SKOS. COMP60421 Sean Bechhofer

Transcription

1 SKOS COMP60421 Sean Bechhofer

2 Ontologies Metadata Resources marked-up with descriptions of their content. No good unless everyone speaks the same language; Terminologies Provide shared and common vocabularies of a domain, so search engines, agents, authors and users can communicate. No good unless everyone means the same thing; Ontologies Provide a shared and common understanding of a domain that can be communicated across people and applications, and will play a major role in supporting information exchange and discovery. 2

3 Ontology A representation of the shared background knowledge for a community Providing the intended meaning of a formal vocabulary used to describe a certain conceptualisation of objects in a domain of interest A vocabulary of terms plus explicit characterisations of the assumptions made in interpreting those terms Nearly always includes some notion of hierarchical classification (is-a) Richer languages allow the definition of classes through description of their characteristics 3

4 A Spectrum of Representation Catalogue Thesauri Formal is-a Value Restrictions Terms/ glossary Informal is-a Frames Expressive Logics Formal representations are not always the most appropriate for applications 4

5 COHSE Conceptual driven navigation around documents Simple text processing + vocabulary + open hypermedia architecture Separating link and document Explicit navigation around a domain vocabulary DLS agent adds links to documents based on the occurrence of concepts in those documents. 5

6 Demo 6

7 COHSE s Architecture HTML Document in Ontology DLS Agent Knowledge Service Resource Service SKOS Search Engine Linked HTML Document out Annotation DB 7

8 Generic Links Generic Links in Open Hypermedia are based on words. Linkbase Link Service Document Linked Document 8

9 Generic Links + Thesaurus A thesaurus can bridge gaps between terms. Thesaurus Linkbase Link Service Document Linked Document 9

10 Generic Links + Ontology An ontology can bridge gaps between concepts. Ontology Linkbase Link Service Document Linked Document 10

11 Reflection Our original approach involved the use of OWL ontologies to support the conceptual models. Over time, we came to see this as a mistake -- looser vocabularies were perhaps more appropriate. The timely appearance of SKOS. S. Bechhofer, Y. Yesilada, R. Stevens, S. Jupp, and B. Horan. Using Ontologies and Vocabularies for Dynamic Linking IEEE Internet Computing12(3), p /MIC

12 SKOS SKOS: Simple Knowledge Organisation Scheme Used to represent term lists, controlled vocabularies and thesauri Lexical labelling Simple broader/narrower hierarchies (with no formal semantics) W3C Recommendation 12

13 Primary Use Cases/Scenarios A. Single controlled vocabulary used to index and then retrieve objects Query/retrieval may make use of some structure in the vocabulary B. Different controlled vocabularies used to index and retrieve objects Mappings required between the vocabularies Also other possible uses (e.g. navigation) 13

14 SKOS Goals to provide a simple, machine-understandable, representation framework for Knowledge Organisation Systems (KOS) that has the flexibility and extensibility to cope with the variation found in KOS idioms that is fully capable of supporting the publication and use of KOS within a decentralised, distributed, information environment such as the world wide (semantic) web. 14

15 SKOS A model for expressing basic structure of concept schemes Thesauri, classification schemes, taxonomies and other controlled vocabularies Many of these already exist and are in use in cultural heritage, library sciences, medicine etc. A wide range of knowledge sources that can potentially provide value for Semantic Web applications SKOS aims to provide an RDF vocabulary for the representation of such schemes. A migration path bringing such resources into the Semantic Web. 15

16 Concept Schemes A concept scheme is a set of concepts, potentially including statements about relationships between those concepts Semantic Relationships Broader/Narrower Terms Related Terms Lexical Labels Preferred, alternative and hidden labels Additional documentation Notes, comments, descriptions 16

17 Knowledge Organisation Thesaurus: Controlled vocabulary in which concepts are represented by preferred terms, formally organised so that paradigmatic relationships between the concepts are made explicit, and the preferred terms are accompanied by lead-in entries for synonyms or quasi-synonyms. Thesaurus Taxonomy Authority File Synonym Ring Controlled Vocabulary Related Terms Hierarchy Preferred Terms Equivalent Terms Collection of Terms Controlled vocabularies: designed for use in classifying or indexing documents and for searching them. 17

