PLearn::NeuralNet Class Reference

#include <NeuralNet.h>

Inheritance diagram for PLearn::NeuralNet:

[legend]Collaboration diagram for PLearn::NeuralNet:

[legend]List of all members.


Public Types
typedef Learner	inherited
Public Member Functions
	NeuralNet ()
virtual	~NeuralNet ()
	PLEARN_DECLARE_OBJECT (NeuralNet)
virtual void	build ()
	** SUBCLASS WRITING: ** This method should be redefined in subclasses, to just call inherited::build() and then build_()
virtual void	forget ()
void	initializeParams ()
virtual void	train (VMat training_set)
virtual void	use (const Vec &inputvec, Vec &prediction)
virtual int	costsize () const
	**** SUBCLASS WRITING: should be re-defined if user re-defines computeCost default version returns
virtual Array< string >	costNames () const
virtual Array< string >	testResultsNames ()
virtual void	useAndCost (const Vec &inputvec, const Vec &targetvec, Vec outputvec, Vec costvec)
	By default this function calls use(input, output) and then computeCost(input, target, output, cost) So you can overload computeCost to change cost computation.
void	computeCost (const Vec &inputvec, const Vec &targetvec, const Vec &outputvec, const Vec &costvec)
	computes the cost vec, given input, target and output The default version applies the declared CostFunc's on the (output,target) pair, putting the cost computed for each CostFunc in an element of the cost vector.
virtual void	makeDeepCopyFromShallowCopy (CopiesMap &copies)
Public Attributes
Func	f
Func	costf
Func	output_and_target_to_cost
int	nhidden
int	nhidden2
real	weight_decay
real	bias_decay
real	layer1_weight_decay
real	layer1_bias_decay
real	layer2_weight_decay
real	layer2_bias_decay
real	output_layer_weight_decay
real	output_layer_bias_decay
real	direct_in_to_out_weight_decay
bool	global_weight_decay
bool	direct_in_to_out
string	output_transfer_func
int	iseed
Array< string >	cost_funcs
	a list of cost functions to use in the form "[ cf1; cf2; cf3; ... ]"
real	semisupervised_flatten_factor
Vec	semisupervised_prior
PP< Optimizer >	optimizer
int	batch_size
int	nepochs
string	saveparams
Array< Vec >	normalization
Static Protected Member Functions
void	declareOptions (OptionList &ol)
	redefine this in subclasses: call declareOption(...) for each option, and then call inherited::declareOptions(options) ( see the declareOption function further down)
Protected Attributes
Var	input
Var	target
Var	costweights
Var	target_and_weights
Var	w1
Var	w2
Var	wout
Var	wdirect
Var	output
VarArray	costs
Var	cost
VarArray	params
Vec	paramsvalues
Vec	initial_paramsvalues
Private Member Functions
void	build_ ()

Member Typedef Documentation

typedef Learner PLearn::NeuralNet::inherited

Reimplemented from PLearn::Learner.
Definition at line 81 of file NeuralNet.h.

Constructor & Destructor Documentation

PLearn::NeuralNet::NeuralNet ( )

Definition at line 84 of file NeuralNet.cc.

PLearn::NeuralNet::~NeuralNet ( ) [virtual]

Definition at line 105 of file NeuralNet.cc.

Member Function Documentation

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

**** SUBCLASS WRITING: **** This method should be redefined in subclasses, to just call inherited::build() and then build_()

Reimplemented from PLearn::Learner.
Definition at line 200 of file NeuralNet.cc.
References build_().

