org.joone.engine

Class ContextLayer

java.lang.Object

org.joone.engine.Layer

org.joone.engine.SimpleLayer

org.joone.engine.ContextLayer

All Implemented Interfaces:

java.io.Serializable, java.lang.Runnable, Learnable, LearnableLayer, NeuralLayer, Inspectable

public class ContextLayer
extends SimpleLayer

The context layer is similar to the linear layer except that it has an auto-recurrent connection between its output and input. MG: Added ability to save and restore the initial state. Also changed so that initial state is used when first pattern is requested. TODO: Context neurons are usually used in the input layer of Jordan networks together with non-recursive neurons. So there are two questions: - how can this layer be combined with an non-recursive layer? - is there already a mechanism how to feed the network output back in the neurons of this layer?

Author:

P.Marrone

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
protected Matrix	initialState The initial state

Fields inherited from class org.joone.engine.Layer

bias, gradientInps, gradientOuts, inps, inputPatternListeners, learnable, learning, m_batch, monitor, myLearner, outputPatternListeners, outs, running, step, STOP_FLAG

Constructor Summary

Constructors

Constructor and Description
ContextLayer()
ContextLayer(java.lang.String name)

Method Summary

Methods

Modifier and Type	Method and Description
void	backward(double[] pattern) Reverse transfer function of the component.
void	forward(double[] pattern) Transfer function to recall a result on a trained net
double	getBeta() Getter for property beta.
double	getDefaultState() Return the default state of a node in this layer, such as 0 for a tanh or 0.5 for a sigmoid layer
double	getDerivative(int i) Similar to the backward message and used by RTRL
Matrix	getInitialState() Get the initial state, also use this to change, in place, the initial state
double	getMaximumState() Return maximum value of a node in this layer
double	getMinimumState() Return minimum value of a node in this layer
double	getTimeConstant() Getter for property timeConstant.
java.util.Collection	Inspections() It doesn't make sense to return biases for this layer
void	resetInitialState() Reset the initial state of this context layer.
void	setBeta(double beta) Setter for property beta.
protected void	setDimensions() Use this to also allocate space for the initial state vector
void	setTimeConstant(double timeConstant) Setter for property timeConstant.

Methods inherited from class org.joone.engine.SimpleLayer

getLearningRate, getLrate, getMomentum, setLrate, setMomentum, setMonitor

Methods inherited from class org.joone.engine.Layer

addInputSynapse, addNoise, addOutputSynapse, adjustSizeToFwdPattern, adjustSizeToRevPattern, check, checkInputEnabled, checkInputs, checkOutputs, copyInto, finalize, fireFwdGet, fireFwdPut, fireRevGet, fireRevPut, fwdRun, getAllInputs, getAllOutputs, getBias, getDimension, getLastGradientInps, getLastGradientOuts, getLastInputs, getLastOutputs, getLayerName, getLearner, getMonitor, getRows, getThreadMonitor, hasStepCounter, init, initLearner, InspectableTitle, isInputLayer, isOutputLayer, isRunning, join, randomize, randomizeBias, randomizeWeights, removeAllInputs, removeAllOutputs, removeInputSynapse, removeListener, removeOutputSynapse, resetInputListeners, revRun, run, setAllInputs, setAllOutputs, setBias, setConnDimensions, setInputDimension, setInputSynapses, setLastInputs, setLastOutputs, setLayerName, setOutputDimension, setOutputSynapses, setRows, start, stop, sumBackInput, sumInput, toString

Methods inherited from class java.lang.Object

clone, equals, getClass, hashCode, notify, notifyAll, wait, wait, wait

Field Detail

initialState

protected Matrix initialState

The initial state

Constructor Detail

ContextLayer

public ContextLayer()

ContextLayer

public ContextLayer(java.lang.String name)

Method Detail

backward

public void backward(double[] pattern)

Description copied from class: Layer

Reverse transfer function of the component.

Overrides:

backward in class SimpleLayer

Parameters:

pattern - input pattern on which to apply the transfer function

getDerivative

public double getDerivative(int i)

Similar to the backward message and used by RTRL

Specified by:

getDerivative in class Layer

forward

public void forward(double[] pattern)

Description copied from class: Layer

Transfer function to recall a result on a trained net

Specified by:

forward in class Layer

Parameters:

pattern - input pattern to which to apply the rtransfer function

getBeta

public double getBeta()

Getter for property beta.

Returns:

Value of property beta.

setBeta

public void setBeta(double beta)

Setter for property beta.

Parameters:

beta - New value of property beta.

getTimeConstant

public double getTimeConstant()

Getter for property timeConstant.

Returns:

Value of property timeConstant.

setTimeConstant

public void setTimeConstant(double timeConstant)

Setter for property timeConstant.

Parameters:

timeConstant - New value of property timeConstant.

Inspections

public java.util.Collection Inspections()

It doesn't make sense to return biases for this layer

Specified by:

Inspections in interface Inspectable

Overrides:

Inspections in class Layer

Returns:

null

See Also:

org.joone.Inspection

setDimensions

protected void setDimensions()

Use this to also allocate space for the initial state vector

Overrides:

setDimensions in class SimpleLayer

getInitialState

public Matrix getInitialState()

Get the initial state, also use this to change, in place, the initial state

resetInitialState

public void resetInitialState()

Reset the initial state of this context layer. This also reset all attached synapses and afterwards fires the initial state into them.

getDefaultState

public double getDefaultState()

Description copied from class: Layer

Return the default state of a node in this layer, such as 0 for a tanh or 0.5 for a sigmoid layer

Specified by:

getDefaultState in class Layer

getMinimumState

public double getMinimumState()

Description copied from class: Layer

Return minimum value of a node in this layer

Specified by:

getMinimumState in class Layer

getMaximumState

public double getMaximumState()

Description copied from class: Layer

Return maximum value of a node in this layer

Specified by:

getMaximumState in class Layer