Standard sigmoid function. More...

#include <shark/Models/SigmoidModel.h>

Inheritance diagram for shark::SigmoidModel:

Public Member Functions
SHARK_EXPORT_SYMBOL	SigmoidModel (bool transform_for_unconstrained=true)

std::string	name () const
	From INameable: return the class name. More...

SHARK_EXPORT_SYMBOL RealVector	parameterVector () const
	Return the parameter vector. More...

SHARK_EXPORT_SYMBOL void	setParameterVector (RealVector const &newParameters)

std::size_t	numberOfParameters () const
	Return the number of parameters. More...

SHARK_EXPORT_SYMBOL void	setOffsetActivity (bool enable_offset)

bool	hasOffset () const

bool	slopeIsExpEncoded () const

virtual SHARK_EXPORT_SYMBOL double	sigmoid (double x) const
	activation function \(g_{output}(x)\) More...

virtual double	sigmoidDerivative (double gx) const
	Computes the derivative of the activation function \(g_{output}(x)\) for the output given the last response of the model gx=g(x) More...

boost::shared_ptr< State >	createState () const
	Creates an internal state of the model. More...

SHARK_EXPORT_SYMBOL void	eval (BatchInputType const &pattern, BatchOutputType &output, State &state) const
	Standard interface for evaluating the response of the model to a batch of patterns. More...

SHARK_EXPORT_SYMBOL void	eval (BatchInputType const &pattern, BatchOutputType &output) const
	Standard interface for evaluating the response of the model to a batch of patterns. More...

SHARK_EXPORT_SYMBOL void	weightedParameterDerivative (BatchInputType const &pattern, BatchOutputType const &coefficients, State const &state, RealVector &gradient) const
	calculates the weighted sum of derivatives w.r.t the parameters. More...

SHARK_EXPORT_SYMBOL void	weightedInputDerivative (BatchInputType const &pattern, BatchOutputType const &coefficients, State const &state, BatchInputType &derivative) const
	calculates the weighted sum of derivatives w.r.t the inputs More...

std::size_t	inputSize () const

std::size_t	outputSize () const

void	setMinLogValue (double logvalue=-230.0)

void	read (InArchive &archive)
	From ISerializable, reads a model from an archive. More...

void	write (OutArchive &archive) const
	From ISerializable, writes a model to an archive. More...

Public Member Functions inherited from shark::AbstractModel< RealVector, RealVector >
	AbstractModel ()

virtual	~AbstractModel ()

const Features &	features () const

virtual void	updateFeatures ()

bool	hasFirstParameterDerivative () const
	Returns true when the first parameter derivative is implemented. More...

bool	hasSecondParameterDerivative () const
	Returns true when the second parameter derivative is implemented. More...

bool	hasFirstInputDerivative () const
	Returns true when the first input derivative is implemented. More...

bool	hasSecondInputDerivative () const
	Returns true when the second parameter derivative is implemented. More...

bool	isSequential () const

virtual void	eval (InputType const &pattern, OutputType &output) const
	Standard interface for evaluating the response of the model to a single pattern. More...

Data< OutputType >	operator() (Data< InputType > const &patterns) const
	Model evaluation as an operator for a whole dataset. This is a convenience function. More...

OutputType	operator() (InputType const &pattern) const
	Model evaluation as an operator for a single pattern. This is a convenience function. More...

BatchOutputType	operator() (BatchInputType const &patterns) const
	Model evaluation as an operator for a single pattern. This is a convenience function. More...

virtual void	weightedParameterDerivative (BatchInputType const &pattern, BatchOutputType const &coefficients, Batch< RealMatrix >::type const &errorHessian, State const &state, RealVector &derivative, RealMatrix &hessian) const
	calculates the weighted sum of derivatives w.r.t the parameters More...

virtual void	weightedInputDerivative (BatchInputType const &pattern, BatchOutputType const &coefficients, typename Batch< RealMatrix >::type const &errorHessian, State const &state, RealMatrix &derivative, Batch< RealMatrix >::type &hessian) const
	calculates the weighted sum of derivatives w.r.t the inputs More...

virtual void	weightedDerivatives (BatchInputType const &patterns, BatchOutputType const &coefficients, State const &state, RealVector &parameterDerivative, BatchInputType &inputDerivative) const
	calculates weighted input and parameter derivative at the same time More...

Public Member Functions inherited from shark::IParameterizable
virtual	~IParameterizable ()

Public Member Functions inherited from shark::INameable
virtual	~INameable ()

Public Member Functions inherited from shark::ISerializable
virtual	~ISerializable ()
	Virtual d'tor. More...

void	load (InArchive &archive, unsigned int version)
	Versioned loading of components, calls read(...). More...

void	save (OutArchive &archive, unsigned int version) const
	Versioned storing of components, calls write(...). More...

	BOOST_SERIALIZATION_SPLIT_MEMBER ()

Protected Attributes
RealVector	m_parameters
	the parameter vector More...

bool	m_useOffset
	whether or not to allow non-zero offset values More...

bool	m_transformForUnconstrained
	flag for encoding variant More...

double	m_minLogValue
	what value should be returned as log-encoded slope if the true slope is actually zero More...

