at line 1016 of file loss_index.cpp.

◆ calculate_h()

type calculate_h ( const type & x ) const

Calculates a proper step size for computing the derivatives, as a function of the inputs point value.

Parameters

x	Input value.

Definition at line 1154 of file loss_index.cpp.

◆ calculate_jacobian_numerical_differentiation()

Tensor< type, 2 > calculate_jacobian_numerical_differentiation ( )

Definition at line 1076 of file loss_index.cpp.

◆ calculate_layers_delta()

void calculate_layers_delta	(	const DataSetBatch &	batch,
		NeuralNetworkForwardPropagation &	forward_propagation,
		LossIndexBackPropagation &	back_propagation
	)		const

Definition at line 701 of file loss_index.cpp.

◆ calculate_layers_delta_lm()

void calculate_layers_delta_lm	(	const DataSetBatch &	batch,
		NeuralNetworkForwardPropagation &	forward_propagation,
		LossIndexBackPropagationLM &	back_propagation
	)		const

Definition at line 729 of file loss_index.cpp.

◆ calculate_output_delta_lm()

virtual void calculate_output_delta_lm	(	const DataSetBatch &	,
		NeuralNetworkForwardPropagation &	,
		LossIndexBackPropagationLM &
	)		const

inlinevirtual

Definition at line 200 of file loss_index.h.

◆ calculate_regularization()

type calculate_regularization ( const Tensor< type, 1 > & parameters ) const

It calculates the regularization term using through the use of parameters. Returns the regularization evaluation, according to the respective regularization type used in the loss index expression.

Parameters

parameters vector with the parameters to get the regularization term.

Definition at line 648 of file loss_index.cpp.

◆ calculate_regularization_gradient()

void calculate_regularization_gradient	(	const Tensor< type, 1 > &	parameters,
		Tensor< type, 1 > &	regularization_gradient
	)		const

It calculate the regularization term using the gradient method. Returns the gradient of the regularization, according to the regularization type. That gradient is the vector of partial derivatives of the regularization with respect to the parameters. The size is thus the number of parameters

Parameters

parameters vector with the parameters to get the regularization term.

Definition at line 669 of file loss_index.cpp.

◆ calculate_regularization_hessian()

void calculate_regularization_hessian	(	const Tensor< type, 1 > &	parameters,
		Tensor< type, 2 > &	regularization_hessian
	)		const

It calculate the regularization term using the Hessian. Returns the Hessian of the regularization, according to the regularization type. That Hessian is the matrix of second partial derivatives of the regularization with respect to the parameters. That matrix is symmetric, with size the number of parameters.

Parameters

parameters vector with the parameters to get the regularization term.

Definition at line 688 of file loss_index.cpp.

◆ calculate_squared_errors_jacobian_lm()

void calculate_squared_errors_jacobian_lm	(	const DataSetBatch &	batch,
		NeuralNetworkForwardPropagation &	forward_propagation,
		LossIndexBackPropagationLM &	loss_index_back_propagation_lm
	)		const

Calculates the Jacobian matrix of the error terms from layers. Returns the Jacobian of the error terms function, according to the objective type used in the loss index expression. Note that this function is only defined when the objective can be expressed as a sum of squared terms. The Jacobian elements are the partial derivatives of a single term with respect to a single parameter. The number of rows in the Jacobian matrix are the number of parameters, and the number of columns the number of terms composing the objective.

Parameters

inputs	Tensor with inputs.
layers_activations	vector of tensors with layers activations.
layers_delta	vector of tensors with layers delta.

Definition at line 508 of file loss_index.cpp.

◆ calculate_squared_errors_lm()

void calculate_squared_errors_lm	(	const DataSetBatch &	,
		const NeuralNetworkForwardPropagation &	,
		LossIndexBackPropagationLM &	loss_index_back_propagation_lm
	)		const

virtual

Definition at line 411 of file loss_index.cpp.

◆ check()

void check ( ) const

Checks that there is a neural network associated to the error term. If some of the above conditions is not hold, the method throws an exception.

Definition at line 295 of file loss_index.cpp.

◆ from_XML()

void from_XML ( const tinyxml2::XMLDocument & document )

Loads a default error term from a XML document.

Parameters

document TinyXML document containing the error term members.

Definition at line 979 of file loss_index.cpp.

◆ get_data_set_pointer()

DataSet * get_data_set_pointer ( ) const

inline

Returns a pointer to the data set object associated to the error term.

Definition at line 92 of file loss_index.h.

◆ get_display()

const bool & get_display ( ) const

Returns true if messages from this class can be displayed on the screen, or false if messages from this class can't be displayed on the screen.

Definition at line 61 of file loss_index.cpp.

◆ get_error_type()

string get_error_type ( ) const

virtual

Returns a string with the default type of error term, "USER_PERFORMANCE_TERM".

Reimplemented in CrossEntropyError, MeanSquaredError, MinkowskiError, NormalizedSquaredError, SumSquaredError, and WeightedSquaredError.

