48
Exploiter le Web Sémantique, le comprendre et y contribuer (dans cet ordre) Mathieu d’Aquin KMi, The Open University – [email protected] Le reste des diapos sont principalement en anglais… The other slides are mostly in English

Exploiter le Web Semantic, le comprendre et y contribuer

Embed Size (px)

DESCRIPTION

Talk at the Nancy Unviersity to students of the Master SCA. (Unlike the title, most of the slides are in english).

Citation preview

Page 1: Exploiter le Web Semantic, le comprendre et y contribuer

Exploiter le Web Sémantique, le comprendre et y contribuer(dans cet ordre)

Mathieu d’Aquin KMi, The Open University – [email protected]

Le reste des diapos sont principalement en anglais…

The other slides are mostly in English

Page 2: Exploiter le Web Semantic, le comprendre et y contribuer

Outline of the talk

Exploiter le Web Sémantique, le comprendre et y contribuer

1.2.

3. 4.

Page 3: Exploiter le Web Semantic, le comprendre et y contribuer

Outline of the talk

Exploiter le Web Sémantique, le comprendre et y contribuer

?

Page 4: Exploiter le Web Semantic, le comprendre et y contribuer

The Semantic Web (in theory)

A large scale, heterogenous collection of formal, machine processable, ontology-based statements (semantic metadata) about web resources and other entities in the world, expressed in a standard syntax

<rdf:RDF> <owl:Ontology rdf:about=""> <owl:imports rdf:resource="http://usefulinc.com/ns/doap#"/> </owl:Ontology> <j.1:Organization rdf:ID="KMi"> <rdfs:comment rdf:datatype="http://www.w3.org/2001/XMLSchema#string" >The Knoledge Media Institute of the Open University, Milton Keynes UK</rdfs:comment> </j.1:Organization> <j.1:Document rdf:ID="KMiWebSite"> … <rdf:RDF>

<channel rdf:about=“http://watson.kmi.open.ac.uk/blog”><title>Elementaries - The Watson Blog</title><link>http://watson.kmi.open.ac.uk:8080/blog/</link><description>"Oh dear! Where the Semantic Web is going to go now?" -- imaginary user 23</description><language>en</language><copyright>Watson team</copyright><lastBuildDate>Thu, 01 Mar 2007 13:49:52 GMT</lastBuildDate><generator>Pebble (http://pebble.sourceforge.net)</generator><docs>http://backend.userland.com/rss</docs>…

Page 5: Exploiter le Web Semantic, le comprendre et y contribuer

Metadata

UoD

<rdf:RDF><channel rdf:about=“http://watson.kmi.open.ac.uk/blog”><title>Elementaries - The Watson Blog</title><link>http://watson.kmi.open.ac.uk:8080/blog/</link><description>"Oh dear! Where the Semantic Web is going to go now?" -- imaginary user 23</description><language>en</language><copyright>Watson team</copyright><lastBuildDate>Thu, 01 Mar 2007 13:49:52 GMT</lastBuildDate><generator>Pebble (http://pebble.sourceforge.net)</generator><docs>http://backend.userland.com/rss</docs>…

<rdf:RDF> <foaf:Image rdf:about='http://static.flickr.com/132/400582453_e1e1f8602c.jpg'> <dc:title>Zen wisteria</dc:title> <dc:description></dc:description> <foaf:page rdf:resource='http://www.flickr.com/photos/xcv/400582453/'/> <foaf:topic rdf:resource='http://www.flickr.com/photos/tags/vittelgarden/'/> <foaf:topic rdf:resource='http://www.flickr.com/photos/tags/wisteria/'/> <dc:creator> <foaf:Person><foaf:name>Mathieu d'Aquin</foaf:name> …

<rdf:RDF> <owl:Ontology rdf:about=""> <owl:imports rdf:resource="http://usefulinc.com/ns/doap#"/> </owl:Ontology> <j.1:Organization rdf:ID="KMi"> <rdfs:comment rdf:datatype="http://www.w3.org/2001/XMLSchema#string" >The Knoledge Media Institute of the Open University, Milton Keynes UK</rdfs:comment> </j.1:Organization> <j.1:Document rdf:ID="KMiWebSite"> …

