As DL-Learner contains several different learning algorithms, support for bagging[1] could be an interesting feature. We still would have to figure out how to aggregate several CEs...

[1] An Empirical Evaluation of Bagging in Inductive Logic Programming. Dutra et al. Proceedings of the Twelfth International Conference on Inductive Logic Programming, pp. 48--65, Sydney, Australia.

@JensLehmann @patrickwestphal @SimonBin
In order to have a complete and up-to-date documentation, all available components need to be annotated. For instance, several heuristics do not have an annotation, thus, they do not occur in the current documentation. Moreover, all config options should be described.

For large number of individuals, retrieving complete list may be inefficient for Sparql. Could try Min/Max and then distribution estimation via binary splits & sparql Count

┆Issue is synchronized with this Trello card by Unito

Reported by kurzum on 2010-08-25 13:47 UTC
In general, if you assign the same instance as positive and negative example (e.g. by mistake)
there is no warning or error message, which is direly needed.
It is quite a common copy and paste mistake and takes forever, if there is no warning.

There is currently no implementation for the FuzzyIndividualReasoner interface. Some time ago there was an implementation for fuzzyDL which was removed and should be re-introduced.

┆Issue is synchronized with this Trello card by Unito

Alternatively throw exceptions in all public methods, whether they are pre-init or post-init methods. Or research other options to stop incorrect DL-Learner usage (by which I mean: calling set Methods after .init or calling methods before .init has been done) Sample exception I added for some specific case in RhoDRDown: https://github.com/AKSW/DL-Learner/blob/94da38d7c1e1adc8ee7613c1938a3a550a0d4a23/components-core/src/main/java/org/dllearner/refinementoperators/RhoDRDown.java#L472

Use RDF to configure DL-Learner (Source could be .ttl file or SPARQL endpoint)

Reported by wormsbee on 2012-11-06 10:18 UTC
curl --data-urlencode [email protected] "http://localhost:9099/interfaces/rest"

{"learningresult":{"error":"An error occured: org.springframework.beans.factory.BeanCreationException: Error creating bean with name 'alg': Error setting property values; nested exception is org.springframework.beans.NotWritablePropertyException: Invalid property 'learningProblem' of bean class [org.dllearner.algorithms.ocel.OCEL]: Bean property 'learningProblem' is not writable or has an invalid setter method. Does the parameter type of the setter match the return type of the getter?","stacktrace":org.springframework.beans.NotWritablePropertyException: Invalid property 'learningProblem' of bean class [org.dllearner.algorithms.ocel.OCEL]: Bean property 'learningProblem' is not writable or has an invalid setter method. Does the parameter type of the setter match the return type of the getter?,"success":"0"}}

Reported by jenslehmann on 2010-02-18 10:52 UTC
Currently, only classes can be ignored/allowed in OCEL. It would be good to reintroduce ignoring/allowing certain object properties.

Problem:

Class expression learning algorithms derived from AbstractCELA allow to set classes and object/data properties that should be ignored during learning. Those ignored entities are only considered during the initialization of the hierarchies in
initClassHierarchy()
initObjectPropertyHierarchy()
initDataPropertyHierarchy()

Those hierarchy objects are put into the refinement operator. The problem is that for some tasks only the underlying reasoner is used:

if(useNumericDatatypes) {
            Set<OWLDataProperty> numericDPs = reasoner.getNumericDataProperties();

or

if(useDataHasValueConstructor) {
            Set<OWLDataProperty> stringDPs = reasoner.getStringDatatypeProperties();

and even in the init() method.

This can lead to solutions that can contain ignored entities, thus, should not be returned and also cause a larger search space than necessary.

Reported by kurzum on 2010-02-16 21:53 UTC
run
examples/testCaseSPARQLDescription
add in ROLlearner2
SparqlQueryDescriptionConvertVisitor visit = new SparqlQueryDescriptionConvertVisitor();
String sparqlQuery = visit.getSparqlQuery(ed.getDescription().toKBSyntaxString());
and
String sparqlQuery = visit.getSparqlQuery(ed.getDescription());

Update the OCEL algorithm to be usable with the Pos Only learning problem.

┆Issue is synchronized with this Trello card by Unito

DL-Learner's OWLAPIReasoner class supports a fallback reasoner for reasoning tasks not supported by the main reasoner. The initialization and calls of this reasoner need to be cleaned up.

