Class VectorElementOp

Provides outgoing messages for GetItem<T>(IReadOnlyList<T>, Int32), given random arguments to the function.

Inheritance

Object

VectorElementOp

Inherited Members

Object.Equals(Object)

Object.Equals(Object, Object)

Object.GetHashCode()

Object.GetType()

Object.MemberwiseClone()

Object.ReferenceEquals(Object, Object)

Object.ToString()

Namespace: Microsoft.ML.Probabilistic.Factors

Assembly: Microsoft.ML.Probabilistic.dll

Syntax

[FactorMethod(typeof(Collection), "GetItem<>", new Type[]{typeof(double), typeof(Vector), typeof(int)})]
[Buffers(new string[]{"ArrayMean", "ArrayVariance"})]
[Quality(QualityBand.Preview)]
public static class VectorElementOp

Methods

ArrayAverageConditional(Gaussian, Int32, VectorGaussian)

EP message to array.

Declaration

public static VectorGaussian ArrayAverageConditional(Gaussian item, int index, VectorGaussian result)

Parameters

Type	Name	Description
Gaussian	item	Incoming message from item. Must be a proper distribution. If uniform, the result will be uniform.
Int32	index	Constant value for index.
VectorGaussian	result	Modified to contain the outgoing message.

Returns

Type	Description
VectorGaussian	result

Remarks

The outgoing message is a distribution matching the moments of array as the random arguments are varied. The formula is proj[p(array) sum_(item) p(item) factor(item,array,index)]/p(array).

Exceptions

Type	Condition
ImproperMessageException	item is not a proper distribution.

ArrayAverageConditional(Double, Int32, VectorGaussian)

EP message to array.

Declaration

public static VectorGaussian ArrayAverageConditional(double item, int index, VectorGaussian result)

Parameters

Type	Name	Description
Double	item	Constant value for item.
Int32	index	Constant value for index.
VectorGaussian	result	Modified to contain the outgoing message.

Returns

Type	Description
VectorGaussian	result

Remarks

The outgoing message is the factor viewed as a function of array conditioned on the given values.

ArrayAverageLogarithm(Gaussian, Int32, VectorGaussian)

VMP message to array.

Declaration

public static VectorGaussian ArrayAverageLogarithm(Gaussian item, int index, VectorGaussian result)

Parameters

Type	Name	Description
Gaussian	item	Incoming message from item. Must be a proper distribution. If uniform, the result will be uniform.
Int32	index	Constant value for index.
VectorGaussian	result	Modified to contain the outgoing message.

Returns

Type	Description
VectorGaussian	result

Remarks

The outgoing message is the factor viewed as a function of array with item integrated out. The formula is sum_item p(item) factor(item,array,index).

Exceptions

Type	Condition
ImproperMessageException	item is not a proper distribution.

ArrayAverageLogarithm(Double, Int32, VectorGaussian)

VMP message to array.

Declaration

public static VectorGaussian ArrayAverageLogarithm(double item, int index, VectorGaussian result)

Parameters

Type	Name	Description
Double	item	Constant value for item.
Int32	index	Constant value for index.
VectorGaussian	result	Modified to contain the outgoing message.

Returns

Type	Description
VectorGaussian	result

Remarks

The outgoing message is the factor viewed as a function of array conditioned on the given values.

ArrayMean(VectorGaussian, PositiveDefiniteMatrix, Vector)

Update the buffer ArrayMean.

Declaration

public static Vector ArrayMean(VectorGaussian array, PositiveDefiniteMatrix ArrayVariance, Vector result)

Parameters

Type	Name	Description
VectorGaussian	array	Incoming message from array. Must be a proper distribution. If any element is uniform, the result will be uniform.
PositiveDefiniteMatrix	ArrayVariance	Buffer ArrayVariance.
Vector	result	Modified to contain the outgoing message.

Returns

Type	Description
Vector	result

Remarks

Exceptions

Type	Condition
ImproperMessageException	array is not a proper distribution.

ArrayMeanInit(VectorGaussian)

Initialize the buffer ArrayMean.

Declaration

public static Vector ArrayMeanInit(VectorGaussian array)

Parameters

Type	Name	Description
VectorGaussian	array	Incoming message from array.

Returns

Type	Description
Vector	Initial value of buffer ArrayMean.

Remarks

ArrayVariance(VectorGaussian, PositiveDefiniteMatrix)

Update the buffer ArrayVariance.

Declaration

public static PositiveDefiniteMatrix ArrayVariance(VectorGaussian array, PositiveDefiniteMatrix result)

Parameters

Type	Name	Description
VectorGaussian	array	Incoming message from array. Must be a proper distribution. If any element is uniform, the result will be uniform.
PositiveDefiniteMatrix	result	Modified to contain the outgoing message.

Returns

Type	Description
PositiveDefiniteMatrix	result

Remarks

Exceptions

Type	Condition
ImproperMessageException	array is not a proper distribution.

ArrayVarianceInit(VectorGaussian)

Initialize the buffer ArrayVariance.

Declaration

public static PositiveDefiniteMatrix ArrayVarianceInit(VectorGaussian array)

Parameters

Type	Name	Description
VectorGaussian	array	Incoming message from array.

Returns

Type	Description
PositiveDefiniteMatrix	Initial value of buffer ArrayVariance.

Remarks

AverageLogFactor(Gaussian, VectorGaussian)

Evidence message for VMP.