FOAF

DCRSS TAP

WORDNET

NCI Galen

Music

…… …

Page 6: Exploiter le Web Semantic, le comprendre et y contribuer

<rdf:RDF> <foaf:Image rdf:about='http://static.flickr.com/132/400582453_e1e1f8602c.jpg'> <dc:title>Zen wisteria</dc:title> <dc:description></dc:description> <foaf:page rdf:resource='http://www.flickr.com/photos/xcv/400582453/'/> <foaf:topic rdf:resource='http://www.flickr.com/photos/tags/vittelgarden/'/> <foaf:topic rdf:resource='http://www.flickr.com/photos/tags/wisteria/'/> <dc:creator> <foaf:Person><foaf:name>Mathieu d'Aquin</foaf:name> …

<rdf:RDF> <owl:Ontology rdf:about=""> <owl:imports rdf:resource="http://usefulinc.com/ns/doap#"/> </owl:Ontology> <j.1:Organization rdf:ID="KMi"> <rdfs:comment rdf:datatype="http://www.w3.org/2001/XMLSchema#string" >The Knoledge Media Institute of the Open University, Milton Keynes UK</rdfs:comment> </j.1:Organization> <j.1:Document rdf:ID="KMiWebSite"> …

Ontology alignmentData integration

Data analysisReasoning

Etc.

Smart Application

Page 7: Exploiter le Web Semantic, le comprendre et y contribuer

Therefore…

• Many research and development efforts in – Supporting the design of ontologies (methodologies,

toolkits, editors, etc.)

– Supporting the annotation Web resources (natural language processing, information extraction, etc.)

– Supporting the publication of semantic data and information online (linking open data, `semantification’ of legacy information systems)

– …

• Resulting in an explosion of the amount of machine processable knowledge online.

Page 8: Exploiter le Web Semantic, le comprendre et y contribuer

OK, nice… but what’s the reality?

2007 2008 2009

Page 9: Exploiter le Web Semantic, le comprendre et y contribuer

Slide 9

And for ontologies?

Page 10: Exploiter le Web Semantic, le comprendre et y contribuer

Slide 10

And for ontologies?

Page 11: Exploiter le Web Semantic, le comprendre et y contribuer

So, the Semantic Web in reality?

• Millions of Semantic Web documents (data), containing billions of RDF triples

• Thousands of ontologies online in OWL and RDFs, covering many different domains (will talk about that later)

• But, distributed and heterogeneous in representation, meaning, quality…

• So, what do we really do with it?

Page 12: Exploiter le Web Semantic, le comprendre et y contribuer

Outline of the talk

Exploiter le Web Sémantique, le comprendre et y contribuer

?

Page 13: Exploiter le Web Semantic, le comprendre et y contribuer

Next Generation Semantic Web Applications

NG SW Application Semantic WebSmart Features

Able to exploit the Semantic Web at large – Dynamically retrieving the relevant semantic resources – Combining at run-time heterogeneous and distributed

Ontologies

Page 14: Exploiter le Web Semantic, le comprendre et y contribuer

Next Generation Semantic Web Applications

Dynamically retrieving, exploiting and combining relevant semantic resources from the SW, at large

Need for a Gateway to the Semantic Web

Page 15: Exploiter le Web Semantic, le comprendre et y contribuer

Watson: a Gateway to the Semantic Web

Page 16: Exploiter le Web Semantic, le comprendre et y contribuer

Architecture

Page 17: Exploiter le Web Semantic, le comprendre et y contribuer

Interface

http://watson.kmi.open.ac.uk

Page 18: Exploiter le Web Semantic, le comprendre et y contribuer

But the important part is: the APIs

• Provide Semantic Web application developers with the ability to efficiently:

– Locate (find) Semantic Web documents online using advanced search functions

– Explore the documents, automatically extracted metadata and content

– Query the documents, to exploit online knowledge in an homogeneous way

• In a set of lightweight APIs, and without having to download the data or use any other dedicated infrastructure.

