MoWGLI Mathematics in the Semantic Web

1
MoWGLI - Mathematics in the Semantic Web

• An Approach to Machine-Understandable
  Representation of the Information in Digital
  Documents
• Bernd Wegner
• Technical University, Berlin
• NA-MKM 2004, Phoenix, AZ

2
MoWGLI

• Mathematics on the Web - Get it by Logic and
  Interfaces
• http//www.cs.unibo.it/mowgli
• Funded by the 5. Framework Programme of the EU
• Start March 2002
• Duration 3 years
• Intermediate progress report in March 2003
  successful

3
MoWGLI - objectives

• Search and use mathematical content in the web
• Context related interpretation of formulas, links
  to proof assistants
• Content mark-up of electronic documents
• Links between machine-readable and
  human-understandable presentation

4
MoWGLI - objectives

• Exploitation of mark-up potentialities in
  documents for
• interoperability,
• automation,
• sophisticated search mechanisms,
• intelligent applications

5
MoWGLI - partners

• Department of Computer Science Bologna,
• project co-ordinator Andrea Asperti
• HELM Hypertextual Electronic Library of
  Mathematics
• http//www.cs.unibo.it/helm
• INRIA, Nice and Rocquencourt
• Lemme -project for writing software in scientific
  computing
• LogiCal -Project (Coq proof assistant) common
  features for proof checking and software security

6
MoWGLI - partners

• German Research Centre for Artificial
  Intelligence, DFKI, Saarbrücken/Kaiserslautern
• ActiveMath - a web based learning environment
• Relations to OmDoc and OpenMath
• KU Nijmegen
• automated theorem proving projects like
• TYPES, FTA,
• Calculemus,
• OpenMath

7
MoWGLI - partners

• Albert-Einstein-Institut (Max Planck Institute
  for Gravitational Physics), Golm/Potsdam
• Living Reviews in Relativity Theory
• test bed for mark-up
• development of LATEX-tools
• Trusted Logic, France
• Software quality and security aspects
• Technical University Berlin (associated to AEI)
• Exploitation and information dissemination

8
MoWGLI - initial actions

• Requirement analysis - done
• Investigations to determine the searchable items
  in a document, development of metadata schemes
  for their description - done
• Preparation of the test bed Living Reviews in
  Relativity - done, voluntary participation from
  other free journals - postponed
• Delivery of an exploitation and dissemination
  plan - done

9
MoWGLI general requirements

• Tools should hook on developments from third
  parties (open sources)
• Authors should be able to apply them easily
• No dependence from commercially protected
  software
• Possibilities for applications should be as wide
  as possible
• Chosen formats should be easily readable

10
MoWGLI problems and solutions

• Application of XML-technology is standard
• DFKI and Nijmegen/Eindhoven use OpenMath and
  OmDoc for the encoding of mathematical objects
• DFKI has created an OmDoc based mathematical
  database
• Presentation done by HTML/MathML
• MathML is supported by the main web-browsers
  Mozilla, Netscape, Internet-explorer

11
MoWGLI problems and solutions

• Scheme of translations
• XML to OmDoc, OmDoc to HTML/MathML via
  presentation XSLT
• http//www.w3.org/TR/xslt
• Conversion from Coq to OmDoc available
• Interactive course notes from Eindhoven, which
  are textual mathematical documents, enhanced with
  the possibility to interact with a computer
  algebra system through OpenMath objects

12
MoWGLI problems and solutions

• Representation and selection problems
• MathML can be produced using appropriate editors
  only
• Selection schemes for content mark-up have to be
  developed
• The same for formulas
• Unique and coordinated search facilities for
  formulas are required (different TEX-codings)

13
MoWGLI problems and solutions

• Proof assistants have to deal with two versions
• machine understandable vs.
• human understandable
• INRIA (Nice) provides applications (and
  conversions) for applying Coq to proofs in
  elementary geometry and incidence geometry (using
  TexMacs)
• AEI has delivered a LaTeX conversion to
  presentation MathML

14
MoWGLI problems and solutions

• Possibly TEX- and MathML extensions are requested
  to comply with further requirements
• First release for metadata schemes is available
  and compliant with other systems like OAI, EULER
  etc.
• An alternative approach to some problems is
  provided by MAYA from DFKI (with strong relations
  to OmDoc)
• OmDoc content dictionaries are available

15
MoWGLI open ends

• Controlled vocabulary still is missing, but
  essential for further progress
• Nothing has been and could be done with respect
  to graphics within the project. This is beyond
  the expertise of most of the partners in the
  current group.
• As far as presentation will be concerned the
  project will be able to deliver good results.
  Development of interactive input components will
  be beyond the capacity of the current project.

16
MoWGLI relations to the mathematical community

• Project is pursued by experts in knowledge
  management and artificial intelligence
• Comments from mathematicians are needed
  concerning
• Relevance of the ideas for mathematical
  publications
• Background information for semantic mark-up
  (names, notions, identification, keywords, etc.)
• User-friendliness of the developed tools
• Extension of the test bed Living Reviews in
  Relativity is highly desirable.

17
Adresses

• Bernd Wegner
• Mathematisches Institut, TU Berlin, Sekr. MA 8-1
• Straße des 17. Juni 135, D-10623 Berlin
• E-Mail
• Wegner_at_math.tu-berlin.de
• http//www.cs.unibo.it/mowgli
• http//www.emis.de