RecurrentNetwork Class Reference

RN constructed from a genotype. More...

#include <RecurrentNetwork.h>

Inheritance diagram for RecurrentNetwork:

[legend]Collaboration diagram for RecurrentNetwork:

[legend]List of all members.

Public Member Functions
	RecurrentNetwork (uint initialInputCount=0, uint initialRecurrenCount=0)
	If the counts are specified, default neurons are added.
void	addInput (real initialValue=0)
void	addInput (InputNeuron *in)
void	addNeuron (real bias=0, real activation=0)
	add a recurrent neuron
void	addNeuron (RecurrentNeuron *n)
virtual const char *	getClassName () const
	Returns "RecurrentNetwork".
virtual uint	getInputCount () const
	number of input "neurons"
virtual uint	getOutputCount () const
	number of recurrent neurons
virtual void	setInput (const Vector &input)
	Actualise the input "neuron's" vector.
RecurrentNeuron &	getNeuron (uint index)
	get a recurrent neuron
RecurrentNeuron &	operator[] (uint index)
InputNeuron &	getInputNeuron (uint index)
virtual void	resetActivations (const Vector &activations)
	Force recurrent neuron's activations (usually done at startup) TODO: activations || outputs ?
virtual void	step ()
	Steps the network in time (asynchronously).
virtual real	getNeuronOutput (uint neuron)
	get the output of particular neuron (usually, we're only interested in some neurons, not all)
virtual Vector	getOutput () const
	Just get the current state of recurrent neurons.
virtual Vector	getOutput (const Vector &input)
	Set input neurons, step and return all recurrent neuron's activations.
virtual void	save (const std::string &filename)
	NOT IMPLEMENTED by default.
Static Public Attributes
const int	INPUT_NEURONS_LABEL_OFFSET
Protected Attributes
InputLayer	_inputs
RecurrentLayer	_neurons
uint	_time

---

Detailed Description

RN constructed from a genotype.

Outputs are read from some (or all) of the recurrent neurons

---

Constructor & Destructor Documentation

RecurrentNetwork (  uint  initialInputCount = 0, uint  initialRecurrenCount = 0 )

If the counts are specified, default neurons are added. You can add new neurons anytime you want - it's your newtork, anyway..

---

Member Function Documentation

void addInput (  InputNeuron *  in  )  [inline]

Here is the call graph for this function:

void addInput (  real  initialValue = 0  )

void addNeuron (  RecurrentNeuron *  n  )  [inline]

Here is the call graph for this function:

void addNeuron (  real  bias = 0, real  activation = 0 )

add a recurrent neuron

virtual const char* getClassName (   )  const [inline, virtual]

Returns "RecurrentNetwork". Implements Network.

virtual uint getInputCount (   )  const [inline, virtual]

number of input "neurons" Reimplemented from Network.

Here is the call graph for this function:

InputNeuron& getInputNeuron (  uint  index  )  [inline]

RecurrentNeuron& getNeuron (  uint  index  )  [inline]

get a recurrent neuron

virtual real getNeuronOutput (  uint  neuron  )  [inline, virtual]

get the output of particular neuron (usually, we're only interested in some neurons, not all)

Here is the call graph for this function:

virtual Vector getOutput (  const Vector &  input  )  [virtual]

Set input neurons, step and return all recurrent neuron's activations. Implements Network.

virtual Vector getOutput (   )  const [virtual]

Just get the current state of recurrent neurons.

virtual uint getOutputCount (   )  const [inline, virtual]

number of recurrent neurons Reimplemented from Network.

Here is the call graph for this function:

RecurrentNeuron& operator[] (  uint  index  )  [inline]

Here is the call graph for this function:

virtual void resetActivations (  const Vector &  activations  )  [virtual]

Force recurrent neuron's activations (usually done at startup) TODO: activations || outputs ?

virtual void save (  const std::string &  filename  )  [inline, virtual]

NOT IMPLEMENTED by default. Implements Network.

virtual void setInput (  const Vector &  input  )  [virtual]

Actualise the input "neuron's" vector. This is usually called before each step(), but it's not neccesary.

virtual void step (   )  [virtual]

Steps the network in time (asynchronously).

---

Member Data Documentation

InputLayer _inputs [protected]

RecurrentLayer _neurons [protected]

uint _time [protected]

const int INPUT_NEURONS_LABEL_OFFSET [static]

---

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

• RecurrentNetwork.h

---