(reproducible on a private ontology)

Caused by: java.util.ConcurrentModificationException
        at java.util.HashMap$HashIterator.nextNode(HashMap.java:1429)
        at java.util.HashMap$KeyIterator.next(HashMap.java:1453)
        at com.clarkparsia.pellet.rules.ContinuousRulesStrategy.applyRuleBindings(ContinuousRulesStrategy.java:147)
        at com.clarkparsia.pellet.rules.ContinuousRulesStrategy.complete(ContinuousRulesStrategy.java:254)
        at org.mindswap.pellet.ABox.isConsistent(ABox.java:1423)
        at org.mindswap.pellet.ABox.isConsistent(ABox.java:1260)
        at org.mindswap.pellet.KnowledgeBase.consistency(KnowledgeBase.java:2017)
        at org.mindswap.pellet.KnowledgeBase.isConsistent(KnowledgeBase.java:2089)
        at com.clarkparsia.pellet.owlapiv3.PelletReasoner.isConsistent(PelletReasoner.java:849)
        at org.dllearner.reasoning.OWLAPIReasoner.init(OWLAPIReasoner.java:246)

changing for( PartialBinding ruleBinding : partialBindings ) to for( PartialBinding ruleBinding : new HashSet(partialBindings) ) on com.clarkparsia.pellet.rules.ContinuousRulesStrategy line 147 only results in

Caused by: java.lang.OutOfMemoryError: GC overhead limit exceeded
        at org.mindswap.pellet.EdgeList.<init>(EdgeList.java:80)
        at org.mindswap.pellet.EdgeList.<init>(EdgeList.java:76)
        at org.mindswap.pellet.EdgeList.getEdges(EdgeList.java:202)
        at org.mindswap.pellet.Individual.getRNeighborEdges(Individual.java:678)
        at com.clarkparsia.pellet.rules.TrivialSatisfactionHelpers$TestProperty.check(TrivialSatisfactionHelpers.java:317)
        at com.clarkparsia.pellet.rules.TrivialSatisfactionHelpers$TestIndividualProperty.check(TrivialSatisfactionHelpers.java:295)
        at com.clarkparsia.pellet.rules.TrivialSatisfactionHelpers.isAtomTrue(TrivialSatisfactionHelpers.java:402)
        at com.clarkparsia.pellet.rules.ContinuousRulesStrategy.createDisjunctionsFromBinding(ContinuousRulesStrategy.java:410)
        at com.clarkparsia.pellet.rules.ContinuousRulesStrategy.applyRuleBindings(ContinuousRulesStrategy.java:224)
        at com.clarkparsia.pellet.rules.ContinuousRulesStrategy.complete(ContinuousRulesStrategy.java:331)
        at org.mindswap.pellet.ABox.isConsistent(ABox.java:1423)
        at org.mindswap.pellet.ABox.isConsistent(ABox.java:1260)
        at org.mindswap.pellet.KnowledgeBase.consistency(KnowledgeBase.java:1989)
        at org.mindswap.pellet.KnowledgeBase.isConsistent(KnowledgeBase.java:2061)
        at com.clarkparsia.pellet.owlapiv3.PelletReasoner.isConsistent(PelletReasoner.java:849)
        at org.dllearner.reasoning.OWLAPIReasoner.init(OWLAPIReasoner.java:246)

Currently, there is only one option to set the frequency threshold of properties values used e.g. in hasValue hypotheses, i.e. we cannot set different values for the values of object properties (OWL individuals) and data properties (OWL literals).

Suggestion:

frequencyThreshold =>

1. dataHasValueFrequencyThreshold
2. objectHasValueFrequencyThreshold

or something else

Important: We have to adapt examples/conf files too after renaming was done

Running Java 1.8 the test of LGGTest.testClassEntailment fails.

Reported by jenslehmann on 2011-08-18 08:31 UTC
There should be a unit test for the various implementations of the PropertyEditor interface in DL-Learner, which checks whether setAsText and getAsText are consistent, i.e. first calling setAsTest(String) and then getAsText should result in the same string for all editors and all valid strings.

Adding unions in class expression like

r some (A or B or ...)
(A or B or ...)

should be pruned resp. avoided if they do not have any extensional impact.

It looks like getting individuals for a class expression with a max. cardinality restriction doesn't work correctly.

