Implementation of a discrete-continuous EM factor. More...

#include <DCEMFactor.h>

Inheritance diagram for dcsam::DCEMFactor< DCFactorType >:

Collaboration diagram for dcsam::DCEMFactor< DCFactorType >:

Public Types
using	Base = DCFactor

Public Types inherited from dcsam::DCFactor
using	Base = gtsam::Factor

Public Member Functions
	DCEMFactor ()=default

	DCEMFactor (const gtsam::KeyVector &continuousKeys, const gtsam::DiscreteKeys &discreteKeys, const std::vector< DCFactorType > factors, const std::vector< double > weights, const bool normalized)

	DCEMFactor (const gtsam::KeyVector &continuousKeys, const gtsam::DiscreteKeys &discreteKeys, const std::vector< DCFactorType > factors, const bool normalized)

DCEMFactor &	operator= (const DCEMFactor &rhs)

virtual	~DCEMFactor ()=default

double	error (const gtsam::Values &continuousVals, const DiscreteValues &discreteVals) const override

std::vector< double >	computeComponentLogProbs (const gtsam::Values &continuousVals, const DiscreteValues &discreteVals) const

size_t	getActiveFactorIdx (const gtsam::Values &continuousVals, const DiscreteValues &discreteVals) const

size_t	dim () const override

bool	equals (const DCFactor &other, double tol=1e-9) const override

boost::shared_ptr< gtsam::GaussianFactor >	linearize (const gtsam::Values &continuousVals, const DiscreteValues &discreteVals) const override

gtsam::DecisionTreeFactor	toDecisionTreeFactor (const gtsam::Values &continuousVals, const DiscreteValues &discreteVals) const override

gtsam::FastVector< gtsam::Key >	getAssociationKeys (const gtsam::Values &continuousVals, const DiscreteValues &discreteVals) const

void	updateWeights (const std::vector< double > &weights)

Public Member Functions inherited from dcsam::DCFactor
	DCFactor ()=default

	DCFactor (const gtsam::KeyVector &continuousKeys, const gtsam::DiscreteKeys &discreteKeys)

	DCFactor (const gtsam::DiscreteKeys &discreteKeys)

DCFactor &	operator= (const DCFactor &rhs)

virtual	~DCFactor ()=default

gtsam::DiscreteKeys	discreteKeys () const

virtual double	logNormalizingConstant (const gtsam::Values &values) const

template<typename NonlinearFactorType >
double	nonlinearFactorLogNormalizingConstant (const NonlinearFactorType &factor, const gtsam::Values &values) const

std::vector< double >	evalProbs (const gtsam::DiscreteKey &dk, const gtsam::Values &continuousVals) const

gtsam::DecisionTreeFactor	conditionalTimes (const gtsam::DecisionTreeFactor &f, const gtsam::Values &continuousVals, const DiscreteValues &discreteVals) const

Additional Inherited Members
Protected Attributes inherited from dcsam::DCFactor
gtsam::DiscreteKeys	discreteKeys_

Detailed Description

template<class DCFactorType>
class dcsam::DCEMFactor< DCFactorType >

Implementation of a discrete-continuous EM factor.

The error function is defined as: r(x) = sum_i w'_i * r_i(x),

where w'_i = w_i * p(z | x, h_i); sum_i w'_i = 1. and h_i here represents the "i-th" hypothesis

The error returned from this factor is a weighted combination of the component factor errors. "x" can be comprised jointly of discrete and continuous values. Prior hypothesis weights can be included if desired via the weights parameter.

Member Typedef Documentation

◆ Base

template<class DCFactorType >

using dcsam::DCEMFactor< DCFactorType >::Base = DCFactor

Constructor & Destructor Documentation

◆ DCEMFactor() [1/3]

template<class DCFactorType >

dcsam::DCEMFactor< DCFactorType >::DCEMFactor ( )

default

◆ DCEMFactor() [2/3]

template<class DCFactorType >

dcsam::DCEMFactor< DCFactorType >::DCEMFactor	(	const gtsam::KeyVector &	continuousKeys,
		const gtsam::DiscreteKeys &	discreteKeys,
		const std::vector< DCFactorType >	factors,
		const std::vector< double >	weights,
		const bool	normalized
	)

inlineexplicit

◆ DCEMFactor() [3/3]

template<class DCFactorType >

dcsam::DCEMFactor< DCFactorType >::DCEMFactor	(	const gtsam::KeyVector &	continuousKeys,
		const gtsam::DiscreteKeys &	discreteKeys,
		const std::vector< DCFactorType >	factors,
		const bool	normalized
	)

inlineexplicit

◆ ~DCEMFactor()

template<class DCFactorType >

virtual dcsam::DCEMFactor< DCFactorType >::~DCEMFactor ( )

virtualdefault

Member Function Documentation

◆ computeComponentLogProbs()

template<class DCFactorType >

std::vector<double> dcsam::DCEMFactor< DCFactorType >::computeComponentLogProbs	(	const gtsam::Values &	continuousVals,
		const DiscreteValues &	discreteVals
	)		const

inline

◆ dim()

template<class DCFactorType >

size_t dcsam::DCEMFactor< DCFactorType >::dim ( ) const

inlineoverridevirtual

Returns the number of rows in the Jacobian with respect to the continuous variables for this factor. Internally this is used in the conversion to a gtsam::NonlinearFactor

TODO(kevin): not sure if needed???

