Class RotateOp

Provides outgoing messages for Rotate(Double, Double, Double), given random arguments to the function.

Inheritance

Object

RotateOp

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(Factor), "Rotate", new Type[]{typeof(double), typeof(double), typeof(double)})]
[Quality(QualityBand.Experimental)]
public static class RotateOp

Methods

AngleAverageLogarithm(VectorGaussian, Gaussian, Gaussian)

VMP message to angle.

Declaration

public static WrappedGaussian AngleAverageLogarithm(VectorGaussian rotate, Gaussian x, Gaussian y)

Parameters

Type	Name	Description
VectorGaussian	rotate	Incoming message from rotate. Must be a proper distribution. If any element is uniform, the result will be uniform.
Gaussian	x	Incoming message from x. Must be a proper distribution. If uniform, the result will be uniform.
Gaussian	y	Incoming message from y. Must be a proper distribution. If uniform, the result will be uniform.

Returns

Type	Description
WrappedGaussian	The outgoing VMP message to the angle argument.

Remarks

The outgoing message is the exponential of the average log-factor value, where the average is over all arguments except angle. Because the factor is deterministic, rotate is integrated out before taking the logarithm. The formula is exp(sum_(x,y) p(x,y) log(sum_rotate p(rotate) factor(rotate,x,y,angle))).

Exceptions

Type	Condition
ImproperMessageException	rotate is not a proper distribution.
ImproperMessageException	x is not a proper distribution.
ImproperMessageException	y is not a proper distribution.

AngleAverageLogarithm(VectorGaussian, Double, Double)

VMP message to angle.

Declaration

public static WrappedGaussian AngleAverageLogarithm(VectorGaussian rotate, double x, double y)

Parameters

Type	Name	Description
VectorGaussian	rotate	Incoming message from rotate. Must be a proper distribution. If any element is uniform, the result will be uniform.
Double	x	Constant value for x.
Double	y	Constant value for y.

Returns

Type	Description
WrappedGaussian	The outgoing VMP message to the angle argument.

Remarks

The outgoing message is the factor viewed as a function of angle with rotate integrated out. The formula is sum_rotate p(rotate) factor(rotate,x,y,angle).

Exceptions

Type	Condition
ImproperMessageException	rotate is not a proper distribution.

AverageLogFactor(VectorGaussian)

Evidence message for VMP.

Declaration

public static double AverageLogFactor(VectorGaussian rotate)

Parameters

Type	Name	Description
VectorGaussian	rotate	Incoming message from rotate.

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.

RotateAverageLogarithm(Gaussian, Gaussian, WrappedGaussian, VectorGaussian)

VMP message to rotate.

Declaration

public static VectorGaussian RotateAverageLogarithm(Gaussian x, Gaussian y, WrappedGaussian angle, VectorGaussian result)

Parameters

Type	Name	Description
Gaussian	x	Incoming message from x. Must be a proper distribution. If uniform, the result will be uniform.
Gaussian	y	Incoming message from y. Must be a proper distribution. If uniform, the result will be uniform.
WrappedGaussian	angle	Incoming message from angle. Must be a proper distribution. If uniform, the result will be uniform.
VectorGaussian	result	Modified to contain the outgoing message.

Returns

Type	Description
VectorGaussian	result

Remarks

The outgoing message is a distribution matching the moments of rotate as the random arguments are varied. The formula is proj[sum_(x,y,angle) p(x,y,angle) factor(rotate,x,y,angle)].

Exceptions

Type	Condition
ImproperMessageException	x is not a proper distribution.
ImproperMessageException	y is not a proper distribution.
ImproperMessageException	angle is not a proper distribution.

RotateAverageLogarithm(Gaussian, Gaussian, Double, VectorGaussian)

VMP message to rotate.

Declaration

public static VectorGaussian RotateAverageLogarithm(Gaussian x, Gaussian y, double angle, VectorGaussian result)

Parameters

Type	Name	Description
Gaussian	x	Incoming message from x. Must be a proper distribution. If uniform, the result will be uniform.
Gaussian	y	Incoming message from y. Must be a proper distribution. If uniform, the result will be uniform.
Double	angle	Constant value for angle.
VectorGaussian	result	Modified to contain the outgoing message.

Returns

Type	Description
VectorGaussian	result

Remarks

The outgoing message is a distribution matching the moments of rotate as the random arguments are varied. The formula is proj[sum_(x,y) p(x,y) factor(rotate,x,y,angle)].

