Class DampForwardOp<T>

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

Inheritance

Object

DampForwardOp<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(Damp), "Forward<>", new Type[]{})]
[Quality(QualityBand.Preview)]
public static class DampForwardOp<T>

Type Parameters

Name	Description
T

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(forward,value,stepsize)). Adding up these values across all factors and variables gives the log-evidence estimate for VMP.

ForwardAverageConditional<Distribution>(Distribution, Double, Distribution)

EP message to forward.

Declaration

public static Distribution ForwardAverageConditional<Distribution>(Distribution value, double stepsize, Distribution to_forward)
    where Distribution : SettableToPower<Distribution>, SettableToProduct<Distribution>, HasPoint<T>, SettableToWeightedSum<Distribution>, SettableToUniform

Parameters

Type	Name	Description
Distribution	value	Incoming message from value. Must be a proper distribution. If uniform, the result will be uniform.
Double	stepsize	Constant value for stepsize.
Distribution	to_forward	Previous outgoing message to forward.

Returns

Type	Description
Distribution	The outgoing EP message to the forward argument.

Type Parameters

Name	Description
Distribution	The type of the distribution over the damped variable.

Remarks

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

Exceptions

Type	Condition
ImproperMessageException	value is not a proper distribution.

ForwardAverageLogarithm<Distribution>(Distribution, Double, Distribution)

Declaration

public static Distribution ForwardAverageLogarithm<Distribution>(Distribution value, double stepsize, Distribution to_forward)
    where Distribution : SettableToPower<Distribution>, SettableToProduct<Distribution>, HasPoint<T>, SettableToWeightedSum<Distribution>, SettableToUniform

Parameters

Type	Name	Description
Distribution	value
Double	stepsize
Distribution	to_forward

Returns

Type	Description
Distribution

Type Parameters

Name	Description
Distribution	The type of the distribution over the damped variable.

LogEvidenceRatio()

Evidence message for EP.

Declaration

public static double LogEvidenceRatio()

Returns

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

Remarks

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

ValueAverageConditional<Distribution>(Distribution)

EP message to value.

Declaration

public static Distribution ValueAverageConditional<Distribution>(Distribution forward)

Parameters

Type	Name	Description
Distribution	forward	Incoming message from forward.

Returns

Type	Description
Distribution	The outgoing EP message to the value argument.

Type Parameters

Name	Description
Distribution	The type of the distribution over the damped variable.

Remarks

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

ValueAverageLogarithm<Distribution>(Distribution)

VMP message to value.

Declaration

public static Distribution ValueAverageLogarithm<Distribution>(Distribution forward)

Parameters

Type	Name	Description
Distribution	forward	Incoming message from forward.

Returns

Type	Description
Distribution	The outgoing VMP message to the value argument.

Type Parameters

Name	Description
Distribution	The type of the distribution over the damped variable.

Remarks

The outgoing message is the factor viewed as a function of value with forward integrated out. The formula is sum_forward p(forward) factor(forward,value,stepsize).