annie Namespace Reference

Command-line (and possibly config file) parameter handling. More...

Classes

class AbstractNeuron

Implementation helper for common types of Neurons. More...

class ArgParser

class CenterNeuron

A center-neuron, the building block of a Radial basis function network. More...

class DynamicKohonenParameters

class EucledianTopology

Standard eucledian topology OPT: zounds. More...

class Exception

A common exception class used by all classes in the annie library. More...

class ExKohonenDraw

class File

The idea is that this class will be used to read in information from text files created by the "save" function in various annie classes. More...

class GABNeuron

Neuron for Gain Adaptive BP. More...

class HopfieldBase

A Hopfield network - content addressable memory and energy. More...

struct HopfieldBase::NData

class HopfieldNetwork

A "standard" incarnation of the Hopfield network usable form most purposes. More...

class Image

represents one image (actually a SDL_Context wrapper ..) More...

struct Image::Color

the color components should be in [0, 1] More...

class ImageCompressor

(usually lossy) image compressor. More...

class ImageCompressor::PartialListener

class InputLayer

A layer of input neurons. More...

class InputNeuron

Input neurons are slightly special. More...

class KohonenAnalyzer

struct KohonenAnalyzer::Outs

class KohonenDraw

Draws examples and progress of the Kohonen clustering using OpenGL --> Should be called from the redraw thread. More...

class KohonenNetwork

Todo: getOutput (&& getWin*) doesn't give anything sensible in this impl. More...

class KohonenNeuron

class KohonenParameters

Note: it is closely tied to Topology - the meaning NeighborhoodSize depends on the Topology used. More...

class Layer

Abstraction for a "layer" of neurons, i.e., a group of neurons not connected to each other. More...

class Link

Abstraction of a connection between two neurons. More...

class Matrix

A class for 2-dimensional matrices. More...

class MLPCompressor

yeah, yeah - ValueUpdateListener should have been inner class More...

class MLPKohonenAnalyzer

Analyzes MLP using Kohonen. More...

class MultiLayerNetwork

Abstraction of a multi-layer perceptron network. More...

struct MultiLayerNetwork::Error

Error summator will be moved up for general use.. More...

class Network

Another core class of the annie library, a generic template for a neural network. More...

class Neuron

One of the fundamental annie classes - the basic Neuron. More...

struct NumberParameter

class ParametrizedNeuron

Neuron with a parametrized activation function. More...

class RadialBasisNetwork

A Radial Basis Function Neural Network. More...

class RecurrentNetwork

RN constructed from a genotype. More...

class RecurrentNeuron

A neuron used for recurrent networks. More...

struct Redrawer

struct RigidVisualiser

Simple text-mode visualiser, which prints updated value after each change (synchronous). More...

struct Selector

Select features from the TS vectors. More...

struct Shrinker

struct Shuffler

Shuffles the given training set every nth epoch. More...

class SimpleNeuron

A simple perceptron - i.e., it takes as input the weighted sum of the outputs of the neurons connected to it. More...

struct SimpleVisualiser

Simple text-mode visualiser, which prints updated value after each change (synchronous), it it didn't occur too soon. More...

class StandardKohonenParameters

learningParam = e ^ (- slope * ( ln (time) ^ 2 ) ) neighborHoodSize = (MAX_NB_SIZE) * learningParam More...

class StaticKohonenParameters

Static kohonen parameters - don't change unless changed by the user. More...

struct Stepper

Waits for keypress on each "epoch" if "stepped" is true. More...

struct StringParameter

also holds the destination More...

class TerminatedException

class TileCompressor

Compress image by dividing it to squares. More...

class TLayer

nice getters... More...

class Topology

Class used as a template to the KohonenNetwork network. More...

class TrainingSet

This is an abstraction for the set of patterns which are used to "train" a network. More...

struct TSTransformer

2 vectors -> 2 vectors More...

class TwoLayerNetwork

Two layered networks are very commonly used. More...

class Vector

class Video

Video (and input) system maintaince. More...

class Video::Redrawer

User redraw class. draw() will be called from the video thread. More...

