CCSW: The Competence Center Semantic Web

1
CCSWThe Competence Center Semantic Web

• Harold Boley, DFKI GmbH
• Presentation in Course Rule Markup Languages
• Univ. Kaiserslautern, April 26th, 2002

2
General Overview

• Semantic Web W3C Activity on machine-interpreted
  documents that can be used (not just for display
  but) for automation, integration, and reuse
  across applications (http//www.w3.org/2001/sw/ac
  tivity)
• DFKI has long been working in Semantic Web
  technologiesDescription logics, ontologies,
  metadata, rule systems, agents,NL parsing,
  information extraction, knowledge management,
  etc.
• Current CCSW focus at DFKI Robust Web-document
  authoring annotation for agent-based
  information management with webizedobject
  representations, ontologies rule systems
• CCSWs Semantic Web view Higher-level system
  emerging from increasingly structured subwebs,
  each serving needs of specific community

Co-Heads Dr. Harold Boley (Kaiserslautern), Dr.
Paul Buitelaar (Saarbrücken)
URL http//ccsw.dfki.de
Services Consulting, Studies Projects
3
Semantic Web and Web ServicesUse Databases and
Rule Systems
4
General DFKI SemWeb Areas

• Content Ontology Development
• Manual, Semi-Automatic Ontology Learning and
  Adaptation
• Specific for a Task, Organisation (IntraNet),
  Domain (ExtraNet)

• Applications Intelligent and Dynamic Information
  Integration and Access
• Intelligent Information Integration
• Intelligent, Cooperative Agents
• Content-Based Information Access
• Cross-Lingual and Multimedia Information Access
• Company- and User-Adaptive Information Systems
• Distributed Agent-Based Organizational Memories

5
Some SemWeb Applications_at_DFKI (I)

• Intelligent Information Integration
  Intelligent, Cooperative Agents
• SmartKOM Combination of User Modeling and Plan
  Recognition to Integrate Knowledge from
  Multimodal Sources

• Intelligent Information Integration
• MUMIS Ontology-Based Information Integration
  from Multilingual Sources

• Content-Based, Cross-Lingual Multimedia
  Information Access
• Combinations of Ontology-Based Information
  Extraction, Text Mining and Semantic Annotation
  for Knowledge Markup of Text or Multimedia
  Documents with Metadata for Content-Based,
  Cross-Lingual, Multimedia Information Access
• GETESS (Information Extraction, Text Mining),
  MuchMore (Semantic Annotation, Text Mining),
  MUMIS (Information Extraction, Multimedia)

6
Some SemWeb Applications_at_DFKI (II)

• Company- and User-Adaptive Information Systems
• Adaptive READ Document Retrieval on the Basis of
  Machine Learning
• Algorithms for Automatic IR-Parameter
  Optimization

• Distributed Agent-Based Organizational Memories
• FRODO Ontology Acquisition from Texts and User
  Interaction
• for Workflow Enactment and Information Access

7
The Semantic Web Layered Architecture
Tim Berners-Lee Axioms, Architecture and
Aspirations W3C all-working group plenary
Meeting 28 February 2001
(http//www.w3.org/2001/Talks/0228-tbl/slide5-0.ht
ml)
8
Present SemWeb Challenges

• Can we make W3Cs original Semantic Web notion
  more
• precise (Semantic) content data vs. metadata
  semantics?
• specific (Web) some intranets vs. the
  Internet?
• What techniques will semantic webs use from
  Information Retrieval, Databases, Ontologies,
  (Description, Horn) Logics, W3C Markup Languages
  (XML, RDF, XSLT), Knowledge Management, Agents,
  Web Services (WSDL), ...?
• Which semweb success stories (killer apps)
  exist (dmoz.org UNSPSC, eCl_at_ss , ECCnet)?
• How to rank candidate semweb applications for
  showing the semweb potentials in our own
  organizations and for our customers?

