Semantic Web for the Working Ontologist Dean Allemang Jim Hendler SNU IDB laboratory

Semantic Web for the Working OntologistDean Allemang, Jim Hendler Working Ontology Contents 1 ■ Chapter 1 What is the Semantic Web? ■ Chapter 2 Semantic Modeling ■ Chapter 3 RDF-The Basis of the semantic Web ■ Chapter 4 Semantic Web Application Architecture ■ Chapter 5 RDF and Inferencing ■ Chapter 6 RDF Schema ■ Chapter 7 RDFS-Plus ■ Chapter 8 Using RDFS-Plus in the Wild ■ Chapter 9 Basic OWL ■ Chapter 10 Counting and Sets in OWL ■ Chapter 11 Using OWL in the Wild ■ Chapter 12 Good and Bad Modeling Practices ■ Chapter 13 OWL Levels and Logic

Semantic Web for the Working OntologistDean Allemang, Jim Hendler Chapter 6 RDF Schema ■ Schema Languages and Their Functions ■ What Does It Mean? Semantics as Inference ■ The RDF Schema Language ■ RDFS Modeling Combinations and Patterns ■ Challenges ■ Modeling with Domains and Ranges ■ Nonmodeling Properties in RDFS 2

Semantic Web for the Working OntologistDean Allemang, Jim Hendler Introduction ■ Semantic Web modeling in RDF: about graphs (to represent data) in the RDF Schema Language (RDFS): about sets ■ RDFS provides a way to talk about the vocabulary that will be used in an RDF graph 3

Semantic Web for the Working OntologistDean Allemang, Jim Hendler Schema Languages and Their Functions [1/3] ■ What is a schema language? Helping us to interpret particular data (RDB schema) ■ Schema tells us something about the information that is expressed in the system 4

Semantic Web for the Working OntologistDean Allemang, Jim Hendler Schema Languages and Their Functions [2/3] ■ RDFS help provide some sense of meaning to the data ■ The mechanism by which it does this is the concept of inference 5

Semantic Web for the Working OntologistDean Allemang, Jim Hendler Schema Languages and Their Functions [3/3] ■ In most modeling systems, there is a clear division between the data and its schema XML DTD (not a valid XML document) RDB schema (not typically expressed in a table) ■ In the case of RDF, the schema language was defined in RDF from the very beginning. That is, all schema information in RDFS is defined with RDF triples 6 :IDB rdf:type rdfs:Class. How a set is defined in RDFS

Semantic Web for the Working OntologistDean Allemang, Jim Hendler What Does It Mean? Semantics as Inference ■ What does it mean for something to be a member of a class? It means that it is also a member of any superclass 7 A rdfs:subClassOf B. r rdf:type A. r rdf:type B. inference

Semantic Web for the Working OntologistDean Allemang, Jim Hendler The RDF Schema Language [1/4] - Relationship Propagation through rdfs:subPropertyOf ■ P rdfs:subPropertyOf R Whenever we have the triple A P B. We can infer the triple A R B. The relationship brother is more specific than the relationship sibling ; if someone is my brother, then he is also my sibling. 8 brother rdfs:subPropertyOf sibling.

Semantic Web for the Working OntologistDean Allemang, Jim Hendler The RDF Schema Language [2/4] - Relationship Propagation through rdfs:subPropertyOf (Cont’d) 1:freeLancesTo rdfs:subPropertyOf :contractsTo. 2:indirectlyContractsTo rdfs:subPropertyOf :contractsTo. 3:isEmployedBy rdfs:subPropertyOf :worksFor. 4:contractsTo rdfs:subPropertyOf :worksFor. Spence freelancesTo TheFirm. Spence contractsTo TheFirm. Spence worksFor TheFirm. 9 a b worksFor contractsTo isEmployedBy freeLancesTo indirectlyContractsTo a b