Data

p(s, o0)
p(s, o1)
p(s, o2)
p(s, o3)
p(s, o4)
A(s)
C(o0)
C(o1)
C(o2)
C(o3)
C(o4)

Query

A ⊓ (≤ 2 p.C)

Result of reasoner.getIndividuals

[s]

In OWL2 EL only the following constructs are supported(see http://www.w3.org/TR/owl2-profiles/#OWL_2_EL):

• existential quantification to a class expression (ObjectSomeValuesFrom) or a data range (DataSomeValuesFrom)
• existential quantification to an individual (ObjectHasValue) or a literal (DataHasValue)
• self-restriction (ObjectHasSelf)
• enumerations involving a single individual (ObjectOneOf) or a single literal (DataOneOf)
• intersection of classes (ObjectIntersectionOf) and data ranges (DataIntersectionOf)

The basic EL learning algorithm ELTL only supports a subset of the constructs above, i.e. everything seems to be correct. On the other hand, we should add support for DataSomeValuesFrom and DataHasValue.

A special case is (ELTL + disjunctive) which also uses disjunction of class expression, i.e. solution might not be fit to OWL2 EL. Moreover, support for DataSomeValuesFrom was recently added. The problem here is that only a particular set of datatypes is supported (which is necessary to obtain the desired computational properties):

• rdf:PlainLiteral
• rdf:XMLLiteral
• rdfs:Literal
• owl:real
• owl:rational
• xsd:decimal
• xsd:integer
• xsd:nonNegativeInteger
• xsd:string
• xsd:normalizedString
• xsd:token
• xsd:Name
• xsd:NCName
• xsd:NMTOKEN
• xsd:hexBinary
• xsd:base64Binary
• xsd:anyURI
• xsd:dateTime
• xsd:dateTimeStamp

This means, that e.g. xsd:double can not be used and it's not clear how to handle such cases.

it did not find some facts when loading from config (ontology+4sources)

Reported by kurzum on 2010-08-19 13:17 UTC
in:
public class TokenMgrError extends Error
This eror is sometimes not caught and transformed to the checked Exception.
So effectively, if an Application uses
try{} catch (ParseException e) {}
the error can still run through and wreck the application

[ERROR] org.dllearner.parser.TokenMgrError: Lexical error at line 1, column 8. Encountered: "a" (97), after : "F"
[ERROR] at org.dllearner.parser.KBParserTokenManager.getNextToken(parser:KBParserTokenManager.java):1309)
[ERROR] at org.dllearner.parser.KBParser.jj_ntk(parser:KBParser.java):1219)
[ERROR] at org.dllearner.parser.KBParser.Concept(parser:KBParser.java):425)
[ERROR] at org.dllearner.parser.KBParser.TBoxEquiv(parser:KBParser.java):267)
[ERROR] at org.dllearner.parser.KBParser.parseConcept(parser:KBParser.java):30)
[ERROR] at org.nlp2rdf.navigator.server.store.LogicalRelationStrategy.add(LogicalRelationStrategy.java:132)
[ERROR] at org.nlp2rdf.navigator.server.store.LogicalRelationStrategy.update(LogicalRelationStrategy.java:154)
[ERROR] at org.nlp2rdf.navigator.server.store.Store.saveConcept(Store.java:39)

Due to internal changes, the Spring test needs to be updated (no bug, just a new test needed):

https://github.com/AKSW/DL-Learner/blob/release-1.0/interfaces/src/test/java/org/dllearner/test/SpringTest.java

check whether http://ns.softwiki.de/req/averagePriorityRate should have datatype xsd:float? reasoner.getDatatype(dp) return null

MILESTest

Add generalized IR measures, e.g. match-, commission-, omission- und inductionrate (see [1]) as heuristics.

[1] A Note on the Evaluation of Inductive Concept Classification Procedures

┆Issue is synchronized with this Trello card by Unito

Since several dependencies (e.g. com/davidsoergel/dsutils/1.051/dsutils-1.051, com/davidsoergel/conja/1.061/conja-1.061) were introduced with the SemKernel component that require com.google.collections:google-collections calling e.g. bin/cli after running the buildRelease.sh script will cause the following error:

patrick@oklasos:/tmp/dll_trials/dllearner-1.0 $ bin/cli examples/father.conf 
DL-Learner command line interface
Refreshing org.springframework.context.support.ClassPathXmlApplicationContext@7494e528: startup date [Sat Mar 28 18:11:26 CET 2015]; root of context hierarchy
Loading XML bean definitions from URL [jar:file:/tmp/dll_trials/dllearner-1.0/lib/interfaces-1.1-SNAPSHOT.jar!/org/dllearner/configuration/spring/bean-post-processor-configuration.xml]
Exception in thread "main" java.lang.NoSuchMethodError: com.google.common.base.Platform.systemNanoTime()J
    at com.google.common.base.Ticker$1.read(Ticker.java:60)
    at com.google.common.base.Stopwatch.start(Stopwatch.java:162)
    at com.google.common.cache.LocalCache$LoadingValueReference.loadFuture(LocalCache.java:3523)
    at com.google.common.cache.LocalCache$Segment.loadSync(LocalCache.java:2319)
    at com.google.common.cache.LocalCache$Segment.lockedGetOrLoad(LocalCache.java:2282)
    at com.google.common.cache.LocalCache$Segment.get(LocalCache.java:2197)
    at com.google.common.cache.LocalCache.get(LocalCache.java:3937)
    at com.google.common.cache.LocalCache.getOrLoad(LocalCache.java:3941)
    at com.google.common.cache.LocalCache$LocalLoadingCache.get(LocalCache.java:4824)
    at org.semanticweb.owlapi.model.IRI.cache(IRI.java:348)
    at org.semanticweb.owlapi.model.IRI.<init>(IRI.java:369)
    at org.semanticweb.owlapi.model.IRI.create(IRI.java:269)
    at org.semanticweb.owlapi.vocab.OWLRDFVocabulary.<init>(OWLRDFVocabulary.java:356)
    at org.semanticweb.owlapi.vocab.OWLRDFVocabulary.<clinit>(OWLRDFVocabulary.java:62)
    at uk.ac.manchester.cs.owl.owlapi.OWLObjectImpl.<clinit>(OWLObjectImpl.java:87)
    at org.dllearner.core.AbstractCELA.<clinit>(AbstractCELA.java:74)
    at java.lang.Class.forName0(Native Method)
    at java.lang.Class.forName(Class.java:264)
    at org.dllearner.core.AnnComponentManager.<init>(AnnComponentManager.java:117)
    at org.dllearner.core.AnnComponentManager.getInstance(AnnComponentManager.java:142)
    at org.dllearner.confparser3.ConfParserConfiguration.getClass(ConfParserConfiguration.java:76)
    at org.dllearner.configuration.spring.ConfigurationBasedBeanDefinitionRegistryPostProcessor.postProcessBeanDefinitionRegistry(ConfigurationBasedBeanDefinitionRegistryPostProcessor.java:48)
    at org.springframework.context.support.AbstractApplicationContext.invokeBeanFactoryPostProcessors(AbstractApplicationContext.java:619)
    at org.springframework.context.support.AbstractApplicationContext.refresh(AbstractApplicationContext.java:461)
    at org.dllearner.configuration.spring.DefaultApplicationContextBuilder.buildApplicationContext(DefaultApplicationContextBuilder.java:60)
    at org.dllearner.cli.CLI.main(CLI.java:259)

Similar errors were already reported, e.g. here or here.

I am trying to run PCELOE (Parallel version of CELOE) algorithm but it gives java concurrent modification error.

A snippet of the log is:
Caused by: java.util.ConcurrentModificationException
at java.util.TreeMap$NavigableSubMap$SubMapIterator.prevEntry(TreeMap.java:1714)
at java.util.TreeMap$NavigableSubMap$DescendingSubMapKeyIterator.next(TreeMap.java:1818)
at org.dllearner.algorithms.celoe.PCELOE.getNextNodeToExpand(PCELOE.java:503)
at org.dllearner.algorithms.celoe.PCELOE.access$200(PCELOE.java:65)
at org.dllearner.algorithms.celoe.PCELOE$PCELOEWorker.run(PCELOE.java:1138)
at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
at java.util.concurrent.FutureTask.run(FutureTask.java:266)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)

Reported by kurzum on 2011-04-14 06:10 UTC
if no examples are set PosNegLP throws a Nullpointer, but it should actually throw a ComponentInitexception.

Additionally, there is no warning, when the pos and negative Examples overlap:
PosNegLPStandard lp = cm.learningProblem(PosNegLPStandard.class, rc);
lp.setPositiveExamples(posExamples);
lp.setNegativeExamples(posExamples);
silently continues.

