====== Reasoning with Ontologies ======
^ Last verification: | 20180914 |
^ Tools required for this lab: | [[http://protege.stanford.edu/products.php#desktop-protege|Protégé Desktop 5.0]] |
^ ::: | [[http://semanticweb.org/wiki/Pellet|Pellet 2.3.1]] ({{http://krzysztof.kutt.pl/didactics/semweb/pellet-2.3.1.zip|pellet-2.3.1.zip}}) |
===== Introduction =====
* Lecture: {{.:2014:eis2014semweb-ontology-logic.pdf|Ontology engineering and reasoning in Semantic Web}} (2014)
* OPTIONAL: [[https://www.cambridgesemantics.com/blog/semantic-university/intro-semantic-web/semantic-web-misconceptions/|Semantic Web Misconceptions]] (this is our last lab, so this is the right moment for this text :-) )
* [[#if_you_want_to_know_more|If you want to know more...]]
* OWL is based on Description Logics formalism -- if you want to know more about it, you can familiarize yourself with optional [[lab-desc-logic|Intro to Description Logics]] lab.
===== Lab instructions =====
==== - TBox and ABox [10 minutes] ====
* The knowledge base in OWL (which is based on Description Logics) consists of two types of concepts:
* TBox - terminology, intensional representation
* ABox - assertions about individuals, extensional representation
* [[http://www.mkbergman.com/489/ontology-best-practices-for-data-driven-applications-part-2/|Why should we care about keeping them split?]]
* 8-) Let's check if you understand this difference properly: Take your [[lab-owl|ontology from Protégé lab]]. Identify which statements belong to Terminology (TBox) and World Description (ABox). Put them in the report divided into two sections: TBox and ABox (5 statements for each section).
==== - OWL Reasoning - Class Inference [10 minutes] ====
* In Description Logics on which the OWL language is based there are the following reasoning tasks for classes (terminology, TBox):
- Class subsumption (Structuring the knowledge base)
- Class equivalence (Are two classes in fact the same class?)
- Class disjointness (Do classes A and B have common members?)
- Class consistency (Isn't class A empty?)
- Model the following ontology axioms in Protege:
Class: bus_driver
EquivalentTo:
person
that drives some bus
Class: driver
EquivalentTo:
person
that drives some vehicle
Class: bus
SubClassOf:
vehicle
\\ Hints:
- To build the first axiom:
- Create appropriate classes \\ {{.:protege-reasoning1.png?w200|}} \\
- and ObjectProperties: \\ {{.:protege-reasoning2.png?w200|}} \\
- Define the BusDriver class with ''Equivalent classes'' button \\ {{.:protege-reasoning3.png?w200|}}
- Build the second axiom analogously: \\ {{.:protege-reasoning4.png?w200|}} \\
- Build the third axiom using the ''SubClass of'' button: \\ {{.:protege-reasoning5.png?w200|}} \\
- Start the reasoner (''Reasoner -> start reasoner'') and observe the inferred class hierarchy. What conclusions have been drawn? \\ {{.:protege-reasoning6.png?w200|}} \\
- 8-) Include the modified ontology in the report.
- 8-) Explain the inferred relations and conclusions.
- Analyze the reasoning examples given [[http://mowl-power.cs.man.ac.uk/2009/07/sssw/people.html#people-class|here]] in "Class Inferences" section.
==== - OWL Reasoning - Instance Inferences [10 minutes] ====
* In Description Logics on which the OWL language is based there are the following reasoning tasks for instances (world description, ABox):
- Class membership (is instance //a// contained in class //C//?)
- Instance retrieval (find all (known) individuals belonging to a given class)
* and for the whole Knowledge Base (ABox+TBox):
- Global consistency of a knowledge base (Is the knowledge base meaningful?)
- Download the {{http://krzysztof.kutt.pl/didactics/semweb/people-prep.owl|pre-prepared ontology}} and load it in Protege (**NB:** the ontology classes and those on screenshots may differ a little).
- Model the following ontology axioms in Protege:
Individual: Daily_Mirror
Types:
owl:Thing
Individual: Q123ABC
Types:
van,
white_thing
Individual: Mick
Types:
man
Facts:
reads Daily_Mirror,
drives Q123ABC
Class: white_van_man
EquivalentTo:
man
that drives some (van
and white_thing)
SubClassOf:
reads only tabloid
\\ //For those not familiar with UK culture, White Van Man is a stereotype used to describe a particular kind of driver. [[http://en.wikipedia.org/wiki/White_van_man|Wikipedia]] provides an entry with some additional information and references.// \\ Hints:
- Create appropriate instances in ''Individuals'' tab \\ {{.:protege-reasoning7.png?w200|}}
- If the instance is of 2 types, create it only once and then add the second type: \\ {{.:protege-reasoning8.png?w200|}}
- Add object properties to connect individuals: \\ {{.:protege-reasoning9.png?w200|}}
- Define the class: \\ {{.:protege-reasoning10.png?w200|}}
- Start the reasoner (''Reasoner -> start reasoner'') and observe the inferred class hierarchy. Where is the ''white_van_man''? What conclusions have been drawn?
- 8-) Include the modified ontology in the report.
- 8-) Explain the inferred relations and conclusions.
