L1PenaltyForTG类 参考

详细描述

The is the base class for L1 penalty/regularization within the FirstOrderMinimizer framework.

For L1 penalty, \(L1(w)\)

\[ L1(w)=\| w \|_1 = \sum_i \| w_i \| \]

This class implements the truncated gradient method.

Reference: Stochastic Gradient Descent Training for L1-regularized Log-linear Models with Cumulative Penalty

在文件 L1PenaltyForTG.h 第 53 行定义.

类 L1PenaltyForTG 继承关系图:

Public 成员函数
	L1PenaltyForTG ()
virtual	~L1PenaltyForTG ()
virtual void	update_variable_for_proximity (SGVector< float64_t > variable, float64_t proximal_weight)
virtual void	update_context (CMinimizerContext *context)
virtual void	load_from_context (CMinimizerContext *context)
virtual float64_t	get_penalty (float64_t variable)
virtual float64_t	get_penalty_gradient (float64_t variable, float64_t gradient_of_variable)
virtual void	set_rounding_epsilon (float64_t epsilon)
virtual float64_t	get_sparse_variable (float64_t variable, float64_t penalty_weight)

Protected 属性
float64_t	m_u
SGVector< float64_t >	m_q
float64_t	m_rounding_epsilon

构造及析构函数说明

L1PenaltyForTG

(

)

在文件 L1PenaltyForTG.h 第 57 行定义.

virtual ~L1PenaltyForTG ( )

virtual

在文件 L1PenaltyForTG.h 第 60 行定义.

成员函数说明

virtual float64_t get_penalty ( float64_t  variable )

virtualinherited

Given the value of a target variable, this method returns the penalty of the variable

参数

variable

value of the variable

返回

penalty of the variable

实现了 Penalty.

在文件 L1Penalty.h 第 66 行定义.

virtual float64_t get_penalty_gradient ( float64_t  variable, float64_t  gradient_of_variable  )

virtualinherited

Return the gradient of the penalty wrt a target variable

Note that for L1 penalty we do not compute the gradient/sub-gradient in our implementation. Instead, we do a proximal projection.

参数

variable	value of a target variable
gradient_of_variable	unregularized/unpenalized gradient of the variable

返回

the gradient of the penalty wrt the variable

实现了 Penalty.

在文件 L1Penalty.h 第 79 行定义.

virtual float64_t get_sparse_variable ( float64_t  variable, float64_t  penalty_weight  )

virtualinherited

Get the sparse variable

参数

variable	the raw variable
penalty_weight	weight of the penalty

返回

sparse value of the variable

实现了 SparsePenalty.

在文件 L1Penalty.h 第 128 行定义.

virtual void load_from_context ( CMinimizerContext *  context )

virtual

Load the given context object to restore mutable variables

参数

context

a context object

重载 L1Penalty .

在文件 L1PenaltyForTG.h 第 115 行定义.

virtual void set_rounding_epsilon ( float64_t  epsilon )

virtualinherited

Set the rounding epsilon

参数

epsilon

rounding epsilon

在文件 L1Penalty.h 第 87 行定义.

virtual void update_context ( CMinimizerContext *  context )

virtual

Update a context object to store mutable variables used in learning rate

参数

context

a context object

重载 L1Penalty .

在文件 L1PenaltyForTG.h 第 96 行定义.

virtual void update_variable_for_proximity ( SGVector< float64_t >  variable, float64_t  proximal_weight  )

virtual

Do proximal projection/operation in place

参数

variable	the raw variable
proximal_weight	weight of the penalty

重载 L1Penalty .

在文件 L1PenaltyForTG.h 第 66 行定义.

类成员变量说明

SGVector<float64_t> m_q

protected

q is defined in Figure 2 of the reference

在文件 L1PenaltyForTG.h 第 132 行定义.

float64_t m_rounding_epsilon

protectedinherited

rounding epsilon

在文件 L1Penalty.h 第 148 行定义.

float64_t m_u

protected

u is defined in Figure 2 of the reference

在文件 L1PenaltyForTG.h 第 130 行定义.

---

该类的文档由以下文件生成:

• L1PenaltyForTG.h