Class ConcatOp

Provides outgoing messages for Concat(Vector, Vector), given random arguments to the function.

Inheritance

Object

ConcatOp

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(Vector), "Concat", new Type[]{typeof(Vector), typeof(Vector)})]
[Quality(QualityBand.Stable)]
public static class ConcatOp

Methods

AverageLogFactor()

Evidence message for VMP.

Declaration

public static double AverageLogFactor()

Returns

Type	Description
Double	Zero.

Remarks

The formula for the result is log(factor(concat,first,second)). Adding up these values across all factors and variables gives the log-evidence estimate for VMP.

ConcatAverageConditional(VectorGaussian, VectorGaussian, VectorGaussian)

EP message to concat.

Declaration

[SkipIfAllUniform]
public static VectorGaussian ConcatAverageConditional(VectorGaussian first, VectorGaussian second, VectorGaussian result)

Parameters

Type	Name	Description
VectorGaussian	first	Incoming message from first.
VectorGaussian	second	Incoming message from second.
VectorGaussian	result	Modified to contain the outgoing message.

Returns

Type	Description
VectorGaussian	result

Remarks

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

ConcatAverageConditional(VectorGaussian, Vector, VectorGaussian)

EP message to concat.

Declaration

public static VectorGaussian ConcatAverageConditional(VectorGaussian first, Vector second, VectorGaussian result)

Parameters

Type	Name	Description
VectorGaussian	first	Incoming message from first.
Vector	second	Constant value for second.
VectorGaussian	result	Modified to contain the outgoing message.

Returns

Type	Description
VectorGaussian	result

Remarks

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

ConcatAverageConditional(Vector, VectorGaussian, VectorGaussian)

EP message to concat.

Declaration

public static VectorGaussian ConcatAverageConditional(Vector first, VectorGaussian second, VectorGaussian result)

Parameters

Type	Name	Description
Vector	first	Constant value for first.
VectorGaussian	second	Incoming message from second.
VectorGaussian	result	Modified to contain the outgoing message.

Returns

Type	Description
VectorGaussian	result

Remarks

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

ConcatAverageConditionalInit(VectorGaussian, VectorGaussian)

Declaration

public static VectorGaussian ConcatAverageConditionalInit(VectorGaussian first, VectorGaussian second)

Parameters

Type	Name	Description
VectorGaussian	first	Incoming message from first.
VectorGaussian	second	Incoming message from second.

Returns

Type	Description
VectorGaussian

Remarks

ConcatAverageConditionalInit(VectorGaussian, Vector)

Declaration

public static VectorGaussian ConcatAverageConditionalInit(VectorGaussian first, Vector second)

Parameters

Type	Name	Description
VectorGaussian	first	Incoming message from first.
Vector	second	Constant value for second.

Returns

Type	Description
VectorGaussian

Remarks

ConcatAverageConditionalInit(Vector, VectorGaussian)

Declaration

public static VectorGaussian ConcatAverageConditionalInit(Vector first, VectorGaussian second)

Parameters

Type	Name	Description
Vector	first	Constant value for first.
VectorGaussian	second	Incoming message from second.

Returns

Type	Description
VectorGaussian

Remarks

ConcatAverageLogarithm(VectorGaussian, VectorGaussian, VectorGaussian)

VMP message to concat.

Declaration

[SkipIfAllUniform]
public static VectorGaussian ConcatAverageLogarithm(VectorGaussian first, VectorGaussian second, VectorGaussian result)

Parameters

Type	Name	Description
VectorGaussian	first	Incoming message from first.
VectorGaussian	second	Incoming message from second.
VectorGaussian	result	Modified to contain the outgoing message.

Returns

Type	Description
VectorGaussian	result

Remarks

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

ConcatAverageLogarithm(VectorGaussian, Vector, VectorGaussian)

VMP message to concat.

Declaration

public static VectorGaussian ConcatAverageLogarithm(VectorGaussian first, Vector second, VectorGaussian result)

Parameters

Type	Name	Description
VectorGaussian	first	Incoming message from first.
Vector	second	Constant value for second.
VectorGaussian	result	Modified to contain the outgoing message.

Returns

Type	Description
VectorGaussian	result

Remarks

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

ConcatAverageLogarithm(Vector, VectorGaussian, VectorGaussian)

VMP message to concat.