Protected Attributes inherited from shark::AbstractModel< RealVector, RealVector >
Features	m_features

Additional Inherited Members
Public Types inherited from shark::AbstractModel< RealVector, RealVector >
enum	Feature

typedef RealVector	InputType
	Defines the input type of the model. More...

typedef RealVector	OutputType
	Defines the output type of the model. More...

typedef Batch< InputType >::type	BatchInputType
	defines the batch type of the input type. More...

typedef Batch< OutputType >::type	BatchOutputType
	defines the batch type of the output type More...

typedef TypedFlags< Feature >	Features

typedef TypedFeatureNotAvailableException< Feature >	FeatureNotAvailableException

Detailed Description

Standard sigmoid function.

: This model maps a real-valued input to the unit interval by the sigmoid function \( f(x) = \frac{1}{1 + \exp(-Ax+B))} \), where the real-valued model parameter A controls the slope, and the real-valued offset model parameter B controls the position of the symmetry point. This is a special case of a feed forward neural network consisting of a single sigmoid layer. Note that the parameter A is expected to be non-negative (and hence does not incorporate the minus sign in the sigmoid's equation). Also, the offset parameter can be disabled using the setOffsetActivity() member function.

See also: TanhSigmoidModel SimpleSigmoidModel

Definition at line 51 of file SigmoidModel.h.

Constructor & Destructor Documentation

◆ SigmoidModel()

SHARK_EXPORT_SYMBOL shark::SigmoidModel::SigmoidModel ( bool transform_for_unconstrained = true )

default ctor

Parameters

transform_for_unconstrained when a new paramVector is set, should the exponent of the first parameter be used as the sigmoid's slope?

Member Function Documentation

◆ createState()

boost::shared_ptr<State> shark::SigmoidModel::createState ( ) const

inlinevirtual

Creates an internal state of the model.

The state is needed when the derivatives are to be calculated. Eval can store a state which is then reused to speed up the calculations of the derivatives. This also allows eval to be evaluated in parallel!

Reimplemented from shark::AbstractModel< RealVector, RealVector >.

Definition at line 102 of file SigmoidModel.h.

References eval(), SHARK_EXPORT_SYMBOL, weightedInputDerivative(), and weightedParameterDerivative().

◆ eval() [1/2]

SHARK_EXPORT_SYMBOL void shark::SigmoidModel::eval	(	BatchInputType const &	patterns,
		BatchOutputType &	outputs,
		State &	state
	)		const

virtual

Standard interface for evaluating the response of the model to a batch of patterns.

Parameters

patterns	the inputs of the model
outputs	the predictions or response of the model to every pattern
state	intermediate results stored by eval which can be reused for derivative computation.

Implements shark::AbstractModel< RealVector, RealVector >.

Referenced by createState().

◆ eval() [2/2]

SHARK_EXPORT_SYMBOL void shark::SigmoidModel::eval	(	BatchInputType const &	patterns,
		BatchOutputType &	outputs
	)		const

virtual

Standard interface for evaluating the response of the model to a batch of patterns.

Parameters

patterns	the inputs of the model
outputs	the predictions or response of the model to every pattern

Reimplemented from shark::AbstractModel< RealVector, RealVector >.

◆ hasOffset()

bool shark::SigmoidModel::hasOffset ( ) const

inline

Definition at line 83 of file SigmoidModel.h.

References m_useOffset.

◆ inputSize()

std::size_t shark::SigmoidModel::inputSize ( ) const

inline

Definition at line 118 of file SigmoidModel.h.

◆ name()

std::string shark::SigmoidModel::name ( ) const

inlinevirtual

From INameable: return the class name.

Reimplemented from shark::INameable.

Reimplemented in shark::TanhSigmoidModel, and shark::SimpleSigmoidModel.

Definition at line 68 of file SigmoidModel.h.

References parameterVector(), setParameterVector(), and SHARK_EXPORT_SYMBOL.

◆ numberOfParameters()

std::size_t shark::SigmoidModel::numberOfParameters ( ) const

inlinevirtual

Return the number of parameters.

Reimplemented from shark::IParameterizable.

Definition at line 77 of file SigmoidModel.h.

References setOffsetActivity(), and SHARK_EXPORT_SYMBOL.

◆ outputSize()

std::size_t shark::SigmoidModel::outputSize ( ) const

inline

Definition at line 121 of file SigmoidModel.h.

References read(), setMinLogValue(), and write().

◆ parameterVector()

SHARK_EXPORT_SYMBOL RealVector shark::SigmoidModel::parameterVector ( ) const

virtual

Return the parameter vector.

Reimplemented from shark::IParameterizable.

Referenced by shark::SvmLogisticInterpretation< InputType >::evalDerivative(), and name().

◆ read()

void shark::SigmoidModel::read ( InArchive & archive )

virtual

From ISerializable, reads a model from an archive.

Reimplemented from shark::AbstractModel< RealVector, RealVector >.

Referenced by outputSize().

