Ontologies, the semantic web and RDF Lecture 2 at Masaryk University Nils Pharo Content –bibliographic languages –document languages and work languages –subject languages - the LIS way –ontologies - the CS way 19 NOVEMBER, 2014 –An ontology is an explicit specification of a conceptualization. – –What does this mean? 19 NOVEMBER, 2014 Bibliographic languages –Document languages –"A document is a particular space-time embodiment of information: a document language describes and provides access to this embodiment." (p. 107) –Work languages –"describe information entities, their intellectual (as opposed to physical) attributes, and relationships among them." (p. 87) – –Svenonius (2000). The Intellectual Foundation of Information Organization 19 NOVEMBER, 2014 Document languages –Production language –Carrier language –Location language 19 NOVEMBER, 2014 Purpose of document languages –for describing the material embodiment - the manifestation of the work – –its physical and carrier attributes –its publication attributes –its external access attributes 19 NOVEMBER, 2014 FRBR – 19 NOVEMBER, 2014 Work languages –Author languages –Title languages –Edition languages –Subject languages –Classification languages –Index languages 19 NOVEMBER, 2014 Author, title and edition languages –controlled and uncontrolled vocabularies –normalized name forms for authority files –uncontrolled names for descriptive cataloging 19 NOVEMBER, 2014 Subject languages –organized with respect to semantic strongness –"free keywords" –keyword lists –taxonomies –thesauri –faceted classification 19 NOVEMBER, 2014 Keyword lists –the most primitive form of controlled vocabulary –biology –horses –primates –psychology –wars 19 NOVEMBER, 2014 Taxonomy –hierarchical keyword list where terms are organized as subtypes/supertypes – –Animals –Cats –Dogs –Horses – –Food –Bread –Butter –Vegetables 19 NOVEMBER, 2014 Taxonomy – 19 NOVEMBER, 2014 Thesauri –CHEFS –UF Cooks –BT Catering personnel –RT Food preparation – –Aitchison, Gilchrist & Bawden (2000). Thesaurus construction and use: a practical manual. (p. 164) 19 NOVEMBER, 2014 Thesaurus – 19 NOVEMBER, 2014 UF/USE UF/USE RT Thesaurus construction rules –three types of relationships: hierarchical, equivalence and associative –scope notes are used to provide definitions, restrict use, clarify content of term etc –standards (ISO 2788 and ISO 5964) that prescribe the implementation of the relationships –recommendations for what associative relationships to be realised 19 NOVEMBER, 2014 Facetted classification –Wine –by region –France –Germany –Italy –by colour –Red –White –Rose –by price –less than 100 NOK –between 100 and 200 NOK –more than 200 NOK 19 NOVEMBER, 2014 – 19 NOVEMBER, 2014 Colour Country Price Charateristics of facetted classification –no "standard" –guidelines, e.g. Spiteri (1998). A simplified model for facet analysis –Ranganathan's Colon classification 19 NOVEMBER, 2014 Enter ontologies –Original definition (from philosophy): the branch of metaphysics dealing with the nature of being. – –Adapted by computer scientists to facilitate artificial intelligence: "An ontology is an explicit specification of a conceptualization.[...] For AI systems, what "exists" is that which can be represented. When the knowledge of a domain is represented in a declarative formalism, the set of objects that can be represented is called the universe of discourse." Gruber (1993). Toward Principles for the Design of Ontologies Used for Knowledge Sharing 19 NOVEMBER, 2014 Ontologies –Eye –Synonym: –Orbital part of face –Orbital region –Part: –Upper eyelid –Lower eyelid –From: Digital Anatomist Foundational Model of Anatomy ontology 19 NOVEMBER, 2014 Domain ontologies and top ontologies –Domain ontologies models a specific domain, e.g. the human body, libraries, bread etc. –Top ontologies describe concepts that are sharable across many domains. 19 NOVEMBER, 2014 Ontology components –instances (individuals, entities, things) –classes (types) –properties (attributes, characteristics) –relationships (relations) –rules and constraints 19 NOVEMBER, 2014 Classes and instances –instances represent concrete individuals or objects –classes represent the collection of objects or individuals –classes may contain other classes – –Nils Pharo is an instance of the class person 19 NOVEMBER, 2014 Properties –used to denote aspects of the classes, instances and relationships – – Fido – $1000 – 5 years – boxer – – – –etc. 19 NOVEMBER, 2014 Relationships –specify how objects are related to other objects in the ontology, the most prominent being –hierarchical superclass/subclass-relationships –dog mammal –part-relationships –tail dog – –However, other forms of hierarchical relationships as well as relationships representing associative relations can be implemented 19 NOVEMBER, 2014 The structure of ontologies –a hierarchical basic structure –properties can be inherited –from superclass to subclass –Mammal has hair –Dog has hair –instances can belong to multiple classes –Fido dog –Fido brown thing 19 NOVEMBER, 2014 Rules and constraints –to secure against illogical inferences, specify cardinality, and clarify the kinds of statements than can be used for specific classes, e.g. – –an animal cannot be both a carnivore and a herbivore –an employee needs to be at a certain level of authencity to get access to high-security information –a month cannot have more than 31 days 19 NOVEMBER, 2014 Purpose of ontologies –model (a restricted part of) the world –to make it possible for computers to infer things about the world –needs to be explicit! –Open world assumption; a statement may be true irrespective of whether or not it is known to be true (Wikipedia article) 19 NOVEMBER, 2014 Ontology questions –how to model? –are instances part of the ontology? –what is the appropriate level of abstraction? –what do the classes/instances refer to? 19 NOVEMBER, 2014 Ontologies compared to old knowledge organization systems –more flexible –less standardized – –solution for merging and sharing –solutions for identity 19 NOVEMBER, 2014 Ontology standard languages –Topic maps (ISO) –RDF/OWL (W3C) 19 NOVEMBER, 2014 Group work –Create a simple ontology on a topic of your own choice! 