WARN [main] (DBpediaClassLearnerCELOE.java:94) -
java.lang.NullPointerException
at org.dllearner.utilities.Helper.union(Helper.java:263)
at org.dllearner.learningproblems.PosNegLP.init(PosNegLP.java:137)
at org.dllearner.learningproblems.PosNegLPStandard.init(PosNegLPStandard.java:87)
at org.dllearner.scripts.improveWikipedia.DBpediaClassLearnerCELOE.learnClass(DBpediaClassLearnerCELOE.java:144)
at org.dllearner.scripts.improveWikipedia.DBpediaClassLearnerCELOE.learnAllClasses(DBpediaClassLearnerCELOE.java:83)
at org.dllearner.scripts.improveWikipedia.DBpediaClassLearnerCELOE.main(DBpediaClassLearnerCELOE.java:74)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:616)
at com.intellij.rt.execution.application.AppMain.main(AppMain.java:115)

Running the enrichment executable on DBpedia live gives me several errors like this one:

$ bin/enrichment -e http://live.dbpedia.org/sparql -g http://dbpedia.org -r http://dbpedia.org/ontology/currency
<http://dbpedia.org/ontology/currency> appears to be an object property. Running appropriate algorithms.

Processing property http://dbpedia.org/ontology/currency
Learning of EquivalentObjectProperties axioms ...
Learning 
of EquivalentObjectProperties axioms 
failed.
 ... EquivalentObjectProperties axioms finished
ERROR - An error occurred while generating EquivalentObjectProperties axioms for object property http://dbpedia.org/ontology/currency
java.lang.RuntimeException: Query Execution failed, even with retries.
    at org.aksw.jena_sparql_api.retry.core.QueryExecutionRetry.doTry(QueryExecutionRetry.java:57)
    at org.aksw.jena_sparql_api.retry.core.QueryExecutionRetry.execSelect(QueryExecutionRetry.java:83)
    at org.dllearner.core.AbstractAxiomLearningAlgorithm.executeSelectQuery(AbstractAxiomLearningAlgorithm.java:469)
    at org.dllearner.algorithms.properties.ObjectPropertyHierarchyAxiomLearner.run(ObjectPropertyHierarchyAxiomLearner.java:114)
    at org.dllearner.algorithms.properties.PropertyAxiomLearner.learnAxioms(PropertyAxiomLearner.java:97)
    at org.dllearner.core.AbstractAxiomLearningAlgorithm.start(AbstractAxiomLearningAlgorithm.java:251)
    at org.dllearner.algorithms.properties.MultiPropertyAxiomLearner.applyAlgorithm(MultiPropertyAxiomLearner.java:300)
    at org.dllearner.algorithms.properties.MultiPropertyAxiomLearner.lambda$start$0(MultiPropertyAxiomLearner.java:157)
    at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
    at java.util.concurrent.FutureTask.run(FutureTask.java:266)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
    at java.lang.Thread.run(Thread.java:745)
Caused by: java.util.concurrent.ExecutionException: java.lang.RuntimeException: java.lang.RuntimeException: org.apache.jena.query.QueryExecException: HTTP QueryExecution has been closed
    at com.google.common.util.concurrent.AbstractFuture.getDoneValue(AbstractFuture.java:476)
    at com.google.common.util.concurrent.AbstractFuture.get(AbstractFuture.java:455)
    at com.google.common.util.concurrent.AbstractFuture$TrustedFuture.get(AbstractFuture.java:79)
    at org.aksw.jena_sparql_api.retry.core.QueryExecutionRetry.doTry(QueryExecutionRetry.java:54)
    ... 12 more
Caused by: java.lang.RuntimeException: java.lang.RuntimeException: org.apache.jena.query.QueryExecException: HTTP QueryExecution has been closed
    at org.aksw.jena_sparql_api.cache.core.QueryExecutionCacheEx.doCacheResultSet(QueryExecutionCacheEx.java:98)
    at org.aksw.jena_sparql_api.cache.core.QueryExecutionCacheEx.execSelect(QueryExecutionCacheEx.java:210)
    at org.aksw.jena_sparql_api.core.QueryExecutionDecoratorBase.execSelect(QueryExecutionDecoratorBase.java:140)
    at org.aksw.jena_sparql_api.delay.core.QueryExecutionDelay.execSelect(QueryExecutionDelay.java:42)
    at org.aksw.jena_sparql_api.core.QueryExecutionDecoratorBase.execSelect(QueryExecutionDecoratorBase.java:140)
    at org.aksw.jena_sparql_api.retry.core.QueryExecutionRetry.access$101(QueryExecutionRetry.java:23)
    at org.aksw.jena_sparql_api.retry.core.QueryExecutionRetry$2.call(QueryExecutionRetry.java:86)
    at org.aksw.jena_sparql_api.retry.core.QueryExecutionRetry$2.call(QueryExecutionRetry.java:83)
    at com.nurkiewicz.asyncretry.AsyncRetryExecutor$2.call(AsyncRetryExecutor.java:72)
    at com.nurkiewicz.asyncretry.SyncRetryJob.run(SyncRetryJob.java:22)
    at com.nurkiewicz.asyncretry.RetryJob.run(RetryJob.java:96)
    at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
    at java.util.concurrent.FutureTask.run(FutureTask.java:266)
    at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$201(ScheduledThreadPoolExecutor.java:180)
    at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293)
    ... 3 more
Caused by: java.lang.RuntimeException: org.apache.jena.query.QueryExecException: HTTP QueryExecution has been closed
    at org.aksw.jena_sparql_api.http.HttpExceptionUtils.makeHumanFriendly(HttpExceptionUtils.java:17)
    at org.aksw.jena_sparql_api.http.QueryExecutionHttpWrapper.wrapException(QueryExecutionHttpWrapper.java:69)
    at org.aksw.jena_sparql_api.http.QueryExecutionHttpWrapper.execSelect(QueryExecutionHttpWrapper.java:26)
    at org.aksw.jena_sparql_api.cache.core.QueryExecutionCacheEx.doCacheResultSet(QueryExecutionCacheEx.java:76)
    ... 17 more
Caused by: org.apache.jena.query.QueryExecException: HTTP QueryExecution has been closed
    at org.apache.jena.sparql.engine.http.QueryEngineHTTP.checkNotClosed(QueryEngineHTTP.java:534)
    at org.apache.jena.sparql.engine.http.QueryEngineHTTP.execSelect(QueryEngineHTTP.java:350)
    at org.aksw.jena_sparql_api.http.QueryExecutionHttpWrapper$1.get(QueryExecutionHttpWrapper.java:29)
    at org.aksw.jena_sparql_api.http.QueryExecutionHttpWrapper$1.get(QueryExecutionHttpWrapper.java:26)
    at org.aksw.jena_sparql_api.http.QueryExecutionHttpWrapper.wrapException(QueryExecutionHttpWrapper.java:55)
    ... 19 more