Page 19: Exploiter le Web Semantic, le comprendre et y contribuer

Some Applications We Developed

Ontology Reuse:The Watson Plugin

Question Answering:PowerAqua

Semantic Browsing:PowerMagpie

Semantic Relation Discovery:Scarlet

Folksonomy Enrichment

And also:Word sense disambiguationQuery ExpansionSynonym DiscoveryWeb Service Annotation…

Page 20: Exploiter le Web Semantic, le comprendre et y contribuer

Example: The Watson Plugin

Chose an entity to search

Get entities from online ontologies

Integrate statements Into the edited ontology

Page 21: Exploiter le Web Semantic, le comprendre et y contribuer

Example: Scarlet

ka2.rdf

Researcher AcademicStaff

Sem

anti

c W

eb

Researcher

AcademicStaff

ISWC SWRCHam SeaFood

Sem

anti

c W

eb

HamSeaFood

Meat

Meat

SeaFood

Agrovoc NALT

pizza-to-go

wine.owl

NALT

Page 22: Exploiter le Web Semantic, le comprendre et y contribuer

Example: Scarlethttp://scarlet.open.ac.uk/

Page 23: Exploiter le Web Semantic, le comprendre et y contribuer

Example: Wahoo http://watson.kmi.open.ac.uk/wahoo

Page 24: Exploiter le Web Semantic, le comprendre et y contribuer

Example: PowerAqua

Natural language question

Answers from online semantic data

Page 25: Exploiter le Web Semantic, le comprendre et y contribuer

Example: FLOR

Can the Semantic Web provide the structure needed to improve search and navigation of tagged spaces?

Page 26: Exploiter le Web Semantic, le comprendre et y contribuer

Search in Tag Spaces

5/24 ≈ 21% relevant

Dog Dog

DogDog

Bird

Bird

Bird

Bird

Bird

Bird

Bird

Tiger

Tiger

Tiger

Tiger

CatLandscape

Landscape

Landscape

Let’s find photos of “animals which live in the water”

Query: Animal Water

Page 27: Exploiter le Web Semantic, le comprendre et y contribuer

Bring in the SW…

Dolphin

Seal

Marine Mammal

Mammal

Sea

livesIn

Whale

Body of Water

Ocean

Sea Elephant

FishlivesIn

Animal

FreshwaterFish SaltwaterFish

livesIn

Animal Water

<Animal livesIn Water>

<Dolphin>or<Seal>or<“Sea Elephant”>or<Whale>

Page 28: Exploiter le Web Semantic, le comprendre et y contribuer

Results

dolphin

seal

whale

sea elephant

18/24 ≈ 75% relevant

Page 29: Exploiter le Web Semantic, le comprendre et y contribuer

SWEET: Semantic Annotation of REST services

Page 30: Exploiter le Web Semantic, le comprendre et y contribuer

And so?

• These are only a few of the applications developed in KMi (i.e., us, the people who are doing Watson)

• Many other people are developing such applications (and we don’t necessarily know all of them)

• Many other tools exists that help building applications (triple stores, query engines, other Semantic Web search engines)…

• But what does that tell us about the Semantic Web?

Page 31: Exploiter le Web Semantic, le comprendre et y contribuer

Outline of the talk

Exploiter le Web Sémantique, le comprendre et y contribuer

?

Page 32: Exploiter le Web Semantic, le comprendre et y contribuer

Watson as a Research Platform

• Watson collects a lot of ontologies and Semantic Web documents that are created by different people for different purposes

• In addition to being a gateway for the development of applications exploiting this knowledge, it can be used to better understand how knowledge is published online, how the Semantic Web looks like, and how it evolves

Page 33: Exploiter le Web Semantic, le comprendre et y contribuer

Characterizing and subset of the Watson Collection (2007)

Number of entities

Domain covered

Underlying description logic

Page 34: Exploiter le Web Semantic, le comprendre et y contribuer

Understanding relations between ontologies online

• Ontologies are naturally related with each other, some are equivalent to others, some are versions, some are similar, some are incompatible to each other

• These relations generally stay implicit

Better understanding these relations is useful to support the use of the Semantic Web

Page 35: Exploiter le Web Semantic, le comprendre et y contribuer

DOOR: An Ontology of Ontology Relations

Page 36: Exploiter le Web Semantic, le comprendre et y contribuer

Example Relation: Different Versions

• Ontologies evolve on the Web, there are many different versions of the same ontology are available

