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

Stores and learns the parameters between one quadratic layer and one linear layer of an RBM. More...

#include <RBMLQParameters.h>

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

List of all members.

Public Member Functions

 RBMLQParameters (real the_learning_rate=0)
 Default constructor.
 RBMLQParameters (string down_types, string up_types, real the_learning_rate=0)
 Constructor from two string prototymes.
virtual void accumulatePosStats (const Vec &down_values, const Vec &up_values)
 Accumulates positive phase statistics to *_pos_stats.
virtual void accumulateNegStats (const Vec &down_values, const Vec &up_values)
 Accumulates negative phase statistics to *_neg_stats.
virtual void update ()
 Updates parameters according to contrastive divergence gradient.
virtual void clearStats ()
 Clear all information accumulated during stats.
virtual void computeUnitActivations (int start, int length, const Vec &activations) const
 Computes the vectors of activation of "length" units, starting from "start", and concatenates them into "activations".
virtual void bpropUpdate (const Vec &input, const Vec &output, Vec &input_gradient, const Vec &output_gradient)
 Adapt based on the output gradient: this method should only be called just after a corresponding fprop; it should be called with the same arguments as fprop for the first two arguments (and output should not have been modified since then).
virtual void forget ()
 reset the parameters to the state they would be BEFORE starting training.
virtual int nParameters (bool share_up_params, bool share_down_params) const
 optionally perform some processing after training, or after a series of fprop/bpropUpdate calls to prepare the model for truly out-of-sample operation.
virtual Vec makeParametersPointHere (const Vec &global_parameters, bool share_up_params, bool share_down_params)
 Make the parameters data be sub-vectors of the given global_parameters.
virtual string classname () const
virtual OptionListgetOptionList () const
virtual OptionMapgetOptionMap () const
virtual RemoteMethodMapgetRemoteMethodMap () const
virtual RBMLQParametersdeepCopy (CopiesMap &copies) const
virtual void build ()
 Post-constructor.
virtual void makeDeepCopyFromShallowCopy (CopiesMap &copies)
 Transforms a shallow copy into a deep copy.

Static Public Member Functions

static string _classname_ ()
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 ()

Public Attributes

Mat weights
 Matrix containing unit-to-unit weights (output_size × input_size)
Vec down_units_bias
 Element i contains the bias of linear down unit i.
TVec< Vecup_units_params
 Element 1,i contains the bias of up unit i, element (2,i) contains the.
Mat weights_pos_stats
 Accumulates positive contribution to the weights' gradient.
Mat weights_neg_stats
 Accumulates negative contribution to the weights' gradient.
Vec down_units_bias_pos_stats
 Accumulates positive contribution to the gradient of down_units_bias.
Vec down_units_bias_neg_stats
 Accumulates negative contribution to the gradient of down_units_params.
TVec< Vecup_units_params_pos_stats
 Accumulates positive contribution to the gradient of down_units_params.
TVec< Vecup_units_params_neg_stats
 Accumulates negative contribution to the gradient of down_units_params.

Static Public Attributes

static StaticInitializer _static_initializer_

Static Protected Member Functions

static void declareOptions (OptionList &ol)
 Declares the class options.

Private Types

typedef RBMParameters inherited

Private Member Functions

void build_ ()
 This does the actual building.

Detailed Description

Stores and learns the parameters between one quadratic layer and one linear layer of an RBM.

The lower layer is quadratic and the upper one is linear. See RBMLQParameters.* when the layers are switched.

Todo:
: yes
Deprecated:
Use ../RBMMatrixConnection.h instead

Definition at line 56 of file RBMLQParameters.h.


Member Typedef Documentation

Reimplemented from PLearn::RBMParameters.

Definition at line 58 of file RBMLQParameters.h.


Constructor & Destructor Documentation

PLearn::RBMLQParameters::RBMLQParameters ( real  the_learning_rate = 0)

Default constructor.

Definition at line 52 of file RBMLQParameters.cc.

                                                         :
    inherited(the_learning_rate)
{
}
PLearn::RBMLQParameters::RBMLQParameters ( string  down_types,
string  up_types,
real  the_learning_rate = 0 
)

Constructor from two string prototymes.