Semantic Web for the Working OntologistDean Allemang, Jim Hendler The RDF Schema Language [3/4] - Typing Data by Usage – rdfs:domain and rdfs:range ■ P rdfs:domain D. P rdfs:range R. ■ Domain and range tell us how P is to be used. 10 IF P rdfs:domain D. and x P y. THEN x rdf:type D. IF P rdfs:range R. and x P y. THEN y rdf:type R. DR P

Semantic Web for the Working OntologistDean Allemang, Jim Hendler The RDF Schema Language [4/4] - Combination of Domain and Range with rdfs:subClassOf ■ X :maidenName Y. X rdf:type :MarriedWoman X rdf:type :Woman IF P rdfs:domain D. and D rdfs:subClassOf C. THEN P rdfs:domain C. :MarriedWoman rdfs:subClassOf :Woman. :maidenName rdfs:domain :MarriedWoman. MarriedWoman Woman rdfs:subClassOf rdfs:domain maidenName 1 2

Semantic Web for the Working OntologistDean Allemang, Jim Hendler RDFS Modeling Combinations and Patterns [1/7] - Set Intersection ■ There is no explicit modeling construct in RDFS for set intersection (or for set union) ■ However, don’t always need to model them explicitly. ■ C rdfs:subClassOf A. C rdfs:subClassOf B. RDFS cannot actually express set intersection; it can only approximate it by supporting the inferencing in one direction 12 x rdf:type C.x rdf:type A. x rdf:type B. C ⊆ A ∩ B (O) C ⊇ A ∩ B (X) AB C

Semantic Web for the Working OntologistDean Allemang, Jim Hendler RDFS Modeling Combinations and Patterns [2/7] - Set Intersection (Cont’d) ■ Kildare is both a member of the staff and a physician, but it is inappropriate to conclude that he must be a surgeon. 13 :Surgeon rdfs:subClassOf :Staff. :Surgeon rdfs:subClassOf :Physician. :Kildare rdf:type :Surgeon. :Kildare rdf:type :Staff. :Kildare rdf:type :Physician. Staff Physician Surgeon rdfs:subClassOf Kildare rdf:type

Semantic Web for the Working OntologistDean Allemang, Jim Hendler RDFS Modeling Combinations and Patterns [3/7] - Property Intersection ■ Cannot make the inference in the other direction 14 :lodgedIn rdfs:subPropertyOf :billedFor. :lodgedIn rdfs:subPropertyOf :assignedTo. :Marcus :lodgedIn :Room101. :Marcus :billedFor :Room101. :Marcus :assignedTo :Room101. P ⊆ A ∩ B x P y.x R y. x S y.

Semantic Web for the Working OntologistDean Allemang, Jim Hendler RDFS Modeling Combinations and Patterns [4/7] - Set Union ■ A ∩ B ⊆ C A rdfs:subClassOf C. B rdfs:subClassOf C. 15 x rdf:type A. orx rdf:type C. x rdf:type B. AB C

Semantic Web for the Working OntologistDean Allemang, Jim Hendler RDFS Modeling Combinations and Patterns [5/7] - Set Union (Cont’d) ■ We can only draw the inference in one direction 16 :MVP rdfs:subClassOf :AllStarCandidate. :TopScorer rdfs:subClassOf :AllStarCandidate. :Reilly rdf:type :MVP. :Kaneda rdf:type :TopScorer. :Reilly rdf:type :AllStarCandidate. :Kaneda rdf:type :AllStarCandidate. C ⊆ A ∩ B (by making C rdfs:subClassOf both A and B ) C ⊇ A ∩ B (by making both A and B rdfs:subClassOf C )

Semantic Web for the Working OntologistDean Allemang, Jim Hendler RDFS Modeling Combinations and Patterns [6/7] - Property Union ■ P rdfs:subPropertyOf R. Q rdfs:subPropertyOf R. If the two properties have exactly the same meaning Else, but want to treat them as the same 17 x P y.x R y. x Q y. Library1:borrows rdfs:subPropertyOf Library2:checkedOut. Library2:checkedOut rdfs:subPropertyOf Library1:borrows. Library1:borrows rdfs:subPropertyOf :hasPossession. Library2:checkedOut rdfs:subPropertyOf :hasPossession. Common Superproperty E.g. Merging Library Records

Semantic Web for the Working OntologistDean Allemang, Jim Hendler RDFS Modeling Combinations and Patterns [7/7] - Property Transfer ■ When modeling the relationship between information that comes from multiple sources ■ Wish to state that all uses of P should be considered as uses of Q 18 X P Y.X Q Y. P rdfs:subPropertyOf Q. X P Y.X Q Y.

