Logics for Data and Knowledge Representation: midterm Exam 2013

Transcription

1 1. [6 PT] Say (mark with an X) whether the following statements are true (T) or false (F). a) In a lightweight ontology there are is-a and part-of relations T F b) Semantic matching is a technique to compute a mapping between T F nodes whose labels have similar meaning in two ontologies c) DERA is a methodology based on the faceted approach for the T F development of classification ontologies d) The Semantic Web infrastructure provides a data model supporting T F a single entity can be distributed over the Web e) RDF allows relationship propagation through rdfs:subclassof T F f) OWL 2 RL (profile) was developed to be implemented using rule based technologies such as rule extended DBMSs T F 2. [6 PT] Translate the following natural language sentences in the DL language with lowest expressiveness possible (e.g. AL, ALC, FL0 ) and say which of the languages you used: a. A parent is a person having at least one child b. Birds have 2 wings c. Male and Female are disjoint d. Germans do not have Italian friends and friends having Italian friends e. The color of a banana can be only yellow or red f. Facebook users can only post photos about their friends PARENT PERSON haschild. (AL) BIRD 2 WING 2 WING (ALN) MALE FEMALE (AL) GERMAN FRIEND-OF. ( ITALIAN FRIEND-OF.ITALIAN) (ALCE) BANANA COLOR.{yellow, red} (FL0) FACEBOOK-USER USER POST.FRIEND-PHOTO (FL0)

2 3. [2 PT] Formally explain the separation of duties RelBAC rule with an example in Description Logic See slides 4. [3 PT] List and provide a brief description of the four basic ABox reasoning services See slides

3 5. [4 PT] Using the tableau calculus, say whether the DL formula below is satisfiable: person ( person eats. plant) eats.(plant dairy) Motivate your answer with a proof. If satisfiable, provide a possible ABox. By -rule we put into the ABox: person(z), ( person eats. plant)(z), eats.(plant dairy)(z), (1) person(z) is already an ABox assertion. (2) ( person eats. plant)(z) by -rule has to be split into: (2.1) person(z) that is clearly in contradiction with (1), therefore we backtrack; (2.2) eats. plant(z) by -rule we add into the ABox: eats(z, y), plant(y) (3) eats.(plant dairy) (z) by -rule we add into the ABox: eats(z, plant(x) dairy(x)) (in fact person(z) is already in the ABox), that by -rule has to be split into: (3.1) plant(y) given that eats(z, y) is in the ABox because of (2.2), that is clearly in contradiction with plant(y) (2.2) (3.2) dairy(x) Thus, there is at least a path which proves the satisfiability of the formula, for instance: (1) (2) (2.1) (3) (3.2) This path generates the ABox A = { person(z), eats(z, y), plant(y), dairy(x) }

4 6. [4 PT] Represent the following statements in RDF: a) If Einstein is a researcher, he is also an investigator, a manofscience and a scientist. b) If John is either an experimenter or a fieldworker or a postdoc, he is also a researcher. Consult RDF modeling in slide 21 of the Resource Description Framework lecture 7. [4 PT] Suppose that in a family tree, relations such as the following ones are functional. a) :haspaternalgrandfather b) :haspaternalgrandmother Represent them in a suitable Semantic Web language and demonstrate their use with necessary entity-entity axioms. See functional property in slide 8 of the OWL lecture

5 8. [4 PT] Suppose that an RDF model represents information about various entities including books. The model is created using standard vocabularies (e.g., Dublic Core). Write a SPARQL query that separates information about books i.e. title, author, date of publication and publisher (if any) and creates another RDF model that is a subset of the original one. PREFIX dc: <dc namespace> CONSTRUCT {?book dc:creator?author.?book dc:title?booktitle.?book dc:date?dateofpublication.?book dc:publisher?pub } WHERE {?book dc:creator?author.?book dc:title?booktitle.?book dc:date?dateofpublication. OPTIONAL {?book dc:publisher?pub} }