18 SKOS Example 18

19 SKOS and OWL SKOS and OWL are intended for different (but related) purposes SKOS Concept schemes are not formal ontologies in the way that, e.g. OWL ontologies are formal ontologies. There is no formal semantics given for the conceptual hierarchies (broader/narrower)represented in SKOS. Contrast with OWL subclass hierarchies which have a formal interpretation (in terms of sets of instances). A weaker ontological commitment. 19

20 SKOS and OWL SKOS Concepts not intended for instantiation in the same way that OWL Classes are instantiated Leo is an instance of Lion Born Free is a book about Lions Concept Schemes allow us to capture general statements about things that aren t necessarily strictly true of everything It s useful to be able to navigate from Cell to Nucleus, even though it s not the case that all Cells have a Nucleus Relationships between Polio and Polio virus, Polio vaccine, Polio disease Relationships between Accident and Accident Prevention, Accidents in the Home, Radiation Accidents But we can t necessarily draw the same kinds of inferences about SKOS hierarchies. Broader hierarchy is not transitive. Although mechanisms are available which allow us to query the transitive closure of the hierarchy. 20

21 SKOS and OWL SKOS itself is defined as an OWL ontology. A particular SKOS vocabulary is an instantiation of that ontology/schema E.g. SKOS Concept is a Class, particular concepts are instances of that class Allows us to use some of the mechanisms of OWL to define properties of SKOS (e.g. the querying of the transitive closure of broader). Allows us to use generic tooling to construct/maintain our vocabularies 21

22 Annotation in OWL OWL data and object properties allow us to define the characteristics of classes Necessary/sufficient conditions etc. Model theory/semantics provides interpretations of the assertions involving the properties Ontology engineering (and use) also requires annotation Decoration of concepts/properties/individuals with information which is useful, but does not impact on the formal semantics or logical interpretations Separation of the concept from its concrete label is usually seen as a Good Thing. 22

23 Annotation General Labels Human readable Textual Definitions Scope notes DC style metadata authorship Change History Provenance information Application Specific Entry points for forms Driving User interaction Syntax round-tripping Hiding engineering aspects of the model Methodological support, e.g. OntoClean 23

24 SKOS as Annotation SKOS labelling and documentation properties are defined as OWL Annotation Properties Preferred/Alternate/Hidden Labels Documentation/Notes SKOS then provides a standardised vocabulary for annotating OWL ontologies Leverage existing tooling. OWL API Protégé 24

25 SKOS and OWL SKOS and OWL are intended for different purposes. OWL allows the explicit modelling/description of a domain SKOS provides vocabulary and navigational structure Interaction between representations is ongoing work. Presenting OWL ontologies as SKOS vocabularies Principled dumbing down Enriching SKOS vocabularies as OWL ontologies. How to handle related Use of SKOS as annotation vocabulary 25

26 Mapping Concept Schemes SKOS also provides a collection of mapping properties that express relationships between concepts in different schemes broadmatch/narrowmatch closematch exactmatch Support alignment of different concept schemes Indiscriminate use of properties such as owl:sameas can lead to undesirable consequences. 26

27 SKOS and Linked Data Linked Data standardised guidelines for publishing data URIs for identification Provide useful information when dereferenced Link to other URIs SKOS as lightweight semantics for LD Facilitating publication of existing KOS/data. Mapping relationships SKOS Indexing/Retrieval Semantic Relations Mapping LD Discovery Navigation Linking and Integration beyond URI matching 27

28 Tooling: SKOSEd Editor supporting construction of SKOS vocabularies Native SKOS implementation Protégé 4 plugin exploiting OWL definition of SKOS vocabulary Reasoning support for classification Lexical labelling Alternate language support Extension points for domain relationships 28

29 Examples Astronomy thesauri: AGROVOC (FAO) IPSV UMBEL E-Culture Europeana ONKI (Finland) LCSH 29

30 Resources SKOS:! SKOSEd:! COHSE:! 30

31 Linked Data and RDF Sean Bechhofer

32 Building a Semantic Web Annotation Associating metadata with resources Integration Integrating information sources Inference Reasoning over the information we have. Could be light-weight (taxonomy) Could be heavy-weight (logic-style) Interoperation and Sharing are key goals 6

33 Linked Data* Linked Data or the Data Web is about using the Web to connect related data that wasn t previously linked. The intention is that we move from a web of documents to a web of data The Web as database The Linked Data approach builds heavily on RDF. *Linked data slides based on material from Ian Davis and Tom Heath: 8