This will be specific to the model being implemented, so it is a pure virtual function and must be overridden

Returns: the dimension (number of rows) in the Jacobian of the error function with respect to continuous variables for

Implements dcsam::DCFactor.

◆ equals()

template<class DCFactorType >

bool dcsam::DCEMFactor< DCFactorType >::equals	(	const DCFactor &	other,
		double	tol = `1e-9`
	)		const

inlineoverridevirtual

Returns true when the DCFactor is equal to other

This will be specific to the model being implemented, so it is a pure virtual function and must be overridden.

Parameters

other	- the DCFactor for which to test equality.
tol	= 1e-9 - numerical tolerance for any floating point comparisons.

Implements dcsam::DCFactor.

◆ error()

template<class DCFactorType >

double dcsam::DCEMFactor< DCFactorType >::error	(	const gtsam::Values &	continuousVals,
		const DiscreteValues &	discreteVals
	)		const

inlineoverridevirtual

Return the error for the DCFactor given an assignment to the variables referenced by the discrete keys and the continuous keys (keys_).

This will be specific to the model being implemented, so it is a pure virtual function and must be overridden.

NOTE: Depending on the application, it may be important that the error function is normalized, i.e. it corresponds to a proper negative log-likelihood, rather than the negative log-likelihood with normalization constants discarded.

Parameters

continuousVals	- the values assigned to the continuous variables, must contain keys in keys_ member variable
discreteVals	- likewise, the values assigned to discrete variables, must contain keys in discreteKeys_

Returns: error (usually the negative log-likelihood) for the measurement model as a double.

Implements dcsam::DCFactor.

◆ getActiveFactorIdx()

template<class DCFactorType >

size_t dcsam::DCEMFactor< DCFactorType >::getActiveFactorIdx	(	const gtsam::Values &	continuousVals,
		const DiscreteValues &	discreteVals
	)		const

inline

◆ getAssociationKeys()

template<class DCFactorType >

gtsam::FastVector<gtsam::Key> dcsam::DCEMFactor< DCFactorType >::getAssociationKeys	(	const gtsam::Values &	continuousVals,
		const DiscreteValues &	discreteVals
	)		const

inline

◆ linearize()

template<class DCFactorType >

boost::shared_ptr<gtsam::GaussianFactor> dcsam::DCEMFactor< DCFactorType >::linearize	(	const gtsam::Values &	continuousVals,
		const DiscreteValues &	discreteVals
	)		const

inlineoverridevirtual

Linearize the error function with respect to the continuous variables (given in keys_) at the point specified by continuousVals. Since this Jacobian can be dependent on the assignment to discrete variables as well, they are required as a parameter.

This will be specific to the model being implemented, so it is a pure virtual function and must be overridden.

Parameters

continuousVals	- Assignment to the continuous variables in `keys_`
discreteVals	- Likewise, assignment to the discrete variables in `discreteKeys__`.

Implements dcsam::DCFactor.

◆ operator=()

template<class DCFactorType >

DCEMFactor& dcsam::DCEMFactor< DCFactorType >::operator= ( const DCEMFactor< DCFactorType > & rhs )

inline

◆ toDecisionTreeFactor()

template<class DCFactorType >

gtsam::DecisionTreeFactor dcsam::DCEMFactor< DCFactorType >::toDecisionTreeFactor	(	const gtsam::Values &	continuousVals,
		const DiscreteValues &	discreteVals
	)		const

inlineoverridevirtual

Converts the DCFactor to a gtsam::DecisionTreeFactor. Internally, this will be used to generate a gtsam::DiscreteFactor type, which itself requires a conversion function to gtsam::DecisionTreeFactor for inference using GTSAM.

Performing this conversion can be problem specific, so we allow for the option to override, but try to implement a sensible default: we assume the error function can be called setting each discrete key's variable individually, and the overall DCFactor can itself be factored as a product of unary gtsam::DecisionTreeFactors.

Alternative implementations might consider something like the gtsam::AllDiff approach here: https://github.com/borglab/gtsam/blob/43e8f1e5aeaf11890262722c1e5e04a11dbf9d75/gtsam_unstable/discrete/AllDiff.cpp#L43

Parameters

continuousVals	- an assignment to the continuous variables
discreteVals	-

Returns: a gtsam::DecisionTreeFactor implementing this DCFactor.

Reimplemented from dcsam::DCFactor.

◆ updateWeights()

template<class DCFactorType >

void dcsam::DCEMFactor< DCFactorType >::updateWeights ( const std::vector< double > & weights )

inline

The documentation for this class was generated from the following file:

include/dcsam/DCEMFactor.h

Public Types

Public Member Functions

Additional Inherited Members

Detailed Description

template<class DCFactorType> class dcsam::DCEMFactor< DCFactorType >

Member Typedef Documentation

◆ Base

Constructor & Destructor Documentation

◆ DCEMFactor() [1/3]

◆ DCEMFactor() [2/3]

◆ DCEMFactor() [3/3]

◆ ~DCEMFactor()

Member Function Documentation

◆ computeComponentLogProbs()

◆ dim()

◆ equals()

◆ error()

◆ getActiveFactorIdx()

◆ getAssociationKeys()

◆ linearize()

◆ operator=()

◆ toDecisionTreeFactor()

◆ updateWeights()

template<class DCFactorType>
class dcsam::DCEMFactor< DCFactorType >