Class ConstrainEqualOp<T>

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

Inheritance

Object

ConstrainEqualOp<T>

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(Constrain), "Equal<>", new Type[]{})]
[Quality(QualityBand.Mature)]
public static class ConstrainEqualOp<T>

Type Parameters

Name	Description
T	The type of the constrained variables.

Methods

AAverageConditional<TDistribution>(TDistribution, TDistribution)

EP message to A.

Declaration

public static TDistribution AAverageConditional<TDistribution>(TDistribution B, TDistribution result)
    where TDistribution : SettableTo<TDistribution>

Parameters

Type	Name	Description
TDistribution	B	Incoming message from B.
TDistribution	result	Modified to contain the outgoing message.

Returns

Type	Description
TDistribution	result

Type Parameters

Name	Description
TDistribution	The distribution over the constrained variables.

Remarks

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

AAverageConditional<TDistribution>(T, TDistribution)

EP message to A.

Declaration

public static TDistribution AAverageConditional<TDistribution>(T B, TDistribution result)
    where TDistribution : HasPoint<T>

Parameters

Type	Name	Description
T	B	Incoming message from B.
TDistribution	result	Modified to contain the outgoing message.

Returns

Type	Description
TDistribution	result

Type Parameters

Name	Description
TDistribution	The distribution over the constrained variables.

Remarks

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

AAverageLogarithm<TDistribution>(TDistribution, TDistribution)

VMP message to A.

Declaration

[NotSupported("VMP does not support Constrain.Equal between random variables")]
public static TDistribution AAverageLogarithm<TDistribution>(TDistribution B, TDistribution result)

Parameters

Type	Name	Description
TDistribution	B	Incoming message from B.
TDistribution	result	Modified to contain the outgoing message.

Returns

Type	Description
TDistribution	result

Type Parameters

Name	Description
TDistribution	The distribution over the constrained variables.

Remarks

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

AAverageLogarithm<TDistribution>(T, TDistribution)

VMP message to A.

Declaration

public static TDistribution AAverageLogarithm<TDistribution>(T B, TDistribution result)
    where TDistribution : HasPoint<T>

Parameters

Type	Name	Description
T	B	Incoming message from B.
TDistribution	result	Modified to contain the outgoing message.

Returns

Type	Description
TDistribution	result

Type Parameters

Name	Description
TDistribution	The distribution over the constrained variables.

Remarks

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

AMaxConditional<TDistribution>(TDistribution, TDistribution)

Declaration

public static TDistribution AMaxConditional<TDistribution>(TDistribution B, TDistribution result)
    where TDistribution : SettableTo<TDistribution>

Parameters

Type	Name	Description
TDistribution	B	Incoming message from B.
TDistribution	result	Modified to contain the outgoing message.

Returns

Type	Description
TDistribution	result

Type Parameters

Name	Description
TDistribution	The distribution over the constrained variables.

Remarks

AverageLogFactor()

Evidence message for VMP.

Declaration

public static double AverageLogFactor()

Returns

Type	Description
Double	Zero.

Remarks

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

BAverageConditional<TDistribution>(TDistribution, TDistribution)

EP message to B.

Declaration

public static TDistribution BAverageConditional<TDistribution>(TDistribution A, TDistribution result)
    where TDistribution : SettableTo<TDistribution>

Parameters

Type	Name	Description
TDistribution	A	Incoming message from A.
TDistribution	result	Modified to contain the outgoing message.

Returns

Type	Description
TDistribution	result

Type Parameters

Name	Description
TDistribution	The distribution over the constrained variables.

Remarks

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

BAverageConditional<TDistribution>(T, TDistribution)

EP message to B.

Declaration

public static TDistribution BAverageConditional<TDistribution>(T A, TDistribution result)
    where TDistribution : HasPoint<T>

Parameters

Type	Name	Description
T	A	Incoming message from A.
TDistribution	result	Modified to contain the outgoing message.

Returns

Type	Description
TDistribution	result

Type Parameters

Name	Description
TDistribution	The distribution over the constrained variables.

Remarks

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

BAverageLogarithm<TDistribution>(TDistribution, TDistribution)

VMP message to B.

Declaration

[NotSupported("VMP does not support Constrain.Equal between random variables")]
public static TDistribution BAverageLogarithm<TDistribution>(TDistribution A, TDistribution result)

Parameters

Type	Name	Description
TDistribution	A	Incoming message from A.
TDistribution	result	Modified to contain the outgoing message.

Returns

Type	Description
TDistribution	result

Type Parameters

Name	Description
TDistribution	The distribution over the constrained variables.

