A utility class template that, when instatiated with a MarginFunction, produces the corresponding linear classifier loss. More...

#include <generic_linear.h>

Inheritance diagram for dismec::objective::GenericMarginClassifier< MarginFunction >:

Public Member Functions
	GenericMarginClassifier (std::shared_ptr< const GenericFeatureMatrix > X, std::unique_ptr< Objective > regularizer, MarginFunction phi)

void	calculate_loss (const DenseRealVector &scores, const BinaryLabelVector &labels, DenseRealVector &out) const override
	Calculates the loss for each instance. More...

void	calculate_derivative (const DenseRealVector &scores, const BinaryLabelVector &labels, DenseRealVector &out) const override
	Calculates the derivative of the loss with respect to the scores for each instance. More...

void	calculate_2nd_derivative (const DenseRealVector &scores, const BinaryLabelVector &labels, DenseRealVector &out) const override
	Calculates the 2nd derivative of the loss with respect to the scores for each instance. More...

Public Member Functions inherited from dismec::objective::GenericLinearClassifier
	GenericLinearClassifier (std::shared_ptr< const GenericFeatureMatrix > X, std::unique_ptr< Objective > regularizer)

Public Member Functions inherited from dismec::objective::LinearClassifierBase
	LinearClassifierBase (std::shared_ptr< const GenericFeatureMatrix > X)

long	num_instances () const noexcept

long	num_variables () const noexcept override

BinaryLabelVector &	get_label_ref ()

void	update_costs (real_t positive, real_t negative)

Public Member Functions inherited from dismec::objective::Objective
	Objective ()

virtual	~Objective () noexcept=default

real_t	value (const HashVector &location)
	Evaluate the objective at the given `location`. More...

void	diag_preconditioner (const HashVector &location, Eigen::Ref< DenseRealVector > target)
	Get precondition to be used in CG optimization. More...

void	project_to_line (const HashVector &location, const DenseRealVector &direction)
	creates a function g such that `g(a) = objective(location + a * direction)` Use `lookup_on_line()` to evaluate `g`. More...

void	gradient_at_zero (Eigen::Ref< DenseRealVector > target)
	Gets the gradient for location zero. More...

void	gradient (const HashVector &location, Eigen::Ref< DenseRealVector > target)
	Evaluate the gradient at `location`. More...

void	hessian_times_direction (const HashVector &location, const DenseRealVector &direction, Eigen::Ref< DenseRealVector > target)
	Calculates the product of the Hessian matrix at `location` with `direction`. More...

void	gradient_and_pre_conditioner (const HashVector &location, Eigen::Ref< DenseRealVector > gradient, Eigen::Ref< DenseRealVector > pre)
	Combines the calculation of gradient and pre-conditioner, which may be more efficient in some cases. More...

Public Member Functions inherited from dismec::stats::Tracked
	Tracked ()
	Default constructor, creates the internal `stats::StatisticsCollection`. More...

void	register_stat (const std::string &name, std::unique_ptr< Statistics > stat)
	Registers a tracker for the statistics `name`. More...

std::shared_ptr< StatisticsCollection >	get_stats () const
	Gets an ownership-sharing reference to the `StatisticsCollection`. More...

Public Attributes
MarginFunction	Phi

Additional Inherited Members
Protected Member Functions inherited from dismec::objective::LinearClassifierBase
const DenseRealVector &	x_times_w (const HashVector &w)
	Calculates the vector of feature matrix times weights `w` More...

template<class Derived >
void	update_xtw_cache (const HashVector &new_weight, const Eigen::MatrixBase< Derived > &new_result)
	Updates the cached value for x_times_w. More...

void	project_linear_to_line (const HashVector &location, const DenseRealVector &direction)
	Prepares the cache variables for line projection. More...

auto	line_interpolation (real_t t) const

void	declare_vector_on_last_line (const HashVector &location, real_t t) override
	State that the given vector corresponds to a certain position on the line of the last line search. More...

const GenericFeatureMatrix &	generic_features () const

const DenseFeatures &	dense_features () const

const SparseFeatures &	sparse_features () const

const DenseRealVector &	costs () const

const BinaryLabelVector &	labels () const

Protected Member Functions inherited from dismec::stats::Tracked
	~Tracked ()
	Non-virtual destructor. Declared protected, so we don't accidentally try to do a polymorphic delete. More...

