Language Grid Project

  • What is the purpose of Language Grid?

    The aim of Language Grid is to improve the accessibility and usability of the variety of language services currently available on the Internet. This enables for a mutual exchange between multiple languages by linking together the standard languages of every country. This also allows for community language resources to be gathered and used in combination at the sites of collaboration between different cultures. The ultimate goal is to overcome language barriers using language resources and language processing functions created by a large number of people, including specialists.

  • I would like to collaborate in the Language Grid Project. Can I participate?

    Please feel free to join us. A variety of different styles of collaboration are available. Contact the Language Grid Project (langrid [at] khn.nict.go.jp) for details.

  • What is the universal communication technology strategy of NICT?

    The universal communication technology strategy of NICT is a technical strategy for implementing the intellectual development program, which is part of the next generation research plan created by NICT based on the Ubiquitous Network Society (USN) strategy program of the Ministry of Internal Affairs and Communications. The aim of this strategy is to develop technology that can overcome various boundaries, such as language, culture, and physical abilities.

  • What is the reason for using the term “intercultural collaboration” instead of “intercultural communication”?

    In the 21st century it is important for us to prevent global cultural conflicts. In order to assist with international activities aimed at promoting collaboration, we wanted to develop an environment that makes it possible to build a specialized language service for specific action groups.

  • Although the language barrier seems to be an important issue in terms of intercultural collaboration, do you also work on other issues?

    This project is run as a collaborative effort between universities, companies, and NPOs using an open laboratory. While this project focuses on the Language Grid for providing language services, the universities and NPOs are undertaking a variety of research into intercultural collaboration using the Language Grid.

  • This is a project that would seem to require international cooperation. Which foreign institutions are planning to collaborate in this project?

    In Europe, we are planning to collaborate on pursing a project within the language triangle of French, German, and Italian, starting with the German Research Center for Artificial Intelligence (DFKI). In Asia, we are searching for future collaboration partners. Candidates include Tsinghua University and Shanghai Jiao Tong University from China, and KAIST from Korea. We are also working together with the international WordNet activities.

Functionality

  • Although around 10 languages from Asia and around 20 languages from all over the world are envisioned, which specific languages are being implemented?

    The 10 Asian languages being investigated are Japanese, Chinese, Korean, Thai, Vietnamese, Indonesian, Tagalog, Hindi, Arabic, and Persian. The European languages being investigated include English, Portuguese, French, German, Italian, Spanish, Russian, Greek, and Dutch.

  • What are the benefits of combining a variety of language services?

    This will allow people to make better use of the large quantity of language resources that have accumulated on the Internet. It will enable a language service to be built that is optimal for the actual field of activity of the intercultural collaboration.

  • Can anyone easily create a Language Grid workflow?

    At the present time, knowledge related to the specific workflows (particularly BPEL4WS) is required. In the future, workflows will become easy to define with the introduction of a scenario definition language and GUI.

  • Will this provide even better results than the existing machine translation services?

    Language Grid will not improve the accuracy of existing machine translations, because it uses them. However, there are many parallel text translations in specific fields, and using Language Grid to combine these services with the machine translation can form a more accurate translation service than just a machine translation alone.

  • Can language resources, machine translations, and other language processing functions that a person owns be registered with Language Grid?

    Yes. If it is available as a web service, then it can be registered with Language Grid. We are presently researching a wrapper construction assistance system so that in the future it will be possible to easily make these functions available as a web service.

  • What is the difference between the horizontal form of Language Grid and a multilingual machine translation system?

    The horizontal form of Language Grid is a single system for developing multilingual machine translation systems in an open environment. It is not only able to perform machine translations, but can also attempt to convert language pairs where no machine translation exists by using a bilingual dictionary, because the horizontal form of Language Grid is able to freely combine language services.

  • What is the difference between the dictionaries used in the horizontal form of Language Grid and the dictionaries used in the vertical form of Language Grid?

    This is a difference between a general-purpose dictionary created by a specialist and a special field dictionary created in the field similar to the actual application.

  • Does Language Grid have any functionality for overcoming cultural barriers?

    Although no efforts have been made towards developing functions for overcoming cultural barriers in Language Grid itself, if the translations of gestures, etc. are defined as language resources then cultural-specific expressive actions can be handled to some extent within the Language Grid framework.