Declaration

public static VectorGaussian ConcatAverageLogarithm(Vector first, VectorGaussian second, VectorGaussian result)

Parameters

Type	Name	Description
Vector	first	Constant value for first.
VectorGaussian	second	Incoming message from second.
VectorGaussian	result	Modified to contain the outgoing message.

Returns

Type	Description
VectorGaussian	result

Remarks

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

ConcatAverageLogarithmInit(VectorGaussian, VectorGaussian)

Declaration

public static VectorGaussian ConcatAverageLogarithmInit(VectorGaussian first, VectorGaussian second)

Parameters

Type	Name	Description
VectorGaussian	first	Incoming message from first.
VectorGaussian	second	Incoming message from second.

Returns

Type	Description
VectorGaussian

Remarks

ConcatAverageLogarithmInit(VectorGaussian, Vector)

Declaration

public static VectorGaussian ConcatAverageLogarithmInit(VectorGaussian first, Vector second)

Parameters

Type	Name	Description
VectorGaussian	first	Incoming message from first.
Vector	second	Constant value for second.

Returns

Type	Description
VectorGaussian

Remarks

ConcatAverageLogarithmInit(Vector, VectorGaussian)

Declaration

public static VectorGaussian ConcatAverageLogarithmInit(Vector first, VectorGaussian second)

Parameters

Type	Name	Description
Vector	first	Constant value for first.
VectorGaussian	second	Incoming message from second.

Returns

Type	Description
VectorGaussian

Remarks

FirstAverageConditional(VectorGaussian, VectorGaussian, VectorGaussian)

EP message to first.

Declaration

public static VectorGaussian FirstAverageConditional(VectorGaussian concat, VectorGaussian second, VectorGaussian result)

Parameters

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

Returns

Type	Description
VectorGaussian	result

Remarks

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

Exceptions

Type	Condition
ImproperMessageException	concat is not a proper distribution.

FirstAverageConditional(VectorGaussian, Vector, VectorGaussian)

EP message to first.

Declaration

public static VectorGaussian FirstAverageConditional(VectorGaussian concat, Vector second, VectorGaussian result)

Parameters

Type	Name	Description
VectorGaussian	concat	Incoming message from concat. Must be a proper distribution. If any element is uniform, the result will be uniform.
Vector	second	Constant value for second.
VectorGaussian	result	Modified to contain the outgoing message.

Returns

Type	Description
VectorGaussian	result

Remarks

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

Exceptions

Type	Condition
ImproperMessageException	concat is not a proper distribution.

FirstAverageConditional(Vector, VectorGaussian)

EP message to first.

Declaration

public static VectorGaussian FirstAverageConditional(Vector concat, VectorGaussian result)

Parameters

Type	Name	Description
Vector	concat	Constant value for concat.
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 first conditioned on the given values.

FirstAverageLogarithm(VectorGaussian, VectorGaussian, VectorGaussian)

VMP message to first.

Declaration

public static VectorGaussian FirstAverageLogarithm(VectorGaussian concat, VectorGaussian second, VectorGaussian result)

Parameters

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

Returns

Type	Description
VectorGaussian	result

Remarks

The outgoing message is the exponential of the average log-factor value, where the average is over all arguments except first. Because the factor is deterministic, concat is integrated out before taking the logarithm. The formula is exp(sum_(second) p(second) log(sum_concat p(concat) factor(concat,first,second))).

Exceptions

Type	Condition
ImproperMessageException	concat is not a proper distribution.

FirstAverageLogarithm(VectorGaussian, Vector, VectorGaussian)

VMP message to first.

Declaration

public static VectorGaussian FirstAverageLogarithm(VectorGaussian concat, Vector second, VectorGaussian result)

Parameters

Type	Name	Description
VectorGaussian	concat	Incoming message from concat. Must be a proper distribution. If any element is uniform, the result will be uniform.
Vector	second	Constant value for second.
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 first with concat integrated out. The formula is sum_concat p(concat) factor(concat,first,second).

Exceptions

Type	Condition
ImproperMessageException	concat is not a proper distribution.

FirstAverageLogarithm(Vector, VectorGaussian)

VMP message to first.

Declaration

public static VectorGaussian FirstAverageLogarithm(Vector concat, VectorGaussian result)

Parameters

Type	Name	Description
Vector	concat	Constant value for concat.
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 first conditioned on the given values.