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

**** SUBCLASS WRITING: **** The build_ and build methods should be redefined in subclasses build_ should do the actual building of the Learner according to build options (member variables) previously set. (These may have been set by hand, by a constructor, by the load method, or by setOption) As build() may be called several times (after changing options, to "rebuild" an object with different build options), make sure your implementation can handle this properly.
Reimplemented from PLearn::Learner.
Definition at line 206 of file NeuralNet.cc.
References PLearn::affine_transform(), PLearn::affine_transform_weight_penalty(), PLearn::TVec< Var >::append(), bias_decay, c_str(), PLearn::classification_loss(), cost, cost_funcs, costf, costs, costweights, PLearn::cross_entropy(), direct_in_to_out, direct_in_to_out_weight_decay, PLearn::exp(), f, PLearn::TVec< T >::fill(), PLearn::hconcat(), PLearn::ifThenElse(), initial_paramsvalues, initializeParams(), input, PLearn::Learner::inputsize(), PLearn::isMissing(), PLearn::TVec< Var >::isNull(), k, layer1_bias_decay, layer1_weight_decay, layer2_bias_decay, layer2_weight_decay, PLearn::TVec< T >::length(), PLearn::TVec< Vec >::length(), PLearn::log_softmax(), PLearn::VarArray::makeSharedValue(), MISSING_VALUE, PLearn::multiclass_loss(), PLearn::neg_log_pi(), PLearn::VarArray::nelems(), PLearn::newObject(), nhidden, nhidden2, normalization, PLearn::onehot_squared_loss(), output, output_and_target_to_cost, output_layer_bias_decay, output_layer_weight_decay, output_transfer_func, PLearn::Learner::outputsize(), params, paramsvalues, PLERROR, PLWARNING, PLearn::TVec< T >::resize(), PLearn::TVec< Var >::resize(), semisupervised_flatten_factor, semisupervised_prior, PLearn::sigmoid(), PLearn::TVec< T >::size(), PLearn::TVec< Var >::size(), PLearn::TVec< string >::size(), PLearn::softmax(), PLearn::softplus(), PLearn::sum(), PLearn::sumsquare(), PLearn::tanh(), target, target_and_weights, PLearn::Learner::targetsize(), PLearn::transposeProduct(), PLearn::var(), w1, w2, wdirect, weight_decay, PLearn::weighted_sumsquare(), PLearn::Learner::weightsize(), and wout.
Referenced by build().

void PLearn::NeuralNet::computeCost ( const Vec & inputvec,

const Vec & targetvec,

const Vec & outputvec,

const Vec & costvec

) [virtual]

computes the cost vec, given input, target and output The default version applies the declared CostFunc's on the (output,target) pair, putting the cost computed for each CostFunc in an element of the cost vector.
If you overload this method in subclasses (e.g. to compute a cost that depends on the internal elements of the model), you must also redefine costsize() and costNames() accordingly.
Reimplemented from PLearn::Learner.
Definition at line 494 of file NeuralNet.cc.
References output_and_target_to_cost.

Array< string > PLearn::NeuralNet::costNames ( ) const [virtual]

returns an Array of strings for the names of the components of the cost. Default version returns the info() strings of the cost functions in test_costfuncs
Reimplemented from PLearn::Learner.
Definition at line 410 of file NeuralNet.cc.
References cost_funcs.
Referenced by testResultsNames().

int PLearn::NeuralNet::costsize ( ) const [virtual]

**** SUBCLASS WRITING: should be re-defined if user re-defines computeCost default version returns

Reimplemented from PLearn::Learner.
Definition at line 415 of file NeuralNet.cc.
References cost.

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

redefine this in subclasses: call declareOption(...) for each option, and then call inherited::declareOptions(options) ( see the declareOption function further down)
ex: static void declareOptions(OptionList& ol) { declareOption(ol, "inputsize", &MyObject::inputsize_, OptionBase::buildoption, "the size of the input\n it must be provided"); declareOption(ol, "weights", &MyObject::weights, OptionBase::learntoption, "the learnt model weights"); inherited::declareOptions(ol); }
Reimplemented from PLearn::Learner.
Definition at line 109 of file NeuralNet.cc.
References PLearn::declareOption(), and PLearn::OptionList.

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

*** SUBCLASS WRITING: *** This method should be called AFTER or inside the build method, e.g. in order to re-initialize parameters. It should put the Learner in a 'fresh' state, not being influenced by any past call to train (everything learned is forgotten!).
Reimplemented from PLearn::Learner.
Definition at line 499 of file NeuralNet.cc.
References initial_paramsvalues, initializeParams(), and params.

void PLearn::NeuralNet::initializeParams ( )

Definition at line 441 of file NeuralNet.cc.
References direct_in_to_out, PLearn::fill_random_normal(), PLearn::Learner::inputsize(), iseed, PLearn::manual_seed(), nhidden, nhidden2, PLearn::seed(), w1, w2, wdirect, and wout.
Referenced by build_(), and forget().

