PLearn 0.1
Public Member Functions | Static Public Member Functions | Public Attributes | Static Public Attributes | Protected Member Functions | Private Types
PLearn::ThresholdBpropVariable Class Reference

#include <ThresholdBpropVariable.h>

Inheritance diagram for PLearn::ThresholdBpropVariable:
Inheritance graph
[legend]
Collaboration diagram for PLearn::ThresholdBpropVariable:
Collaboration graph
[legend]

List of all members.

Public Member Functions

 ThresholdBpropVariable ()
 Default constructor for persistence.
 ThresholdBpropVariable (Variable *input, real the_gradient_threshold_factor=0.0)
virtual string classname () const
virtual OptionListgetOptionList () const
virtual OptionMapgetOptionMap () const
virtual RemoteMethodMapgetRemoteMethodMap () const
virtual ThresholdBpropVariabledeepCopy (CopiesMap &copies) const
virtual void build ()
 Post-constructor.
virtual void recomputeSize (int &l, int &w) const
 Recomputes the length l and width w that this variable should have, according to its parent variables.
virtual void fprop ()
 Nothing to do by default.
virtual void bprop ()
 Nothing to do by default.

Static Public Member Functions

static string _classname_ ()
 ThresholdBpropVariable.
static OptionList_getOptionList_ ()
static RemoteMethodMap_getRemoteMethodMap_ ()
static Object_new_instance_for_typemap_ ()
static bool _isa_ (const Object *o)
static void _static_initialize_ ()
static const PPathdeclaringFile ()
static void declareOptions (OptionList &ol)
 Declare options (data fields) for the class.

Public Attributes

real gradient_threshold_factor

Static Public Attributes

static StaticInitializer _static_initializer_

Protected Member Functions

void build_ ()
 Object-specific post-constructor.

Private Types

typedef UnaryVariable inherited

Detailed Description

* ThresholdBpropVariable *

Variable that copies its unique input, and can completely block the flow of gradient backward, or optionally let if flow partially (with an optional coefficient), i.e. fprop: output = input bprop: dC/dinput = sign(dC/doutput) min(abs(input*gradient_threshold_factor),abs(dC/doutput))

Definition at line 61 of file ThresholdBpropVariable.h.


Member Typedef Documentation

Reimplemented from PLearn::UnaryVariable.

Definition at line 63 of file ThresholdBpropVariable.h.


Constructor & Destructor Documentation

PLearn::ThresholdBpropVariable::ThresholdBpropVariable ( ) [inline]

Default constructor for persistence.

Definition at line 70 of file ThresholdBpropVariable.h.

PLearn::ThresholdBpropVariable::ThresholdBpropVariable ( Variable input,
real  the_gradient_threshold_factor = 0.0 
)

Definition at line 61 of file ThresholdBpropVariable.cc.

References build_().

    : inherited(input, input->length(), input->width()),
      gradient_threshold_factor(the_gradient_threshold_factor)
{
    build_();
}

Here is the call graph for this function:


Member Function Documentation

string PLearn::ThresholdBpropVariable::_classname_ ( ) [static]

ThresholdBpropVariable.

Reimplemented from PLearn::UnaryVariable.

Definition at line 59 of file ThresholdBpropVariable.cc.

OptionList & PLearn::ThresholdBpropVariable::_getOptionList_ ( ) [static]

Reimplemented from PLearn::UnaryVariable.

Definition at line 59 of file ThresholdBpropVariable.cc.

RemoteMethodMap & PLearn::ThresholdBpropVariable::_getRemoteMethodMap_ ( ) [static]

Reimplemented from PLearn::UnaryVariable.

Definition at line 59 of file ThresholdBpropVariable.cc.

bool PLearn::ThresholdBpropVariable::_isa_ ( const Object o) [static]

Reimplemented from PLearn::UnaryVariable.

Definition at line 59 of file ThresholdBpropVariable.cc.

Object * PLearn::ThresholdBpropVariable::_new_instance_for_typemap_ ( ) [static]

Reimplemented from PLearn::UnaryVariable.

Definition at line 59 of file ThresholdBpropVariable.cc.

StaticInitializer ThresholdBpropVariable::_static_initializer_ & PLearn::ThresholdBpropVariable::_static_initialize_ ( ) [static]

Reimplemented from PLearn::UnaryVariable.

Definition at line 59 of file ThresholdBpropVariable.cc.

void PLearn::ThresholdBpropVariable::bprop ( ) [virtual]

Nothing to do by default.

Reimplemented from PLearn::UnaryVariable.

Definition at line 105 of file ThresholdBpropVariable.cc.

References PLearn::abs(), gradient_threshold_factor, PLearn::Variable::gradientdata, and PLearn::UnaryVariable::input.

{
    if(gradient_threshold_factor!=0)
    {
        for(int k=0; k<input->nelems(); k++)
            if(abs(gradientdata[k]) > abs(gradient_threshold_factor * input->valuedata[k]))
                input->gradientdata[k] += abs(gradient_threshold_factor * input->valuedata[k]) * (gradientdata[k]>0 ? 1 : -1);
            else
                input->gradientdata[k] += gradientdata[k];
    }
}

Here is the call graph for this function:

void PLearn::ThresholdBpropVariable::build ( ) [virtual]

Post-constructor.