Definition at line 608 of file loss_index.cpp.

◆ get_error_type_text()

string get_error_type_text ( ) const

virtual

Returns a string with the default type of error term in text format, "USER_PERFORMANCE_TERM".

Reimplemented in CrossEntropyError, MeanSquaredError, MinkowskiError, NormalizedSquaredError, SumSquaredError, and WeightedSquaredError.

Definition at line 616 of file loss_index.cpp.

◆ get_neural_network_pointer()

NeuralNetwork * get_neural_network_pointer ( ) const

inline

Returns a pointer to the neural network object associated to the error term.

Definition at line 70 of file loss_index.h.

◆ get_regularization_method()

LossIndex::RegularizationMethod get_regularization_method ( ) const

Returns the regularization method.

Definition at line 101 of file loss_index.cpp.

◆ get_regularization_weight()

const type & get_regularization_weight ( ) const

Returns regularization weight.

Definition at line 52 of file loss_index.cpp.

◆ has_data_set()

bool has_data_set ( ) const

Returns true if this error term object has a data set pointer associated, and false otherwise.

Definition at line 86 of file loss_index.cpp.

◆ has_neural_network()

bool has_neural_network ( ) const

Returns true if this error term object has a neural nework class pointer associated, and false otherwise

Definition at line 70 of file loss_index.cpp.

◆ has_selection()

bool has_selection ( ) const

Returns true if there are selection samples and false otherwise.

Definition at line 279 of file loss_index.cpp.

◆ regularization_from_XML()

void regularization_from_XML ( const tinyxml2::XMLDocument & document )

Definition at line 891 of file loss_index.cpp.

◆ set() [1/5]

void set ( )

Sets all the member pointers to nullptr(neural network, data set). It also initializes all the rest of class members to their default values.

Definition at line 110 of file loss_index.cpp.

◆ set() [2/5]

void set ( const LossIndex & other_error_term )

Sets to this error term object the members of another error term object.

Parameters

other_error_term Error term to be copied.

Definition at line 163 of file loss_index.cpp.

◆ set() [3/5]

void set ( DataSet * new_data_set_pointer )

Sets all the member pointers to nullptr, but the data set, which set to a given pointer. It also initializes all the rest of class members to their default values.

Parameters

new_data_set_pointer Pointer to a data set object.

Definition at line 136 of file loss_index.cpp.

◆ set() [4/5]

void set ( NeuralNetwork * new_neural_network_pointer )

Sets all the member pointers to nullptr, but the neural network, which set to a given pointer. It also initializes all the rest of class members to their default values.

Parameters

new_neural_network_pointer Pointer to a neural network object.

Definition at line 123 of file loss_index.cpp.

◆ set() [5/5]

void set	(	NeuralNetwork *	new_neural_network_pointer,
		DataSet *	new_data_set_pointer
	)

Sets new neural network and data set pointers. Finally, it initializes all the rest of class members to their default values.

Parameters

new_neural_network_pointer	Pointer to a neural network object.
new_data_set_pointer	Pointer to a data set object.

Definition at line 150 of file loss_index.cpp.

◆ set_data_set_pointer()

void set_data_set_pointer ( DataSet * new_data_set_pointer )

virtual

Sets a new data set on which the error term is to be measured.

Reimplemented in WeightedSquaredError, and NormalizedSquaredError.

Definition at line 196 of file loss_index.cpp.

◆ set_default()

void set_default ( )

Sets the members of the error term to their default values:

Definition at line 204 of file loss_index.cpp.

◆ set_display()

void set_display ( const bool & new_display )

Sets a new display value. If it is set to true messages from this class are to be displayed on the screen; if it is set to false messages from this class are not to be displayed on the screen.

Parameters

new_display Display value.

Definition at line 271 of file loss_index.cpp.

◆ set_neural_network_pointer()

void set_neural_network_pointer ( NeuralNetwork * new_neural_network_pointer )

Sets a pointer to a neural network object which is to be associated to the error term.

Parameters

new_neural_network_pointer Pointer to a neural network object to be associated to the error term.

Definition at line 188 of file loss_index.cpp.

◆ set_normalization_coefficient()

virtual void set_normalization_coefficient ( )

inlinevirtual

Reimplemented in NormalizedSquaredError, and WeightedSquaredError.

Definition at line 147 of file loss_index.h.

◆ set_regularization_method() [1/2]

void set_regularization_method ( const RegularizationMethod & new_regularization_method )

Sets the object with the regularization method.

Parameters

new_regularization_method String with method.

Definition at line 251 of file loss_index.cpp.

◆ set_regularization_method() [2/2]

void set_regularization_method ( const string & new_regularization_method )

Sets the object with the regularization method.

Parameters

new_regularization_method String with method.

Definition at line 221 of file loss_index.cpp.

◆ set_regularization_weight()

void set_regularization_weight ( const type & new_regularization_weight )

Sets the object with the regularization weights.

Parameters

new_regularization_method New regularization weight.

Definition at line 260 of file loss_index.cpp.