Remarks

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

BAverageLogarithm<TDistribution>(T, TDistribution)

VMP message to B.

Declaration

public static TDistribution BAverageLogarithm<TDistribution>(T A, TDistribution result)
    where TDistribution : HasPoint<T>

Parameters

Type	Name	Description
T	A	Incoming message from A.
TDistribution	result	Modified to contain the outgoing message.

Returns

Type	Description
TDistribution	result

Type Parameters

Name	Description
TDistribution	The distribution over the constrained variables.

Remarks

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

BMaxConditional<TDistribution>(TDistribution, TDistribution)

Declaration

public static TDistribution BMaxConditional<TDistribution>(TDistribution A, TDistribution result)
    where TDistribution : SettableTo<TDistribution>

Parameters

Type	Name	Description
TDistribution	A	Incoming message from A.
TDistribution	result	Modified to contain the outgoing message.

Returns

Type	Description
TDistribution	result

Type Parameters

Name	Description
TDistribution	The distribution over the constrained variables.

Remarks

LogAverageFactor(T, T)

Evidence message for EP.

Declaration

public static double LogAverageFactor(T a, T b)

Parameters

Type	Name	Description
T	a	Incoming message from A.
T	b	Incoming message from B.

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_(A,B) p(A,B) factor(A,B)).

LogAverageFactor<TDistribution>(TDistribution, TDistribution)

Evidence message for EP.

Declaration

public static double LogAverageFactor<TDistribution>(TDistribution a, TDistribution b)
    where TDistribution : CanGetLogAverageOf<TDistribution>

Parameters

Type	Name	Description
TDistribution	a	Incoming message from A.
TDistribution	b	Incoming message from B.

Returns

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

Type Parameters

Name	Description
TDistribution	The distribution over the constrained variables.

Remarks

The formula for the result is log(sum_(A,B) p(A,B) factor(A,B)).

LogAverageFactor<TDistribution>(TDistribution, T)

Evidence message for EP.

Declaration

public static double LogAverageFactor<TDistribution>(TDistribution a, T b)
    where TDistribution : CanGetLogProb<T>

Parameters

Type	Name	Description
TDistribution	a	Incoming message from A.
T	b	Incoming message from B.

Returns

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

Type Parameters

Name	Description
TDistribution	The distribution over the constrained variables.

Remarks

The formula for the result is log(sum_(A,B) p(A,B) factor(A,B)).

LogAverageFactor<TDistribution>(T, TDistribution)

Evidence message for EP.

Declaration

public static double LogAverageFactor<TDistribution>(T a, TDistribution b)
    where TDistribution : CanGetLogProb<T>

Parameters

Type	Name	Description
T	a	Incoming message from A.
TDistribution	b	Incoming message from B.

Returns

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

Type Parameters

Name	Description
TDistribution	The distribution over the constrained variables.

Remarks

The formula for the result is log(sum_(A,B) p(A,B) factor(A,B)).

LogEvidenceRatio(T, T)

Evidence message for EP.

Declaration

public static double LogEvidenceRatio(T a, T b)

Parameters

Type	Name	Description
T	a	Incoming message from A.
T	b	Incoming message from B.

Returns

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

Remarks

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

LogEvidenceRatio<TDistribution>(TDistribution, TDistribution)

Evidence message for EP.

Declaration

public static double LogEvidenceRatio<TDistribution>(TDistribution a, TDistribution b)
    where TDistribution : CanGetLogAverageOf<TDistribution>

Parameters

Type	Name	Description
TDistribution	a	Incoming message from A.
TDistribution	b	Incoming message from B.

Returns

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

Type Parameters

Name	Description
TDistribution	The distribution over the constrained variables.

Remarks

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

LogEvidenceRatio<TDistribution>(TDistribution, T)

Evidence message for EP.

Declaration

public static double LogEvidenceRatio<TDistribution>(TDistribution a, T b)
    where TDistribution : CanGetLogProb<T>

Parameters

Type	Name	Description
TDistribution	a	Incoming message from A.
T	b	Incoming message from B.

Returns

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

Type Parameters

Name	Description
TDistribution	The distribution over the constrained variables.

Remarks

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

LogEvidenceRatio<TDistribution>(T, TDistribution)

Evidence message for EP.

Declaration

public static double LogEvidenceRatio<TDistribution>(T a, TDistribution b)
    where TDistribution : CanGetLogProb<T>

Parameters

Type	Name	Description
T	a	Incoming message from A.
TDistribution	b	Incoming message from B.

Returns

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

Type Parameters

Name	Description
TDistribution	The distribution over the constrained variables.

Remarks

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