LogAverageFactor(VectorGaussian, VectorGaussian)

Evidence message for EP.

Declaration

public static double LogAverageFactor(VectorGaussian concat, VectorGaussian to_concat)

Parameters

Type	Name	Description
VectorGaussian	concat	Incoming message from concat.
VectorGaussian	to_concat	Outgoing message to concat.

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_(concat) p(concat) factor(concat,first,second)).

LogAverageFactor(VectorGaussian, Vector, Vector)

Evidence message for EP.

Declaration

public static double LogAverageFactor(VectorGaussian concat, Vector first, Vector second)

Parameters

Type	Name	Description
VectorGaussian	concat	Incoming message from concat.
Vector	first	Constant value for first.
Vector	second	Constant value for second.

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_(concat) p(concat) factor(concat,first,second)).

LogAverageFactor(Vector, VectorGaussian, VectorGaussian)

Evidence message for EP.

Declaration

public static double LogAverageFactor(Vector concat, VectorGaussian first, VectorGaussian second)

Parameters

Type	Name	Description
Vector	concat	Constant value for concat.
VectorGaussian	first	Incoming message from first.
VectorGaussian	second	Incoming message from second.

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_(first,second) p(first,second) factor(concat,first,second)).

LogAverageFactor(Vector, VectorGaussian, Vector)

Evidence message for EP.

Declaration

public static double LogAverageFactor(Vector concat, VectorGaussian first, Vector second)

Parameters

Type	Name	Description
Vector	concat	Constant value for concat.
VectorGaussian	first	Incoming message from first.
Vector	second	Constant value for second.

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_(first) p(first) factor(concat,first,second)).

LogAverageFactor(Vector, Vector, VectorGaussian)

Evidence message for EP.

Declaration

public static double LogAverageFactor(Vector concat, Vector first, VectorGaussian second)

Parameters

Type	Name	Description
Vector	concat	Constant value for concat.
Vector	first	Constant value for first.
VectorGaussian	second	Incoming message from second.

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_(second) p(second) factor(concat,first,second)).

LogAverageFactor(Vector, Vector, Vector)

Evidence message for EP.

Declaration

public static double LogAverageFactor(Vector concat, Vector first, Vector second)

Parameters

Type	Name	Description
Vector	concat	Constant value for concat.
Vector	first	Constant value for first.
Vector	second	Constant value for second.

Returns

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

Remarks

The formula for the result is log(factor(concat,first,second)).

LogEvidenceRatio(VectorGaussian)

Evidence message for EP.

Declaration

public static double LogEvidenceRatio(VectorGaussian concat)

Parameters

Type	Name	Description
VectorGaussian	concat	Incoming message from concat.

Returns

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

Remarks

The formula for the result is log(sum_(concat) p(concat) factor(concat,first,second) / sum_concat p(concat) messageTo(concat)). Adding up these values across all factors and variables gives the log-evidence estimate for EP.

LogEvidenceRatio(Vector, VectorGaussian, VectorGaussian)

Evidence message for EP.

Declaration

public static double LogEvidenceRatio(Vector concat, VectorGaussian first, VectorGaussian second)

Parameters

Type	Name	Description
Vector	concat	Constant value for concat.
VectorGaussian	first	Incoming message from first.
VectorGaussian	second	Incoming message from second.

Returns

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

Remarks

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

LogEvidenceRatio(Vector, VectorGaussian, Vector)

Evidence message for EP.

Declaration

public static double LogEvidenceRatio(Vector concat, VectorGaussian first, Vector second)

Parameters

Type	Name	Description
Vector	concat	Constant value for concat.
VectorGaussian	first	Incoming message from first.
Vector	second	Constant value for second.

Returns

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

Remarks

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

LogEvidenceRatio(Vector, Vector, VectorGaussian)

Evidence message for EP.

Declaration

public static double LogEvidenceRatio(Vector concat, Vector first, VectorGaussian second)

Parameters

Type	Name	Description
Vector	concat	Constant value for concat.
Vector	first	Constant value for first.
VectorGaussian	second	Incoming message from second.

Returns

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

Remarks

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

LogEvidenceRatio(Vector, Vector, Vector)

Evidence message for EP.

Declaration

public static double LogEvidenceRatio(Vector concat, Vector first, Vector second)

Parameters

Type	Name	Description
Vector	concat	Constant value for concat.
Vector	first	Constant value for first.
Vector	second	Constant value for second.

