SWT Lecture Session 8 - Rules
Download as PPTX, PDF•0 likes•817 views
This document discusses rules and the Semantic Web Rule Language (SWRL). It defines rules as a means of representing knowledge similar to if-then statements. SWRL combines OWL and rule-based languages by allowing users to write rules that can refer to OWL classes, properties, individuals and datatypes. SWRL has an abstract and XML syntax and supports built-in predicates for manipulating data types. Rules provide more expressivity than RDFS and OWL in some cases, such as defining application behaviors, but rule-based reasoning is less performant so they should not be overused when RDFS/OWL suffice.
![+
What are rules
Means of representing knowledge
An ontology language
Can be seen as conditions statements:
if []
then []
if x is a Man and x hasChild y Father(x)](https://image.slidesharecdn.com/8-rules-131012114345-phpapp01/85/SWT-Lecture-Session-8-Rules-4-320.jpg)
![+
7
SWRL
How can datalog and OWL be combined?
SWRL – Semantic Web Rule Language [swirl]
•
Proposal for a rule based Semantic Web Language (W3C member
submission)
http://www.w3.org/Submission/SWRL/
•
Idea: datalog rules referring to RDF classes/properties
Symbols in rules can be OWL identifiers or new symbols
•
Various further features and syntactic forms
•
Support in inference engines (very wide spread support)](https://image.slidesharecdn.com/8-rules-131012114345-phpapp01/85/SWT-Lecture-Session-8-Rules-7-320.jpg)
SWT Lecture Session 8 - Rules
- 1. + Rules Mariano Rodriguez-Muro, Free University of Bozen-Bolzano
- 2. + Disclaimer License This work is licensed under a Creative Commons Attribution-Share Alike 3.0 License (http://creativecommons.org/licenses/by-sa/3.0/)
- 3. + Reading material Chapter 7. Semantic Web Programming Chapter 6. Foundations of Semantic Web.
- 4. + What are rules Means of representing knowledge An ontology language Can be seen as conditions statements: if [] then [] if x is a Man and x hasChild y Father(x)
- 5. + Why rules Easy to understand RDFS (and sometimes OWL) is not enough: If a person x is the brother of somebody y, and there is z such that z is the uncle of x, then z is the uncle of y too. for all x,y,z, if hasBrother(x,y) & hasUncle(x,z) then hasUncle(y,z)
- 6. + SWRL
- 7. + 7 SWRL How can datalog and OWL be combined? SWRL – Semantic Web Rule Language [swirl] • Proposal for a rule based Semantic Web Language (W3C member submission) http://www.w3.org/Submission/SWRL/ • Idea: datalog rules referring to RDF classes/properties Symbols in rules can be OWL identifiers or new symbols • Various further features and syntactic forms • Support in inference engines (very wide spread support)
- 8. + SWRL Syntaxes Two syntaxes: 1. Abstract syntax based on “Extended Backus-Naur Form notation 1. Directly translatable to a Human readable syntax (SPARQL lie) 2. XML Concrete Syntax 1. Parsable by existing OWL/RDF/XML parsers 2. Suitable as a exchange language for the semantic web
- 10. + Abstract Syntax Example For the rule: The corresponding abstract syntax:
- 11. + Example 1 A very common use for rules is to move property values from one individual to a related individual Express the fact that the style of an art object is the same as the style of the creator. Artist(?x) & artistStyle(?x,?y) & Style(?y) & creator(?z,?x) ⇒ style/period(?z,?y) Implies(Antecedent(Artist(I-variable(x)) artistStyle(I-variable(x) I-variable(y)) Style(I-variable(y)) creator(I-variable(z) I-variable(x))) Consequent(style/period(I-variable(z) I-variable(y))))
- 12. + RDFS as rules Provide a translation of all RDFS axiom as rules subClassOf subPropertyOf domain range
- 13. + Beyond RDFS Transitivity of a property P Reflexivity of a property P Inverse of a property P Equivalence of properties/classes Sufficient conditions for class membership
- 14. + Built ins One of the main motivations for SWRL rules Provide means to manipulate data Comparisons Mathematical transformations List operators Modifiers for strings, dates and times Boolean and URI checks URI construction
- 15. + Built-ins examples hasAddress(?creditCardMachine, ?ccAddress) ^ hasAddress(?customer, ?custAddress) ^ hasState(?ccAddress, ?ccState) ^ hasState(?custAddress, ?custState) ^ swrlb:equal(?custState, ?ccState) ^ hasBirthday(?customer, ?bDate) ^ swrlb:subtractYearMonthDurations(?diff, ?today, ?bDate) ^ swrlb:greaterThanOrEqual(?diff, "P18Y0M") → LegalCigaretteBuyer(?customer)
- 16. + Built-ins examples foaf:Person(?person) ^ foaf:gender(?person, "female") ^ foaf:name(?person, ?name) → swrlb:stringConcat(?s, "Dear Ms. ", ?name, ":") ^ hasFormalGreeting(?person, ?s)
- 17. + Empty antecedents/consequent Empty antecedent indicate the consequent is ALWAYS true State class/property disjointness Empty consequents indicate the antecedent is a contradiction State a FK-like constraint State number-restriction constraints Allow to define FACTs and CONSTRAINTS
- 18. + XML Concrete Syntax • Based on OWL XML Presentation Syntax and RuleML • The SWRL XML syntax uses the OWL XML Ontology root element and some of its subelements: • VersionInfo • PriorVersion • BackwardCompatibleWith • IncompatibleWith • Imports • Annotation • Class • EnumeratedClass •SubClassOf •EquivalentClasses •DisjointClasses •DatatypeProperty •ObjectProperty •SubPropertyOf •EquivalentProperties •Individual •SameIndividual •DifferentIndividuals
- 27. + XML Elements (cont) • builtInAtom Built in atoms include: Comparison, Mathematical transformations, List operators, Modifiers for strings, dates and times, Boolean and URI checks, URI constructions
- 28. + When RDFS/Rules RDFS/OWL Expressivity for RDFS/OWL is enough Use is knowledge sharing Application requires high performance/uses specialized reasoner Application requires standard behavior Rules Expressivity only captures by rules Use is application behavior Application requires complex reasoning Application relies on a particular inference engine only
- 29. + Conclusion Rules are a VERY expressive (powerful) ontology language Rules can capture most inferences provided by RDFS (part of OWL too) Rules shouldn’t be abused. Performance of rule based reasoning is suboptimal w.r.t. RDFS/OWL specialized reasoning. Rules can be a powerful mean to capture application behiavor. Note, although SWRL is the defacto rule language for the semantic web, there are new standards, i.e. RIF, that might take over in the future (further information in FSW Chapter 6)