Multilingual Semantic Wiki - Review 2012


Multilingual Semantic Wiki

  • multilingual
    • multilingual interface for editing and querying
    • synchronized multilingual content
  • semantic
    • notions of consistency, entailment, automatic Q&A
  • wiki
    • collaborative
    • fast and accessible

Existing background technologies

  • Grammatical Framework and its Resource Grammar Library
  • Attempto Controlled English (ACE): first-order language with English syntax
  • AceWiki

Main tasks of WP11

  • port the ACE grammar to 15 MOLTO languages (assisted by UGOT)
  • develop AceWiki into a multilingual controlled language wiki-system
    • modularize and extend AceWiki
    • integrate existing GF tools for translation and smart text input
  • evaluate the new multilingual wiki-system (assisted by UHEL)

AceWiki

  • goal: user-friendly expressive semantic wiki system
  • collaborative editing (web-based tool)
  • background reasoning language: OWL
    • expressive fragment of first-order logic
    • decidable reasoning tasks: consistency checking, question answering, ...
    • complex syntax
  • front-end language: ACE
    • subset of natural English
    • well-defined translation into first-order logic
    • end-user documentation: construction and interpretation rules
  • see more: http://attempto.ifi.uzh.ch/acewiki/

Bidirectional ACE-to-OWL mapping

Every country that does not border a sea is a landlocked-country.

SubClassOf(
   ObjectIntersectionOf(
      :country
      ObjectComplementOf(
         ObjectSomeValuesFrom(
            :border
            :sea
         )
      )
   )
   :landlocked-country
)

Which country is a landlocked-country?

ObjectIntersectionOf(
    :country
    :landlocked-country
)

Article (screenshot)


Look-ahead editor (screenshot)


Reasoning (screenshot)


AceWiki in MOLTO

  • preserve the features of the AceWiki
    • controlled natural language
    • Attempto Controlled English
    • notions of consistency, automatic question answering, ...
    • wiki aspects (collaborative editing, multiple interlinked articles)
    • look-ahead editing
  • extend AceWiki to be multilingual
    • base the implementation on GF

Generalization of AceWiki

  • multiple languages
    • natural: English, German, ACE, ...
    • formal: ACE, Sage, ...
    • languages for content, UI, meta information
  • multiple grammars
    • ACE (or its subsets)
    • Math Grammar Library
    • Phrasebook
    • ...
  • multiple reasoners
    • ACE-based (RACE, OWL-based, ...)
    • math reasoners, e.g. Sage, WolframAlpha
    • ...

Multiple languages

  • multiple languages for
    • content
    • UI (labels etc.)
    • meta queries (authors, edits)
  • the content
    • viewable/editable/queryable in multiple languages
    • automatically kept in sync
  • some languages are formal, i.e. they are (mainly) meant for the reasoners
  • implemented using Grammatical Framework (GF)
    • single abstract syntax corresponds to multiple concrete syntaxes
    • Resource Grammar Library

(Stable AceWiki: single language: OWL-compatible subset of ACE)


Multiple grammars

  • different grammars in different sections of the wiki
  • interaction between grammars
  • support for changing the grammars
    • full grammar editor
    • UI for adding/editing words and their forms
  • implemented by integrating multiple GF grammars and a GF editor

(Stable AceWiki: single grammar: OWL-compatible subset of ACE)


Multiple reasoners

Syntactic vs semantic reasoning

  • syntactic
    • queries over GF trees to answer e.g. "for a given sentence, show me all the syntactically similar sentences"
  • semantic (e.g. ACE-based, math reasoning)
    • consistency checking
    • Q&A
    • ...
    • explanation of reasoning results
  • implemented by using external reasoning tools

(Stable AceWiki: single reasoner: OWL-reasoner with multiple implementations)


Current integration with GF

is based on the GF Webservice and the GF online editor.

  • storage is now based on (the language-neutral) abstract trees. (In case of standard AceWiki, the storage is based on ACE.)
  • access to multiple grammars (PGF)
  • multilingual viewing and editing of wiki content
  • look-ahead editing of sentences
  • presentation of the GF-analysis of sentences (translations, word alignment diagrams, GF syntax trees, ...)
  • preliminary grammar editing using the online GF editor
  • GF webservice accessed via a new Java API

AceWiki and the online GF editor (screenshot)


Immediate future work

  • structuring the content
    • multiple articles
    • sentence order
  • viewing/querying the content
    • dynamic views based on queries over the GF abstract trees

ACE in GF

  • starting point: Angelov and Ranta (2009)
    • based on ACE v6.0 (early 2009)
    • covers 7 languages (multilinguality via the GF RGL)
    • languages: Eng, Ger, Ita, Swe, Fin, Fre, Urd
  • issues with Angelov and Ranta (2009)
    • coverage and precision not perfect with respect to ACE v6.6
    • no AceWiki-oriented subset
    • small number of languages
  • new developments (https://github.com/Attempto/ACE-in-GF)
    • minor updates and fixes
    • new AceWiki-oriented subset (less overgeneration in the AceWiki context)
    • work on a precise AceWiki-oriented subset in GF continues as a collaboration between UZH and UGOT

Dissemination

  • paper or poster at the CNL 2012
    • August 2012 in Zurich, Switzerland
  • AceWiki repository on GitHub
  • Attempto mailing list