◆ setMinLogValue()

void shark::SigmoidModel::setMinLogValue ( double logvalue = -230.0 )

set the minimum log value that should be returned as log-encoded slope if the true slope is actually zero. default in ctor sets -230. param logvalue the new minimum log value

Referenced by outputSize().

◆ setOffsetActivity()

SHARK_EXPORT_SYMBOL void shark::SigmoidModel::setOffsetActivity ( bool enable_offset )

Referenced by numberOfParameters().

◆ setParameterVector()

SHARK_EXPORT_SYMBOL void shark::SigmoidModel::setParameterVector ( RealVector const & newParameters )

virtual

note that the parameters are not expected to incorporate the minus sign in the sigmoid's equation

Parameters

newParameters the new parameter vector A and offset B concatenated

Reimplemented from shark::IParameterizable.

Referenced by name().

◆ sigmoid()

virtual SHARK_EXPORT_SYMBOL double shark::SigmoidModel::sigmoid ( double x ) const

virtual

activation function \(g_{output}(x)\)

Reimplemented in shark::TanhSigmoidModel, and shark::SimpleSigmoidModel.

Referenced by slopeIsExpEncoded().

◆ sigmoidDerivative()

virtual double shark::SigmoidModel::sigmoidDerivative ( double gx ) const

virtual

Computes the derivative of the activation function \(g_{output}(x)\) for the output given the last response of the model gx=g(x)

Reimplemented in shark::TanhSigmoidModel, and shark::SimpleSigmoidModel.

Referenced by slopeIsExpEncoded().

◆ slopeIsExpEncoded()

bool shark::SigmoidModel::slopeIsExpEncoded ( ) const

inline

Definition at line 87 of file SigmoidModel.h.

References m_transformForUnconstrained, SHARK_EXPORT_SYMBOL, sigmoid(), and sigmoidDerivative().

◆ weightedInputDerivative()

SHARK_EXPORT_SYMBOL void shark::SigmoidModel::weightedInputDerivative	(	BatchInputType const &	pattern,
		BatchOutputType const &	coefficients,
		State const &	state,
		BatchInputType &	derivative
	)		const

virtual

calculates the weighted sum of derivatives w.r.t the inputs

Parameters

pattern	the patterns to evaluate
coefficients	the coefficients which are used to calculate the weighted sum for every pattern
state	intermediate results stored by eval to sped up calculations of the derivatives
derivative	the calculated derivative for every pattern

Reimplemented from shark::AbstractModel< RealVector, RealVector >.

Referenced by createState().

◆ weightedParameterDerivative()

SHARK_EXPORT_SYMBOL void shark::SigmoidModel::weightedParameterDerivative	(	BatchInputType const &	pattern,
		BatchOutputType const &	coefficients,
		State const &	state,
		RealVector &	derivative
	)		const

virtual

calculates the weighted sum of derivatives w.r.t the parameters.

Parameters

pattern	the patterns to evaluate
coefficients	the coefficients which are used to calculate the weighted sum for every pattern
state	intermediate results stored by eval to speed up calculations of the derivatives
derivative	the calculated derivative as sum over all derivates of all patterns

Reimplemented from shark::AbstractModel< RealVector, RealVector >.

Referenced by createState().

◆ write()

void shark::SigmoidModel::write ( OutArchive & archive ) const

virtual

From ISerializable, writes a model to an archive.

Reimplemented from shark::AbstractModel< RealVector, RealVector >.

Referenced by outputSize().

Member Data Documentation

◆ m_minLogValue

double shark::SigmoidModel::m_minLogValue

protected

what value should be returned as log-encoded slope if the true slope is actually zero

Definition at line 140 of file SigmoidModel.h.

◆ m_parameters

RealVector shark::SigmoidModel::m_parameters

protected

the parameter vector

Definition at line 136 of file SigmoidModel.h.

◆ m_transformForUnconstrained

bool shark::SigmoidModel::m_transformForUnconstrained

protected

flag for encoding variant

Definition at line 139 of file SigmoidModel.h.

Referenced by slopeIsExpEncoded().

◆ m_useOffset

bool shark::SigmoidModel::m_useOffset

protected

whether or not to allow non-zero offset values

Definition at line 137 of file SigmoidModel.h.

Referenced by hasOffset().

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

include/shark/Models/SigmoidModel.h

Public Member Functions

Protected Attributes

Additional Inherited Members

Detailed Description

Constructor & Destructor Documentation

◆ SigmoidModel()

Member Function Documentation

◆ createState()

◆ eval() [1/2]

◆ eval() [2/2]

◆ hasOffset()

◆ inputSize()

◆ name()

◆ numberOfParameters()

◆ outputSize()

◆ parameterVector()

◆ read()

◆ setMinLogValue()

◆ setOffsetActivity()

◆ setParameterVector()

◆ sigmoid()

◆ sigmoidDerivative()

◆ slopeIsExpEncoded()

◆ weightedInputDerivative()

◆ weightedParameterDerivative()

◆ write()

Member Data Documentation

◆ m_minLogValue

◆ m_parameters

◆ m_transformForUnconstrained

◆ m_useOffset