34 Database tables isbn title author publisherid pages Q&A Vikas Swarup X The Rotters Club Jonathan Coe

35 Rows represent things isbn title author publisherid pages Q&A Vikas Swarup X The Rotters Club Jonathan Coe

36 Columns represent properties isbn title author publisherid pages Q&A Vikas Swarup X The Rotters Club Jonathan Coe

37 Intersections represent properties of things isbn title author publisherid pages Q&A Vikas Swarup X The Rotters Club Jonathan Coe

38 Graphical Representation book title The Rotters Club more generally: subject property value 13

39 Selecting multiple properties isbn title author publisherid pages Q&A Vikas Swarup X The Rotters Club Jonathan Coe

40 Multiple properties graphically title The Rotters Club book isbn X author Jonathan Coe 15

41 Relations between things title The Rotters Club book isbn X publisher author Jonathan Coe publisher name Penguin Books 16

42 Identification We need to be able to identify things globally and uniquely. URIs provide this capability Key to Linked Data is the use of URIs, specifically URIs. 17

43 URIs in graphs URIs as names for nodes Jonathan Coe URIs as names for relations 18

44 URIs and naming URIs identify the things we are describing. If two people create data using the same URI, the assumption is that they are describing the same thing. Merging/integrating data then becomes easy Although introduces issues of URI control. 19

45 Graph Merging Jonathan Coe Birmingham 20

46 Graph Merging Jonathan Coe Birmingham 21

47 URIs are active URIs can be more than just names -- they can be dereferenced, and information can be retrieved. In particular, we can lookup the URIs in a graph and potentially retrieve more information about the URI. Follow your nose navigation Information should be returned in appropriate, machine readable formats (e.g. another graph) 22

48 Linked Data Principles 1. Use URIs as names for things 2. Use http URIs so that those names can be dereferenced. 3. When a URI is looked up, provide useful information 4. Include statements that link to other URIs so that more information can be discovered. Common infrastructure facilitates construction of applications. Largely browsers up to now. Other guidelines relating to connecting documents with the data that describes them. Use of content negotiation to supply appropriate representations Use of microformats/rdfa to publish data 23

49 RDF RDF stands for Resource Description Framework It is a W3C Recommendation RDF is a graphical formalism ( + concrete syntax) for representing metadata for describing the semantics of information in a machineaccessible way Provides a simple data model based on triples. Allows us to represent relationships between things. 24

50 The RDF Data Model Statements are <subject, predicate, object> triples:! <Sean,hasColleague,Uli> Can be represented as a graph: Sean Statements describe properties of resources A resource is any object that can be pointed to by a URI Properties themselves are also resources (URIs) hascolleague Uli 25

51 Linking Statements The subject of one statement can be the object of another Such collections of statements form a directed, labeled graph hasname Sean K. Bechhofer Sean hascolleague hascolleague Uli hashomepage Carole Note that the object of a triple can also be a literal (a string) 26

52 RDF Syntax RDF has an XML syntax that has a specific meaning: Every Description element describes a resource Every attribute or nested element inside a Description is a property of that Resource We can refer to resources by URIs <Description about="some.uri/person/sean_bechhofer"> <hascolleague resource="some.uri/person/uli_sattler"/> <hasname rdf:datatype="&xsd;string">sean K. Bechhofer</hasName> </Description> <Description about="some.uri/person/uli_sattler"> <o:hashomepage> </Description> <Description about="some.uri/person/carole_goble"> <o:hascolleague resource="some.uri/person/uli_sattler"/> </Description> 27

53 What does RDF give us? A mechanism for annotating data and resources. Single (simple) data model. Syntactic consistency between names (URIs). Low level integration of data. The Linked Data/Web of Data approach. 28

54 RDF(S): RDF Schema RDF gives a formalism for meta data annotation, and a way to write it down in XML, but it doesn t give any special meaning to vocabulary such as subclassof or type Interpretation is an arbitrary binary relation RDF Schema extends RDF with a schema vocabulary that allows you to define basic vocabulary terms and the relations between those terms Class, Property type, subclassof range, domain 29

55 RDF(S) These terms are the RDF Schema building blocks (constructors) used to create vocabularies: <Person,type,Class> <hascolleague,type,property> <Professor,subClassOf,Person> <Carole,type,Professor> <hascolleague,range,person> <hascolleague,domain,person> Semantics gives extra meaning to particular RDF predicates and resources specifies how terms should be interpreted 30