Typedefs

typedef double real

Use this instead of double/float for real numbers pertaining to annie.

typedef real(* TimeSeriesCreatingFunction )(unsigned time)

Function that can be sampled in discrete time steps.

typedef std::vector< real > SamplesContainer

Store for samples.

typedef real(* SamplingFunction )()

generates one sample at each call see...

typedef real(* ActivationFunction )(real)

These functions are used as activation functions of the neurons, i.e., the function applied to the raw output of the neurons.

typedef std::vector< Link * > LINKS

A set of Links, ie, weighted connections between two neurons.

typedef real(* ParametrizedActivationFunction )(real, real parameter)

typedef TLayer< RecurrentNeuron > RecurrentLayer

implementation detail, don't bother :)

typedef Vector(* XformFunction )(const Vector &in)

transforms one vector to another.

typedef std::vector< annie::Image * > Images

container for images

typedef std::vector< real > VFather

A vector of real numbers.

Functions

void parseArgs (int argc, char *argv[], const NumberParameter *numeric, StringParameter *strin)

shortcut - you don't have to use the ArgParser yourself

char getLastKeyPressed ()

char waitForKey ()

void initControlThread (uint winW=600, uint winH=600)

also initializes video

void goodbyeControlThread ()

nicely terminates CT Must be called even if you don't want to terminate the thread prematurely

void changeWindowSize (uint winW, uint winH, bool wait=true)

change the screen size

void forceRedraw (bool wait=true)

redraw the screen

Video * getVideo ()

void assert_failed (const char *cond, const char *function, const char *file, int line)

void sampleFunction (TimeSeriesCreatingFunction f, unsigned from, unsigned to, SamplesContainer &out)

Sample the given function.

TrainingSet * makeExamples (unsigned inputs, unsigned outputs, const SamplesContainer &samples, int from=0, int to=-1)

TrainingSet * randomSamples (uint samples, uint dim, SamplingFunction *samplingFunctions)

Produce set of points randomly drawn from the N-d cube.

TrainingSet * randomSamples (uint samples, uint dim, SamplingFunction samplingFunction)

sugar for randomSamples - apply the same sampling function to all dimensions

TrainingSet * transformInputs (TrainingSet &source, XformFunction f, uint res_ins) throw ()

Transform the input component of training examples by given function.

TrainingSet * transformInputs (TrainingSet &source, XformFunction f) throw ()

sugar

Vector toPolar (const Vector &in)

map input to sphere [0] = r [1] = angle ( <- [0, 1], not radians!!!)

Vector normalize (const Vector &in)

Vector vectorIdentity (const Vector &in)

template<int T_multiplier1000> Vector Xscale (const Vector &in)

Heck, float cannot be used as a template parameter...

template<int T_add1000> Vector Xadd (const Vector &in)

TrainingSet * uniformCube (uint samples, uint dim)

some shortcuts these are good examples of samplers usage

TrainingSet * uniformSphere (uint samples, uint dim)

each spike is equally covered => center is denser

real isPositive (real x)

Similar to the signum function, but used for non-bipolar Hopfield networks.

std::ostream & operator<< (std::ostream &o, const Image &i)

std::ostream & operator<< (std::ostream &o, const Image::Color &c)

void runKohonenAnalyzer (KohonenAnalyzer &ka)

real identity (real x)

The identity function, ie, f(x) = x.

real didentity (real x)

Derivative of the identity function, thus always returns 1.

real sigmoid (real x)

Sigmoidal activation f(x).

real dsigmoid (real x)

Derivative of the sigmoidal function.

real gaussian (real x)

The guassian function. Sigma is taken to be 1.0.

real dgaussian (real x)

Derivative of the gaussian function (sigma is taken to be 1.0).

real signum (real x)

