The “language barrier” separates country from country and people from people. It has also been a cause of misunderstandings and disputes. Although we are considered to be in an age of globalization, the language barrier remains high. Toru Ishida, of the Department of Social Informatics at Kyoto University, aims to cross over the language barrier using so called the Language Grid. In the Language Grid, the machine translation systems and multilingual dictionary databases (language resources) used by universities, research institutes, corporations and other bodies in different countries will be linked on the Internet to maximize their benefit and utility. Ishida aims to develop the Grid into a powerful tool for realization of a multicultural society, in which anyone can engage easily in international exchange and people of different cultures can coexist in harmony.
Language resources (machine translation systems, multilingual dictionaries, etc.) on the Internet are linked in a network (grid). Optimal combinations of resources result in the multilingual environment needed for diverse worksites. The Language Grid provides a multilingual foundation for communication that crosses over the language barrier, and for intercultural collaboration.
Participation by 115 organizations in 17 countries; accessible all over the world
“Our first step was to build a system for translating French into the Osaka dialect. We translated from French into English, from English into standard Japanese, and from there into the Osaka dialect. News reports focused chiefly on how “Je t’aime” translated into “I love you” in Osaka slang. The real objective of this research was not yet understood,” said Ishida, laughing.
Development began in April 2006, jointly with Japan’s National Institute of Information and Communications Technology (NICT). The Grid began operating in December 2007. At the start, 30 organizations in 5 countries took part. By October 2009, participation had grown to 115 organizations in 17 countries. In spring 2009, Google of the USA decided to take part, providing the company’s machine translation facilities. The Wikimedia Foundation, which runs Wikipedia, is also considering participation.
Universities, research institutes, corporations and other bodies possessing language resources such as machine translation systems will be able to provide their resources as Web services. These can be used anywhere, from any terminal with Internet access. Each machine translation system or dictionary DB uses a slightly different method for calling up information (the user interface). “It took us two years to build an interface common to all (using lapping technology) and develop procedures for calling up and executing processes step by step (workflow).” At present, translations can be undertaken for numerous languages, including English, German, French, Spanish, Chinese and Korean. Diverse specialized dictionaries and collections of illustrative examples are also available.
Different services have different interfaces. Wrapping technology uses a program known as a wrapper to absorb interface components unique to each service, and present an easily accessible unified interface.
A process diagram that sets out the order and methods of processes necessary for a series of services to be executed efficiently and smoothly. This enables non-professional users who know nothing of the details of processes to use the services without difficulty.
Providing support for foreign patients in hospitals, international NPOs, etc.
Dictionaries affect the accuracy of translations. “Technical terminology (jargon)” is of course essential in such specialized areas as science and disaster prevention. Furthermore, there is no lack of jargon in hospitals, schools and other areas of daily life. Mistranslation of jargon, with serious consequences, is not uncommon.
“Language Grid participants can register their custom-made dictionary databases. If the same dictionaries are then shared by all users, translation accuracy improves. The online encyclopedia Wikipedia is written and amended by users, and has thus become a huge knowledge database. Similarly, translation dictionaries can rapidly evolve through ‘combined strength,’ or ‘collective intelligence.’ The software source code is open, and users can therefore independently develop new tools.”
Language Grid usage is already beginning to grow in Japan and abroad. For example, it is being used by Kyoto City Hospital and Kyoto University Hospital to support the registration of foreign patients at reception desks. In Kamigyo Ward, Kyoto City, a real estate agency dealing in rental of housing and business premises is planning to use machine translation when dealing with foreign clients. Kyoto Information Card System (KICS), of which about 1,300 stores are members, has begun using the Language Grid for multilingual translation of pamphlets, for introducing Kyoto specialty products, etc. Additionally, schools with many foreign students can use the Grid to enhance communication between teachers and students. Universities, also, can use the system in responding to enquiries from foreign students and others.
Opening the grid to enterprises in the near future is under discussion
The international NPO Pangea is using the grid to smooth exchange of information and ideas among volunteer staff. Pangea is running an Internet exchange site called the Universal Playground, for children in seven countries, including Japan, Korea, Australia and Kenya. When a child enters a comment in his or her native language, it is automatically translated into Japanese, Korean, English, German, etc., and displayed on the websites in the corresponding countries. Using the Language Grid, Pangea is enabling the collaboration of volunteers all over the world.
At the moment the Grid is limited to non-profit uses. However, opening the Grid to business research activities and profit-making activities in the near future is under consideration. That will turn the Grid into an indispensable part of infrastructure in a multiculturally coexistent society.
Ishida concluded: “For over 10 years, my laboratory has taken various roads in searching for points of contact between IT and society, like digital cities. The experience we have gained to date is an integral part of the Language Grid.”
(Akira Miki December 11, 2009)