Upload
alexandra-thomson
View
213
Download
0
Embed Size (px)
Citation preview
Institut für Wissenstechnologien
1
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
Introduction toSemantic Technologies
Viktoria Pammer-Schindler, Wolfgang Kienreich
October 11, 2012
Institut für Wissenstechnologien
2
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
Syntax versus Semantics
Syntax
Greek συνταξις – arrangement, ordering
„The study of the principles and processes by which sentences are constructed in particular languages“ [Chomsky 1957]
Semantics
(grch. σημαντικος – about characters
“The study of interpretation of signs or symbols as used (…) within particular contexts” [Neurath et al. 1995]
Institut für Wissenstechnologien
3
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
Do you pme#?
Symbols
p,m,e,#
Rules
A sequence of # starts and ends every sentence
Each out of p,m,e is always preceeded by a sequence of #
Each out of p,m,e is always followed by a sequence of #
The e is always the last out of p,m,e to be used
The e is always used exactly once
#p#e#, ###m#p#e##, ##p#e#, #m#m#m#p##p##e###
p#p#e#, #p##e, ###
Institut für Wissenstechnologien
4
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
Semantics of pme#
#p#e##
#p##m#e##
###m##e#
#p##e#
##m##e####
###m##e#####
Institut für Wissenstechnologien
5
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
Another interpretation of pme#
#p#m#e##
##m##e####
###m##e######
#m#e##
#p##m#e##
###m##e#
Institut für Wissenstechnologien
6
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
contextualized information applicable for problem solving
„an ontology is a formal, explicit specification of a shared conceptualisation”
„knowledge is information in (contextualized) action (application)”
Semantics and Knowledge
[Gruber 1993]
[Kuhlen 1995]
[Kienreich 2012]
Institut für Wissenstechnologien
7
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
Semantic TechnologiesTechnologies for working with
contextualized information (knowledge?)
A priori
Explicitly model knowledge
A posteriori
Extract implicit knowledge
Institut für Wissenstechnologien
8
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
Semantic TechnologiesA Posteriori Peak
▼A Priori Summit
▼Mount NLP
▼
Machine Learning Ridge
▼Point
Statistic
▼
Analytical Cliff
▼
Knowledge Base
▼
First Order Ridge
▼Second Order Valley
▼Ontology hills
▼
Data Flow
▼
Institut für Wissenstechnologien
9
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
Semantic TechnologiesA Posteriori Peak
▼A Priori Summit
▼Mount NLP
▼
Machine Learning Ridge
▼Point
Statistic
▼
Analytical Cliff
▼
Knowledge Base
▼
First Order Ridge
▼Second Order Valley
▼Ontology hills
▼
Data Flow
▼
Institut für Wissenstechnologien
10
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
Chris Welty on Semantic Technologies as Bridge
Download video
Institut für Wissenstechnologien
11
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
Semantic Technologies
Working with contextualized information (knowledge?)
Common understanding: Semantic Technologies are about about encoding meaning and working with it
But there is more:
Semantic Search
Semantic Web
Semantic Mediation
…
Many areas (search…) are saturated with standard (statistical, non context-aware) methods. New, semantic methods are required for the next step
Institut für Wissenstechnologien
12
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
Semantic Technologies
Example: Semantic Search
Information retrieval, search solutions
Indexing of documents based on terms
Fast term lookup to find documents
Spell Checking, Query Expansion, Related Terms, Did You Mean, Result facetation …
Þ Need more than just terms in documentÞ Thessaurus approach to erichment and retrievalÞ Thessaurus is modelled knowledge
Institut für Wissenstechnologien
13
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
Semantic Technologies
Example: Identify semantics in news articles
„Wegen einer Presseaussendung der Staatsanwaltschaft Wien zu einer Hausdurchsuchung klagt der frühere Finanzminister Karl-Heinz Grasser die Republik Österreich. Mit der Klage will er seinen Anspruch auf Schadenersatz feststellen.
Die Staatsanwaltschaft Wien habe durch ihre Aussendung anlässlich der Hausdurchsuchungen beim Ex-Minister am 26. Mai 2011 Grassers Persönlichkeitsrechte verletzt und insbesondere seinen „wirtschaftlichen Ruf“ beschädigt, heißt es in der Klage beim Landesgericht für Zivilrechtssachen.
Konkret klagt Grassers Anwalt Michael Rami auf Feststellung eines Schadenersatzanspruches im Rahmen der Amtshaftung für die Justizorgane. Die Klage richtet sich gegen den Bund, der durch die Finanzprokuratur vertreten wird.“
Propose some methods, consider a priori and a posteriori
Institut für Wissenstechnologien
14
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
Semantic Web?
The WWW has problems
Content formatted to be human-readable
Many visions require machine understanding
Search
Cross-Language
Fact Validation
Contextualization
=> Semantic Web
Institut für Wissenstechnologien
15
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
Semantic Systems
The WWW is not alone with its problems
Enterprise Information Systems
Smart mobile systems
…Þ Semantic Systems
Ballance between
a-priori modelling to ease ai tasks
a-posteriori methods for problem solving
Institut für Wissenstechnologien
16
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
What do we need to do?
Extract knowledge (manually or automatic)
Model knowledge
Reason about knowledge (Expand, Verfiy, Proof…)
Make knowledge available (Query …)
Display knowledge (Visualization)
Institut für Wissenstechnologien
17
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
What do we have?
*lots* of technologies
A priori
knowledge representation: skos, rdf, rdfs, (xtm), owl
triple stores: jena, sesame
querying: sparql
A posteriori
Information retrieval
Machine learning
Social software, tagging, crowdsourcing…
Institut für Wissenstechnologien
18
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
What will we talk about
Focus on A priori
knowledge representation
data modelling
semantic technologies
storage
retrieval
related topics
Institut für Wissenstechnologien
19
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
Semantic (Web) Stack
http://en.wikipedia.org/wiki/File:Semantic-web-stack.png
Institut für Wissenstechnologien
20
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
Representing Semantics
S OP
With Triples
Institut für Wissenstechnologien
21
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
Modelling Standards
<Image removed from public slides – see http://
www.mkbergman.com/wp-content/themes/ai3/images/2008Posts/080606_SemanticTechnologies.jpg on site http://
www.mkbergman.com/445/no-semantic-technologies />
Institut für Wissenstechnologien
22
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
Storage Standards
<Image removed from public slides, see http://www.garshol.priv.no/blog/231.html >
Institut für Wissenstechnologien
23
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
Semantic (Web) Stack
<Image removed from public slides – see http://
bnode.org/media/2009/07/08/semantic_web_technology_stack.png >
Institut für Wissenstechnologien
24
V. Pammer-Schindler, W. Kienreich Oct 11, 2012 VU SemTech 2012 - Intro
Questions?