The normal implementation should call simply inherited::build(), then this class's build_(). This method should be callable again at later times, after modifying some option fields to change the "architecture" of the object.

Reimplemented from PLearn::UnaryVariable.

Definition at line 69 of file ThresholdBpropVariable.cc.

References PLearn::UnaryVariable::build(), and build_().

Here is the call graph for this function:

void PLearn::ThresholdBpropVariable::build_ ( ) [protected]

Object-specific post-constructor.

This method should be redefined in subclasses and do the actual building of the object according to previously set option fields. Constructors can just set option fields, and then call build_. This method is NOT virtual, and will typically be called only from three places: a constructor, the public virtual build() method, and possibly the public virtual read method (which calls its parent's read). build_() can assume that its parent's build_() has already been called.

Reimplemented from PLearn::UnaryVariable.

Definition at line 76 of file ThresholdBpropVariable.cc.

Referenced by build(), and ThresholdBpropVariable().

{
}

Here is the caller graph for this function:

string PLearn::ThresholdBpropVariable::classname ( ) const [virtual]

Reimplemented from PLearn::UnaryVariable.

Definition at line 59 of file ThresholdBpropVariable.cc.

void PLearn::ThresholdBpropVariable::declareOptions ( OptionList ol) [static]

Declare options (data fields) for the class.

Redefine this in subclasses: call declareOption(...) for each option, and then call inherited::declareOptions(options). Please call the inherited method AT THE END to get the options listed in a consistent order (from most recently defined to least recently defined).

  static void MyDerivedClass::declareOptions(OptionList& ol)
  {
      declareOption(ol, "inputsize", &MyObject::inputsize_,
                    OptionBase::buildoption,
                    "The size of the input; it must be provided");
      declareOption(ol, "weights", &MyObject::weights,
                    OptionBase::learntoption,
                    "The learned model weights");
      inherited::declareOptions(ol);
  }
Parameters:
olList of options that is progressively being constructed for the current class.

Reimplemented from PLearn::UnaryVariable.

Definition at line 81 of file ThresholdBpropVariable.cc.

References PLearn::OptionBase::buildoption, PLearn::declareOption(), PLearn::UnaryVariable::declareOptions(), and gradient_threshold_factor.

{
    declareOption(ol, "gradient_threshold_factor", &ThresholdBpropVariable::gradient_threshold_factor, 
                  OptionBase::buildoption, "Factor by which the value of the variable is multiplied to obtain the absolute threshold\n"
                  "on the gradient.");
    inherited::declareOptions(ol);
}

Here is the call graph for this function:

static const PPath& PLearn::ThresholdBpropVariable::declaringFile ( ) [inline, static]

Reimplemented from PLearn::UnaryVariable.

Definition at line 75 of file ThresholdBpropVariable.h.

:
    void build_();
ThresholdBpropVariable * PLearn::ThresholdBpropVariable::deepCopy ( CopiesMap copies) const [virtual]

Reimplemented from PLearn::UnaryVariable.

Definition at line 59 of file ThresholdBpropVariable.cc.

void PLearn::ThresholdBpropVariable::fprop ( ) [virtual]

Nothing to do by default.

Reimplemented from PLearn::UnaryVariable.

Definition at line 98 of file ThresholdBpropVariable.cc.

References PLearn::UnaryVariable::input, and PLearn::Variable::valuedata.

{
    for(int k=0; k<input->nelems(); k++)
        valuedata[k] = input->valuedata[k];
}
OptionList & PLearn::ThresholdBpropVariable::getOptionList ( ) const [virtual]

Reimplemented from PLearn::UnaryVariable.

Definition at line 59 of file ThresholdBpropVariable.cc.

OptionMap & PLearn::ThresholdBpropVariable::getOptionMap ( ) const [virtual]

Reimplemented from PLearn::UnaryVariable.

Definition at line 59 of file ThresholdBpropVariable.cc.

RemoteMethodMap & PLearn::ThresholdBpropVariable::getRemoteMethodMap ( ) const [virtual]

Reimplemented from PLearn::UnaryVariable.

Definition at line 59 of file ThresholdBpropVariable.cc.

void PLearn::ThresholdBpropVariable::recomputeSize ( int l,
int w 
) const [virtual]

Recomputes the length l and width w that this variable should have, according to its parent variables.

This is used for ex. by sizeprop() The default version stupidly returns the current dimensions, so make sure to overload it in subclasses if this is not appropriate.

Reimplemented from PLearn::Variable.

Definition at line 89 of file ThresholdBpropVariable.cc.

References PLearn::UnaryVariable::input, PLearn::Var::length(), and PLearn::Var::width().

{
    if (input) {
        l = input->length();
        w = input->width();
    } else
        l = w = 0;
}

Here is the call graph for this function:


Member Data Documentation

Reimplemented from PLearn::UnaryVariable.

Definition at line 75 of file ThresholdBpropVariable.h.

Definition at line 66 of file ThresholdBpropVariable.h.

Referenced by bprop(), and declareOptions().


The documentation for this class was generated from the following files:
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends Defines