GradientFunction Class Reference

#include <GradientFunction.h>

Inheritance diagram for GradientFunction:

Collaboration diagram for GradientFunction:

Public Member Functions
	GradientFunction ()
virtual	~GradientFunction ()
virtual void	clearDelta ()
	Clears the derivatives. More...
virtual unsigned int	nParams () const =0
	Returns the number of parameters. More...
virtual void	backpropagate (real *outputError)=0
	Backpropagates the error, updating the derivatives. More...
virtual void	update ()=0
	Updates the weights according to the derivatives. More...
virtual void	save (XFile *file)
	Saves the model to a file. More...
virtual void	load (XFile *file)
	Loads the model from a file. More...
Public Member Functions inherited from Function
	Function ()
virtual	~Function ()
virtual void	init ()
	Initializes the function. More...
virtual unsigned	nInputs () const =0
	Returns the number of inputs. More...
virtual unsigned	nOutputs () const =0
	Returns the number of outputs. More...
virtual void	setInputs (const real *input)
	Sets the value of the inputs. More...
virtual void	getOutputs (real *output) const
	Get the value of the outputs. More...
virtual void	setInput (int i, real x)=0
	Sets input i to value x. More...
virtual float	getOutput (int i) const =0
	Get output i. More...
virtual void	propagate ()=0
	Propagates inputs to outputs. More...

Public Attributes
real *	weights
	The weights (parameters) of the gradient function. More...
real *	dWeights
	The derivatives of the weights. More...

Detailed Description

Abstract class for gradient functions, such as a NeuralNetwork. A GradientFunction has a set of parameters (weights) and error derivatives. It can back-propagate the errors to compute the derivatives and udpate its weights accordingly.

Constructor & Destructor Documentation

GradientFunction::GradientFunction ( )

inline

virtual GradientFunction::~GradientFunction ( )

inlinevirtual

Member Function Documentation

virtual void GradientFunction::backpropagate ( real *  outputError )

pure virtual

Backpropagates the error, updating the derivatives.

Implemented in NeuralNetwork, and QFunction.

virtual void GradientFunction::clearDelta ( )

inlinevirtual

Clears the derivatives.

virtual void GradientFunction::load ( XFile *  file )

inlinevirtual

Loads the model from a file.

Implements Function.

Reimplemented in NeuralNetwork, and QFunction.

virtual unsigned int GradientFunction::nParams ( ) const

pure virtual

Returns the number of parameters.

Implemented in NeuralNetwork, and QFunction.

virtual void GradientFunction::save ( XFile *  file )

inlinevirtual

Saves the model to a file.

Implements Function.

Reimplemented in NeuralNetwork, and QFunction.

virtual void GradientFunction::update ( )

pure virtual

Updates the weights according to the derivatives.

Implemented in NeuralNetwork, and QFunction.

Member Data Documentation

real* GradientFunction::dWeights

The derivatives of the weights.

real* GradientFunction::weights

The weights (parameters) of the gradient function.

---

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

• src/qualia/learning/GradientFunction.h