Definition at line 57 of file RBMLQParameters.cc.

References build().

                                                           :
    inherited( down_types, up_types, the_learning_rate )
{
    // We're not sure inherited::build() has been called
    build();
}

Here is the call graph for this function:


Member Function Documentation

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

Reimplemented from PLearn::RBMParameters.

Definition at line 50 of file RBMLQParameters.cc.

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

Reimplemented from PLearn::RBMParameters.

Definition at line 50 of file RBMLQParameters.cc.

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

Reimplemented from PLearn::RBMParameters.

Definition at line 50 of file RBMLQParameters.cc.

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

Reimplemented from PLearn::RBMParameters.

Definition at line 50 of file RBMLQParameters.cc.

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

Reimplemented from PLearn::Object.

Definition at line 50 of file RBMLQParameters.cc.

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

Reimplemented from PLearn::RBMParameters.

Definition at line 50 of file RBMLQParameters.cc.

void PLearn::RBMLQParameters::accumulateNegStats ( const Vec down_values,
const Vec up_values 
) [virtual]

Accumulates negative phase statistics to *_neg_stats.

Implements PLearn::RBMParameters.

Definition at line 193 of file RBMLQParameters.cc.

References down_units_bias_neg_stats, PLearn::externalProductAcc(), i, PLearn::RBMParameters::neg_count, PLearn::RBMParameters::up_layer_size, up_units_params, up_units_params_neg_stats, and weights_neg_stats.

{
    // weights_pos_stats += up_values * down_values'
    externalProductAcc( weights_neg_stats, up_values, down_values );

    down_units_bias_neg_stats += down_values;
    up_units_params_neg_stats[0] += up_values;

    for(int i=0 ; i<up_layer_size ; ++i) { 
        up_units_params_neg_stats[1][i] += 2 * up_units_params[1][i] *
            up_values[i] *up_values[i];
    }
    
    neg_count++;

    
}

Here is the call graph for this function:

void PLearn::RBMLQParameters::accumulatePosStats ( const Vec down_values,
const Vec up_values 
) [virtual]

Accumulates positive phase statistics to *_pos_stats.

Implements PLearn::RBMParameters.

Definition at line 176 of file RBMLQParameters.cc.

References down_units_bias_pos_stats, PLearn::externalProductAcc(), i, PLearn::RBMParameters::pos_count, PLearn::RBMParameters::up_layer_size, up_units_params, up_units_params_pos_stats, and weights_pos_stats.

{
    // weights_pos_stats += up_values * down_values'
    externalProductAcc( weights_pos_stats, up_values, down_values );

    down_units_bias_pos_stats += down_values;
    up_units_params_pos_stats[0] += up_values;

    for(int i=0 ; i<up_layer_size ; ++i) { 
        up_units_params_pos_stats[1][i] += 2 * up_units_params[1][i] *
            up_values[i] * up_values[i];
    }
    
    pos_count++;
}

Here is the call graph for this function:

void PLearn::RBMLQParameters::bpropUpdate ( const Vec input,
const Vec output,
Vec input_gradient,
const Vec output_gradient 
) [virtual]

Adapt based on the output gradient: this method should only be called just after a corresponding fprop; it should be called with the same arguments as fprop for the first two arguments (and output should not have been modified since then).

this version allows to obtain the input gradient as well

Since sub-classes are supposed to learn ONLINE, the object is 'ready-to-be-used' just after any bpropUpdate. N.B. A DEFAULT IMPLEMENTATION IS PROVIDED IN THE SUPER-CLASS, WHICH JUST CALLS bpropUpdate(input, output, input_gradient, output_gradient) AND IGNORES INPUT GRADIENT. this version allows to obtain the input gradient as well N.B. THE DEFAULT IMPLEMENTATION IN SUPER-CLASS JUST RAISES A PLERROR.

Definition at line 303 of file RBMLQParameters.cc.