Signum function Returns real instead of int because this function can be used as an activation function too, so wanted to keep the signature the same as other activation functions.

real tansig (real x)

The tansig activation function. Courtesy Jose Ramos (warta@users.sourceforge.net).

real dtansig (real x)

Derivative of the tansign activation function. Courtesy Jose Ramos (warta@users.sourceforge.net).

real psigmoid (real x, real lambda)

Sigmoidal activation with slope modifier lambda f(x) = 1/(1+e^(lambda*x)).

real dpsigmoid (real x, real lambda)

Derivative of the sigmoidal function.

real uniformRandom (real min, real max)

returns a number drawn uniformly from min, max

Vector uniformRandomVector (real min, real max, uint size)

real random ()

Generates a random real number between -1.0 and 1.0.

real random01 ()

Generates a random real number between 0.0 and 1.0.

int randomInt (int low, int high)

Generates a random integer between given bounds.

void sysInit ()

system dependent.

Images makeSubimages (const annie::Image &i, unsigned xSize, unsigned ySize)

TrainingSet * makeSubImageExamples (const Images &outImages, std::vector< Vector > &outExamples)

make vector from each tile.

void runCompressor (ImageCompressor &ic, const char *image, bool graay=false)

generic ImageCompressor utilization uses ["input"] as the input file

std::ostream & operator<< (std::ostream &os, const annie::Vector &v)

bool operator== (const Vector &o1, const Vector &o2)

Variables

const real REAL_MAX = std::numeric_limits<double>::max() REAL_MIN=std::numeric_limits<double>::min()

const int BINARY_FILE = 0

The TrainingSet can be saved as a binary file or a text file, the latter allowing users to create a training set without using annie.

const int TEXT_FILE = 1

const real SIGMOID_APROX_THRESHOLD = 30

SDL_PixelFormat defaultRGBformat

the only format of surface we create

Detailed Description

Command-line (and possibly config file) parameter handling.

Author:

Id: args.h,v 1.1 2004/06/16 10:53:30 opx Exp

Typedef Documentation

typedef real(* ActivationFunction)(real)

These functions are used as activation functions of the neurons, i.e., the function applied to the raw output of the neurons.

See also:
identity
sigmoid
gaussian

typedef std::vector<annie::Image *> Images

container for images

typedef std::vector<Link *> LINKS

A set of Links, ie, weighted connections between two neurons.

typedef real(* ParametrizedActivationFunction)(real, real parameter)

typedef double real

Use this instead of double/float for real numbers pertaining to annie.

typedef TLayer<RecurrentNeuron> RecurrentLayer

implementation detail, don't bother :)

typedef std::vector< real > SamplesContainer

Store for samples.

typedef real(* SamplingFunction)()

generates one sample at each call see...
we have random01 :)

typedef real(* TimeSeriesCreatingFunction)(unsigned time)

Function that can be sampled in discrete time steps.

typedef std::vector<real> VFather

A vector of real numbers.

typedef Vector(* XformFunction)(const Vector &in)

transforms one vector to another.
Doesn't have to come up with the same size. look in annie/examples.h for common xform functions

Function Documentation

void assert_failed ( const char * cond,

const char * function,

const char * file,

int line

)

void changeWindowSize ( uint winW,

uint winH,

bool wait = true

)

change the screen size

real dgaussian ( real x )

Derivative of the gaussian function (sigma is taken to be 1.0).

real didentity ( real x )

Derivative of the identity function, thus always returns 1.

real dpsigmoid ( real x,

real lambda

)

Derivative of the sigmoidal function.

real dsigmoid ( real x )

Derivative of the sigmoidal function.

real dtansig ( real x )

Derivative of the tansign activation function. Courtesy Jose Ramos (warta@users.sourceforge.net).

void forceRedraw ( bool wait = true )

redraw the screen

real gaussian ( real x )

The guassian function. Sigma is taken to be 1.0.

char getLastKeyPressed ( )

Video* getVideo ( )

void goodbyeControlThread ( )

nicely terminates CT Must be called even if you don't want to terminate the thread prematurely

real identity ( real x )

The identity function, ie, f(x) = x.

void initControlThread ( uint winW = 600,

uint winH = 600

)

also initializes video

real isPositive ( real x )

Similar to the signum function, but used for non-bipolar Hopfield networks.

