of 33 lecture 1: introduction. of 33 the semantic web vision today’s web (1) web content – for...
TRANSCRIPT
of 33
lecture 1: introduction
of 33
the semantic web visiontoday’s web (1)
web content – for human consumption (no structural information)
people search for information, keep in touch with other people, make purchases, look for entertainment … minimal support by software tools
ece 720, winter'12 2
of 33
the semantic web visiontoday’s web (2)
keyword-based search engines … with some problems:
high recall & low precision – too many documents retrieved
low/no recall – no relevant documents found results sensitive to vocabulary – similar words not considered
results are single web pages – “distributed” information not retrieved
human needed for interpretation, no other software tools use the result
ece 720, winter'12 3
of 33
the semantic web visiontoday’s web (3)
a problem with semantics …so… a need for different representation of information, and intelligence that takes advantage of that
ece 720, winter'12 4
of 33
the semantic web vision… the future (1)
knowledge management key activity of large businesses (knowledge seen as intellectual asset)
information is in a weakly structured form
problems with: sharing, extracting, marinating, viewing
Semantic Web -> knowledge organized, tools for knowledge maintenance, query answering tools
ece 720, winter'12 5
of 33
the semantic web vision… the future (2)
business-to-customer electronic commerce browsing on-line shops, selecting the best alternative/offer -> time and effort consuming activity
decisions making based on limited/partial information
Semantic Web -> software “shopping” agents able to interpret the product information and the terms of service, and conduct negotiations
ece 720, winter'12 6
of 33
the semantic web vision… the future (3)
business-to-business electronic commerce data exchange using the Electronic Data Interchange (EDI) approach (complex, isolated)
lack of standards
Semantic Web -> differences in terminology resolved (abstract domain models), automatic translation of services, auctioning, negotiations, and drafting of contracts
ece 720, winter'12 7
of 33
the semantic web vision… the future (4)
wikis collections of pages that allow users to add content (collaborative knowledge creation)
Semantic Web -> usage of semantic technologies (enriching structured text and untyped hyperlinks with semantic annotation referring to an underlying model of the captured knowledge)
ece 720, winter'12 8
of 33
the semantic web vision… the future (5)
personal agents: a future scenario
Michael has some neck pain, he asked his Semantic Web agent to arrange for him physical therapy sessions which were suggested by Michael’s physician …
ece 720, winter'12 9
of 33
semantic web technologiesmetadata (1)
HTML is the predominant language used for writing web pages
humans have no problem reading itmachines are not able to distinguish different people, things, follow links to find out more about topics – simple they have no clue what a given text is about
ece 720, winter'12 10
of 33
semantic web technologiesmetadata (2)
<company> <companyName>… </companyName><staff>
<director> … </director><viceDirector> …
</viceDirector></staff>
</company>
ece 720, winter'12 11
of 33
semantic web technologiesmetadata (3)
HTML is replaced by a language that carries information about a contentmetadata – data about data, capture part of the meaning of data (tags that are “defined” in a systematic way)
ece 720, winter'12 12
of 33
semantic web technologiesontology (1)study of nature of existence, how to describe things that actually existworld is made up of specific objects that can be grouped into abstract classes based on shared propertiesit describes a domain of interest
ece 720, winter'12 13
of 33
semantic web technologiesontology (2)consists of a finite list of terms and the relationships between themterms denote concepts (classes of objects) of the domainrelationships include hierarchies of classes + properties of objects, value restrictions, logical relationships between objects
ece 720, winter'12 14
of 33
semantic web technologiesontology (3)
provide “a shared understanding of a domain”overcome differences in terminology
ece 720, winter'12 15
of 33
semantic web technologiesontology languagesRDF – is a data model for objects and relations between themRDF Schema – a description language for describing properties and classes of RDF resources, provides a semantics for generalization hierarchiesOWL – a rich language for describing properties and classes, and relations between them
ece 720, winter'12 16
of 33
semantic web versus AIlogic (1)provides principles of reasoningprovides well-understood formal semantics“unlocks” automated reasoners
leads to uncovering ontological knowledge given implicitly
ece 720, winter'12 17
of 33
semantic web versus AIlogic (2)
used by intelligent agents for making decisions and selecting courses of action
provides explanations for conclusions
ece 720, winter'12 18
of 33
semantic web versus AIlogic (3)
SWRL – semantic web rule language for constructing rules based on ontology classes and relationships
ece 720, winter'12 19
of 33
semantic web versus AIagents (1)
software agents work autonomously and proactively
ece 720, winter'12 20
of 33
semantic web versus AIagents (2)tasks of personal software agent in SW: receive tasks and preferences from users seek information on the web, communicate with other agents
compare information from different sources, make choices
give answer to users
ece 720, winter'12 21
of 33
semantic web versus AIagents (3)metadata identify and extract information
ontology web search, information interpretation communication with other agents
logic information retrieval, conclusion derivation
ece 720, winter'12 22
of 33
a layered approach
development of SW in steps
each step building a layer on the top of another
ece 720, winter'12 23
of 33
a layered approach
ece 720, winter'12 24
of 33
a layered approach
ece 720, winter'12 25
alternative stack
of 33
a layered approachXML layer syntactic basis
RDF layer basic data model for facts RDF Schema simple ontology language
ontology layer more expressive languages - OWL
ece 720, winter'12 26
of 33
a layered approachlogic layer enhance ontology, application specific knowledge
proof layer proof generation, validation
trust layer digital signature, rating agents …
ece 720, winter'12 27
of 33
the semantic web vision… the present (1)
linked data
a new paradigm built on one of the layers of SW – RDF core
ece 720, winter'12 28
of 33
the semantic web vision… the present (2)
linked data
each piece of information is linked to other pieces
different types of linksdifferent sources of data
ece 720, winter'12 29
of 33
the semantic web vision… the present (3)
linked data
every piece of information presented as a triple:subject-property-object
ece 720, winter'12 30
of 33
the semantic web vision… the present (4)
linked data
a vast, interconnected web of triples (pieces of information)easy to query with SPARQL – a special query language for RDF
ece 720, winter'12 31
of 33
the semantic web vision… linked data
linked data
ece 720, winter'12 32
of 33
future web …
intelligent representationand processing of data
ece 720, winter'12 33