Package org.tribuo.math.optimisers.util

Class ShrinkingVector

java.lang.Object

org.tribuo.math.la.DenseVector

org.tribuo.math.optimisers.util.ShrinkingVector

All Implemented Interfaces:

Serializable, Iterable<VectorTuple>, SGDVector, Tensor, ShrinkingTensor, ProtoSerializable<org.tribuo.math.protos.TensorProto>

public class ShrinkingVector
extends DenseVector
implements ShrinkingTensor

A subclass of DenseVector which shrinks the value every time a new value is added.

Be careful when modifying this or DenseVector.

See Also:

• Serialized Form

Field Summary

Fields inherited from class org.tribuo.math.la.DenseVector

CURRENT_VERSION, elements

Fields inherited from interface org.tribuo.protos.ProtoSerializable

DESERIALIZATION_METHOD_NAME, PROVENANCE_SERIALIZER

Fields inherited from interface org.tribuo.math.optimisers.util.ShrinkingTensor

tolerance

Constructor Summary

Constructors

Constructor	Description
ShrinkingVector(DenseVector v, double baseRate, boolean scaleShrinking)	Constructs a shrinking vector copy of the supplied dense matrix.
ShrinkingVector(DenseVector v, double baseRate, double lambda)	Constructs a shrinking vector copy of the supplied dense vector.

Method Summary

Modifier and Type	Method	Description
DenseVector	convertToDense()	Converts the tensor into a dense tensor.
ShrinkingVector	copy()	Returns a deep copy of this vector.
static ShrinkingVector	deserializeFromProto(int version, String className, com.google.protobuf.Any message)	Deserialization factory.
double	dot(SGDVector other)	Calculates the dot product between this vector and other.
double	get(int index)	Gets an element from this vector.
int	indexOfMax()	Returns the index of the maximum value.
void	intersectAndAddInPlace(Tensor other, DoubleUnaryOperator f)	Updates this Tensor by adding all the values from the intersection with other.
VectorIterator	iterator()
double	maxValue()	Returns the maximum value.
double	minValue()	Returns the minimum value.
void	scaleInPlace(double value)	Scales each element of this Tensor by coefficient.
org.tribuo.math.protos.TensorProto	serialize()	Serializes this object to a protobuf.
double	sum()	Calculates the sum of this vector.
double[]	toArray()	Generates a copy of the values in this DenseVector.
double	twoNorm()	Calculates the euclidean norm for this vector.

Methods inherited from class org.tribuo.math.la.DenseVector

add, add, createDenseVector, createDenseVector, equals, euclideanDistance, expNormalize, fill, foreachIndexedInPlace, foreachInPlace, getShape, hadamardProductInPlace, hashCode, l1Distance, meanVariance, normalize, numActiveElements, oneNorm, outer, reduce, reduce, reshape, scale, set, setElements, size, sparsify, sparsify, subtract, sum, toString, unpackProto, variance

Methods inherited from class java.lang.Object

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Methods inherited from interface java.lang.Iterable

forEach, spliterator

Methods inherited from interface org.tribuo.math.la.SGDVector

cosineDistance, cosineSimilarity, l2Distance, variance

Methods inherited from interface org.tribuo.math.la.Tensor

hadamardProductInPlace, intersectAndAddInPlace, scalarAddInPlace

Constructor Details

ShrinkingVector

public ShrinkingVector(DenseVector v,
 double baseRate,
 boolean scaleShrinking)

Constructs a shrinking vector copy of the supplied dense matrix.

This vector shrinks during each call to intersectAndAddInPlace(Tensor, DoubleUnaryOperator).

Parameters:

v - The vector to copy.

baseRate - The base amount of shrinking to apply after each update.

scaleShrinking - If true reduce the shrinking value over time proportionally to the number of updates.

ShrinkingVector

public ShrinkingVector(DenseVector v,
 double baseRate,
 double lambda)

Constructs a shrinking vector copy of the supplied dense vector.

This vector shrinks during each call to intersectAndAddInPlace(Tensor, DoubleUnaryOperator), and then reprojects the vector so it has the same twoNorm.

Parameters:

v - The vector to copy.

baseRate - The base rate of shrinkage.

lambda - The lambda value (see Pegasos).

