Language Technologies and the Semantic Web: An Essential Relationship.

1
Language Technologies and the Semantic Web An
Essential Relationship.

• Enrico Motta
• Professor of Knowledge Technologies
• Knowledge Media Institute
• The Open University

2
Content of the Talk

• Update on the Semantic Web
• Beyond the hype
• What it is
• Why it is interesting
• Whats its status?
• Semantic Web and AI
• Semantic Web Applications
• Key features
• Reasoning on the Semantic Web
• Key role of Language Technologies
• Conclusions

3
The Semantic Web in 2 minutes
4
(No Transcript)
5
(No Transcript)
6
ltfoafPerson rdfabout"http//identifiers.kmi.ope
n.ac.uk/people/enrico-motta/"gt
ltfoafnamegtEnrico Mottalt/foafnamegt
ltfoaffirstNamegtEnricolt/foaffirstNamegt
ltfoafsurnamegtMottalt/foafsurnamegt ltfoafphone
rdfresource"tel[phone]"/gt
ltfoafhomepage rdfresource"http//kmi.open.ac.uk
/people/motta/"/gt ltfoafworkplaceHomepage
rdfresource"http//kmi.open.ac.uk/"/gt
ltfoafdepiction rdfresource"http//kmi.open.ac.u
k/img/members/enrico.jpg"/gt ltfoaftopic_interest
gtKnowledge Technologieslt/foaftopic_interestgt
ltfoaftopic_interestgtSemantic Weblt/foaftopic_inte
restgt ltfoaftopic_interestgtOntologieslt/foaftopi
c_interestgt ltfoaftopic_interestgtProblem
Solving Methodslt/foaftopic_interestgt
ltfoaftopic_interestgtKnowledge Modellinglt/foaftop
ic_interestgt ltfoaftopic_interestgtKnowledge
Managementlt/foaftopic_interestgt
ltfoafbased_neargt ltgeoPointgt
ltgeolatgt52.024868lt/geolatgt
ltgeolonggt-0.707143lt/geolonggt
ltcontactnearestAirportgt
ltairportnamegtLondon Luton Airportlt/airportnamegt
ltairportiataCodegtLTNlt/airportiataCodegt
ltairportlocationgtLuton, United
Kingdomlt/airportlocationgt
ltgeolatgt51.866666666667lt/geolatgt
ltgeolonggt-0.36666666666667lt/geolonggt
ltrdfsseeAlso rdfresource"http//www.daml.org/cg
i-bin/airport?LTN"/gt ltfoafcurrentProjectgt ltf
oafProjectgt ltfoafnamegtAquaLoglt/foafnamegt
lt/foafcurrentProjectgt
7
The foaf ontology
8
The SW as Web of Data
9
Current status of the semantic web

• 10-20 million semantic web documents
• Expressed in RDF, OWL, DAMLOIL
• 7K-10K ontologies
• These cover a variety of domains - multimedia,
  computing, management, bio-medical sciences,
  geography, entertainment, upper level concepts,
  etc

The above figures refer to resources which are
publicly accessible on the web
10
The Semantic Web today

• To a significant extent the Semantic Web is
  already in place and is characterized by a
  widespread production of formalized knowledge
  models (ontologies and metadata), from a variety
  of different groups and individuals
• The Next Knowledge Medium - An information
  network with semi-automated services for the
  generation, distribution, and consumption of
  knowledge
• Stefik, 1986
• Knowledge modelling to become a new form of
  literacy?
• Stutt and Motta, 1997
• Still primarily a research enterprise, however
  interest is rapidly increasing in both
  governmental and business organizations
• early adopters phase
• The result is slowly emerging as an unprecedented
  knowledge resource, which can enable a new
  generation of intelligent applications on the web

11
Semantic Web Applications

• What can you do with the Semantic Web?

