Dagstuhl Seminar on the Multilingual Semantic Web

http://dragoman.org/dagstuhl
Notes of M.T. Carrasco Benitez

• The Spanish government uses MT
• Seminar number 12362, 3-7 September 2012
• Schedule
• Wiki
• Participants
• Abstracts
• On multilingual web sites, position paper for this event
• Multilingual Web Sites Community Group
• Open architecture for multilingual web sites, working draft at the Community Group
• Standards for multilingual web sites, presentation at the MultilingualWeb, 4-5 April 2011, Pisa, Italy
• MultilingualWeb Project
• Open architecture for multilingual parallel texts, 2008 e-print at arXiv
• Multilingual Dataset Format (muset)
• W3C Internationalization (I18n) Activity
• Memento - Adding Time to the Web
• Linport
• Transparent Content Negotiation in HTTP
  • 4.7 Dimensions of negotiation
  • 4.8 Feature negotiation
• TMX 1.4b Specification
• Europa
• Official Journal of the European Union
• Location

• URI: A Uniform Resource Identifier (URI) is a compact sequence of characters that identifies an abstract or physical resource. RFC-3986 - URI syntax
• Resource:
  • whatever might be identified by a URI ... a collection of other resources. RFC-3986
  • A network data object or service that can be identified by a URI, as defined in section 3.2. Resources may be available in multiple representations (e.g. multiple languages, data formats, size, and resolutions) or vary in other ways. RFC-2616 - HTTP

Introduction

Brainstorming

• Users perspective
• Interaction & integration NLP/SW
• Open source OS
• Standard for Multilingual Web Sites (MWS)
• Corpora - linked open data (LOD)
• Unification theory
• SW & LR - infrastructure - institutional governance
• Multicultiralism
• Representation of formal languages for multilinguality
• Multiple views on data
• Standards
• Localization of SW resources
• Knowledge in language representation
• Business models for using SW technology
• Scalability of everything
• Cross-lingual links
• Underresource languages
• Benchmarking

• Extension of the LOD by a multilingual layer - language should be a dimension
• Language labeling - Primary Language in HTML
• Size of the data - access to a whole set (1T) and to a record (1K)
• Structure of the data - complex - simple pair
• Language symbol Ϣ - COPTIC CAPITAL LETTER SHEI - U+03E2
• My domain place: Multilingual web and resources (corpus, LOD)

Panel 1

• Roberto: interaction between NLP and semantic web communities - 7 - A
• Graere: linguistic representation issues - 8 - B
• Sebastian: development perspectives - 6 - C
• Sergei: content - 2 - B
• Hans: bootstrapping translingual web (translation, concept search) - monolingual, multilingual, translingual (crosslingual) - 8 - C
• Marko: economically viable transtion language to logic - 2 - B
• MultiJEDI - Roberto Navigli
• META-FORUM 2012 - Report
• META-NET Strategic Research Agenda for Multilingual Europe 2020
• WordNet

Group C: Sebastian + Hans - bootstrapping a translingual web

• Use the existing web standards
• Develop missing standards
• Multilingual Linked Open Data for Enterprises
• Multilingual Linked Open Data (mLOD) vs. Linguistic Linked Open Data Cloud (LLOD)
• Wikidata
• Infobox

Kimmo Rossi

• DG Connect - G.3 Data value chain - Language Technologies
• Public Sector Information (PSI)
• Big data
• CELLAR
• EU Open Data Portal
• epSOS
• e-CODEX

Reports of Panel 1

• Group C: Sebastian
• Group A: Roberto
• Group B: Josef van Genabith

• DBpedia
• YAGO

Panel 2

• Nicolectta - SW4RL & LR4SW - big data - Fostering Language Resources Network (FLaReNet)
• Christian - Interoperability - The TreeAligner Project - Linguistic Linked Open Data
• Asunción Gómez Pérez - transformations of LRs into LD - Monnet Project
• Nancy - MultiMASC/standardization - American National Corpus (ANC) - GrAF
• Jonh McCrae - lexicalization of LD - Lexicon Model for Ontologies (Lemon) - Lemon resource
• Martin Volk - parallel corpora - Sumat Project - Text+Berg Corpus - Bilingwis - database
• Ernesto William De Luca - user orientation and licensing
• Gerard de Melo - contribute to the LLOD - shared identifiers - Lexvo

Group C: multilingual parallel texts

• Raw
• Organised - Multilingual Dataset Format (muset)
• URI
• http://example.com/foo
• http://example.com/foo/de
• TaX - Tabularisation of XML data - tree to table
• Grammatical

Industrial & Application Perspective

• Jeroen van Grondelle, be informed - MOLTO
• Christian Lieske, Felix Sasaki

Reports of Panel 2

• Group A: Resources for semantic web - standards - OpenAnnotation
• Group B: LEMON
• Group C: Parallel corpora

Panel 3

• Dimitra Anastasiou - multi-modal dialog - XLIFF - Bremen Ambient Assisted Living Lab (BAALL)
• Bo Fu - collaborative support
• Huang Chu Ren - Telic semantic web - Kyoto Project
• Antoine Isaac - interoperability - Europeana - language codes:
  - Multilingual : mm - mul
  - Undetermined: un - und
  - No linguistic content: xx - zxx
• Laurette Pretorius - under-resource languages
• Gabriele Sauberer - call for diversity
• Gerhard Budin - cross-domain terminology adaption
• Key-Sun Choi - multilingual sync

Group: Interoperability

• Telic - taxonomy and more dimension(s)
• Scope - emotions
• Hermeneutic web
• dict scheme - http://dict.org - curl

Reports of Panel 3

• Collaborative issues. Cross-domain adaptation of terminologies and ontologies. Pervasive multimodality and multilinguality of the web (of data) Gerhard Budin
• Under-resource languages.
• Antoine Isaac

Panel 4

• Multilingual projection - Pushpak Bhattacharyya
• Multilingual web standards - Web as a medium: Common ground, MLOD, MWS - M. T. Carrasco Benitez
• Scalability to language domains tasks - Iryna Gurevych
• Use cases for precision-oriented MSW - MOLTO - Aarne Ranta
• Multilingual web content processing - Josef van Genabith

Group multilingual web

• Linked data
• URI - Axioms
Cool URIs
• Linked data Madrid - photo

Reports of Panel 4

• Multilingual web standards
• Iryna
• Sergei

Demo

• BabelNet - Cytoscape
• Grammatical Framework - Phrasebook
• Indowordnet

Wrap-up

Dimensions

• Low-hanging fruit
• Emerging directions
• Vision

Fields

• Standardization
• Scalability
• Precision-oriented MSW
• Under-resourced languages
• Lexicalization of linked data