Parameters:

x A real value

Returns:
0 if x<=0, 1 otherwise.

TrainingSet * makeExamples ( unsigned inputs,

unsigned outputs,

const SamplesContainer & samples,

int from = 0,

int to = -1

)

TrainingSet* makeSubImageExamples ( const Images & outImages,

std::vector< Vector > & outExamples

)

make vector from each tile.
Only the 1st byte of each color component is taken into account (to reduce dimensionalyty for MLP) retuens the outExamples and the same stored as identity in TrainingSet

Images makeSubimages ( const annie::Image & i,

unsigned xSize,

unsigned ySize

)

Vector normalize ( const Vector & in )

std::ostream& operator<< ( std::ostream & os,

const annie::Vector & v

)

std::ostream& operator<< ( std::ostream & o,

const Image::Color & c

) [inline]

std::ostream& operator<< ( std::ostream & o,

const Image & i

) [inline]

Here is the call graph for this function:

bool operator== ( const Vector & o1,

const Vector & o2

)

void parseArgs ( int argc,

char * argv[],

const NumberParameter * numeric,

StringParameter * strin

)

shortcut - you don't have to use the ArgParser yourself

real psigmoid ( real x,

real lambda

)

Sigmoidal activation with slope modifier lambda f(x) = 1/(1+e^(lambda*x)).

real random ( )

Generates a random real number between -1.0 and 1.0.

real random01 ( )

Generates a random real number between 0.0 and 1.0.

int randomInt ( int low,

int high

)

Generates a random integer between given bounds.

Parameters:

low Lower bound

high Upper bound (exclusively)

Returns:
A random integer i, low<=i<high

TrainingSet* randomSamples ( uint samples,

uint dim,

SamplingFunction samplingFunction

)

sugar for randomSamples - apply the same sampling function to all dimensions

TrainingSet* randomSamples ( uint samples,

uint dim,

SamplingFunction * samplingFunctions

)

Produce set of points randomly drawn from the N-d cube.
Each I-O pair will only have the "I" component - dimension of output will be 0. Each dimension is supposed to be a [0, 1] interval This function is useful for Kohonen Maps, for example

Parameters:

samples - how many samples do you want?

samplingFunctions - array [dim] of sample generators for each dimension

Returns:
a newly created training set. Caller is responsible for deletion!!

void runCompressor ( ImageCompressor & ic,

const char * image,

bool graay = false

)

generic ImageCompressor utilization uses ["input"] as the input file

Parameters:

gray whether to gray the input image before processing

void runKohonenAnalyzer ( KohonenAnalyzer & ka )

void sampleFunction ( TimeSeriesCreatingFunction f,

unsigned from,

unsigned to,

SamplesContainer & out

)

Sample the given function.

real sigmoid ( real x )

Sigmoidal activation f(x).

real signum ( real x )

Signum function Returns real instead of int because this function can be used as an activation function too, so wanted to keep the signature the same as other activation functions.

Parameters:

x Real value as input

Returns:
-1 if x<0, +1 if x>=0

void sysInit ( )

system dependent.
May install some goodies like segfault handler, so better call it prior to anything else

real tansig ( real x )

The tansig activation function. Courtesy Jose Ramos (warta@users.sourceforge.net).

Vector toPolar ( const Vector & in )

map input to sphere [0] = r [1] = angle ( <- [0, 1], not radians!!!)
actually, the output is scaled to [0, 1] ( .5 + .5 * )

TrainingSet* transformInputs ( TrainingSet & source,

XformFunction f

) throw ()

sugar

TrainingSet* transformInputs ( TrainingSet & source,

XformFunction f,

uint res_ins

) throw ()

Transform the input component of training examples by given function.