19 NOVEMBER, 2014 Content –the semantic web –why do we need it? –the RDF standard –interoperability recapitulated –the data silo problem –linked data 20 NOVEMBER, 2014 The semantic web –"The Semantic Web provides a common framework that allows data to be shared and reused across application, enterprise, and community boundaries. It is a collaborative effort led by W3C with participation from a large number of researchers and industrial partners. It is based on the Resource Description Framework (RDF)." (http://www.w3.org/2001/sw/) 20 NOVEMBER, 2014 Challenges for the World Wide Web –The current Web is challenged on several areas –too much noise –internal systems with bad communication capabilities (data silos) –large costs of communication 20 NOVEMBER, 2014 The noise problem –more sophisticated IR-systems might help a bit –needs more sophisticated mark-up –e.g. –Cantilever bridge 20 NOVEMBER, 2014 Internal systems problem –difficult to share data –difficult to compare data –difficult to reuse data 20 NOVEMBER, 2014 Costs of communication –End-users need to –collect –interpret –compare –connect – –information themselves 20 NOVEMBER, 2014 Is the semantic web the solution? –partial solution –problem/domain dependent 20 NOVEMBER, 2014 The web for computer applications –the SW is not intended to be interpreted by humans –data semantically marked up and structured to be processed by intelligent agents –SW is an extension of the Web –SW is a web of data 20 NOVEMBER, 2014 – – – – – – – – – – –From Tim Berners-Lee’s 1989 proposal 20 NOVEMBER, 2014 Ontologies, modelling domains for the semantic web –"An ontology is an explicit representation of a conceptualization" (Gruber, 1992) –concepts and concretes modelled as classes (man) –relationships ( mammal) –properties ( –constraint rules – –to provide a: "shared and common understanding of a domain that can be communicated between people and application systems" Towards the semantic web, 2003 20 NOVEMBER, 2014 Technologies for developing the semantic Web –W3C standard technologies –XML –RDF –RDF Schema (RDFS) –OWL (Web Ontology Language) 20 NOVEMBER, 2014 XML –XML represents internal metadata to the item/document – –Example: Tim Berners-Lee 20 NOVEMBER, 2014 RDF –W3C standard (recommendation, 22.02.99) –http://www.w3.org/RDF –semantic Web - http://www.w3.org/2001/sw/ –tool for embedding metadata in digital documents 20 NOVEMBER, 2014 RDF describes –things (subjects) –properties (predicate) –values (objects) – –preferably identified by URIs 20 NOVEMBER, 2014 Domain independent –RDF is a domain independent data model –RDF describes triples representing things that have properties with values – –Nils Pharo is a teacher of Digital knowledge organization 20 NOVEMBER, 2014 Relational database model Books Isbn Author Title 1-932394-20-6 Thomas B. Passin Explorer's guide to the Semantic Web 0-262-19433-3 Elaine Svenonius The intellectual foundation of information organization 0-8050-8043-8 David Weinberger Everything is miscellaneous 20 NOVEMBER, 2014 RDF model 20 NOVEMBER, 2014 RDF example –http://www.jbi.hio.no/bibin/dig_korg/sem_web.htm has a creator whose value is Nils Pharo – –In RDF/XML syntax: – Nils Pharo – 20 NOVEMBER, 2014 Bibliographic RDF example – Information Architecture for the World Wide Web Peter Morville –Louis Rosenfeld O'Reilly 2006 en – 20 NOVEMBER, 2014 Bibliographic example 2 – A review of Information Architecture for the World Wide Web, 3rd edition Lee McKusick PenLUG 2006-23-12 20 NOVEMBER, 2014 Notation 3 –A simpler syntax for human readiability – Nils Pharo Semantic Web –equals: –<@prefix dc: . <@prefix rdf: . dc:Creator "Nils Pharo" ; dc:Subject "Semantic Web" . – 20 NOVEMBER, 2014 RDF describes instances –The rdf:type property can be used to state that a resource is an instance of a class –RDF schema is a simple ontology language –OWL is a full ontology language – 20 NOVEMBER, 2014 RDF schema –RDF schema is used for defining RDF terminologies –RDF schema is a type system for RDF –RDF schema makes semantic information machine-accessible –RDF schema is a simple ontology language –Example: the statement "Nils Pharo is a teacher of Digital knowledge organization" can be used to deduce that "Nils Pharo is a member of the academic staff" and that "Nils Pharo is involved with Digital knowledge organization" –key components: class, subclass relations, property, subproperty relations, domain and range constraints – 20 NOVEMBER, 2014 RDFS example – – 20 NOVEMBER, 2014 RDFS example 2 – – 20 NOVEMBER, 2014 RDFS example 3 (Notation 3-format) –@prefix rdfs: . @prefix rdf: < http://www.w3.org/1999/02/22-rdf-syntax-ns# >. a rdfs:Class . a rdfs:Class ; rdfs:subClassOf 20 NOVEMBER, 2014 OWL -Web ontology language –funded on DAML+OIL –OWL is a richer ontology language than RDF schema –3 versions supporting different levels of complexity; Full, DL, and Lite – –can be used to specify that: "academic staff members must teach at least one course" or "every book must have a title" – 20 NOVEMBER, 2014 OWL elements –OWL uses RDF, RDF schema and its own terminology to define ontologies: –Web Ontology Language: OWL by Grigoris Antoniou and Frank van Harmelen which includes an example of an OWL-defined ontology –W3 org's OWL guide – 20 NOVEMBER, 2014 Assignment –Model the hierarchical parts of the ontology you constructed previously with the use of RDF and RDFS 20 NOVEMBER, 2014