Semantic Web for the Working OntologistDean Allemang, Jim Hendler Challenges [1/10] - Term Reconciliation > Challenge 2 ■ How do we enforce the assertion that any member of the one class will automatically be treated as a member of the other? For example, we determine that a researcher is actually a special case of an analyst 19 :Researcher rdfs:subClassOf :Analyst. :Wenger rdf:type :Researcher. :Wenger rdf:type :Analyst.

Semantic Web for the Working OntologistDean Allemang, Jim Hendler Challenges [2/10] - Term Reconciliation > Challenge 3 ■ What if the relationship is more subtle? Suppose there is considerable semantic overlap between the two concepts analyst and researcher, but neither concept is defined in a sharp, formal way → Want to treat these two entities as the same 20 :Analyst rdfs:subClassOf :Investigator. :Researcher rdf:subClassOf :Investigator.

Semantic Web for the Working OntologistDean Allemang, Jim Hendler Challenges [3/10] - Term Reconciliation > Challenge 4 ■ Suppose that we determine that the two class really are identical in every way. In terms of inference, we would like any member of one class to be a member of the other, and vice versa RDFS does not provide a primitive statement of class equivalence 21 :Analyst rdfs:subClassOf :Researcher. :Researcher rdfs:subClassOf :Analyst. :Reilly rdf:type :Researcher. :Kaneda rdf:type :Analyst. :Reilly rdf:type :Analyst. :Kaneda rdf:type :Researcher.

Semantic Web for the Working OntologistDean Allemang, Jim Hendler Challenges [4/10] - Instance-Level Data Integration > Challenge 5 ■ Suppose you have contributions to a single question coming from multiple sources. In the case where the question determines which instances are of interest, there is a simple way to integrate them using rdfs:subClassOf Any instance from either the facility descriptions or the airspace descriptions that have been identified as restricted will be inferred to have cc:OffLimits 22 fc:CivilianFacility rdfs:subClassOf cc:OffLimits. space:NoFlyZone rdfs:subClassOf cc:OffLimits.

Semantic Web for the Working OntologistDean Allemang, Jim Hendler Challenges [5/10] - Readable Labels with rdfs:label > Challenge 6 ■ RDFS provides a built-in property, rdfs:label, whose intended use is to provide a printable name for any resource. We would like to use a generic display mechanism, which uses the standard property rdfs:label to display information about these people and movies 23 :Person1 :personName ”James Dean”. :Person2 :personName ”Elizabeth Taylor”. :Person3 :personName ”Rock Hudson”. :Movie1 :movieTitle “Rebel Without a Cause”. :Movie2 :movieTitle “Giant”. :Movie3 :movieTitle “East of Eden”. :personName rdfs:subPropertyOf rdfs:label. :movieTitle rdfs:subPropertyOf rdfs:label.

Semantic Web for the Working OntologistDean Allemang, Jim Hendler Challenges [6/10] - Data Typing Based on Use > Challenge 7 ■ How can we automatically classify each vessel into more specific subclasses depending on the information we have about it in table 24 NameMaiden VoyageNext DepartureDecommission DateDestruction DateCommander BerengariaJune 16, Johnson QEIIMay 2, 1969March 4, 2010Warwick TitanicApril 10, 1912April 14, 1912Smith ConstitutionJuly 22, 1798January 12, 2009Prebel ship:DeployedVessel rdfs:subClassOf ship:Vessel. ship:InServiceVessel rdfs:subClassOf ship:Vessel. ship:OutOfServiceVessel rdfs:subClassOf ship:Vessel. ship:maidenVoyage rdfs:domain ship:DeployedVessel. ship:nextDeparture rdfs:domain ship:InserviceVessel. ship:decommissionedDate rdfs:domain ship:OutOfServiceVessel. ship:destructionDate rdfs:domain ship:OutOfServiceVessel. Given