template<class T >
void	record (stat_id_t stat, T &&value)
	Record statistics. This function just forwards all its arguments to the internal `StatisticsCollection`. More...

void	declare_stat (stat_id_t index, StatisticMetaData meta)
	Declares a new statistics. This function just forwards all its arguments to the internal `StatisticsCollection`. More...

void	declare_tag (tag_id_t index, std::string name)
	Declares a new tag. This function just forwards all its arguments to the internal `StatisticsCollection`. More...

template<class... Args>
void	set_tag (tag_id_t tag, long value)
	Set value of tag. This function just forwards all its arguments to the internal `StatisticsCollection`. More...

template<class... Args>
auto	make_timer (stat_id_t id, Args... args)
	Creates a new `ScopeTimer` using `stats::record_scope_time`. More...

Detailed Description

template<class MarginFunction>
struct dismec::objective::GenericMarginClassifier< MarginFunction >

A utility class template that, when instatiated with a MarginFunction, produces the corresponding linear classifier loss.

Template Parameters

MarginFunction A class that needs to provide three scalar functions: value, grad, and quad that give the value, first order, and second order approximation to the function at the given point.

Definition at line 120 of file generic_linear.h.

Constructor & Destructor Documentation

◆ GenericMarginClassifier()

template<class MarginFunction >

dismec::objective::GenericMarginClassifier< MarginFunction >::GenericMarginClassifier	(	std::shared_ptr< const GenericFeatureMatrix >	X,
		std::unique_ptr< Objective >	regularizer,
		MarginFunction	phi
	)

inline

Definition at line 121 of file generic_linear.h.

Member Function Documentation

◆ calculate_2nd_derivative()

template<class MarginFunction >

void dismec::objective::GenericMarginClassifier< MarginFunction >::calculate_2nd_derivative	(	const DenseRealVector &	scores,
		const BinaryLabelVector &	labels,
		DenseRealVector &	out
	)		const

inlineoverridevirtual

Calculates the 2nd derivative of the loss with respect to the scores for each instance.

Parameters

[in]	scores	The scores, i.e. the product weights times features, for each instance.
[in]	labels	The binary labels as a vector of plus and minus ones.
[out]	out	This vector will be filled with the instance-wise loss 2nd derivatives.

Implements dismec::objective::GenericLinearClassifier.

Definition at line 149 of file generic_linear.h.

References dismec::objective::LinearClassifierBase::labels(), and dismec::objective::GenericMarginClassifier< MarginFunction >::Phi.

◆ calculate_derivative()

template<class MarginFunction >

void dismec::objective::GenericMarginClassifier< MarginFunction >::calculate_derivative	(	const DenseRealVector &	scores,
		const BinaryLabelVector &	labels,
		DenseRealVector &	out
	)		const

inlineoverridevirtual

Calculates the derivative of the loss with respect to the scores for each instance.

Parameters

[in]	scores	The scores, i.e. the product weights times features, for each instance.
[in]	labels	The binary labels as a vector of plus and minus ones.
[out]	out	This vector will be filled with the instance-wise loss derivatives.

Implements dismec::objective::GenericLinearClassifier.

Definition at line 138 of file generic_linear.h.

References dismec::objective::LinearClassifierBase::labels(), and dismec::objective::GenericMarginClassifier< MarginFunction >::Phi.

◆ calculate_loss()

template<class MarginFunction >

void dismec::objective::GenericMarginClassifier< MarginFunction >::calculate_loss	(	const DenseRealVector &	scores,
		const BinaryLabelVector &	labels,
		DenseRealVector &	out
	)		const

inlineoverridevirtual

Calculates the loss for each instance.

Parameters

[in]	scores	The scores, i.e. the product weights times features, for each instance.
[in]	labels	The binary labels as a vector of plus and minus ones.
[out]	out	This vector will be filled with the instance-wise loss value.

Implements dismec::objective::GenericLinearClassifier.

Definition at line 128 of file generic_linear.h.

References dismec::objective::LinearClassifierBase::labels(), and dismec::objective::GenericMarginClassifier< MarginFunction >::Phi.