void PLearn::NeuralNet::makeDeepCopyFromShallowCopy ( CopiesMap & copies ) [virtual]

Does the necessary operations to transform a shallow copy (this) into a deep copy by deep-copying all the members that need to be. Typical implementation:
void CLASS_OF_THIS::makeDeepCopyFromShallowCopy(CopiesMap& copies) { SUPERCLASS_OF_THIS::makeDeepCopyFromShallowCopy(copies); member_ptr = member_ptr->deepCopy(copies); member_smartptr = member_smartptr->deepCopy(copies); member_mat.makeDeepCopyFromShallowCopy(copies); member_vec.makeDeepCopyFromShallowCopy(copies); ... }
Reimplemented from PLearn::Learner.
Definition at line 508 of file NeuralNet.cc.
References PLearn::CopiesMap, PLearn::deepCopyField(), and optimizer.

PLearn::NeuralNet::PLEARN_DECLARE_OBJECT ( NeuralNet )

virtual Array<string> PLearn::NeuralNet::testResultsNames ( ) [inline, virtual]

Definition at line 144 of file NeuralNet.h.
References costNames().

void PLearn::NeuralNet::train ( VMat training_set ) [virtual]

*** SUBCLASS WRITING: *** Does the actual training. Subclasses must implement this method. The method should upon entry, call setTrainingSet(training_set); Make sure that a if(measure(step, objective_value)) is done after each training step, and that training is stopped if it returned true
Implements PLearn::Learner.
Definition at line 418 of file NeuralNet.cc.
References batch_size, cost, costf, input, PLearn::VMat::length(), PLearn::meanOf(), nepochs, optimizer, output_and_target_to_cost, params, paramsvalues, PLearn::save(), saveparams, and target_and_weights.

void PLearn::NeuralNet::use ( const Vec & inputvec,

Vec & prediction

) [virtual]

*** SUBCLASS WRITING: *** Uses a trained decider on input, filling output. If the cost should also be computed, then the user should call useAndCost instead of this method.
Implements PLearn::Learner.
Definition at line 484 of file NeuralNet.cc.
References f.

void PLearn::NeuralNet::useAndCost ( const Vec & inputvec,

const Vec & targetvec,

Vec outputvec,

Vec costvec

) [virtual]

By default this function calls use(input, output) and then computeCost(input, target, output, cost) So you can overload computeCost to change cost computation.

Reimplemented from PLearn::Learner.
Definition at line 489 of file NeuralNet.cc.
References costf.

Member Data Documentation

int PLearn::NeuralNet::batch_size

Definition at line 116 of file NeuralNet.h.
Referenced by train().

real PLearn::NeuralNet::bias_decay

Definition at line 91 of file NeuralNet.h.
Referenced by build_().

Var PLearn::NeuralNet::cost [protected]

Definition at line 67 of file NeuralNet.h.
Referenced by build_(), costsize(), and train().

Array<string> PLearn::NeuralNet::cost_funcs

a list of cost functions to use in the form "[ cf1; cf2; cf3; ... ]"

Definition at line 109 of file NeuralNet.h.
Referenced by build_(), and costNames().

Func PLearn::NeuralNet::costf

Definition at line 76 of file NeuralNet.h.
Referenced by build_(), train(), and useAndCost().

VarArray PLearn::NeuralNet::costs [protected]

Definition at line 66 of file NeuralNet.h.
Referenced by build_().

Var PLearn::NeuralNet::costweights [protected]

Definition at line 58 of file NeuralNet.h.
Referenced by build_().

bool PLearn::NeuralNet::direct_in_to_out

Definition at line 101 of file NeuralNet.h.
Referenced by build_(), and initializeParams().

real PLearn::NeuralNet::direct_in_to_out_weight_decay

Definition at line 98 of file NeuralNet.h.
Referenced by build_().

Func PLearn::NeuralNet::f

Definition at line 75 of file NeuralNet.h.
Referenced by build_(), and use().

bool PLearn::NeuralNet::global_weight_decay

Definition at line 100 of file NeuralNet.h.

Vec PLearn::NeuralNet::initial_paramsvalues [protected]

Definition at line 72 of file NeuralNet.h.
Referenced by build_(), and forget().