12
Corporate Semantic Webs

• A corporate ontology is used to provide a
  homogeneous view over heterogeneous data sources
• Often tackle Enterprise Information Integration
  scenarios
• Hailed by Gartner as one of the key emerging
  strategic technology trends
• E.g., see personal information management in
  Garlik

13
(No Transcript)
14
Exploiting large scale semantics
Next GenerationSW Applications
SemanticWeb
Semantic Web Gateway
15
Exploiting large scale semantics
Next GenerationSW Applications
SemanticWeb
16
NGSW Applications in the context of AI research
17
Knowledge-Based Systems
Today there has been a shift in paradigm. The
fundamental problem of understanding intelligence
is not the identification of a few powerful
techniques, but rather the question of how to
represent large amounts of knowledge in a fashion
that permits their effective use
Goldstein and Papert, 1977
Intelligent Behaviour
18
The Knowledge Acquisition Bottleneck
KA Bottleneck
Intelligent Behaviour
19
SW as Enabler of Intelligent Behaviour
Both a platform for knowledge publishing and a
large scale source of knowledge
Intelligent Behaviour
20
KBS vs SW Systems
Classic KBS SW Systems
Provenance Centralized Distributed
Size Small/Medium Extra Huge
Repr. Schema Homogeneous Heterogeneous
Quality High Very Variable
Degree of trust High Very Variable
21
Key Paradigm Shift
Classic KBS SW Systems
Intelligent Behaviour A function of sophisticated, logical, task-centric problem solving A side-effect of being able to integrate different types of reasoning to handle size and heterogeneous quality and representation
22
Next Generation SW Applications Examples

• Case Study 1 Automatic Alignment of Thesauri in
  the Agricultural/Fishery Domain

23
Method

• SCARLET - matching by Harvesting the SW
• Automatically select and combine multiple online
  ontologies to derive a relation

Access
Semantic Web
Scarlet
Deduce
Concept_A (e.g., Supermarket)
Concept_B (e.g., Building)
Semantic Relation ( )
24
Two strategies
Building
OrganicChemical
PublicBuilding
Lipid
Shop
Steroid
Steroid
Supermarket
Cholesterol
Semantic Web
Scarlet
Scarlet
Building
Cholesterol
OrganicChemical
Supermarket
(A)
(B)
Deriving relations from (A) one ontology and (B)
across ontologies.
25
Experiment

• Matching
• AGROVOC
• UNs Food and Agriculture
• Organisation (FAO) thesaurus
• 28.174 descriptor terms
• 10.028 non-descriptor terms
• NALT
• US National Agricultural
• Library Thesaurus
• 41.577 descriptor terms
• 24.525 non-descriptor terms

26
226 Used Ontologies
http//139.91.183.309090/RDF/VRP/Examples/tap.rdf
http//reliant.teknowledge.com/DAML/SUMO.daml
http//reliant.teknowledge.com/DAML/Mid-level-onto
logy.daml
http//reliant.teknowledge.com/DAML/Economy.daml
http//gate.ac.uk/projects/ htechsight/Technologie
s.daml
27
Evaluation 1 - Precision

• Manual assessment of 1000 mappings (15)
• Evaluators
• Researchers in the area of the Semantic Web
• 6 people split in two groups
• Results
• Comparable to best results for background
  knowledge based matchers.

28
Evaluation 2 Error Analysis
29
Other Case Studies
30
Giving meaning to tags
31
Example
Cluster_1 college commerce corporate course education high instructing learn learning lms school student
1http//gate.ac.uk/projects/htechsight/Employment.
daml. 2http//reliant.teknowledge.com/DAML/Mid-lev
el-ontology.daml. 3http//www.mondeca.com/owl/mos
es/ita.owl. 4http//www.cs.utexas.edu/users/mfkb/R
KF/tree/CLib-core-office.owl.
32
(No Transcript)
33
(No Transcript)
34
(No Transcript)
35
Conclusions
36
Typical misconceptions

• The SW is a long-term vision
• Ehmactually it already exists
• The SW will never work because nobody is going
  to annotate their web pages
• The SW is not about annotating web pages, the SW
  is a web of data, most of which are generated
  from DBs, or from web mining software, or from
  applications which produce SW data as a side
  effect of supporting users tasks
• The idea of a universal ontology has failed
  before and will fail again. Hence the SW is
  doomed
• The SW is not about a single universal ontology.
  Already there are around 10K ontologies and the
  number is growing
• SW applications may use 1, 2, 3, or even hundreds
  of ontologies.

37
SW and Language Technologies

• All the applications mentioned here combine
  language, web, statistical and semantic
  technologies
• Heterogeneity and sloppy modelling implies that
  language and statistical technologies are almost
  always needed when building NGSW apps
• In contrast with traditional KBS, intelligent
  behaviour is more a side-effect of intg. multiple
  techniques to handle scale and heterogeneity,
  rather than a function of powerful deductive
  reasoning

38
(No Transcript)