org.aika.neuron

Class Neuron

java.lang.Object

org.aika.neuron.Neuron

All Implemented Interfaces:

Comparable<Neuron>, Writable

Direct Known Subclasses:

InputNeuron

public class Neuron
extends Object
implements Comparable<Neuron>, Writable

The Neuron class represents a neuron in Aikas neural network and is connected to other neurons through input synapses and output synapses. The activation value of a neuron is calculated by computing the weighted sum (input act. value * synapse weight) of the input synapses, adding the bias to it and sending the resulting value through a transfer function (the upper part of tanh). The neuron does not store its activations by itself. The activation objects are stored within the logic nodes. To access the activations of this neuron simply use the member variable node or use the method getFinalActivations(Document doc) to ge the final activations of this neuron.

Author:

Lukas Molzberger

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	Neuron.NormWeight

Field Summary

Fields

Modifier and Type	Field and Description
double	activationSum
double	bias
static AtomicInteger	currentNeuronId
int	id
boolean	initialized
TreeSet<Synapse>	inputSynapses
TreeSet<Synapse>	inputSynapsesByWeight
boolean	isBlocked
String	label
static double	LEARN_RATE
ReadWriteLock	lock
Model	m
static int	MAX_SELF_REFERENCING_DEPTH
double	maxRecurrentSum
double	negDirSum
double	negRecSum
Node	node
boolean	noTraining
int	numberOfActivations
TreeMap<Synapse.Key,InputNode>	outputNodes
TreeSet<Synapse>	outputSynapses
double	posRecSum
static double	TOLERANCE
static double	WEIGHT_TOLERANCE

Constructor Summary

Constructors

Constructor and Description
Neuron()
Neuron(String label)
Neuron(String label, boolean isBlocked, boolean noTraining)

Method Summary

Modifier and Type	Method and Description
double	avgActivation()
int	compareTo(Neuron n)
void	computeBounds(Activation act)
void	computeErrorSignal(Document doc, Activation act)
Activation.State	computeWeight(int round, Activation act, SearchNode en, Document doc)
void	count(Document doc)
static Neuron	create(Document doc, Neuron n, double bias, double negDirSum, double negRecSum, double posRecSum, Set<Synapse> inputs)
Collection<Activation>	getFinalActivations(Document doc) getFinalActivations is a convenience method to retrieve all activations of the given neuron that are part of the final interpretation.
void	propagateAddedActivation(Document doc, Activation act)
void	propagateRemovedActivation(Document doc, Activation act)
void	publish(Document doc)
static Neuron	read(DataInput in, Document doc)
void	readFields(DataInput in, Document doc) Deserialize the fields of this object from in.
void	remove(Document doc)
static double	sigmoid(double x)
String	toString()
String	toStringWithSynapses()
void	train(Document doc, Activation act)
void	train(Document doc, Activation iAct, Integer rid, double x, long v)
static double	transferFunction(double x)
void	unpublish(Document doc)
void	write(DataOutput out) Serialize the fields of this object to out.

Methods inherited from class java.lang.Object

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

Field Detail

LEARN_RATE

public static final double LEARN_RATE

See Also:

Constant Field Values

WEIGHT_TOLERANCE

public static final double WEIGHT_TOLERANCE

See Also:

Constant Field Values

TOLERANCE

public static final double TOLERANCE

See Also:

Constant Field Values

MAX_SELF_REFERENCING_DEPTH

public static final int MAX_SELF_REFERENCING_DEPTH

See Also:

Constant Field Values

m

public Model m

currentNeuronId

public static AtomicInteger currentNeuronId

id

public int id

label

public String label

bias

public volatile double bias

negDirSum

public volatile double negDirSum

negRecSum

public volatile double negRecSum

posRecSum

public volatile double posRecSum

maxRecurrentSum

public volatile double maxRecurrentSum

outputSynapses

public TreeSet<Synapse> outputSynapses

inputSynapses

public TreeSet<Synapse> inputSynapses

inputSynapsesByWeight

public TreeSet<Synapse> inputSynapsesByWeight

outputNodes

public TreeMap<Synapse.Key,InputNode> outputNodes

node

public Node node

initialized

public boolean initialized

isBlocked

public boolean isBlocked

noTraining

public boolean noTraining

activationSum

public volatile double activationSum

numberOfActivations

public volatile int numberOfActivations

lock

public ReadWriteLock lock

Constructor Detail

Neuron

public Neuron()

Neuron

public Neuron(String label)

Neuron

public Neuron(String label,
              boolean isBlocked,
              boolean noTraining)

Method Detail

avgActivation

public double avgActivation()

create

public static Neuron create(Document doc,
                            Neuron n,
                            double bias,
                            double negDirSum,
                            double negRecSum,
                            double posRecSum,
                            Set<Synapse> inputs)

publish

public void publish(Document doc)

unpublish

public void unpublish(Document doc)

remove

public void remove(Document doc)

propagateAddedActivation

public void propagateAddedActivation(Document doc,
                                     Activation act)

propagateRemovedActivation

public void propagateRemovedActivation(Document doc,
                                       Activation act)

computeBounds

public void computeBounds(Activation act)

computeWeight

public Activation.State computeWeight(int round,
                                      Activation act,
                                      SearchNode en,
                                      Document doc)

computeErrorSignal

public void computeErrorSignal(Document doc,
                               Activation act)

train

public void train(Document doc,
                  Activation act)

train

public void train(Document doc,
                  Activation iAct,
                  Integer rid,
                  double x,
                  long v)

transferFunction

public static double transferFunction(double x)

sigmoid

public static double sigmoid(double x)

count

public void count(Document doc)

write

public void write(DataOutput out)
           throws IOException

Description copied from interface: Writable

Serialize the fields of this object to out.

Specified by:

write in interface Writable

Parameters:

out - DataOuput to serialize this object into.

Throws:

IOException

readFields

public void readFields(DataInput in,
                       Document doc)
                throws IOException

Description copied from interface: Writable

Deserialize the fields of this object from in.

For efficiency, implementations should attempt to re-use storage in the existing object where possible.

Specified by:

readFields in interface Writable

Parameters:

in - DataInput to deseriablize this object from.

Throws:

IOException

read

public static Neuron read(DataInput in,
                          Document doc)
                   throws IOException

Throws:

IOException

compareTo

public int compareTo(Neuron n)

Specified by:

compareTo in interface Comparable<Neuron>

toString

public String toString()

Overrides:

toString in class Object

toStringWithSynapses

public String toStringWithSynapses()

getFinalActivations

public Collection<Activation> getFinalActivations(Document doc)

getFinalActivations is a convenience method to retrieve all activations of the given neuron that are part of the final interpretation. Before calling this method, the doc.process() needs to be called first.

Parameters:

doc - The current document

Returns:

A collection with all final activations of this neuron.