- Analyze the reasoning examples given [[http://mowl-power.cs.man.ac.uk/2009/07/sssw/people.html#people-instance|here]] in "Instance Inferences" section.
==== - OWL Reasoning - Reasoners [5 minutes] ====
DL reasoners may be integrated with other tools - as in the case of Protege presented before - as well as run independently via various interfaces. Popular DL reasoners include: FaCT++, Pellet, HermiT, RacerPro and many others.
We will use the {{http://krzysztof.kutt.pl/didactics/semweb/pellet-2.3.1.zip|Pellet}} reasoner.
- Unpack the zip.
- Run ''pellet.sh help'' to get familiar with available commands.
- Try ''pellet.sh consistency '' where the '''' is:
- Your ontology from [[.:lab-owl|previous lab]]
- ''people+pets.owl'' ontology provided with Pellet in ''examples/data/'' directory \\ and observe the results.
- Try ''pellet.sh classify '' with the two above mentioned ontologies and observe the results.
- 8-) What are the results? Write them down or provide a screenshot of the answer in the report.
==== - Advanced OWL reasoning [50 minutes] ====
These instructions are based on the [[http://www.cs.man.ac.uk/~stevensr/menupages/fhkb.php|The Family History Knowledge Base Tutorial]]((Updated version of the tutorial is available [[http://owl.cs.manchester.ac.uk/publications/talks-and-tutorials/fhkbtutorial/|here]].)) by The University of Manchester, School of Computer Science.
- Download the {{.:ontologies.zip|}} pack.
- Open Protege (type ''protege'' in command line on lab316 laptop).
=== - Warm up ===
- Open (''Family-tree-0.owl'')
- It contains ''Sex'', ''Female'', ''Male'', ''Person'', ''Man'' and ''Woman'' Classes.
- Create ''hasSex'' property, Set domain and ranges on ''hasSex'' (Person/Sex), make it //functional//.
- Create restrictions on ''Person'' ''hasSex some Sex''.
- Make Man and Woman **defined** (use ''Equivalent classes''): ''Man EquivalentTo Person that hasSex some Male'' etc. \\ {{.:owl-man.png?180|}}
- Import plain-individuals, observe that none of them have any assertions on them. (''Active Ontology'' tab -> ''Imported ontologies'' -> ''Direct imports'').
- Choose an individual, add ''Types'': ''Person and (hasSex some Male)'' \\ {{.:owl-robert-individual.png?300|}}
- Using the ''Reasoner'' menu in Protege (''Reasoner -> Start/Synchronize reasoner''), classify the ontology. What inferences are made? Write them in the report 8-)
=== - Characteristics of properties ===
- Close the previous ontology. Load ''Family-tree-1.owl''. Check how ''Man'' and ''Woman'' classes are defined.
- Create object properties: ''hasFather'', ''hasMother'', make them //functional//, add //domain// and //range//: ''Person'', ''Man/Woman'', add //inverse properties//: \\ {{.:owl-hasmother.png?200|}}
- Create restrictions on Person: ''has Mother some Woman'', ''hasFather some Man'' \\ {{.:owl-hasparents.png?200|}}
- Load ''individuals-1-parantage.owl'' file.
- Classify the ontology. What inferences are made? Write them in the report 8-)
=== - Datatype properties ===
- Close the previous ontology. Load ''Family-tree-2.owl''.
- Create or find datatype properties: ''birthYear'', ''deathYear'', make them //functional//. \\ {{.:owl-datatypes.png?200|}}
- Create ''Victorian'' class: ''Person and birthYear some integer[>=1837, <=1901]''
- Classify the ontology. How many individuals have been classified as Victorians? 8-)
=== - Transitive properties ===
- Close the previous ontology. Load ''Family-tree-3.owl''.
- Create a new super-property of ''hasParent'' called ''hasAncestor'',
- Make it transitive and create inverse property.
- Clasify, look at property hierarchy. What do we know about Robert? 8-)
- Close the previous ontology. Load ''Family-tree-4.owl''.
- Model the information about David: \\ {{.:owl-david.png?300|}}
- Classify the ontology. Answer the questions:
- What type does David_bright_1934 have? 8-)
- How many children does he have? 8-)
=== - Property chains ===
- Close the previous ontology. Load Family-tree-5.owl.
- Add hasGrandParent sub prop of hasAncestor (inverse grandparentOf)
- Add domain and ranges
- Add sub properties ''hasGrandMother/Father'' + domains and ranges
- Add property chain ''hasParent o hasFather/Mother -> hasGrandFather/Mother'' \\ {{.:owl-grandpa.png?300|}}
- Classify the ontology. What do we learn about Robert? 8-)
=== - Classes with Individuals, Open World Assumption ===
- Close the previous ontology. Load Family-tree-6.owl.
- Make class of all ancestors and class of robert's ancestors. \\ {{.:owl-class-individuals.png?400|}}
- Classify. What are the results?
- Make class of people with min 3 children.
- Classify the ontology, do we find parents with 3 children? Why? 8-)
- How to really find parents that have 3 children? 8-)
- How to make individuals different from each other? 8-)
===== Materials =====
{{section>.:lab-ontologies#if_you_want_to_know_more}}