Reported by lorenz_b on 2010-02-21 22:33 UTC
There is need to refactor the updateCWAOntology() method in PelletReasoner.

Hi @JensLehmann @LorenzBuehmann @SimonBin , I am using DL-Learner for my research. I found a bug in the heuristic computation in the above line: the accuracy gain is wrong, it shall be conversely computed as:

 score += (score - parentAccuracy) * gainBonusFactor; 

Reported by moahmedsherif on 2012-10-09 14:52 UTC
I run Jena-eyeball to check the enriched ontology "dbpedia_3.8_enrichment.owl", but it generate some errors.
like undeclared datatype' URIs (http://dbpedia.org/datatype/litre, http://dbpedia.org/datatype/metre ...)
also badURI: "owlapi:ontology:ont1"
I attached the complete log file generated by eyeball.

Having a configuration like

prefixes = [ ("ex","http://ex.com/") ]
ks.type = "OWL File"
ks.fileName = "kb.owl"

reasoner.type = "cwr"
reasoner.sources = {ks}

lp.type = "posNegStandard"
lp.reasoner = reasoner
lp.positiveExamples = {"ex:p1"}
lp.negativeExamples = {"ex:n1"}

op.type = "rho"
op.useHasValueConstructor = true

alg.type = "celoe"
alg.maxExecutionTimeInSeconds = 60
alg.reasoner = reasoner

will result in a RuntimeException:

Caused by: java.lang.NullPointerException
	at org.dllearner.refinementoperators.RhoDRDown.init(RhoDRDown.java:356)
	at org.dllearner.algorithms.celoe.CELOE.init(CELOE.java:313)
	at org.dllearner.configuration.spring.ComponentInitializationBeanPostProcessor.postProcessBeforeInitialization(ComponentInitializationBeanPostProcessor.java:35)