56 RDF(S) Inference rdf:type rdfs:class rdfs:subclassof rdfs:subclassof Person Academic rdf:type rdf:type rdf:subclassof Lecturer 31

57 RDF(S) Inference rdf:type rdfs:class Academic rdfs:subclassof rdf:type rdfs:type Lecturer rdf:type Sean 32

58 RDF/RDF(S) Liberality No distinction between classes and instances (individuals) <Species,type,Class> <Lion,type,Species> <Leo,type,Lion> No distinction between language constructors and ontology vocabulary, so constructors can be applied to themselves/each other <type,range,class> <Property,type,Class> <type,subpropertyof,subclassof> In order to cope with this, RDF(S) has a particular non-standard model theory. 33

59 What does RDF(S) give us? Ability to use simple schema/vocabularies when describing our resources. Consistent vocabulary use and sharing. Basic inference 34

60 Problems with RDF(S) RDF(S) is too weak to describe resources in sufficient detail No localised range and domain constraints! Can t say that the range of haschild is Person when applied to Persons and Elephant when applied to Elephants No existence/cardinality constraints! Can t say that all instances of Person have a mother that is also a Person, or that Persons have exactly 2 parents No transitive, inverse or symmetrical properties! Can t say that ispartof is a transitive property, that haspart is the inverse of ispartof or that touches is symmetrical Difficult to provide reasoning support No native reasoners for non-standard semantics May be possible to reason via FO axiomatisation 35

61 OWL OWL: Web Ontology Language Extends existing Web standards Such as XML, RDF, RDFS Is (hopefully) easy to understand and use Based on familiar KR idioms Of adequate expressive power Formally specified Possible to provide automated reasoning support But you already know all this What about SKOS??? 36

62 SKOS and Linked Data Linked Data standardised guidelines for publishing data URIs for identification Provide useful information when dereferenced Link to other URIs SKOS as lightweight semantics for LD Facilitating publication of existing KOS/data. Mapping relationships SKOS Indexing/Retrieval Semantic Relations Mapping LD Discovery Navigation Linking and Integration beyond URI matching 32

63 Microformats, RDFa Microformats: open data formats for embedding data in pages. Fixed schemas for calendars, address books, opinions, licenses etc. RDFa: general mechanism for embedding RDF metadata within pages. No fixed schema Used with Linked Data resources E.g. LCSH: 33

64 Linked Data Benefits Separation of data from formatting and presentational aspects Self-describing data. Applications encountering unfamiliar vocabularies can dereference and access definitions Simplified data access via HTTP and RDF. Heterogeneity of Web APIs Open Applications not implemented against fixed set of data sources. 34

65 Linked Data Examples Universities dbpedia RDFised version of wikipedia Scraping structured information from info-boxes. Quality? BBC Programme catalogues published as Linked Data Crosslinking to resources like dbpedia and MusicBrainz GeoNames Geographical data Lat/long, postal codes etc. LCSH SKOS 37

66 dbpedia 36

67 Ordnance Survey 37

68 Open University 38

69 BBC 39

70 Library of Congress 40

71 fishdelish 41

72 42

73 43

74 data.gov.uk Open Public Data Empowering citizen developers But pushback against formats URI Design Hash vs slash Statistical information Tensions Centralised vs Distributed Rules vs Guidance Trust in the Web! Darwinian Evolution 38

75 LD in Use Five Stars of Open Linked Data Make data available! Make it available as structured data! Use non-proprietary formats! Use URLs to identify things! Link your data Costs and Benefits 45

76 Issues with Linked Data Identity and co-reference How do we handle the fact that different URIs may be used to refer to the same things? Use of owl:sameas may be too strong (can result in all information, including annotations, metadata etc.) being merged. Versioning Version information in URLs? Versioning at architectural level (Memento) How does versioning play with a follow your nose paradigm? Querying Distributed query across data sets LD applications tend to use an extract, transform, load approach. 39

77 Issues with Linked Data Still primarily about data Vocabularies used to facilitate integration Little deep semantics. Big O vs little o! Role of SKOS and RDF(S) Scalability Focus on mechanisms for data publication rather than consumption Lots of work on recipes, mangling relation sources into RDF etc. What do you actually do with the stuff? Applications (so far) mainly browsing/cv generation/introspective apps End user applications? Build it and they will come.??? 40