Project factsheets will no longer be updated. All information relevant to the project can be found on the CORDIS factsheet . This is updated on a regular basis with public deliverables, etc.
MONNET - Multilingual Ontologies for Networked Knowledge
248458 - STREP
At a glance
ICT-2007.2.2 - Cognitive Systems, Interaction, Robotics
The Monnet project will provide a semantics-based solution for integrated information access across language barriers, which is of growing importance to industry as exemplified by the Monnet use cases. A key solution to this problem is to deal with information at the semantic level, i.e. by abstracting away over language and form, allowing for more advanced and uniform: i) integration, ii) aggregation, iii) querying and iv) presentation of information across languages.
Current approaches to cross-lingual information access provide only partial solutions that address the problem in a restricted way, operating only at the document level without addressing either uniform extraction, representation, integration and querying of information across different languages or heterogeneous (textual, semi-structured, structured) data sources. Hence, the state of the art in machine translation is still far removed from providing multilingual services in specific domains.
A key solution to this problem is to deal with information at the semantic level, i.e. by abstracting from language and form, allowing for more advanced and uniform: i) integration, ii) aggregation, iii) querying and iv) presentation of information across languages.
The Monnet project will provide a semantics-based solution for accessing information across language barriers. This is of growing importance to both industry and government, as exemplified by the Monnet use cases:
- The SAP-XBRL use case concerns cross-lingual business intelligence on companies and financial services
- The Be Informed use case concerns multilingual access to government services
The goals of the SAP-XBRL use case are:
- To make financial information and business service descriptions available to be searched and queried in the language of choice of the user.
- To enable the creation of reports that present the search and query results in a chosen language.
The goal of the Be Informed use case is:
- To allow people to find their way through a government web site in their preferred language.
Monnet will advance the state of the art in cross-language information access by:
- definition of an integrated Ontology-Lexicon Model that allows for associating linguistic information with domain semantics as defined by ontology classes and properties
- Multilingual Ontology Localisation services that allow for the automatic localisation of an ontology to different languages
- Cross-Lingual Ontology-based Information Extraction techniques for semantic-level extraction of information from text and (semi) structured data across languages using multilingual localized ontologies defined according to the ontology-lexicon model
- a framework for Cross-Lingual Knowledge Access and Presentation that can be quickly customized to different languages in a principled manner on the basis of the integrated ontology-lexicon model and multilingual localized ontologies
The Monnet project will produce outcomes at the levels of technology, methodology and insight. In particular in providing:
- a formal model for multilingual, lexicalized knowledge representation
- automatic ontology localisation services
- a methodology for developing cross-lingual information access applications
- a tightly integrated approach to multilingual ontology localization, cross-lingual ontology-based information extraction and ontology-based language-independent information access
- novel end-to-end (“in vivo”) evaluation methodologies
- real-life applications that will demonstrate the exploitation potential in areas such as finance services, business services and public services
The Monnet ontology-based solutions to multilingual content access and management will be portable, scalable and easy to integrate. The portability (to other domains) and scalability (to other languages) will stem from the fact that content, presentation and multilingual realization are all clearly separated, which allows for moving to different domains (by exchanging the content in the knowledge base, the ontology, the data etc.) as well as to extend the system to different languages. This is possible as our system does not rely on specific language pairs, but on a language-independent representation. Easy integration of our solutions will be ensured by developing generic modules and technology which will be suitably adapted for and integrated into the use cases under consideration.
Monnet results will be presented at relevant industry conventions and events such as the annual XBRL conferences, Cebit, etc. as well as at a wide range of scientific conferences in different relevant communities such as the International Semantic Web Conference (ISWC), web conferences (WWW, WI), computational linguistics conferences (ACL, EACL, NAACL, COLING, LREC), artificial intelligence conferences (ECAI, EKAW, IJCAI, KCAP, IEEE Semantic Computing), etc. The Monnet consortium will also organize specific events around the main topics of the project, e.g.:
1st International Workshop on the “Multilingual Semantic Web” at the 19th World Wide Web Conference in April 2010: http://msw.deri.ie/
TKE workshop on “Establishing and using ontologies as a basis for terminological and knowledge engineering resources” at the 2010 Terminology and Knowledge Engineering Conference: http://www.dcu.ie/fiontar/conference/home_baile.shtml
Name: Dr Paul Buitelaar
Organisation: National University of Ireland Galway, Ireland
This page is maintained by: Susan Fraser