Returns

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

Remarks

The formula for the result is log(factor(concat,first,second)). Adding up these values across all factors and variables gives the log-evidence estimate for EP.

SecondAverageConditional(VectorGaussian, VectorGaussian, VectorGaussian)

EP message to second.

Declaration

public static VectorGaussian SecondAverageConditional(VectorGaussian concat, VectorGaussian first, VectorGaussian result)

Parameters

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

Returns

Type	Description
VectorGaussian	result

Remarks

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

Exceptions

Type	Condition
ImproperMessageException	concat is not a proper distribution.

SecondAverageConditional(VectorGaussian, Vector, VectorGaussian)

EP message to second.

Declaration

public static VectorGaussian SecondAverageConditional(VectorGaussian concat, Vector first, VectorGaussian result)

Parameters

Type	Name	Description
VectorGaussian	concat	Incoming message from concat. Must be a proper distribution. If any element is uniform, the result will be uniform.
Vector	first	Constant value for first.
VectorGaussian	result	Modified to contain the outgoing message.

Returns

Type	Description
VectorGaussian	result

Remarks

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

Exceptions

Type	Condition
ImproperMessageException	concat is not a proper distribution.

SecondAverageConditional(Vector, VectorGaussian, VectorGaussian)

EP message to second.

Declaration

public static VectorGaussian SecondAverageConditional(Vector concat, VectorGaussian first, VectorGaussian result)

Parameters

Type	Name	Description
Vector	concat	Constant value for concat.
VectorGaussian	first	Incoming message from first.
VectorGaussian	result	Modified to contain the outgoing message.

Returns

Type	Description
VectorGaussian	result

Remarks

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

SecondAverageConditional(Vector, Vector, VectorGaussian)

EP message to second.

Declaration

public static VectorGaussian SecondAverageConditional(Vector concat, Vector first, VectorGaussian result)

Parameters

Type	Name	Description
Vector	concat	Constant value for concat.
Vector	first	Constant value for first.
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 second conditioned on the given values.

SecondAverageLogarithm(VectorGaussian, VectorGaussian, VectorGaussian)

VMP message to second.

Declaration

public static VectorGaussian SecondAverageLogarithm(VectorGaussian concat, VectorGaussian first, VectorGaussian result)

Parameters

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

Returns

Type	Description
VectorGaussian	result

Remarks

The outgoing message is the exponential of the average log-factor value, where the average is over all arguments except second. Because the factor is deterministic, concat is integrated out before taking the logarithm. The formula is exp(sum_(first) p(first) log(sum_concat p(concat) factor(concat,first,second))).

Exceptions

Type	Condition
ImproperMessageException	concat is not a proper distribution.

SecondAverageLogarithm(VectorGaussian, Vector, VectorGaussian)

VMP message to second.

Declaration

public static VectorGaussian SecondAverageLogarithm(VectorGaussian concat, Vector first, VectorGaussian result)

Parameters

Type	Name	Description
VectorGaussian	concat	Incoming message from concat. Must be a proper distribution. If any element is uniform, the result will be uniform.
Vector	first	Constant value for first.
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 second with concat integrated out. The formula is sum_concat p(concat) factor(concat,first,second).

Exceptions

Type	Condition
ImproperMessageException	concat is not a proper distribution.

SecondAverageLogarithm(Vector, VectorGaussian, VectorGaussian)

VMP message to second.

Declaration

public static VectorGaussian SecondAverageLogarithm(Vector concat, VectorGaussian first, VectorGaussian result)

Parameters

Type	Name	Description
Vector	concat	Constant value for concat.
VectorGaussian	first	Incoming message from first.
VectorGaussian	result	Modified to contain the outgoing message.

Returns

Type	Description
VectorGaussian	result

Remarks

The outgoing message is the exponential of the average log-factor value, where the average is over all arguments except second. The formula is exp(sum_(first) p(first) log(factor(concat,first,second))).

SecondAverageLogarithm(Vector, Vector, VectorGaussian)

VMP message to second.

Declaration

public static VectorGaussian SecondAverageLogarithm(Vector concat, Vector first, VectorGaussian result)

Parameters

Type	Name	Description
Vector	concat	Constant value for concat.
Vector	first	Constant value for first.
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 second conditioned on the given values.