References PLearn::cube(), PLearn::RBMParameters::down_layer_size, PLearn::externalProductAcc(), i, j, PLearn::RBMParameters::learning_rate, PLearn::multiplyAcc(), PLASSERT, PLearn::TVec< T >::resize(), PLearn::TVec< T >::size(), PLearn::transposeProduct(), PLearn::RBMParameters::up_layer_size, up_units_params, and weights.

{
    //TODO: clean up the code a bit
    PLASSERT( input.size() == down_layer_size );
    PLASSERT( output.size() == 2 * up_layer_size );
    PLASSERT( output_gradient.size() == 2 * up_layer_size );
    input_gradient.resize( down_layer_size );

    // weights -= learning_rate * output_gradient * input'
//    externalProductAcc( weights, (-learning_rate)*output_gradient, input );

    Vec scaled_out_grad(up_layer_size) ;  
    
    Vec prod_w_input( up_layer_size ) ; 
    
    for(int i=0 ; i<up_layer_size ; ++i) 
    {
        real a_i_square = up_units_params[1][i] * up_units_params[1][i] ; 
        
        scaled_out_grad[i] = -0.5 * output_gradient[2 * i] / a_i_square ; 
        
//        up_units_params[0][i] -= learning_rate * ( -0.5 / a_i_square ) *
//                                 output_gradient[2*i] ; 
        
        for(int j=0 ; j < down_layer_size ; ++j) {             
            prod_w_input[i] += weights[i][j] * input[j] ; 
        }
    }

    // input_gradient = weights' * output_gradient
    transposeProduct( input_gradient, weights, scaled_out_grad );

    externalProductAcc( weights, (-learning_rate)*scaled_out_grad, input );
    
    // (up) bias -= learning_rate * output_gradient
    multiplyAcc( up_units_params[0], scaled_out_grad, -learning_rate );

    for(int i=0 ; i<up_layer_size ; ++i) { 
      up_units_params[1][i] -= learning_rate * ( up_units_params[0][i] +
                prod_w_input[i] ) / (cube(up_units_params[1][i])) *
          output_gradient[2*i] ; 
    }

    // (up) bias -= learning_rate * output_gradient
//    multiplyAcc( up_units_params[0], output_gradient, -learning_rate );

}

Here is the call graph for this function:

void PLearn::RBMLQParameters::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::RBMParameters.

Definition at line 154 of file RBMLQParameters.cc.

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

Referenced by RBMLQParameters().

Here is the call graph for this function:

Here is the caller graph for this function:

void PLearn::RBMLQParameters::build_ ( ) [private]

This does the actual building.

Reimplemented from PLearn::RBMParameters.

Definition at line 95 of file RBMLQParameters.cc.

References clearStats(), PLearn::RBMParameters::down_layer_size, down_units_bias, down_units_bias_neg_stats, down_units_bias_pos_stats, PLearn::RBMParameters::down_units_types, forget(), i, PLearn::TMat< T >::length(), PLearn::OnlineLearningModule::output_size, PLERROR, PLearn::TVec< T >::resize(), PLearn::TMat< T >::resize(), PLearn::RBMParameters::up_layer_size, up_units_params, up_units_params_neg_stats, up_units_params_pos_stats, PLearn::RBMParameters::up_units_types, weights, weights_neg_stats, weights_pos_stats, and PLearn::TMat< T >::width().

Referenced by build().

{
    if( up_layer_size == 0 || down_layer_size == 0 )
        return;

    output_size = 0;
    bool needs_forget = false; // do we need to reinitialize the parameters?