TrainingSet* uniformCube ( uint samples,

uint dim

)

some shortcuts these are good examples of samplers usage

real uniformRandom ( real min,

real max

)

returns a number drawn uniformly from min, max

Vector uniformRandomVector ( real min,

real max,

uint size

) [inline]

Here is the call graph for this function:

TrainingSet* uniformSphere ( uint samples,

uint dim

)

each spike is equally covered => center is denser

Vector vectorIdentity ( const Vector & in )

char waitForKey ( )

Vector Xadd ( const Vector & in ) [static]

Vector Xscale ( const Vector & in ) [static]

Heck, float cannot be used as a template parameter...
probably to decline the piggy things like this one we should use xforming class instead of xforming function (like trhat for mixed i-o's)

Variable Documentation

const int BINARY_FILE = 0

The TrainingSet can be saved as a binary file or a text file, the latter allowing users to create a training set without using annie.
A binary file is referred to as annie::BINARY_FILE and a text file as annie::TEXT_FILE
See also:
TrainingSet
TrainingSet::save
TEXT_FILE

SDL_PixelFormat defaultRGBformat

the only format of surface we create

const real REAL_MAX = std::numeric_limits<double>::max() REAL_MIN=std::numeric_limits<double>::min()

const real SIGMOID_APROX_THRESHOLD = 30

const int TEXT_FILE = 1

See also:
BINARY_FILE

Generated on Fri Jun 18 13:18:50 2004 for Annie by

1.3.5


Classes
class	AbstractNeuron
	Implementation helper for common types of Neurons. More...
class	ArgParser
class	CenterNeuron
	A center-neuron, the building block of a Radial basis function network. More...
class	DynamicKohonenParameters
class	EucledianTopology
	Standard eucledian topology OPT: zounds. More...
class	Exception
	A common exception class used by all classes in the annie library. More...
class	ExKohonenDraw
class	File
	The idea is that this class will be used to read in information from text files created by the "save" function in various annie classes. More...
class	GABNeuron
	Neuron for Gain Adaptive BP. More...
class	HopfieldBase
	A Hopfield network - content addressable memory and energy. More...
struct	HopfieldBase::NData
class	HopfieldNetwork
	A "standard" incarnation of the Hopfield network usable form most purposes. More...
class	Image
	represents one image (actually a SDL_Context wrapper ..) More...
struct	Image::Color
	the color components should be in [0, 1] More...
class	ImageCompressor
	(usually lossy) image compressor. More...
class	ImageCompressor::PartialListener
class	InputLayer
	A layer of input neurons. More...
class	InputNeuron
	Input neurons are slightly special. More...
class	KohonenAnalyzer
struct	KohonenAnalyzer::Outs
class	KohonenDraw
	Draws examples and progress of the Kohonen clustering using OpenGL --> Should be called from the redraw thread. More...
class	KohonenNetwork
	Todo: getOutput (&& getWin*) doesn't give anything sensible in this impl. More...
class	KohonenNeuron
class	KohonenParameters
	Note: it is closely tied to Topology - the meaning NeighborhoodSize depends on the Topology used. More...
class	Layer
	Abstraction for a "layer" of neurons, i.e., a group of neurons not connected to each other. More...
class	Link
	Abstraction of a connection between two neurons. More...
class	Matrix
	A class for 2-dimensional matrices. More...
class	MLPCompressor
	yeah, yeah - ValueUpdateListener should have been inner class More...
class	MLPKohonenAnalyzer
	Analyzes MLP using Kohonen. More...
class	MultiLayerNetwork
	Abstraction of a multi-layer perceptron network. More...
struct	MultiLayerNetwork::Error
	Error summator will be moved up for general use.. More...
class	Network
	Another core class of the annie library, a generic template for a neural network. More...
class	Neuron
