Represents an optimizable object. More...

#include <Optimizable.h>

Inheritance diagram for OpenANN::Optimizable:

Public Member Functions
virtual	~Optimizable ()

virtual void	finishedIteration ()
	This callback is called after each optimization algorithm iteration. More...

Batch Methods
Functions that must be implemented in every Optimizable.
virtual bool	providesInitialization ()=0
	Check if the object knows how to initialize its parameters. More...

virtual void	initialize ()=0
	Initialize the optimizable parameters. More...

virtual unsigned	dimension ()=0
	Request the number of optimizable parameters. More...

virtual const Eigen::VectorXd &	currentParameters ()=0
	Request the current parameters. More...

virtual void	setParameters (const Eigen::VectorXd &parameters)=0
	Set new parameters. More...

virtual double	error ()=0
	Compute error on training set. More...

virtual bool	providesGradient ()=0
	Check if the object provides a gradient of the error function with respect to its parameters. More...

virtual Eigen::VectorXd	gradient ()=0
	Compute gradient of the error function with respect to the parameters. More...

Mini-batch Methods
Functions that should be implemented to speed up optimization and are required by some optimization algorithms.
virtual unsigned	examples ()
	Request number of training examples. More...

virtual double	error (unsigned n)
	Compute error of a given training example. More...

virtual Eigen::VectorXd	gradient (unsigned n)
	Compute gradient of a given training example. More...

virtual void	errorGradient (int n, double &value, Eigen::VectorXd &grad)
	Calculates the function value and gradient of a training example. More...

virtual void	errorGradient (double &value, Eigen::VectorXd &grad)
	Calculates the function value and gradient of all training examples. More...

virtual Eigen::VectorXd	error (std::vector< int >::const_iterator startN, std::vector< int >::const_iterator endN)
	Calculates the errors of given training examples. More...

virtual Eigen::VectorXd	gradient (std::vector< int >::const_iterator startN, std::vector< int >::const_iterator endN)
	Calculates the accumulated gradient of given training examples. More...

virtual void	errorGradient (std::vector< int >::const_iterator startN, std::vector< int >::const_iterator endN, double &value, Eigen::VectorXd &grad)
	Calculates the accumulated gradient and error of given training examples. More...

Detailed Description

Represents an optimizable object.

An optimizable object can be arbitrary as long as it provides some objective function. The objective function in this context is called error function and will be minimized. This could be e.g. the sum of squared error between predictions and targets for a neural network. But the idea behind this is more general. We can e.g. optimize the reward in a reinforcement learning problem or the energy of an unsupervised model like an RBM.

Constructor & Destructor Documentation

virtual OpenANN::Optimizable::~Optimizable ( )

inlinevirtual

Member Function Documentation

virtual const Eigen::VectorXd& OpenANN::Optimizable::currentParameters ( )

pure virtual

Request the current parameters.

Returns: current parameters

Implemented in OpenANN::Net, OpenANN::RBM, OpenANN::IntrinsicPlasticity, NeuroEvolutionAgent, and OpenANN::SparseAutoEncoder.

virtual unsigned OpenANN::Optimizable::dimension ( )

pure virtual

Request the number of optimizable parameters.

Returns: number of optimizable parameters

Implemented in OpenANN::Net, OpenANN::RBM, NeuroEvolutionAgent, OpenANN::IntrinsicPlasticity, and OpenANN::SparseAutoEncoder.

virtual double OpenANN::Optimizable::error ( )

pure virtual

Compute error on training set.

Returns: current error on training set or objective function value

Implemented in OpenANN::Net, OpenANN::RBM, NeuroEvolutionAgent, OpenANN::IntrinsicPlasticity, and OpenANN::SparseAutoEncoder.

virtual double OpenANN::Optimizable::error ( unsigned n)

inlinevirtual

Compute error of a given training example.

Parameters

n	index of the training example in the dataset

Returns: error of the n-th training example

virtual Eigen::VectorXd OpenANN::Optimizable::error	(	std::vector< int >::const_iterator	startN,
		std::vector< int >::const_iterator	endN
	)

virtual

Calculates the errors of given training examples.

