org.cognity.neural.layer

Interface Layer

All Known Subinterfaces:

TransferLayer

All Known Implementing Classes:

GenericTransferLayer, InputLayer

public interface Layer

A layer represents a set of neurons which share the same properties. Every layer stores two vectors: the input vector, which describes input for each neuron and the output vector which describes the output of each neuron.

Author:

Artur Markiewicz

Method Summary

Methods

Modifier and Type	Method and Description
void	activate() Performs calculations for this layer.
double	getInput(int neuron) Returns the value of net input of the specified neuron.
int	getNeuronCount() Returns the number of neurons in this layer.
double	getOutput(int neuron) Returns the value of the output of the specified neuron.
void	init(LayerInitializer initializer) Initializes this layer using the provided initializer.
void	stimulate(int neuron, double value) Stimulates specified neuron with given value.

Method Detail

activate

void activate()

Performs calculations for this layer. Calling this method updates inputs and outputs of neurons.

stimulate

void stimulate(int neuron,
             double value)

Stimulates specified neuron with given value. The exact effect of the stimulation is implementation dependent.

Parameters:

neuron - the index of the neuron

value - the stimulation value

getInput

double getInput(int neuron)

Returns the value of net input of the specified neuron.

Parameters:

neuron - an index of the neuron in this layer

Returns:

the value of the input of the specified neuron

getOutput

double getOutput(int neuron)

Returns the value of the output of the specified neuron.

Parameters:

neuron - an index of the neuron in this layer

Returns:

the value of the output of the specified neuron

getNeuronCount

int getNeuronCount()

Returns the number of neurons in this layer.

Returns:

the number of neurons in this layer.

init

void init(LayerInitializer initializer)

Initializes this layer using the provided initializer. This is a visitor-pattern method, the implementation should only call initializer.init(this).

Parameters:

initializer - the initializer