The main problem seems to be that the operator's init() is called twice in CELOE:

if (operator == null) {
    // we use a default operator and inject the class hierarchy for now
    operator = new RhoDRDown();
    ((CustomStartRefinementOperator) operator).setStartClass(startClass);
    ((ReasoningBasedRefinementOperator) operator).setReasoner(reasoner);
}
if (operator instanceof CustomHierarchyRefinementOperator) {
    ((CustomHierarchyRefinementOperator) operator).setClassHierarchy(classHierarchy);
    ((CustomHierarchyRefinementOperator) operator).setObjectPropertyHierarchy(objectPropertyHierarchy);
    ((CustomHierarchyRefinementOperator) operator).setDataPropertyHierarchy(datatypePropertyHierarchy);
}
operator.init();

We would have to check whether operator is already initialized an use the clone constructor to make the hierarchy settings and use/re-initialize the clone afterwards.

Reported by jenslehmann on 2010-12-13 10:50 UTC
When a user is asked as question about whether a resource r is in the result set, then the query tree corresponding to r should not be the same as some query tree corresponding to a resource r' for which a question has been asked previously.

The learning algorithms should have the option that universal quantifiers for a role r are only used in learned expressions when there is a role filler for r - essentially using "someOnly" i.e. a combination of universal and existential quantifiers.

Update the ELTL algorithm to be usable with the Pos Only learning problem.

When I run mvn clean install on the top level project, the build fails in the components-ext module with the following message (excerpt):

Tests run: 3, Failures: 0, Errors: 2, Skipped: 0, Time elapsed: 30.852 sec <<< FAILURE! - in org.dllearner.algorithms.probabilistic.structure.unife.leap.LEAPTest
testCase1(org.dllearner.algorithms.probabilistic.structure.unife.leap.LEAPTest)  Time elapsed: 10.306 sec  <<< ERROR!
java.lang.ExceptionInInitializerError: null
    at unife.bundle.bdd.BDDFactory2.<clinit>(BDDFactory2.java:61)
    at unife.edge.EDGE.init(EDGE.java:846)
    at org.dllearner.algorithms.probabilistic.parameter.unife.edge.EDGE.start(EDGE.java:96)
    at org.dllearner.algorithms.probabilistic.structure.unife.leap.LEAP.greedySearch(LEAP.java:212)
    at org.dllearner.algorithms.probabilistic.structure.unife.leap.LEAP.start(LEAP.java:103)
    at org.dllearner.algorithms.probabilistic.structure.unife.leap.LEAPTest.testCase1(LEAPTest.java:93)

testCase2(org.dllearner.algorithms.probabilistic.structure.unife.leap.LEAPTest)  Time elapsed: 10.28 sec  <<< ERROR!
java.lang.NoClassDefFoundError: Could not initialize class unife.bundle.bdd.BDDFactory2
    at unife.edge.EDGE.init(EDGE.java:846)
    at org.dllearner.algorithms.probabilistic.parameter.unife.edge.EDGE.start(EDGE.java:96)
    at org.dllearner.algorithms.probabilistic.structure.unife.leap.LEAP.greedySearch(LEAP.java:212)
    at org.dllearner.algorithms.probabilistic.structure.unife.leap.LEAP.start(LEAP.java:103)
    at org.dllearner.algorithms.probabilistic.structure.unife.leap.LEAPTest.testCase2(LEAPTest.java:125)


Results :

Tests in error: 
  LEAPTest.testCase1:93 » ExceptionInInitializer
  LEAPTest.testCase2:125 » NoClassDefFound Could not initialize class unife.bund...

Tests run: 5, Failures: 0, Errors: 2, Skipped: 0

[INFO] ------------------------------------------------------------------------
[INFO] Reactor Summary:
[INFO] 
[INFO] DL Learner Parent Pom .............................. SUCCESS [  0.687 s]
[INFO] DL Learner Core Components ......................... SUCCESS [02:34 min]
[INFO] components-ext ..................................... FAILURE [ 57.257 s]
[INFO] dllearner-interfaces ............................... SKIPPED
[INFO] interfaces-ext ..................................... SKIPPED
[INFO] Executable scripts ................................. SKIPPED
[INFO] ------------------------------------------------------------------------
[INFO] BUILD FAILURE
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 03:32 min
[INFO] Finished at: 2016-10-13T21:14:18+02:00
[INFO] Final Memory: 69M/888M
[INFO] ------------------------------------------------------------------------
[ERROR] Failed to execute goal org.apache.maven.plugins:maven-surefire-plugin:2.18.1:test (default-test) on project components-ext: There are test failures.