Parameters

startN	iterator over index vector
endN	iterator over index vector

Returns: each row contains the error for one training example

virtual void OpenANN::Optimizable::errorGradient	(	int	n,
		double &	value,
		Eigen::VectorXd &	grad
	)

virtual

Calculates the function value and gradient of a training example.

Parameters

n	index of training example
value	function value
grad	gradient of the function, lenght must be dimension()

Reimplemented in OpenANN::Net.

virtual void OpenANN::Optimizable::errorGradient	(	double &	value,
		Eigen::VectorXd &	grad
	)

virtual

Calculates the function value and gradient of all training examples.

Parameters

value	function value
grad	gradient of the function, lenght must be dimension()

Reimplemented in OpenANN::Net, and OpenANN::SparseAutoEncoder.

virtual void OpenANN::Optimizable::errorGradient	(	std::vector< int >::const_iterator	startN,
		std::vector< int >::const_iterator	endN,
		double &	value,
		Eigen::VectorXd &	grad
	)

virtual

Calculates the accumulated gradient and error of given training examples.

Parameters

startN	iterator over index vector
endN	iterator over index vector
value	function value
grad	gradient of the function, lenght must be dimension()

Returns: each row contains the gradient for one training example

Reimplemented in OpenANN::Net, and OpenANN::RBM.

virtual unsigned OpenANN::Optimizable::examples ( )

inlinevirtual

Request number of training examples.

Returns: number of training examples

Reimplemented in OpenANN::Net, OpenANN::RBM, and OpenANN::IntrinsicPlasticity.

virtual void OpenANN::Optimizable::finishedIteration ( )

inlinevirtual

This callback is called after each optimization algorithm iteration.

Reimplemented in OpenANN::Net.

virtual Eigen::VectorXd OpenANN::Optimizable::gradient ( )

pure virtual

Compute gradient of the error function with respect to the parameters.

Returns: gradient

Implemented in OpenANN::Net, OpenANN::RBM, OpenANN::IntrinsicPlasticity, NeuroEvolutionAgent, and OpenANN::SparseAutoEncoder.

virtual Eigen::VectorXd OpenANN::Optimizable::gradient ( unsigned n)

inlinevirtual

Compute gradient of a given training example.

Parameters

n	index of the training example in the dataset

Returns: gradient of the n-th training example

virtual Eigen::VectorXd OpenANN::Optimizable::gradient	(	std::vector< int >::const_iterator	startN,
		std::vector< int >::const_iterator	endN
	)

virtual

Calculates the accumulated gradient of given training examples.

Parameters

startN	iterator over index vector
endN	iterator over index vector

Returns: each row contains the gradient for one training example

virtual void OpenANN::Optimizable::initialize ( )

pure virtual

Initialize the optimizable parameters.

Implemented in OpenANN::Net, OpenANN::RBM, NeuroEvolutionAgent, OpenANN::IntrinsicPlasticity, and OpenANN::SparseAutoEncoder.

virtual bool OpenANN::Optimizable::providesGradient ( )

pure virtual

Check if the object provides a gradient of the error function with respect to its parameters.

Returns: does the optimizable provide a gradient?

Implemented in OpenANN::Net, OpenANN::RBM, NeuroEvolutionAgent, OpenANN::IntrinsicPlasticity, and OpenANN::SparseAutoEncoder.

virtual bool OpenANN::Optimizable::providesInitialization ( )

pure virtual

Check if the object knows how to initialize its parameters.

Returns: does the optimizable object provide a parameter initialization?

Implemented in OpenANN::Net, OpenANN::RBM, NeuroEvolutionAgent, OpenANN::IntrinsicPlasticity, and OpenANN::SparseAutoEncoder.

virtual void OpenANN::Optimizable::setParameters ( const Eigen::VectorXd & parameters)

pure virtual

Set new parameters.

Parameters

parameters new parameters

Implemented in OpenANN::Net, OpenANN::RBM, NeuroEvolutionAgent, OpenANN::IntrinsicPlasticity, and OpenANN::SparseAutoEncoder.

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

OpenANN/optimization/Optimizable.h

Public Member Functions

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation