edu.stanford.nlp.optimization

Class AbstractCachingDiffFunction

java.lang.Object

edu.stanford.nlp.optimization.AbstractCachingDiffFunction

All Implemented Interfaces:

DiffFunction, Function, HasInitial

Direct Known Subclasses:

AbstractStochasticCachingDiffFunction, BiasedLogConditionalObjectiveFunction, BiasedLogisticObjectiveFunction, CRFLogConditionalObjectiveFunctionForLOP, CRFNonLinearLogConditionalObjectiveFunction, CRFNonLinearSecondOrderLogConditionalObjectiveFunction, DVParserCostAndGradient, GeneralizedExpectationObjectiveFunction, LogisticObjectiveFunction, SemiSupervisedLogConditionalObjectiveFunction, SentimentCostAndGradient

public abstract class AbstractCachingDiffFunction
extends Object
implements DiffFunction, HasInitial

A differentiable function that caches the last evaluation of its value and derivative.

Author:

Dan Klein

Field Summary

Fields

Modifier and Type	Field and Description
protected double[]	derivative
protected Random	generator
protected double	value

Constructor Summary

Constructors

Constructor and Description
AbstractCachingDiffFunction()

Method Summary

Modifier and Type	Method and Description
protected abstract void	calculate(double[] x) Calculate the value at x and the derivative and save them in the respective fields.
protected void	clearCache() Clears the cache in a way that doesn't require reallocation :-)
protected static void	copy(double[] copy, double[] orig)
double[]	derivativeAt(double[] x) Returns the first-derivative vector at the input location.
void	ensure(double[] x)
double[]	getDerivative()
boolean	gradientCheck()
boolean	gradientCheck(int numOfChecks, int numOfRandomChecks, double[] x)
double[]	initial() Returns the intitial point in the domain (but not necessarily a feasible one).
double	lastValue()
double[]	randomInitial()
double	valueAt(double[] x) Returns the value of the function at a single point.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Methods inherited from interface edu.stanford.nlp.optimization.Function

domainDimension

Field Detail

derivative

protected double[] derivative

value

protected double value

generator

protected Random generator

Constructor Detail

AbstractCachingDiffFunction

public AbstractCachingDiffFunction()

Method Detail

gradientCheck

public boolean gradientCheck()

gradientCheck

public boolean gradientCheck(int numOfChecks,
                             int numOfRandomChecks,
                             double[] x)

calculate

protected abstract void calculate(double[] x)

Calculate the value at x and the derivative and save them in the respective fields.

Parameters:

x - The point at which to calculate the function

clearCache

protected void clearCache()

Clears the cache in a way that doesn't require reallocation :-)

initial

public double[] initial()

Description copied from interface: HasInitial

Returns the intitial point in the domain (but not necessarily a feasible one).

Specified by:

initial in interface HasInitial

Returns:

a domain point

randomInitial

public double[] randomInitial()

copy

protected static void copy(double[] copy,
                           double[] orig)

ensure

public void ensure(double[] x)

valueAt

public double valueAt(double[] x)

Description copied from interface: Function

Returns the value of the function at a single point.

Specified by:

valueAt in interface Function

Parameters:

x - a double[] input

Returns:

the function value at the input

derivativeAt

public double[] derivativeAt(double[] x)

Description copied from interface: DiffFunction

Returns the first-derivative vector at the input location.

Specified by:

derivativeAt in interface DiffFunction

Parameters:

x - a double[] input vector

Returns:

the vector of first partial derivatives.

lastValue

public double lastValue()

getDerivative

public double[] getDerivative()