• This is rarely made explicit through the appropriate metadata for ontologies (e.g., owl:preVersion)

• But version info is often encoded in the URIs of ontologies, e.g., http://loki.cae.drexel.edu/wbs/ontology/2003/10/iso-metadata http://loki.cae.drexel.edu/wbs/ontology/2004/01/iso-metadata

• Extracting this information can help in studying the evolution of ontologies on the Semantic Web, i.e., the Semantic Web dynamics

Page 37: Exploiter le Web Semantic, le comprendre et y contribuer

Example Relation: Different Versions – Initial Experiment

• We developed simple method based on a few rules recognizing specific patterns in the differences between URIs of ontologies (dates, timestamps, etc.) and ran it on a set of 6,898 ontologies from Watson.

• We found 155,501 (directed) versioning relations between these ontologies, which represent 1,365 evolving ontologies

• A manual evaluation indicates that more than 50% of these are correct

• Next step: improve the method and study evolution patterns on the Semantic Web

Page 38: Exploiter le Web Semantic, le comprendre et y contribuer

Example Relations: Agreement and Disagreement

• Ontologies are knowledge artifacts, they express opinions and beliefs and contradict each others

• Assessing (dis)agreement in ontologies is very useful to understand how to combine knowledge from different sources

• A possible approach would be to check whether inconsistencies and incoherencies appear while combining the ontologies. However we believe that:– There are di erent levels of agreement/disagreement ff

– Covering di erent domains is not agreeing ff

– It is possible to agree and disagree at the same time

• Based on these requirements we define a set of measures for assessing (dis)agreement between statements and ontologies.

Page 39: Exploiter le Web Semantic, le comprendre et y contribuer

Example Relations: Agreement and Disagreement - Measures

• Agreement(st, O) [0..1] and Disagreement(st, O) [0..1] where st is a statement <subject, predicate, object> and O is an ontology

• Based on extracting the part of the ontology that express a relation between subject and object

• (Dis)agreement between ontologies:

• Global (dis)agreement in a repository

• Consensus:

• Controversy:

Page 40: Exploiter le Web Semantic, le comprendre et y contribuer

Example Relations: Agreement and Disagreement – Application?

• Experiment: assessing statements related to the class Seafood in Watson:

a: global agreement, d: global disagreement, cs: consensus, ct: controversy

• More experiments on the Way!

Page 41: Exploiter le Web Semantic, le comprendre et y contribuer

Using 21 ontologies containing a concept SeaFood

Agreement

Disagreement

Camp 1: seaFood disjointWith MeatCamp 2: SeaFood subClassOf Meat

Page 42: Exploiter le Web Semantic, le comprendre et y contribuer

Outline of the talk

Exploiter le Web Sémantique, le comprendre et y contribuer

?

Page 43: Exploiter le Web Semantic, le comprendre et y contribuer

Slide 43

From a Semantic Web search engine…

Page 44: Exploiter le Web Semantic, le comprendre et y contribuer

Slide 44

… to Ontology Repositories?

Ontologies

Ontology Metadata

Versions of

Alignments

Comments and

Reviews

Page 45: Exploiter le Web Semantic, le comprendre et y contribuer

Cu

pb

oard

.op

en.a

c.u

k

Page 46: Exploiter le Web Semantic, le comprendre et y contribuer

Sum

mary

Meta

data

Revie

ws

Page 47: Exploiter le Web Semantic, le comprendre et y contribuer

Final message

• I hope I convinced you that

Using the Semantic Web

Understanding the Semantic Web

Contributing to the Semantic Web

• Through Watson, Cupboard and our applications, our aim is to build an open and efficient platform making the Semantic Web a `playground for research and development’

• There is still a lot to do, and everybody is welcome to comment, help, contribute…

Page 48: Exploiter le Web Semantic, le comprendre et y contribuer

Thank You!

Mathieu d’Aquin ([email protected], http://people.kmi.open.ac.uk/mathieu) With contributions from many people in KMi (http://kmi.open.ac.uk) and the NeOn project (http://neon-project.org)

/* I would normally include a bibliography slide at the end, but all the relevant papers can be found on these 3 websites */