    if( weights.length() != up_layer_size ||
        weights.width() != down_layer_size )
    {
        weights.resize( up_layer_size, down_layer_size );
        needs_forget = true;
    }

    weights_pos_stats.resize( up_layer_size, down_layer_size );
    weights_neg_stats.resize( up_layer_size, down_layer_size );

    down_units_bias.resize( down_layer_size );
    down_units_bias_pos_stats.resize( down_layer_size );
    down_units_bias_neg_stats.resize( down_layer_size );
    for( int i=0 ; i<down_layer_size ; i++ )
    {
        char dut_i = down_units_types[i];
        if( dut_i != 'l' ) // not linear activation unit
            PLERROR( "RBMLQParameters::build_() - value '%c' for"
                     " up_units_types[%d]\n"
                     "should be 'l'.\n",
                     dut_i, i );
    }
    
    up_units_params.resize( 2 ) ;
    up_units_params[0].resize( up_layer_size );
    up_units_params[1].resize( up_layer_size );

    up_units_params_pos_stats.resize( 2 );
    up_units_params_pos_stats[0].resize( up_layer_size );
    up_units_params_pos_stats[1].resize( up_layer_size );
    
    up_units_params_neg_stats.resize( 2 );
    up_units_params_neg_stats[0].resize( up_layer_size );
    up_units_params_neg_stats[1].resize( up_layer_size );

    for( int i=0 ; i<up_layer_size ; i++ )
    {
        char uut_i = up_units_types[i];
        if( uut_i != 'q' ) // not quadratic activation unit
            PLERROR( "RBMLQParameters::build_() - value '%c' for"
                     " down_units_types[%d]\n"
                     "should be 'q'.\n",
                     uut_i, i );
    }

    if( needs_forget )
        forget();

    clearStats();
}

Here is the call graph for this function:

Here is the caller graph for this function:

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

Reimplemented from PLearn::Object.

Definition at line 50 of file RBMLQParameters.cc.

void PLearn::RBMLQParameters::clearStats ( ) [virtual]
void PLearn::RBMLQParameters::computeUnitActivations ( int  start,
int  length,
const Vec activations 
) const [virtual]

Computes the vectors of activation of "length" units, starting from "start", and concatenates them into "activations".

"start" indexes an up unit if "going_up", else a down unit.

Implements PLearn::RBMParameters.

Definition at line 263 of file RBMLQParameters.cc.

References PLearn::TMat< T >::column(), i, PLearn::TVec< T >::length(), PLASSERT, PLearn::product(), PLearn::TVec< T >::toMat(), and PLearn::transposeProduct().

{
    //activations[2 * i] = mu of unit (i - start)
    //activations[2 * i + 1] = sigma of unit (i - start)
    if( going_up )
    {

        // TODO: change it to work with start and length
        PLASSERT( start+length <= down_layer_size );
        Mat activations_mat = activations.toMat( activations.length()/2 , 2);
        Mat mu = activations_mat.column(0) ; 
        Mat sigma = activations_mat.column(1) ; 
        
        product( mu , weights , input_vec.toMat(input_vec.length(),1) );


        // activations[i-start] = sum_j weights(j,i) input_vec[j] + b[i]
        for(int i=0 ; i<length ; ++i) { 
            real a_i = up_units_params[1][i] ; 
            mu[i][0] = - (mu[i][0] + up_units_params[0][i]) / (2 * a_i * a_i)     ; 
            sigma[i][0] = 1 / (2. * a_i * a_i) ; 
        }
        
    }
    else
    {
        PLASSERT( activations.length() == length );
        PLASSERT( start+length <= up_layer_size );
        // mu = activations[i] = -(sum_j weights(i,j) input_vec[j] + b[i])
        //                    / (2 * up_units_params[i][1]^2)
//        product( weights, input_vec , activations) ;
        transposeProduct( activations , weights, input_vec ) ;
        activations += down_units_bias ; 
    }
}

Here is the call graph for this function:

void PLearn::RBMLQParameters::declareOptions ( OptionList ol) [static, protected]

Declares the class options.

Reimplemented from PLearn::RBMParameters.

Definition at line 65 of file RBMLQParameters.cc.

References PLearn::declareOption(), PLearn::RBMParameters::declareOptions(), down_units_bias, PLearn::OptionBase::learntoption, up_units_params, and weights.

