org.simbrain.network.neuron_update_rules

Class LinearRule

java.lang.Object

org.simbrain.network.core.NeuronUpdateRule

org.simbrain.network.neuron_update_rules.LinearRule

All Implemented Interfaces:

BiasedUpdateRule, BoundedUpdateRule, ClippableUpdateRule, DifferentiableUpdateRule, NoisyUpdateRule

public class LinearRule
extends NeuronUpdateRule
implements BiasedUpdateRule, DifferentiableUpdateRule, BoundedUpdateRule, ClippableUpdateRule, NoisyUpdateRule

LinearNeuron is a standard linear neuron.

Nested Class Summary

Nested classes/interfaces inherited from class org.simbrain.network.core.NeuronUpdateRule

NeuronUpdateRule.InputType

Field Summary

Fields inherited from class org.simbrain.network.core.NeuronUpdateRule

DEFAULT_INCREMENT, increment, inputType

Constructor Summary

Constructors

Constructor and Description
LinearRule()

Method Summary

Modifier and Type	Method and Description
double	clip(double val) Clip the current activation.
void	contextualDecrement(Neuron n) Decrement a neuron by increment, respecting neuron specific constraints.
void	contextualIncrement(Neuron n) Increment a neuron by increment, respecting neuron specific constraints.
LinearRule	deepCopy() Returns a deep copy of the update rule.
boolean	getAddNoise() Return true if add noise is turned on.
double	getBias()
double	getDerivative(double val) The inverse of the activation function.
java.lang.String	getDescription() Returns a brief description of this update rule.
double	getLowerBound() Returns the lower bound.
Randomizer	getNoiseGenerator() Return the noise generator.
double	getSlope()
Network.TimeType	getTimeType() Returns the type of time update (discrete or continuous) associated with this neuron.
double	getUpperBound() Returns the upper bound.
boolean	isClipped() Is clipping turned on or not.
void	setAddNoise(boolean addNoise) Set whether noise should be used.
void	setBias(double bias)
void	setClipped(boolean clipping) Turn clipping on and off.
void	setLowerBound(double lowerBound) Sets the lower bound of this neuron update rule's activation.
void	setNoiseGenerator(Randomizer noise) Set the noise generator.
void	setSlope(double slope)
void	setUpperBound(double upperBound) Sets the upper bound of this neuron update rule's activation.
void	update(Neuron neuron) Apply the update rule.

Methods inherited from class org.simbrain.network.core.NeuronUpdateRule

clear, decrementActivation, getDefaultIncrement, getGraphicalLowerBound, getGraphicalUpperBound, getIncrement, getInputType, getRandomValue, getToolTipText, incrementActivation, isSkipsSynapticUpdates, isSpikingNeuron, setIncrement, setInputType

Methods inherited from class java.lang.Object

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

Constructor Detail

LinearRule

public LinearRule()

Method Detail

getTimeType

public Network.TimeType getTimeType()

Returns the type of time update (discrete or continuous) associated with this neuron.

Specified by:

getTimeType in class NeuronUpdateRule

Returns:

the time type

deepCopy

public LinearRule deepCopy()

Returns a deep copy of the update rule.

Specified by:

deepCopy in class NeuronUpdateRule

Returns:

Duplicated update rule

update

public void update(Neuron neuron)

Apply the update rule.

Specified by:

update in class NeuronUpdateRule

Parameters:

neuron - parent neuron

clip

public double clip(double val)

Clip the current activation.

Specified by:

clip in interface ClippableUpdateRule

Parameters:

val - the value to clip

Returns:

the clipped value

contextualIncrement

public void contextualIncrement(Neuron n)

Increment a neuron by increment, respecting neuron specific constraints. Intended to be overriden.

Overrides:

contextualIncrement in class NeuronUpdateRule

Parameters:

n - neuron to be incremented

contextualDecrement

public void contextualDecrement(Neuron n)

Decrement a neuron by increment, respecting neuron specific constraints. Intended to be overriden.

Overrides:

contextualDecrement in class NeuronUpdateRule

Parameters:

n - neuron

getDerivative

public double getDerivative(double val)

The inverse of the activation function.

Specified by:

getDerivative in interface DifferentiableUpdateRule

Parameters:

val - the value being sent through the neuron's derivative

Returns:

the derivative of the neuron's activation function with respect to val.

getBias

public double getBias()

Specified by:

getBias in interface BiasedUpdateRule

Returns:

Returns the bias.

setBias

public void setBias(double bias)

Specified by:

setBias in interface BiasedUpdateRule

Parameters:

bias - The bias to set.

getSlope

public double getSlope()

Returns:

Returns the slope.

setSlope

public void setSlope(double slope)

Parameters:

slope - The slope to set.

getNoiseGenerator

public Randomizer getNoiseGenerator()

Description copied from interface: NoisyUpdateRule

Return the noise generator.

Specified by:

getNoiseGenerator in interface NoisyUpdateRule

Returns:

Returns the noise generator.

setNoiseGenerator

public void setNoiseGenerator(Randomizer noise)

Description copied from interface: NoisyUpdateRule

Set the noise generator.

Specified by:

setNoiseGenerator in interface NoisyUpdateRule

Parameters:

noise - The noise generator to set.

getAddNoise

public boolean getAddNoise()

Description copied from interface: NoisyUpdateRule

Return true if add noise is turned on.

Specified by:

getAddNoise in interface NoisyUpdateRule

Returns:

Returns the addNoise.

setAddNoise

public void setAddNoise(boolean addNoise)

Description copied from interface: NoisyUpdateRule

Set whether noise should be used.

Specified by:

setAddNoise in interface NoisyUpdateRule

Parameters:

addNoise - The addNoise to set.

getDescription

public java.lang.String getDescription()

Description copied from class: NeuronUpdateRule

Returns a brief description of this update rule. Used in combo boxes in the GUI.

Specified by:

getDescription in class NeuronUpdateRule

Returns:

the description.

getUpperBound

public double getUpperBound()

Description copied from interface: BoundedUpdateRule

Returns the upper bound.

Specified by:

getUpperBound in interface BoundedUpdateRule

Returns:

the upper boud

getLowerBound

public double getLowerBound()

Description copied from interface: BoundedUpdateRule

Returns the lower bound.

Specified by:

getLowerBound in interface BoundedUpdateRule

Returns:

the lower bound

setUpperBound

public void setUpperBound(double upperBound)

Description copied from interface: BoundedUpdateRule

Sets the upper bound of this neuron update rule's activation.

Specified by:

setUpperBound in interface BoundedUpdateRule

Parameters:

upperBound - the upper bound

setLowerBound

public void setLowerBound(double lowerBound)

Description copied from interface: BoundedUpdateRule

Sets the lower bound of this neuron update rule's activation.

Specified by:

setLowerBound in interface BoundedUpdateRule

Parameters:

lowerBound - the lower bound

isClipped

public boolean isClipped()

Description copied from interface: ClippableUpdateRule

Is clipping turned on or not.

Specified by:

isClipped in interface ClippableUpdateRule

Returns:

true if clipping is on; false if it's turned off

setClipped

public void setClipped(boolean clipping)

Description copied from interface: ClippableUpdateRule

Turn clipping on and off.

Specified by:

setClipped in interface ClippableUpdateRule

Parameters:

clipping - true if clipping should be on; false otherwise