edu.stanford.nlp.parser.dvparser

Class DVParserCostAndGradient

java.lang.Object

edu.stanford.nlp.optimization.AbstractCachingDiffFunction

edu.stanford.nlp.parser.dvparser.DVParserCostAndGradient

All Implemented Interfaces:

DiffFunction, Function, HasInitial

public class DVParserCostAndGradient
extends AbstractCachingDiffFunction

Field Summary

Fields inherited from class edu.stanford.nlp.optimization.AbstractCachingDiffFunction

derivative, generator, value

Constructor Summary

Constructors

Constructor and Description
DVParserCostAndGradient(java.util.List<Tree> trainingBatch, java.util.IdentityHashMap<Tree,java.util.List<Tree>> topParses, DVModel dvModel, Options op)

Method Summary

Methods

Modifier and Type	Method and Description
void	backpropDerivative(Tree tree, java.util.List<java.lang.String> words, java.util.IdentityHashMap<Tree,SimpleMatrix> nodeVectors, TwoDimensionalMap<java.lang.String,java.lang.String,SimpleMatrix> binaryW_dfs, java.util.Map<java.lang.String,SimpleMatrix> unaryW_dfs, TwoDimensionalMap<java.lang.String,java.lang.String,SimpleMatrix> binaryScoreDerivatives, java.util.Map<java.lang.String,SimpleMatrix> unaryScoreDerivatives, java.util.Map<java.lang.String,SimpleMatrix> wordVectorDerivatives)
void	backpropDerivative(Tree tree, java.util.List<java.lang.String> words, java.util.IdentityHashMap<Tree,SimpleMatrix> nodeVectors, TwoDimensionalMap<java.lang.String,java.lang.String,SimpleMatrix> binaryW_dfs, java.util.Map<java.lang.String,SimpleMatrix> unaryW_dfs, TwoDimensionalMap<java.lang.String,java.lang.String,SimpleMatrix> binaryScoreDerivatives, java.util.Map<java.lang.String,SimpleMatrix> unaryScoreDerivatives, java.util.Map<java.lang.String,SimpleMatrix> wordVectorDerivatives, SimpleMatrix deltaUp)
void	calculate(double[] theta) Calculate the value at x and the derivative and save them in the respective fields.
static SimpleMatrix	concatenate(SimpleMatrix... vectors) Concatenates several column vectors into one large column vector
static SimpleMatrix	concatenateWithBias(SimpleMatrix... vectors) Concatenates several column vectors into one large column vector, adds a 1.0 at the end as a bias term
int	domainDimension() Returns the number of dimensions in the function's domain
static SimpleMatrix	elementwiseApplyNonlinearity(SimpleMatrix input)
java.util.List<DeepTree>	getAllHighestScoringTreesTest(java.util.List<Tree> trees)
DeepTree	getHighestScoringTree(Tree tree, double lambda)
double	getMargin(Tree goldTree, Tree bestHypothesis)
static SimpleMatrix	nonlinearityVectorToDerivative(SimpleMatrix input)
static void	outputSpans(Tree tree)
double	score(Tree tree, java.util.IdentityHashMap<Tree,SimpleMatrix> nodeVectors)
static double	sigmoid(double x)

Methods inherited from class edu.stanford.nlp.optimization.AbstractCachingDiffFunction

clearCache, copy, derivativeAt, getDerivative, gradientCheck, gradientCheck, initial, lastValue, randomInitial, valueAt

Methods inherited from class java.lang.Object

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

Constructor Detail

DVParserCostAndGradient

public DVParserCostAndGradient(java.util.List<Tree> trainingBatch,
                       java.util.IdentityHashMap<Tree,java.util.List<Tree>> topParses,
                       DVModel dvModel,
                       Options op)

Method Detail

sigmoid

public static double sigmoid(double x)

elementwiseApplyNonlinearity

public static SimpleMatrix elementwiseApplyNonlinearity(SimpleMatrix input)

nonlinearityVectorToDerivative

public static SimpleMatrix nonlinearityVectorToDerivative(SimpleMatrix input)

concatenateWithBias

public static SimpleMatrix concatenateWithBias(SimpleMatrix... vectors)

Concatenates several column vectors into one large column vector, adds a 1.0 at the end as a bias term

concatenate

public static SimpleMatrix concatenate(SimpleMatrix... vectors)

Concatenates several column vectors into one large column vector

outputSpans

public static void outputSpans(Tree tree)

score

public double score(Tree tree,
           java.util.IdentityHashMap<Tree,SimpleMatrix> nodeVectors)

domainDimension

public int domainDimension()

Description copied from interface: Function

Returns the number of dimensions in the function's domain

Returns:

the number of domain dimensions

getAllHighestScoringTreesTest

public java.util.List<DeepTree> getAllHighestScoringTreesTest(java.util.List<Tree> trees)

getHighestScoringTree

public DeepTree getHighestScoringTree(Tree tree,
                             double lambda)

calculate

public void calculate(double[] theta)

Description copied from class: AbstractCachingDiffFunction

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

Specified by:

calculate in class AbstractCachingDiffFunction

Parameters:

theta - The point at which to calculate the function

getMargin

public double getMargin(Tree goldTree,
               Tree bestHypothesis)

backpropDerivative

public void backpropDerivative(Tree tree,
                      java.util.List<java.lang.String> words,
                      java.util.IdentityHashMap<Tree,SimpleMatrix> nodeVectors,
                      TwoDimensionalMap<java.lang.String,java.lang.String,SimpleMatrix> binaryW_dfs,
                      java.util.Map<java.lang.String,SimpleMatrix> unaryW_dfs,
                      TwoDimensionalMap<java.lang.String,java.lang.String,SimpleMatrix> binaryScoreDerivatives,
                      java.util.Map<java.lang.String,SimpleMatrix> unaryScoreDerivatives,
                      java.util.Map<java.lang.String,SimpleMatrix> wordVectorDerivatives)

backpropDerivative

public void backpropDerivative(Tree tree,
                      java.util.List<java.lang.String> words,
                      java.util.IdentityHashMap<Tree,SimpleMatrix> nodeVectors,
                      TwoDimensionalMap<java.lang.String,java.lang.String,SimpleMatrix> binaryW_dfs,
                      java.util.Map<java.lang.String,SimpleMatrix> unaryW_dfs,
                      TwoDimensionalMap<java.lang.String,java.lang.String,SimpleMatrix> binaryScoreDerivatives,
                      java.util.Map<java.lang.String,SimpleMatrix> unaryScoreDerivatives,
                      java.util.Map<java.lang.String,SimpleMatrix> wordVectorDerivatives,
                      SimpleMatrix deltaUp)