{
    // ### Declare all of this object's options here.
    // ### For the "flags" of each option, you should typically specify
    // ### one of OptionBase::buildoption, OptionBase::learntoption or
    // ### OptionBase::tuningoption. If you don't provide one of these three,
    // ### this option will be ignored when loading values from a script.
    // ### You can also combine flags, for example with OptionBase::nosave:
    // ### (OptionBase::buildoption | OptionBase::nosave)

    declareOption(ol, "weights", &RBMLQParameters::weights,
                  OptionBase::learntoption,
                  "Matrix containing unit-to-unit weights (output_size ×"
                  " input_size)");

    declareOption(ol, "down_units_bias",
                  &RBMLQParameters::down_units_bias,
                  OptionBase::learntoption,
                  "Element i contains the bias of up unit i");

    declareOption(ol, "up_units_params",
                  &RBMLQParameters::up_units_params,
                  OptionBase::learntoption,
                  "Element 0,i contains the bias of down unit i. Element 1,i"
                  "contains the quadratic term of down unit i ");

    // Now call the parent class' declareOptions
    inherited::declareOptions(ol);
}

Here is the call graph for this function:

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

Reimplemented from PLearn::RBMParameters.

Definition at line 167 of file RBMLQParameters.h.

:

RBMLQParameters * PLearn::RBMLQParameters::deepCopy ( CopiesMap copies) const [virtual]

Reimplemented from PLearn::RBMParameters.

Definition at line 50 of file RBMLQParameters.cc.

void PLearn::RBMLQParameters::forget ( ) [virtual]

reset the parameters to the state they would be BEFORE starting training.

Note that this method is necessarily called from build().

Implements PLearn::OnlineLearningModule.

Definition at line 355 of file RBMLQParameters.cc.

References PLearn::TVec< T >::clear(), PLearn::TMat< T >::clear(), clearStats(), d, PLearn::RBMParameters::down_layer_size, down_units_bias, PLearn::TVec< T >::fill(), PLearn::RBMParameters::initialization_method, PLearn::max(), PLearn::OnlineLearningModule::random_gen, PLearn::sqrt(), PLearn::RBMParameters::up_layer_size, up_units_params, and weights.

Referenced by build_().

{
    if( initialization_method == "zero" )
        weights.clear();
    else
    {
        if( !random_gen )
            random_gen = new PRandom();

        real d = 1. / max( down_layer_size, up_layer_size );
        if( initialization_method == "uniform_sqrt" )
            d = sqrt( d );

        random_gen->fill_random_uniform( weights, -d, d );
    }

    up_units_params[0].clear();    
    up_units_params[1].fill(1.);
    
    down_units_bias.clear();

    clearStats();
}

Here is the call graph for this function:

Here is the caller graph for this function:

OptionList & PLearn::RBMLQParameters::getOptionList ( ) const [virtual]

Reimplemented from PLearn::Object.

Definition at line 50 of file RBMLQParameters.cc.

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

Reimplemented from PLearn::Object.

Definition at line 50 of file RBMLQParameters.cc.

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

Reimplemented from PLearn::Object.

Definition at line 50 of file RBMLQParameters.cc.

void PLearn::RBMLQParameters::makeDeepCopyFromShallowCopy ( CopiesMap copies) [virtual]
Vec PLearn::RBMLQParameters::makeParametersPointHere ( const Vec global_parameters,
bool  share_up_params,
bool  share_down_params 
) [virtual]

Make the parameters data be sub-vectors of the given global_parameters.

The argument should have size >= nParameters. The result is a Vec that starts just after this object's parameters end, i.e. result = global_parameters.subVec(nParameters(),global_parameters.size()-nParameters()); This allows to easily chain calls of this method on multiple RBMParameters.

Implements PLearn::RBMParameters.

Definition at line 404 of file RBMLQParameters.cc.

References PLearn::TVec< T >::data(), down_units_bias, i, PLearn::TVec< T >::length(), m, PLearn::TMat< T >::makeSharedValue(), PLearn::TVec< T >::makeSharedValue(), n, nParameters(), PLERROR, PLearn::TVec< T >::size(), PLearn::TMat< T >::size(), PLearn::TVec< T >::subVec(), up_units_params, and weights.

