de.learnlib.algorithms.adt.learner

Class ADTLearner<I,O>

java.lang.Object

de.learnlib.algorithms.adt.learner.ADTLearner<I,O>

Type Parameters:

I - input alphabet type

O - output alphabet type

All Implemented Interfaces:

PartialTransitionAnalyzer<ADTState<I,O>,I>, ResumableLearner<ADTLearnerState<ADTState<I,O>,I,O>>, SupportsGrowingAlphabet<I>, LearningAlgorithm<net.automatalib.automata.transout.MealyMachine<?,I,?,O>,I,net.automatalib.words.Word<O>>, LearningAlgorithm.MealyLearner<I,O>

@ParametersAreNonnullByDefault
public class ADTLearner<I,O>
extends Object
implements LearningAlgorithm.MealyLearner<I,O>, PartialTransitionAnalyzer<ADTState<I,O>,I>, SupportsGrowingAlphabet<I>, ResumableLearner<ADTLearnerState<ADTState<I,O>,I,O>>

The main learning algorithm.

Author:

frohme

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	ADTLearner.BuilderDefaults

Nested classes/interfaces inherited from interface de.learnlib.api.algorithm.LearningAlgorithm

LearningAlgorithm.DFALearner<I>, LearningAlgorithm.MealyLearner<I,O>

Nested classes/interfaces inherited from interface de.learnlib.algorithms.adt.api.PartialTransitionAnalyzer

PartialTransitionAnalyzer.HypothesisModificationException

Field Summary

Fields inherited from interface de.learnlib.algorithms.adt.api.PartialTransitionAnalyzer

HYPOTHESIS_MODIFICATION_EXCEPTION

Constructor Summary

Constructors

Constructor and Description
ADTLearner(net.automatalib.words.Alphabet<I> alphabet, SymbolQueryOracle<I,O> oracle, LeafSplitter leafSplitter, ADTExtender adtExtender, SubtreeReplacer subtreeReplacer)

Method Summary

Modifier and Type	Method and Description
void	addAlphabetSymbol(I symbol)
void	closeTransition(ADTState<I,O> state, I input) Determine the successor/output of the transition in question (which is usually achieved by sifting the corresponding long-prefix through the ADT).
net.automatalib.automata.transout.MealyMachine<?,I,?,O>	getHypothesisModel() Returns the current hypothesis model.
boolean	isTransitionDefined(ADTState<I,O> state, I input) Check whether the transition in question is defined or not.
boolean	refineHypothesis(DefaultQuery<I,net.automatalib.words.Word<O>> ce) Triggers a refinement of the model by providing a counterexample.
boolean	refineHypothesisInternal(DefaultQuery<I,net.automatalib.words.Word<O>> ceQuery)
void	resume(ADTLearnerState<ADTState<I,O>,I,O> state) Does not get the learner to continue learning.
void	startLearning() Starts the model inference process, creating an initial hypothesis in the provided model object.
ADTLearnerState<ADTState<I,O>,I,O>	suspend() Expose the serializable learner state object.

Methods inherited from class java.lang.Object

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

Constructor Detail

ADTLearner

public ADTLearner(net.automatalib.words.Alphabet<I> alphabet,
                  SymbolQueryOracle<I,O> oracle,
                  LeafSplitter leafSplitter,
                  ADTExtender adtExtender,
                  SubtreeReplacer subtreeReplacer)

Method Detail

startLearning

public void startLearning()

Description copied from interface: LearningAlgorithm

Starts the model inference process, creating an initial hypothesis in the provided model object. Please note that it should be illegal to invoke this method twice.

Specified by:

startLearning in interface LearningAlgorithm<net.automatalib.automata.transout.MealyMachine<?,I,?,O>,I,net.automatalib.words.Word<O>>

refineHypothesis

public boolean refineHypothesis(DefaultQuery<I,net.automatalib.words.Word<O>> ce)

Description copied from interface: LearningAlgorithm

Triggers a refinement of the model by providing a counterexample. A counterexample is a query which exposes different behavior of the real SUL compared to the hypothesis. Please note that invoking this method before an initial invocation of LearningAlgorithm.startLearning() should be illegal.

Specified by:

refineHypothesis in interface LearningAlgorithm<net.automatalib.automata.transout.MealyMachine<?,I,?,O>,I,net.automatalib.words.Word<O>>

Parameters:

ce - the query which exposes diverging behavior, as posed to the real SUL (i.e. with the SULs output).

Returns:

true if the counterexample triggered a refinement of the hypothesis, false otherwise (i.e., it was no counterexample).

refineHypothesisInternal

public boolean refineHypothesisInternal(DefaultQuery<I,net.automatalib.words.Word<O>> ceQuery)

getHypothesisModel

@Nonnull
public net.automatalib.automata.transout.MealyMachine<?,I,?,O> getHypothesisModel()

Description copied from interface: LearningAlgorithm

Returns the current hypothesis model.

N.B.: By the contract of this interface, the model returned by this method may not be modified (i.e., M generally should refer to an immutable interface), and its validity is retained only until the next invocation of LearningAlgorithm.refineHypothesis(DefaultQuery). If older hypotheses have to be maintained, a copy of the returned model must be made.

Please note that it should be illegal to invoke this method before an initial invocation of LearningAlgorithm.startLearning().

Specified by:

getHypothesisModel in interface LearningAlgorithm<net.automatalib.automata.transout.MealyMachine<?,I,?,O>,I,net.automatalib.words.Word<O>>

Returns:

the current hypothesis model.

closeTransition

public void closeTransition(ADTState<I,O> state,
                            I input)
                     throws PartialTransitionAnalyzer.HypothesisModificationException

Description copied from interface: PartialTransitionAnalyzer

Determine the successor/output of the transition in question (which is usually achieved by sifting the corresponding long-prefix through the ADT).

Specified by:

closeTransition in interface PartialTransitionAnalyzer<ADTState<I,O>,I>

Parameters:

state - the (source) state of the transition in question

input - the input symbol of the transition in question

Throws:

PartialTransitionAnalyzer.HypothesisModificationException - if closing the transition (sifting in the ADT) discovered a new hypothesis state and thus potentially invalidates previously observed behavior

isTransitionDefined

public boolean isTransitionDefined(ADTState<I,O> state,
                                   I input)

Description copied from interface: PartialTransitionAnalyzer

Check whether the transition in question is defined or not.

Specified by:

isTransitionDefined in interface PartialTransitionAnalyzer<ADTState<I,O>,I>

Parameters:

state - the (source) state of the transition in question

input - the input symbol of the transition in question

Returns:

true if the transition (and thus the successor/output) is defined, false otherwise

addAlphabetSymbol

public void addAlphabetSymbol(I symbol)

Specified by:

addAlphabetSymbol in interface SupportsGrowingAlphabet<I>

suspend

public ADTLearnerState<ADTState<I,O>,I,O> suspend()

Description copied from interface: ResumableLearner

Expose the serializable learner state object.

Does not stop a running learning process. Since most data structures that are used during learning are mutable, use this method inside of a learning loop with care.

Specified by:

suspend in interface ResumableLearner<ADTLearnerState<ADTState<I,O>,I,O>>

Returns:

The learner state.

resume

public void resume(ADTLearnerState<ADTState<I,O>,I,O> state)

Description copied from interface: ResumableLearner

Does not get the learner to continue learning. Instead, the learner updates its internal state according to the given state object.

Specified by:

resume in interface ResumableLearner<ADTLearnerState<ADTState<I,O>,I,O>>

Parameters:

state - The learner state.