Semantic Web for the Working OntologistDean Allemang, Jim Hendler Challenges [7/10] - Data Typing Based on Use > Challenge 7 (Cont’d) 25 DeployedVessel maidenVoyage nextDeparture destructionDate decommissionDate InServiceVessel OutOfServiceVessel rdf:type rdfs:domain Vessel QEII maidenVoyage = nextDeparture = maidenVoyage = nextDeparture = Constitution destructionDate = maidenVoyage = Titanic decommissionedDate = maidenVoyage = Berengaria rdfs:subClassOf

Semantic Web for the Working OntologistDean Allemang, Jim Hendler Challenges [8/10] - Data Typing Based on Use > Challenge 8 ■ It is also possible to draw inferences about the entities in the other table cells How can we express the fact that the commander of a ship has the rank of Captain ? 26 ship:hasCommander rdfs:range ship:Captain. Captain hasCommander rdfs:range Warwick Preble Smith Johnson QEII Constitution Berengaria Titanic rdf:type hasCommander

Semantic Web for the Working OntologistDean Allemang, Jim Hendler Challenges [9/10] - Filtering Undefined Data > Challenge 9 ■ In the preceding example, the set of vessels for which nextDeparture is defined could be used as input to a scheduling system that plans group tours. Ships for which no nextDeparture is known should not be considered 27 ship:DepartingVessel rdfs:type rdfs:Class. Ship:nextDeparture rdfs:domain ship:DepartingVessel. Vessel DepartingVessel nextDeparture rdfs:domain rdfs:subClassOf QEII maidenVoyage = nextDeparture = maidenVoyage = nextDeparture = Constitution rdf:type

Semantic Web for the Working OntologistDean Allemang, Jim Hendler Challenges [10/10] - RDFS and Knowledge Discovery ■ Domains and ranges are not used to validate information but instead are used to determine new information based on old information Because of the inference-based semantics of RDFS ■ On the Semantic Web, we don’t know in advance how information from somewhere else on the Web should be interpreted in a new context The RDFS definitions of domain and range allow us to discover new things about our data based on its use 28

Semantic Web for the Working OntologistDean Allemang, Jim Hendler Modeling with Domains and Ranges - Multiple Domains/Ranges Any such vessel will be inferred to be in the intersection of the two classes specified in the domain statements. This is something that many people find counterintuitive, even though it is “correct” in RDFS. 29 ship:nextDeparture rdfs:domain DepartingVessel. ship:nextDeparture rdfs:domain InServiceVessel. ship:QEII ship:maidenVoyage “May 2, 1969”. ship:QEII ship:nextDeparture “Mar 4, 2010”. ship:QEII ship:hasCommander Warwick. ship:QEII rdf:type ship:DepartingVessel. ship:QEII rdf:type ship:InServiceVessel.

Semantic Web for the Working OntologistDean Allemang, Jim Hendler Modeling with Domains and Ranges - Multiple Domains/Ranges (Cont’d) 30 PP Union / Or Object Oriented P Intersection / And RDFS

Semantic Web for the Working OntologistDean Allemang, Jim Hendler Nonmodeling Properties in RDFS ■ Cross-Referencing Files: rdfs:seeAlso provides a way to specify the web location of supplementary information ■ Organizing Vocabularies: rdfs:isDefinedBy provides a link to the primary source of information about a resource rdfs:isDefinedBy rdfs:subPropertyOf rdfs:seeAlso ■ Model Documentation: rdfs:comment Since RDFS is implemented entirely in RDF, the comment feature is also implemented in RDF 31 sales:nextDeparture rdfs:comment “This indicates the next planned departure date for a salesperson.”