Language Technologies

back to overview

Organic.Lingua - Demonstrating the potential of a multilingual Web portal for Sustainable Agricultural & Environmental Education

270999 - Pilot B

At a glance

PSP -2009.6.2 - Multilingual online services Duration: 36 months Start date: 1 March 2011 End date: 28 February 2014 Project officer: Aleksandra Wesololwska Flyer Presentation Annual Report 2011 Website

Challenge

Organic.Edunet and other similar specialized information services are not fully exploiting language technologies. This hampers their possibilities to reach wider trans-national audiences. The Organic.Edunet portal currently supports sixteen languages, but the current process makes translation error prone and time consuming, and misses opportunities to enhance the quality and efficiency of cross-language functions with available technology. The existing solution relies completely on human effort in translating and does not use existing linguistic resources and tools to help make resources and descriptions available in different languages or to enable cross-lingual search. In addition, resource retrieval is restricted as a consequence of the fragmentation of resource descriptions in several languages.

Objectives

Organic.Edunet web portal receives visits from more than 180 countries and its visitors speak more than 40 languages. Currently Organic.Edunet web portal supports 16 languages at the interface level and provides content in 10 languages which is quite far from being truly multilingual. To this end the Organic.Lingua project focuses on enhancing the approach to linguistic support and coverage of the existing Organic.Edunet infrastructure. More specifically, it aims to achieve the following objectives:

• To advance in the identification of user needs regarding linguistic support in on-line information services, using the existing features of the Organic.Edunet Web portal.
• To identify and explore new or anticipated user interface, functional and technical requirements related to linguistic technologies, analyzing existing services and competing approaches, perceived gaps and weaknesses of existing solutions.
• To devise and validate innovative and effective methods, processes and workflows for linguistic support in Organic.Edunet, over and beyond mere integration of linguistic technology.
• To enhance, extend and re-engineer the existing Organic.Edunet Web portal towards a wider and cost-effective multilingual support, broadening in this way the current audience and impact of the service.
• To deploy large pilots trials that demonstrate a cost-effective approach for delivering language technologies in existing Web portals based on open standards and open source software, that can be easily transferred to other domains.

Overall, Organic.Lingua aims to make the Organic.Edunet Web portal a service that will support agricultural researchers and educators around Europe, as well as explore new business opportunities for the specific service but also for the developers/providers of language technologies.

Innovation

The key innovative aspects of Organic.Lingua project are:

• The project will cover the full learning resource curation lifecycle, which opens opportunities for innovating both in digital description and preservation practices and also in finding specificities of the use of language technologies.
• Analysis of the user interaction needs for translation of metadata and ontologies in a practical and real-world setting. This represents an opportunity to advance in the knowledge about the application of language technologies with an emphasis on effectiveness and cost-effectiveness that would make multilingual support more effective.
• Language technologies will be selected according to clearly defined user requirements, and their piloting and evaluation will enable a contrast of different technological applications and of the current support of different language pair combinations.
• Machine translation will be used as an aid in the translation of the different elements of the system: metadata, ontologies, descriptions, etc. This enables opportunities for combining some of these elements to improve the overall process of content population.
• Cross-lingual facilities in information retrieval will make resources more accessible by breaking language barriers. The application of cross-lingual search will be evaluated in large pilots, and the broad coverage of the extended collection base of Organic.Edunet represents a significant testbed for the current status of linguistic infrastructures.
  

The result

The main outcome of the Organic.Lingua project will be a number of services that will facilitate the usage, exploitation and extension of digital educational content related to Organic Agriculture and Agroecology.

More specifically, the project will provide the following services and tools:

• Machine translation services for the automatic translation of metadata and documents
• Cross language retrieval technologies that will allow searching of educational resources in the user’s native language
• A wiki based tool that will allow the evolution of the Organic Agriculture and Agro-ecology ontology
• A user centric content evolution approach that will allow users to generate educational content and to contribute to metadata and document translation
• Translation services that will facilitate the extension of Organic.Edunet infrastructure to new language

Since the linguistic limitations of Organic.Edunet are common to most current multilingual Web portals, an important outcome of the Organic.Lingua project will be a set of guidelines, good practice methods and software tools that can be adopted in other portal cases.

Target group of the project

Impact

The Organic.Lingua will devise and validate innovative methods processes and workflow for effective integration of linguistic technologies in the existing Organic.Edunet portal. The project will develop a good practice guide providing information on application of new language technologies fostering the adoption of the tools. Organic.Lingua will be deployed in most EU languages and countries covering a sufficient broad range of EU countries.

What has been already done?

The Organic.Lingua vision
One of the first outcomes of the project was the vision about what the project aims to achieve. The Organic.Lingua vision is published on the project website and it will be constantly updated during the project according to the feedback both from consortium and the interested stakeholders.

User requirements
One of the first goals towards developing the desired solution was the recognition of stakeholders’ requirements regarding multilinguality issues around the Organic.Edunet web portal. Online surveys were conducted in order to identify requirements regarding multilinguality. A number of consultation meetings with content providers and user communities have been organized by the project partners in France, Greece, Turkey, United Kingdom, Estonia and Latvia. The main goal was to define new and anticipated requirements for the development and deployment of multilingual services in the Organic.Edunet portal.

Conceptual modelling framework
In Organic.Edunet portal thematic ontologies and metadata schemas are being used by content providers to annotate, classify and contribute resources. In order to meet the multilingual requirements new versions of metadata schemas and ontology were developed.

Language resources and services
The Organic.Lingua team has studied existing language technologies that can enhance the Organic.Edunet portal, as well as existing multilingual vocabularies, dictionaries, thesauri and multilingual corpora resources that can be used to train and test the various language components of the architecture. In addition, available language services such as Google Translate API and Bing API were studied. These services will be used in the context of the project to support languages additional to the ones supported by the tools provided by project partners.

Extending Content Coverage
The extension of content coverage of Organic.Edunet learning federation consists in a) connection of new collections to the Organic.Edunet learning federation and b) user generated content that will be enabled through services that will allow translation of documents and metadata by users as well as suggestion of learning resources that user finds on the internet.

back to overview