  • Does Language Grid have any functionality for assisting in collaboration?

    The Language Grid Project provides collaboration tools that use Language Grid. Chat, BBS, blackboard, etc. are being developed. These are planned to be released one at a time.

  • Is Language Grid related to grid computing?

    Yes, Language Grid is a language service that is suitable for grid computing. In general, a grid is defined as "something that can provide a high quality service by using open standard protocols to cooperatively operate distributed resources". This proposal is the world's first grid related to languages. The suitability of web services for grid computing has been growing in recent years, and in Language Grid also, various different kinds of language services are implemented as web services.

  • What are the future research problems that are related to the functionality of Language Grid?

    One research issue for the future is estimating and guaranteeing the quality of a service created when a language service is used in cooperation with Language Grid. That is, estimating the quality of a compound service. Another research problem for the future is a creation assistance system for community-developed language services. We are investigating into assisting the creation of not only parallel translation texts and bilingual dictionaries, but also for unique language resources, such as dictionaries of emoticons.

Usage

  • What do I need to do to use Language Grid? Do I need a special tool?

    Since Language Grid itself is provided as a web service, it can be used from a regular web service client. Collaboration assistance tools that use Language Grid will be made available in the near future. A workflow definition is required in order to build a compound service using Language Grid. Workflow definitions can be implemented by using an existing workflow editor that outputs a BPEL document. A specialized workflow editor for combining language services is currently being developed.

  • Could you tell me when general users will be able to use Language Grid?

    The service began operations as a trial limited to users engaged in non-profit activities from the second half of 2006, and is planned to be completely open by 2009.

  • Where can I download the tools, such as chat and blackboard, which were referred to in the demo?

    This is scheduled for publication on the Language Grid website from November 2007, but only for the limited users engaged in non-profit activities. At the same time as the release, a trial version will be made available that can be used for a limited period of one week.

  • Will Language Grid become available free of charge?

    The trial version of Language Grid can be used free of charge. In the future, a fee may be charged depending on the language services being used.

  • Are the usage fees for machine translations, etc. expected to be reduced as a result of Language Grid?

    It is assumed that the operational structure of Language Grid will reduce prices as a result of various language services being supplied in bulk to many users within Language Grid.

  • Is Language Grid an open source?

    Although the system is open, we are currently investigating whether the source will be published.

  • What is the incentive for publishing a language service for free that cost money to build?

    From the point of view of a business, we believe that income will increase as the number of users and amount of use increases. It is assumed that the strategies for improving a businesses product can be produced by obtaining feedback from many users. The presence of a company and its image can also be improved if provided exposure on Language Grid. Furthermore, the language resources developed by an NPO are not originally developed with profits in mind, and can be made in terms of more people by participating in Language Grid. In addition, participation in Language Grid is expected to invite participation in a collaborative framework based on a give and take relationship, because the language resources that were independently created and that have a limited scope can be used in combination. In the future, it is possible that language resources and language processing functions will be built through the combined effort of many people in the same way as Wikipedia.

  • How are you planning to administrate and operate Language Grid?

    We are currently investigating into establishing an operational organization to take over responsibility of everyday operations at the end of 2008 when the research and development of Language Grid is finalized.

  • What kinds of language services do NPOs develop?

    There are a variety of services that correspond to individual NPO activities. The Language Grid Project provides an environment where it is easy to build the services needed for the immediate issues of medical interpreting volunteer assistance, communication assistance for children, and multilingual broadcast program development assistance.

  • Multilingual parallel translation systems have been created in the past in such fields as medicine. Will Language Grid be superior to these?

    Language Grid is not a stand-alone service, but is a framework for combining multiple language services. Therefore, existing parallel translation systems are also able to participate as part of Language Grid as one of the language services. By using Language Grid, it will be possible to, for example, use a parallel translation database created by a specialist in combination with a parallel translation database developed by an NPO or at a medical facility. In other words, Language Grid can be thought of as the infrastructure for building extensible systems.

  • What kinds of future research topics are there related to using Language Grid?

    We still need to investigate into usable business models, such as the pricing system for using charged language services via Language Grid. Furthermore, we are going to investigate into the billing area among others, to enable for-profit businesses to develop multilingual services and provide them as services on Language Grid.