Is there a missing (external) dependency? bundle-2.2.jar is in my Maven local repository.

Reported by lorenz_b on 2010-02-21 22:30 UTC
When loading this ontology, and repairing the inconsistency the programm needs long time and goes to the unsatisfiable panel, although there is no need to do so.

http://owl.cs.manchester.ac.uk/repository/download?ontology=http://lsdis.cs.uga.edu/projects/glycomics/propreo&format=RDF/XML

AbstractRequirement ⊓ Requirement ⊓ Resource

1385 [CELOE] - more accurate (15.62%) class expression found after 384ms: Requirement
Exception in thread "main" java.lang.NullPointerException
    at org.dllearner.reasoning.SPARQLReasoner.getDatatype(SPARQLReasoner.java:2391)
    at org.dllearner.refinementoperators.RhoDRDown.computeM(RhoDRDown.java:1260)
    at org.dllearner.refinementoperators.RhoDRDown.computeTopRefinements(RhoDRDown.java:1058)
    at org.dllearner.refinementoperators.RhoDRDown.computeTopRefinements(RhoDRDown.java:1050)

Hello,

I would like to employ DL-Learner in order to learn class expression of my positive example. To this end, the subset of dbpedia 2016-10 is being used as one may very well see in the below.

I would like to know why it is not possible to assign predicate or object of N-triples as positive examples.

ks.fileName = "spring_KnowledgeBase.owl"
reasoner.type = "closed world reasoner"
reasoner.sources = {ks}
lp.type = "posonlylp"
lp.positiveExamples ={ some URIs}
alg.type = "celoe"
alg.maxExecutionTimeInSeconds = 1

Any suggestion appreciated :)

Both implementations are used to solve a ternary learning problem and seem to share code. We should check whether we only need one of them.

Reported by jenslehmann on 2013-03-07 16:08 UTC
Running the the enrichment script in SVN with the following parameters results in a NullPointer (and also timeouts):

-e http://live.dbpedia.org/sparql -r http://dbpedia.org/ontology/language --maxExecutionTimeInSeconds 30

java.lang.NullPointerException
at org.dllearner.core.owl.ObjectProperty.compareTo(ObjectProperty.java:71)
at org.dllearner.core.owl.ObjectProperty.compareTo(ObjectProperty.java:1)
at java.util.TreeMap.put(TreeMap.java:560)
at java.util.TreeSet.add(TreeSet.java:255)
at org.dllearner.kb.sparql.SPARQLTasks.getAllObjectProperties(SPARQLTasks.java:685)
at org.dllearner.algorithms.properties.DisjointObjectPropertyAxiomLearner.start(DisjointObjectPropertyAxiomLearner.java:102)
at org.dllearner.cli.Enrichment.applyLearningAlgorithm(Enrichment.java:506)
at org.dllearner.cli.Enrichment.runObjectPropertyAlgorithms(Enrichment.java:382)
at org.dllearner.cli.Enrichment.start(Enrichment.java:347)
at org.dllearner.cli.Enrichment.main(Enrichment.java:918)

For learned class expressions like “hasCulture min 6 Thing” it might look more natural if the range of the property is used as filler concept, like e.g. “hasCulture min 6 Culture”. There is a flag in CELOE, but it's not available via conf file option.

May need to use a different replace method

Some of the PHP examples to show the web service feature need to be refactored.

Location: ./php-examples

The problem is in the Utitlities.php file, which includes some HTTP library that does not exist in the examples folder.

A solution would be to replace it by

a) some more common lib which is accessible via standard PHP package manager
b) add the lib file to the php-examples folder

Links to documentation on dl-learner.org and here do not work:

http://dl-learner.org/files/dl-learner-manual.pdf
http://jens-lehmann.org/files/2009_dllearner_sparql.pdf

We already use Java 8 so we should move to OWL API 5 and benefit from its improvements as well.

This means we also have to update the reasoner dependencies. Currently, there are OWL API 5 compatible implementations for:

• Pellet
• HermiT
• JFact

┆Issue is synchronized with this Trello card by Unito