Declaration

public static double AverageLogFactor(Gaussian item, VectorGaussian array)

Parameters

Type	Name	Description
Gaussian	item	Incoming message from item.
VectorGaussian	array	Incoming message from array.

Returns

Type	Description
Double	Zero.

Remarks

In Variational Message Passing, the evidence contribution of a deterministic factor is zero. Adding up these values across all factors and variables gives the log-evidence estimate for VMP.

AverageLogFactor(Double, VectorGaussian, Vector, PositiveDefiniteMatrix, Int32)

Evidence message for VMP.

Declaration

public static double AverageLogFactor(double item, VectorGaussian array, Vector ArrayMean, PositiveDefiniteMatrix ArrayVariance, int index)

Parameters

Type	Name	Description
Double	item	Constant value for item.
VectorGaussian	array	Incoming message from array. Must be a proper distribution. If any element is uniform, the result will be uniform.
Vector	ArrayMean	Buffer ArrayMean.
PositiveDefiniteMatrix	ArrayVariance	Buffer ArrayVariance.
Int32	index	Constant value for index.

Returns

Type	Description
Double	Zero.

Remarks

In Variational Message Passing, the evidence contribution of a deterministic factor is zero. Adding up these values across all factors and variables gives the log-evidence estimate for VMP.

Exceptions

Type	Condition
ImproperMessageException	array is not a proper distribution.

ItemAverageConditional(VectorGaussian, Vector, PositiveDefiniteMatrix, Int32)

EP message to item.

Declaration

public static Gaussian ItemAverageConditional(VectorGaussian array, Vector ArrayMean, PositiveDefiniteMatrix ArrayVariance, int index)

Parameters

Type	Name	Description
VectorGaussian	array	Incoming message from array. Must be a proper distribution. If any element is uniform, the result will be uniform.
Vector	ArrayMean	Buffer ArrayMean.
PositiveDefiniteMatrix	ArrayVariance	Buffer ArrayVariance.
Int32	index	Constant value for index.

Returns

Type	Description
Gaussian	The outgoing EP message to the item argument.

Remarks

The outgoing message is a distribution matching the moments of item as the random arguments are varied. The formula is proj[p(item) sum_(array) p(array) factor(item,array,index)]/p(item).

Exceptions

Type	Condition
ImproperMessageException	array is not a proper distribution.

ItemAverageConditionalInit()

Declaration

public static Gaussian ItemAverageConditionalInit()

Returns

Type	Description
Gaussian

Remarks

ItemAverageLogarithm(VectorGaussian, Vector, PositiveDefiniteMatrix, Int32)

VMP message to item.

Declaration

public static Gaussian ItemAverageLogarithm(VectorGaussian array, Vector ArrayMean, PositiveDefiniteMatrix ArrayVariance, int index)

Parameters

Type	Name	Description
VectorGaussian	array	Incoming message from array. Must be a proper distribution. If any element is uniform, the result will be uniform.
Vector	ArrayMean	Buffer ArrayMean.
PositiveDefiniteMatrix	ArrayVariance	Buffer ArrayVariance.
Int32	index	Constant value for index.

Returns

Type	Description
Gaussian	The outgoing VMP message to the item argument.

Remarks

The outgoing message is a distribution matching the moments of item as the random arguments are varied. The formula is proj[sum_(array) p(array) factor(item,array,index)].

Exceptions

Type	Condition
ImproperMessageException	array is not a proper distribution.

ItemAverageLogarithmInit()

Declaration

public static Gaussian ItemAverageLogarithmInit()

Returns

Type	Description
Gaussian

Remarks

LogAverageFactor(Double, VectorGaussian, Vector, PositiveDefiniteMatrix, Int32)

Evidence message for EP.

Declaration

public static double LogAverageFactor(double item, VectorGaussian array, Vector ArrayMean, PositiveDefiniteMatrix ArrayVariance, int index)

Parameters

Type	Name	Description
Double	item	Constant value for item.
VectorGaussian	array	Incoming message from array. Must be a proper distribution. If any element is uniform, the result will be uniform.
Vector	ArrayMean	Buffer ArrayMean.
PositiveDefiniteMatrix	ArrayVariance	Buffer ArrayVariance.
Int32	index	Constant value for index.

Returns

Type	Description
Double	Logarithm of the factor's average value across the given argument distributions.

Remarks

The formula for the result is log(sum_(array) p(array) factor(item,array,index)).

Exceptions

Type	Condition
ImproperMessageException	array is not a proper distribution.

LogEvidenceRatio(Double, VectorGaussian, Vector, PositiveDefiniteMatrix, Int32)

Evidence message for EP.

Declaration

public static double LogEvidenceRatio(double item, VectorGaussian array, Vector ArrayMean, PositiveDefiniteMatrix ArrayVariance, int index)

Parameters

Type	Name	Description
Double	item	Constant value for item.
VectorGaussian	array	Incoming message from array. Must be a proper distribution. If any element is uniform, the result will be uniform.
Vector	ArrayMean	Buffer ArrayMean.
PositiveDefiniteMatrix	ArrayVariance	Buffer ArrayVariance.
Int32	index	Constant value for index.

Returns

Type	Description
Double	Logarithm of the factor's contribution the EP model evidence.

Remarks

The formula for the result is log(sum_(array) p(array) factor(item,array,index)). Adding up these values across all factors and variables gives the log-evidence estimate for EP.

Exceptions

Type	Condition
ImproperMessageException	array is not a proper distribution.