lucid::LbfgsTuner Class Reference

Optimiser that uses the L-BFGS algorithm. More...

#include <LbfgsTuner.h>

Inheritance diagram for lucid::LbfgsTuner:

Public Member Functions
	LbfgsTuner (const LbfgsParameters &parameters={})
	Construct a new LbfgsTuner with the provided parameters.
	LbfgsTuner (const Eigen::VectorXd &lb, const Eigen::VectorXd &ub, const LbfgsParameters &parameters={})
	Construct a new LbfgsTuner with the provided bounds and parameters.
	LbfgsTuner (const std::vector< std::pair< Scalar, Scalar > > &bounds, const LbfgsParameters &parameters={})
	Construct a new LbfgsTuner with the provided bounds and parameters.
bool	is_bounded () const
	Check whether the optimisation is bounded.
const LbfgsParameters &	parameters () const
	Get read-only access to the parameters of the L-BGS optimisation.
std::string	to_string () const override
	Obtain the string representation of this object.
Public Member Functions inherited from lucid::Tuner
void	tune (Estimator &estimator, ConstMatrixRef training_inputs, ConstMatrixRef training_outputs) const
	Optimise the kernel hyperparameters.
void	tune_online (Estimator &estimator, ConstMatrixRef training_inputs, const OutputComputer &training_outputs) const
	Optimise the kernel hyperparameters.

Private Member Functions
void	tune_impl (Estimator &estimator, ConstMatrixRef training_inputs, const OutputComputer &training_outputs) const override
	Optimise the kernel hyperparameters.

Private Attributes
Eigen::VectorXd	lb_
	Lower bounds for the parameters. If empty, no bounds are applied.
Eigen::VectorXd	ub_
	Upper bounds for the parameters. If empty, no bounds are applied.
LbfgsParameters	parameters_
	Optimization parameters for the L-BFGS algorithm.

Detailed Description

Optimiser that uses the L-BFGS algorithm.

It optimizes the parameters of an estimator by minimizing the loss function. It can operate with or without bounds on the parameters. If bounds are provided, they must match the dimension of the solution space. Note that the solution space for isotropic estimators (i.e., those where the parameters are the same for all dimensions) will always have a dimension of 1.

Constructor & Destructor Documentation

LbfgsTuner() [1/3]

lucid::LbfgsTuner::LbfgsTuner ( const LbfgsParameters & parameters = {} )

explicit

Construct a new LbfgsTuner with the provided parameters.

Parameters

parameters

optimization parameters for the L-BFGS algorithm

LbfgsTuner() [2/3]

lucid::LbfgsTuner::LbfgsTuner ( const Eigen::VectorXd & lb, const Eigen::VectorXd & ub, const LbfgsParameters & parameters = {} )

explicit

Construct a new LbfgsTuner with the provided bounds and parameters.

The solution will be constrained to the provided bounds. The size of the bounds must match the dimension of the solution space.

Precondition

lb and ub must have the same size

Parameters

lb	Lower bounds for the parameters
ub	Upper bounds for the parameters
parameters	optimization parameters for the L-BFGS algorithm

LbfgsTuner() [3/3]

lucid::LbfgsTuner::LbfgsTuner ( const std::vector< std::pair< Scalar, Scalar > > & bounds, const LbfgsParameters & parameters = {} )

explicit

Construct a new LbfgsTuner with the provided bounds and parameters.

The solution will be constrained to the provided bounds. The size of the bounds must match the dimension of the solution space.

Parameters

bounds	vector of pairs of lower and upper bounds
parameters	optimization parameters for the L-BFGS algorithm

Member Function Documentation

is_bounded()

bool lucid::LbfgsTuner::is_bounded ( ) const

nodiscard

Check whether the optimisation is bounded.

Returns

true if the optimisation is bounded

false if the optimisation is not bounded

parameters()

const LbfgsParameters & lucid::LbfgsTuner::parameters ( ) const

inlinenodiscard

Get read-only access to the parameters of the L-BGS optimisation.

Returns

parameters of the L-BGS optimisation

to_string()

std::string lucid::LbfgsTuner::to_string ( ) const

nodiscardoverridevirtual

Obtain the string representation of this object.

Returns

string representation of this object

Reimplemented from lucid::Tuner.

tune_impl()

void lucid::LbfgsTuner::tune_impl ( Estimator & estimator, ConstMatrixRef training_inputs, const OutputComputer & training_outputs ) const

overrideprivatevirtual

Optimise the kernel hyperparameters.

Starting from the initial guess, the optimiser finds the best hyperparameters for the kernel. It is up to the subclass to determine the optimisation algorithm used.

Note

The method must call the consolidate method of the estimator to ensure that the kernel is ready to be used.

Precondition

The number of rows in the training inputs should be equal to the number of rows in the training outputs.

Todo

If the OutputComputer becomes a bottleneck, consider having a separate method for when the outputs are available from the get-go.

Parameters

estimator	estimator to optimise
training_inputs	training input data
training_outputs	training output data. It uses an OutputComputer to compute the outputs when needed

Implements lucid::Tuner.

---

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

• lucid/model/LbfgsTuner.h
• lucid/model/LbfgsTuner.cpp