	One of the fundamental annie classes - the basic Neuron. More...
struct	NumberParameter
class	ParametrizedNeuron
	Neuron with a parametrized activation function. More...
class	RadialBasisNetwork
	A Radial Basis Function Neural Network. More...
class	RecurrentNetwork
	RN constructed from a genotype. More...
class	RecurrentNeuron
	A neuron used for recurrent networks. More...
struct	Redrawer
struct	RigidVisualiser
	Simple text-mode visualiser, which prints updated value after each change (synchronous). More...
struct	Selector
	Select features from the TS vectors. More...
struct	Shrinker
struct	Shuffler
	Shuffles the given training set every nth epoch. More...
class	SimpleNeuron
	A simple perceptron - i.e., it takes as input the weighted sum of the outputs of the neurons connected to it. More...
struct	SimpleVisualiser
	Simple text-mode visualiser, which prints updated value after each change (synchronous), it it didn't occur too soon. More...
class	StandardKohonenParameters
	learningParam = e ^ (- slope * ( ln (time) ^ 2 ) ) neighborHoodSize = (MAX_NB_SIZE) * learningParam More...
class	StaticKohonenParameters
	Static kohonen parameters - don't change unless changed by the user. More...
struct	Stepper
	Waits for keypress on each "epoch" if "stepped" is true. More...
struct	StringParameter
	also holds the destination More...
class	TerminatedException
class	TileCompressor
	Compress image by dividing it to squares. More...
class	TLayer
	nice getters... More...
class	Topology
	Class used as a template to the KohonenNetwork network. More...
class	TrainingSet
	This is an abstraction for the set of patterns which are used to "train" a network. More...
struct	TSTransformer
	2 vectors -> 2 vectors More...
class	TwoLayerNetwork
	Two layered networks are very commonly used. More...
class	Vector
class	Video
	Video (and input) system maintaince. More...
class	Video::Redrawer
	User redraw class. draw() will be called from the video thread. More...
Typedefs
typedef double	real
	Use this instead of double/float for real numbers pertaining to annie.
typedef real(*	TimeSeriesCreatingFunction )(unsigned time)
	Function that can be sampled in discrete time steps.
typedef std::vector< real >	SamplesContainer
	Store for samples.
typedef real(*	SamplingFunction )()
	generates one sample at each call see...
typedef real(*	ActivationFunction )(real)
	These functions are used as activation functions of the neurons, i.e., the function applied to the raw output of the neurons.
typedef std::vector< Link * >	LINKS
	A set of Links, ie, weighted connections between two neurons.
typedef real(*	ParametrizedActivationFunction )(real, real parameter)
typedef TLayer< RecurrentNeuron >	RecurrentLayer
	implementation detail, don't bother :)
typedef Vector(*	XformFunction )(const Vector &in)
	transforms one vector to another.
typedef std::vector< annie::Image * >	Images
	container for images
typedef std::vector< real >	VFather
	A vector of real numbers.
Functions
void	parseArgs (int argc, char argv[], const NumberParameter numeric, StringParameter *strin)
	shortcut - you don't have to use the ArgParser yourself
char	getLastKeyPressed ()
char	waitForKey ()
void	initControlThread (uint winW=600, uint winH=600)
	also initializes video
void	goodbyeControlThread ()
	nicely terminates CT Must be called even if you don't want to terminate the thread prematurely
void	changeWindowSize (uint winW, uint winH, bool wait=true)
	change the screen size
void	forceRedraw (bool wait=true)
	redraw the screen
Video *	getVideo ()
void	assert_failed (const char cond, const char function, const char *file, int line)
void	sampleFunction (TimeSeriesCreatingFunction f, unsigned from, unsigned to, SamplesContainer &out)
	Sample the given function.