9
SemWeb Language Principles

• Existing (database, logic) languages can be
  webized (Tim Berners-Lee) by introducing URIs
  as a new kind of (constant) symbols
• The languages should be scalable to a large
  amount of Web-distributed content, hence should
  use a small, if not minimal, formalism
• A simple formalism doesnt interfere with the
  content
• Relational databases with SQL are a good example
• XML DTDs, the RDF model, the DAMLOIL core, and
  the modularized RuleML are such candidate
  languages (unlike, perhaps, XML Schema, the many
  RDF syntaxes, full DAMLOIL, or a monolithic
  RuleML)

10
SemWeb Core IssueMetadata Ontologies (I)

• For Web-page annotation, browsers should use a
  top-level pane/menu for metadata (cf. Annotea)
• Metadata should be generated interactively from
  content data, via standardized domain ontologies
  (NLP tools/resources for metadata extraction
  annotation)
• Search engines should show same ontologies for
  navigating-searching content with high precision
• Information agents may also use the ontologies
  for retrieving and integrating content for users

11
SemWeb Core IssueMetadata Ontologies (II)

• Instead of a single global ontology for
  metadata there will certainly be several local
  ontologies, which require integration, e.g. by
  alignment on demand or via derivation/transformati
  on rules
• Maintenance of domain ontologies for metadata
  must be machine-supported, e.g. by links and/or
  transformations between versions (cf. MeSH)
• Metadata ontologies can describe heterogeneous
  Web pages in a homogeneous format
• Some ontology queries provide direct answers
  (fact retrieval) others provide relevant Web
  pages (document retrieval) yet others, both

12
Web-Based B2C or B2B Rule Exchange
. . .
translate to standard format (e.g., RuleML)
publish rulebase1
publish rulebasem
compare, instantiate, and run rulebases
13
From Natural Language to Horn Logic
14
RuleML Markup and Tree
''The discount for a customer buying a product is
5.0 percent if the customer is premium and the
product is regular.''
ltimpgt lt_headgt ltatomgt
lt_oprgtltrelgtdiscountlt/relgtlt/_oprgt
ltvargtcustomerlt/vargt ltvargtproductlt/vargt
ltindgt5.0 percentlt/indgt lt/atomgt
lt/_headgt lt_bodygt ltandgt ltatomgt
lt_oprgtltrelgtpremiumlt/relgtlt/_oprgt
ltvargtcustomerlt/vargt lt/atomgt
ltatomgt lt_oprgtltrelgtregularlt/relgtlt/_oprgt
ltvargtproductlt/vargt lt/atomgt
lt/andgt lt/_bodygt lt/impgt
15
Intertranslating RuleML and RFML
''The discount for a customer buying a product is
5.0 percent if the customer is premium and the
product is regular.''
ltimpgt lt_headgt ltatomgt
lt_oprgtltrelgtdiscountlt/relgtlt/_oprgt
ltvargtcustomerlt/vargt ltvargtproductlt/vargt
ltindgt5.0 percentlt/indgt lt/atomgt
lt/_headgt lt_bodygt ltandgt ltatomgt
lt_oprgtltrelgtpremiumlt/relgtlt/_oprgt
ltvargtcustomerlt/vargt lt/atomgt
ltatomgt lt_oprgtltrelgtregularlt/relgtlt/_oprgt
ltvargtproductlt/vargt lt/atomgt
lt/andgt lt/_bodygt lt/impgt
lthngt ltpattopgt ltcongtdiscountlt/congt
ltvargtcustomerlt/vargt ltvargtproductlt/vargt
ltcongt5.0 percentlt/congt lt/pattopgt
ltcallopgt ltcongtpremiumlt/congt
ltvargtcustomerlt/vargt lt/callopgt ltcallopgt
ltcongtregularlt/congt ltvargtproductlt/vargt
lt/callopgt lt/hngt
16
Current Players

• USA W3C, DARPA, NSF, Maryland, Stanford,
  ...
• Canada NRC-IIT-CISTI, ...
• Europe IST
• Netherlands Amsterdam, Twente, ...
• UK Manchester, Newcastle, ...
• France INRIA , ...
• Germany Karlsruhe, DFKI, Hannover, Hamburg,
  Berlin, IW-Köln, ...
• Sweden Linköping
• Switzerland MCM
• Japan INTAP, Keio, CARC, Ricoh, ...
• Korea KAIST
• Australia Melbourne, ...
• . . .

17
Major Funding

• USA DAML, W3C Web Ontology Working Group
• Canada NRC
• Europe OntoWeb, Semantic Web Technologies
• Japan METI
• . . .
• Canada Europe ISTEC
• Japan Europe ?
• . . .

18
SemWeb Courses

• University of Maryland
• Stanford University
• Lehigh University
• Vrije Universiteit Amsterdam
• Universität Karlsruhe
• Universität Kaiserslautern
• Universität Saarbrücken
• ...