Var PLearn::NeuralNet::input [protected]

Definition at line 56 of file NeuralNet.h.
Referenced by build_(), and train().

int PLearn::NeuralNet::iseed

Definition at line 103 of file NeuralNet.h.
Referenced by initializeParams().

real PLearn::NeuralNet::layer1_bias_decay

Definition at line 93 of file NeuralNet.h.
Referenced by build_().

real PLearn::NeuralNet::layer1_weight_decay

Definition at line 92 of file NeuralNet.h.
Referenced by build_().

real PLearn::NeuralNet::layer2_bias_decay

Definition at line 95 of file NeuralNet.h.
Referenced by build_().

real PLearn::NeuralNet::layer2_weight_decay

Definition at line 94 of file NeuralNet.h.
Referenced by build_().

int PLearn::NeuralNet::nepochs

Definition at line 119 of file NeuralNet.h.
Referenced by train().

int PLearn::NeuralNet::nhidden

Definition at line 87 of file NeuralNet.h.
Referenced by build_(), and initializeParams().

int PLearn::NeuralNet::nhidden2

Definition at line 88 of file NeuralNet.h.
Referenced by build_(), and initializeParams().

Array<Vec> PLearn::NeuralNet::normalization

Definition at line 123 of file NeuralNet.h.
Referenced by build_().

PP<Optimizer> PLearn::NeuralNet::optimizer

Definition at line 114 of file NeuralNet.h.
Referenced by makeDeepCopyFromShallowCopy(), and train().

Var PLearn::NeuralNet::output [protected]

Definition at line 65 of file NeuralNet.h.
Referenced by build_().

Func PLearn::NeuralNet::output_and_target_to_cost

Definition at line 77 of file NeuralNet.h.
Referenced by build_(), computeCost(), and train().

real PLearn::NeuralNet::output_layer_bias_decay

Definition at line 97 of file NeuralNet.h.
Referenced by build_().

real PLearn::NeuralNet::output_layer_weight_decay

Definition at line 96 of file NeuralNet.h.
Referenced by build_().

string PLearn::NeuralNet::output_transfer_func

Definition at line 102 of file NeuralNet.h.
Referenced by build_().

VarArray PLearn::NeuralNet::params [protected]

Definition at line 69 of file NeuralNet.h.
Referenced by build_(), forget(), and train().

Vec PLearn::NeuralNet::paramsvalues [protected]

Definition at line 71 of file NeuralNet.h.
Referenced by build_(), and train().

string PLearn::NeuralNet::saveparams

Definition at line 121 of file NeuralNet.h.
Referenced by train().

real PLearn::NeuralNet::semisupervised_flatten_factor

Definition at line 110 of file NeuralNet.h.
Referenced by build_().

Vec PLearn::NeuralNet::semisupervised_prior

Definition at line 111 of file NeuralNet.h.
Referenced by build_().

Var PLearn::NeuralNet::target [protected]

Definition at line 57 of file NeuralNet.h.
Referenced by build_().

Var PLearn::NeuralNet::target_and_weights [protected]

Definition at line 59 of file NeuralNet.h.
Referenced by build_(), and train().

Var PLearn::NeuralNet::w1 [protected]

Definition at line 60 of file NeuralNet.h.
Referenced by build_(), and initializeParams().

Var PLearn::NeuralNet::w2 [protected]

Definition at line 61 of file NeuralNet.h.
Referenced by build_(), and initializeParams().

Var PLearn::NeuralNet::wdirect [protected]

Definition at line 63 of file NeuralNet.h.
Referenced by build_(), and initializeParams().

real PLearn::NeuralNet::weight_decay

Definition at line 90 of file NeuralNet.h.
Referenced by build_().

Var PLearn::NeuralNet::wout [protected]

Definition at line 62 of file NeuralNet.h.
Referenced by build_(), and initializeParams().

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

Generated on Tue Aug 17 16:27:27 2004 for PLearn by

1.3.7

PLearn::NeuralNet Class Reference

Public Types

Public Member Functions

Public Attributes

Static Protected Member Functions

Protected Attributes

Private Member Functions

Member Typedef Documentation

Constructor & Destructor Documentation

Member Function Documentation

Member Data Documentation