TrainingSet *	makeExamples (unsigned inputs, unsigned outputs, const SamplesContainer &samples, int from=0, int to=-1)
TrainingSet *	randomSamples (uint samples, uint dim, SamplingFunction *samplingFunctions)
	Produce set of points randomly drawn from the N-d cube.
TrainingSet *	randomSamples (uint samples, uint dim, SamplingFunction samplingFunction)
	sugar for randomSamples - apply the same sampling function to all dimensions
TrainingSet *	transformInputs (TrainingSet &source, XformFunction f, uint res_ins) throw ()
	Transform the input component of training examples by given function.
TrainingSet *	transformInputs (TrainingSet &source, XformFunction f) throw ()
	sugar
Vector	toPolar (const Vector &in)
	map input to sphere [0] = r [1] = angle ( <- [0, 1], not radians!!!)
Vector	normalize (const Vector &in)
Vector	vectorIdentity (const Vector &in)
template<int T_multiplier1000> Vector	Xscale (const Vector &in)
	Heck, float cannot be used as a template parameter...
template<int T_add1000> Vector	Xadd (const Vector &in)
TrainingSet *	uniformCube (uint samples, uint dim)
	some shortcuts these are good examples of samplers usage
TrainingSet *	uniformSphere (uint samples, uint dim)
	each spike is equally covered => center is denser
real	isPositive (real x)
	Similar to the signum function, but used for non-bipolar Hopfield networks.
std::ostream &	operator<< (std::ostream &o, const Image &i)
std::ostream &	operator<< (std::ostream &o, const Image::Color &c)
void	runKohonenAnalyzer (KohonenAnalyzer &ka)
real	identity (real x)
	The identity function, ie, f(x) = x.
real	didentity (real x)
	Derivative of the identity function, thus always returns 1.
real	sigmoid (real x)
	Sigmoidal activation f(x).
real	dsigmoid (real x)
	Derivative of the sigmoidal function.
real	gaussian (real x)
	The guassian function. Sigma is taken to be 1.0.
real	dgaussian (real x)
	Derivative of the gaussian function (sigma is taken to be 1.0).
real	signum (real x)
	Signum function Returns real instead of int because this function can be used as an activation function too, so wanted to keep the signature the same as other activation functions.
real	tansig (real x)
	The tansig activation function. Courtesy Jose Ramos (warta@users.sourceforge.net).
real	dtansig (real x)
	Derivative of the tansign activation function. Courtesy Jose Ramos (warta@users.sourceforge.net).
real	psigmoid (real x, real lambda)
	Sigmoidal activation with slope modifier lambda *f(x) = 1/(1+e^(lambdax))**.
real	dpsigmoid (real x, real lambda)
	Derivative of the sigmoidal function.
real	uniformRandom (real min, real max)
	returns a number drawn uniformly from min, max
Vector	uniformRandomVector (real min, real max, uint size)
real	random ()
	Generates a random real number between -1.0 and 1.0.
real	random01 ()
	Generates a random real number between 0.0 and 1.0.
int	randomInt (int low, int high)
	Generates a random integer between given bounds.
void	sysInit ()
	system dependent.
Images	makeSubimages (const annie::Image &i, unsigned xSize, unsigned ySize)
TrainingSet *	makeSubImageExamples (const Images &outImages, std::vector< Vector > &outExamples)
	make vector from each tile.
void	runCompressor (ImageCompressor &ic, const char *image, bool graay=false)
	generic ImageCompressor utilization uses ["input"] as the input file
std::ostream &	operator<< (std::ostream &os, const annie::Vector &v)
bool	operator== (const Vector &o1, const Vector &o2)
Variables
const real	REAL_MAX = std::numeric_limits<double>::max() REAL_MIN=std::numeric_limits<double>::min()
const int	BINARY_FILE = 0
	The TrainingSet can be saved as a binary file or a text file, the latter allowing users to create a training set without using annie.
const int	TEXT_FILE = 1
const real	SIGMOID_APROX_THRESHOLD = 30
SDL_PixelFormat	defaultRGBformat
	the only format of surface we create