Exceptions

Type	Condition
ImproperMessageException	x is not a proper distribution.
ImproperMessageException	y is not a proper distribution.

RotateAverageLogarithm(Double, Double, WrappedGaussian, VectorGaussian)

VMP message to rotate.

Declaration

public static VectorGaussian RotateAverageLogarithm(double x, double y, WrappedGaussian angle, VectorGaussian result)

Parameters

Type	Name	Description
Double	x	Constant value for x.
Double	y	Constant value for y.
WrappedGaussian	angle	Incoming message from angle. Must be a proper distribution. If uniform, the result will be uniform.
VectorGaussian	result	Modified to contain the outgoing message.

Returns

Type	Description
VectorGaussian	result

Remarks

The outgoing message is a distribution matching the moments of rotate as the random arguments are varied. The formula is proj[sum_(angle) p(angle) factor(rotate,x,y,angle)].

Exceptions

Type	Condition
ImproperMessageException	angle is not a proper distribution.

RotateAverageLogarithmInit()

Declaration

public static VectorGaussian RotateAverageLogarithmInit()

Returns

Type	Description
VectorGaussian

Remarks

XAverageLogarithm(VectorGaussian, WrappedGaussian)

VMP message to x.

Declaration

public static Gaussian XAverageLogarithm(VectorGaussian rotate, WrappedGaussian angle)

Parameters

Type	Name	Description
VectorGaussian	rotate	Incoming message from rotate. Must be a proper distribution. If any element is uniform, the result will be uniform.
WrappedGaussian	angle	Incoming message from angle. Must be a proper distribution. If uniform, the result will be uniform.

Returns

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

Remarks

The outgoing message is the exponential of the average log-factor value, where the average is over all arguments except x. Because the factor is deterministic, rotate is integrated out before taking the logarithm. The formula is exp(sum_(angle) p(angle) log(sum_rotate p(rotate) factor(rotate,x,y,angle))).

Exceptions

Type	Condition
ImproperMessageException	rotate is not a proper distribution.
ImproperMessageException	angle is not a proper distribution.

XAverageLogarithm(VectorGaussian, Double)

VMP message to x.

Declaration

public static Gaussian XAverageLogarithm(VectorGaussian rotate, double angle)

Parameters

Type	Name	Description
VectorGaussian	rotate	Incoming message from rotate. Must be a proper distribution. If any element is uniform, the result will be uniform.
Double	angle	Constant value for angle.

Returns

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

Remarks

The outgoing message is the factor viewed as a function of x with rotate integrated out. The formula is sum_rotate p(rotate) factor(rotate,x,y,angle).

Exceptions

Type	Condition
ImproperMessageException	rotate is not a proper distribution.

YAverageLogarithm(VectorGaussian, WrappedGaussian)

VMP message to y.

Declaration

public static Gaussian YAverageLogarithm(VectorGaussian rotate, WrappedGaussian angle)

Parameters

Type	Name	Description
VectorGaussian	rotate	Incoming message from rotate. Must be a proper distribution. If any element is uniform, the result will be uniform.
WrappedGaussian	angle	Incoming message from angle. Must be a proper distribution. If uniform, the result will be uniform.

Returns

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

Remarks

The outgoing message is the exponential of the average log-factor value, where the average is over all arguments except y. Because the factor is deterministic, rotate is integrated out before taking the logarithm. The formula is exp(sum_(angle) p(angle) log(sum_rotate p(rotate) factor(rotate,x,y,angle))).

Exceptions

Type	Condition
ImproperMessageException	rotate is not a proper distribution.
ImproperMessageException	angle is not a proper distribution.

YAverageLogarithm(VectorGaussian, Double)

VMP message to y.

Declaration

public static Gaussian YAverageLogarithm(VectorGaussian rotate, double angle)

Parameters

Type	Name	Description
VectorGaussian	rotate	Incoming message from rotate. Must be a proper distribution. If any element is uniform, the result will be uniform.
Double	angle	Constant value for angle.

Returns

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

Remarks

The outgoing message is the factor viewed as a function of y with rotate integrated out. The formula is sum_rotate p(rotate) factor(rotate,x,y,angle).

Exceptions

Type	Condition
ImproperMessageException	rotate is not a proper distribution.