Method Details

deserializeFromProto

public static ShrinkingVector deserializeFromProto(int version,
 String className,
 com.google.protobuf.Any message)
                                            throws com.google.protobuf.InvalidProtocolBufferException

Deserialization factory.

Parameters:

version - The serialized object version.

className - The class name.

message - The serialized data.

Returns:

The deserialized object.

Throws:

com.google.protobuf.InvalidProtocolBufferException - If the protobuf could not be parsed from the message.

serialize

public org.tribuo.math.protos.TensorProto serialize()

Description copied from interface: ProtoSerializable

Serializes this object to a protobuf.

Specified by:

serialize in interface ProtoSerializable<org.tribuo.math.protos.TensorProto>

Overrides:

serialize in class DenseVector

Returns:

The protobuf.

convertToDense

public DenseVector convertToDense()

Description copied from interface: ShrinkingTensor

Converts the tensor into a dense tensor.

Specified by:

convertToDense in interface ShrinkingTensor

Returns:

A dense tensor copy of this shrinking tensor.

copy

public ShrinkingVector copy()

Description copied from interface: SGDVector

Returns a deep copy of this vector.

Specified by:

copy in interface SGDVector

Specified by:

copy in interface Tensor

Overrides:

copy in class DenseVector

Returns:

A copy of this vector.

toArray

public double[] toArray()

Description copied from class: DenseVector

Generates a copy of the values in this DenseVector.

This implementation uses Arrays.copyOf, and should be overridden if the get function has been modified.

Specified by:

toArray in interface SGDVector

Overrides:

toArray in class DenseVector

Returns:

A copy of the values in this DenseVector.

get

public double get(int index)

Description copied from interface: SGDVector

Gets an element from this vector.

Specified by:

get in interface SGDVector

Overrides:

get in class DenseVector

Parameters:

index - The index of the element.

Returns:

The value at that index.

sum

public double sum()

Description copied from interface: SGDVector

Calculates the sum of this vector.

Specified by:

sum in interface SGDVector

Overrides:

sum in class DenseVector

Returns:

The sum.

intersectAndAddInPlace

public void intersectAndAddInPlace(Tensor other,
 DoubleUnaryOperator f)

Description copied from interface: Tensor

Updates this Tensor by adding all the values from the intersection with other.

The function f is applied to all values from other before the addition.

Each value is updated as value += f(otherValue).

Specified by:

intersectAndAddInPlace in interface Tensor

Overrides:

intersectAndAddInPlace in class DenseVector

Parameters:

other - The other Tensor.

f - A function to apply.

indexOfMax

public int indexOfMax()

Description copied from interface: SGDVector

Returns the index of the maximum value. Requires probing the array.

Specified by:

indexOfMax in interface SGDVector

Overrides:

indexOfMax in class DenseVector

Returns:

The index of the maximum value.

dot

public double dot(SGDVector other)

Description copied from interface: SGDVector

Calculates the dot product between this vector and other.

Specified by:

dot in interface SGDVector

Overrides:

dot in class DenseVector

Parameters:

other - The other vector.

Returns:

The dot product.

scaleInPlace

public void scaleInPlace(double value)

Description copied from interface: Tensor

Scales each element of this Tensor by coefficient.

Specified by:

scaleInPlace in interface Tensor

Parameters:

value - The coefficient of scaling.

twoNorm

public double twoNorm()

Description copied from interface: SGDVector

Calculates the euclidean norm for this vector.

Specified by:

twoNorm in interface SGDVector

Specified by:

twoNorm in interface Tensor

Overrides:

twoNorm in class DenseVector

Returns:

The euclidean norm.

maxValue

public double maxValue()

Description copied from interface: SGDVector

Returns the maximum value. Requires probing the array.

Specified by:

maxValue in interface SGDVector

Overrides:

maxValue in class DenseVector

Returns:

The maximum value.

minValue

public double minValue()

Description copied from interface: SGDVector

Returns the minimum value. Requires probing the array.

Specified by:

minValue in interface SGDVector

Overrides:

minValue in class DenseVector

Returns:

The minimum value.

iterator

public VectorIterator iterator()

Specified by:

iterator in interface Iterable<VectorTuple>

Overrides:

iterator in class DenseVector