{
    int n = nParameters(share_up_params,share_down_params);
    int m = global_parameters.size();
    if (m<n)
        PLERROR("RBMLLParameters::makeParametersPointHere: argument has length %d, should be longer than nParameters()=%d",m,n);
    real* p = global_parameters.data();
    weights.makeSharedValue(p,weights.size());
    p+=weights.size();
    if (share_down_params)
    {
        down_units_bias.makeSharedValue(p,down_units_bias.size());
        p+=down_units_bias.size();
    }
    if (share_up_params)
        for (int i=0;i<up_units_params.length();i++)
        {
            up_units_params[i].makeSharedValue(p,up_units_params[i].size());
            p+=up_units_params[i].size();
        }
    return global_parameters.subVec(n,m-n);
}

Here is the call graph for this function:

int PLearn::RBMLQParameters::nParameters ( bool  share_up_params,
bool  share_down_params 
) const [virtual]

optionally perform some processing after training, or after a series of fprop/bpropUpdate calls to prepare the model for truly out-of-sample operation.

return the number of parameters

THE DEFAULT IMPLEMENTATION PROVIDED IN THE SUPER-CLASS DOES NOT DO ANYTHING. return the number of parameters

Implements PLearn::RBMParameters.

Definition at line 390 of file RBMLQParameters.cc.

References down_units_bias, i, PLearn::TVec< T >::length(), m, PLearn::TVec< T >::size(), PLearn::TMat< T >::size(), up_units_params, and weights.

Referenced by makeParametersPointHere().

{
    int m = weights.size() +  (share_down_params?down_units_bias.size():0);
    if (share_up_params)
        for (int i=0;i<up_units_params.length();i++)
            m += up_units_params[i].size();
    return m;
}

Here is the call graph for this function:

Here is the caller graph for this function:

void PLearn::RBMLQParameters::update ( ) [virtual]

Member Data Documentation

Reimplemented from PLearn::RBMParameters.

Definition at line 167 of file RBMLQParameters.h.

Element i contains the bias of linear down unit i.

Definition at line 69 of file RBMLQParameters.h.

Referenced by build_(), declareOptions(), forget(), makeDeepCopyFromShallowCopy(), makeParametersPointHere(), nParameters(), and update().

Accumulates negative contribution to the gradient of down_units_params.

Definition at line 86 of file RBMLQParameters.h.

Referenced by accumulateNegStats(), build_(), clearStats(), makeDeepCopyFromShallowCopy(), and update().

Accumulates positive contribution to the gradient of down_units_bias.

Definition at line 84 of file RBMLQParameters.h.

Referenced by accumulatePosStats(), build_(), clearStats(), makeDeepCopyFromShallowCopy(), and update().

Element 1,i contains the bias of up unit i, element (2,i) contains the.

Definition at line 73 of file RBMLQParameters.h.

Referenced by accumulateNegStats(), accumulatePosStats(), bpropUpdate(), build_(), declareOptions(), forget(), makeDeepCopyFromShallowCopy(), makeParametersPointHere(), nParameters(), and update().

Accumulates negative contribution to the gradient of down_units_params.

Definition at line 90 of file RBMLQParameters.h.

Referenced by accumulateNegStats(), build_(), clearStats(), makeDeepCopyFromShallowCopy(), and update().

Accumulates positive contribution to the gradient of down_units_params.

Definition at line 88 of file RBMLQParameters.h.

Referenced by accumulatePosStats(), build_(), clearStats(), makeDeepCopyFromShallowCopy(), and update().

Matrix containing unit-to-unit weights (output_size × input_size)

Definition at line 66 of file RBMLQParameters.h.

Referenced by bpropUpdate(), build_(), declareOptions(), forget(), makeDeepCopyFromShallowCopy(), makeParametersPointHere(), nParameters(), and update().

Accumulates negative contribution to the weights' gradient.

Definition at line 81 of file RBMLQParameters.h.

Referenced by accumulateNegStats(), build_(), clearStats(), makeDeepCopyFromShallowCopy(), and update().

Accumulates positive contribution to the weights' gradient.

Definition at line 78 of file RBMLQParameters.h.

Referenced by accumulatePosStats(), build_(), clearStats(), makeDeepCopyFromShallowCopy(), and update().


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