org.aika.neuron

Class Activation.Rounds

java.lang.Object

org.aika.neuron.Activation.Rounds

Enclosing class:

Activation

public static class Activation.Rounds
extends Object

Since Aika is a recurrent neural network, it is necessary to compute several rounds of activation values. The computation stops if no further changes occur to the state. Only the recurrent synapses depend on the previous round.

Field Summary

Fields

Modifier and Type	Field and Description
long	modified
TreeMap<Integer,Activation.State>	rounds

Constructor Summary

Constructors

Constructor and Description
Rounds()

Method Summary

Modifier and Type	Method and Description
boolean	compare(Activation.Rounds r)
Activation.Rounds	copy()
Activation.State	get(int r)
Activation.State	getLast()
Integer	getLastRound()
boolean	isQueued(int r)
void	reset()
boolean	set(int r, Activation.State s)
void	setQueued(int r, boolean v)
String	toString()

Methods inherited from class java.lang.Object

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

Field Detail

modified

public long modified

rounds

public TreeMap<Integer,Activation.State> rounds

Constructor Detail

Rounds

public Rounds()

Method Detail

set

public boolean set(int r,
                   Activation.State s)

get

public Activation.State get(int r)

copy

public Activation.Rounds copy()

getLastRound

public Integer getLastRound()

getLast

public Activation.State getLast()

setQueued

public void setQueued(int r,
                      boolean v)

isQueued

public boolean isQueued(int r)

reset

public void reset()

toString

public String toString()

Overrides:

toString in class Object

compare

public boolean compare(Activation.Rounds r)