◆ set_threads_number()

void set_threads_number ( const int & new_threads_number )

Definition at line 175 of file loss_index.cpp.

◆ write_regularization_method()

string write_regularization_method ( ) const

Returns a string with the regularization information of the error term. It will be used by the training strategy to monitor the training process.

Definition at line 625 of file loss_index.cpp.

◆ write_regularization_XML()

void write_regularization_XML ( tinyxml2::XMLPrinter & file_stream ) const

Definition at line 928 of file loss_index.cpp.

◆ write_XML()

void write_XML ( tinyxml2::XMLPrinter & file_stream ) const

virtual

Serializes a default error term object into a XML document of the TinyXML library without keep the DOM tree in memory. See the OpenNN manual for more information about the format of this document.

Reimplemented in CrossEntropyError, MeanSquaredError, MinkowskiError, NormalizedSquaredError, SumSquaredError, and WeightedSquaredError.

Definition at line 881 of file loss_index.cpp.

Member Data Documentation

◆ A_B

const Eigen::array<IndexPair<Index>, 1> A_B = {IndexPair<Index>(1, 0)}

protected

Definition at line 273 of file loss_index.h.

◆ AT_B

const Eigen::array<IndexPair<Index>, 1> AT_B = {IndexPair<Index>(0, 0)}

protected

Definition at line 272 of file loss_index.h.

◆ data_set_pointer

DataSet* data_set_pointer = nullptr

protected

Pointer to a data set object.

Definition at line 258 of file loss_index.h.

◆ display

bool display = true

protected

Display messages to screen.

Definition at line 270 of file loss_index.h.

◆ neural_network_pointer

NeuralNetwork* neural_network_pointer = nullptr

protected

Pointer to a neural network object.

Definition at line 254 of file loss_index.h.

◆ non_blocking_thread_pool

NonBlockingThreadPool* non_blocking_thread_pool = nullptr

protected

Definition at line 249 of file loss_index.h.

◆ regularization_method

RegularizationMethod regularization_method = RegularizationMethod::L2

protected

Pointer to a regularization method object.

Definition at line 262 of file loss_index.h.

◆ regularization_weight

type regularization_weight = static_cast<type>(0.01)

protected

Regularization weight value.

Definition at line 266 of file loss_index.h.

◆ rows_sum

const Eigen::array<int, 1> rows_sum = {Eigen::array<int, 1>({1})}

protected

Definition at line 277 of file loss_index.h.

◆ SSE

const Eigen::array<IndexPair<Index>, 2> SSE = {IndexPair<Index>(0, 0), IndexPair<Index>(1, 1)}

protected

Definition at line 275 of file loss_index.h.

◆ thread_pool_device

ThreadPoolDevice* thread_pool_device = nullptr

protected

Definition at line 250 of file loss_index.h.

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

Public Types

Public Member Functions

Protected Attributes

Detailed Description

Member Enumeration Documentation

◆ RegularizationMethod

Constructor & Destructor Documentation

◆ LossIndex() [1/2]

◆ LossIndex() [2/2]

◆ ~LossIndex()

Member Function Documentation

◆ back_propagate()

◆ back_propagate_lm()

◆ calculate_error()

◆ calculate_error_gradient()

◆ calculate_error_gradient_lm()

◆ calculate_error_hessian_lm()

◆ calculate_error_lm()

◆ calculate_errors()

◆ calculate_errors_lm()

◆ calculate_eta()

◆ calculate_gradient_numerical_differentiation()

◆ calculate_h()

◆ calculate_jacobian_numerical_differentiation()

◆ calculate_layers_delta()

◆ calculate_layers_delta_lm()

◆ calculate_output_delta_lm()

◆ calculate_regularization()

◆ calculate_regularization_gradient()

◆ calculate_regularization_hessian()

◆ calculate_squared_errors_jacobian_lm()

◆ calculate_squared_errors_lm()

◆ check()

◆ from_XML()

◆ get_data_set_pointer()

◆ get_display()

◆ get_error_type()

◆ get_error_type_text()

◆ get_neural_network_pointer()

◆ get_regularization_method()

◆ get_regularization_weight()

◆ has_data_set()

◆ has_neural_network()

◆ has_selection()

◆ regularization_from_XML()

◆ set() [1/5]

◆ set() [2/5]

◆ set() [3/5]

◆ set() [4/5]

◆ set() [5/5]

◆ set_data_set_pointer()

◆ set_default()

◆ set_display()

◆ set_neural_network_pointer()

◆ set_normalization_coefficient()

◆ set_regularization_method() [1/2]

◆ set_regularization_method() [2/2]

◆ set_regularization_weight()

◆ set_threads_number()

◆ write_regularization_method()

◆ write_regularization_XML()

◆ write_XML()

Member Data Documentation

◆ A_B

◆ AT_B

◆ data_set_pointer

◆ display

◆ neural_network_pointer

◆ non_blocking_thread_pool

◆ regularization_method

◆ regularization_weight

◆ rows_sum

◆ SSE

◆ thread_pool_device