多言語情報発信シンポジウム (Symposium on Multilingual Information Sharing and Launch of TKUN)

＜Speaker＞

Dr. Makoto Nagao (Formerly President of Kyoto University)
Career: Dr. Nagao is currently a professor emeritus of Kyoto University. He served as a former president of Kyoto University, and a former director of National Diet Library. He is also a member of the Japan Academy.
Awards: Medal with Purple Ribbon (Medal of Honour by the Cabinet of the Government of Japan) in 1997, Japan Prize in 2005, and a prize for Person of Cultural Merit (prize for a person who has made outstanding cultural contributions in Japan) in 2008.
Dr. Nagao is the founder of machine translation research in Japan, and a devoted and outstanding contributor therefor. He proposed and initiated the example-based machine translation system.

【Overview】
In the globalized era, communication, trade, technology transfer, and the like are tremendously increasing. This leads to increasing demand for mutual translation between multi languages. In consideration of the Olympic and Paralympic to be held in Tokyo in 2020, speech translation is also required in multi languages. I would like to share my view with you today that, under those circumstances, it is quite important to make a system such that those in translation industry and enterprises which need translation services (clients for translation services) mutually cooperate with one another.

＜Speaker＞

Hitoshi Isahara（Professor of Toyohashi Institute of Technology）
Dr. Isahara made his research activities on natural language processing and machine translation in the following institution: Electrotechnical Laboratory in Ministry of International Trade and Industry (currently National Institute of Advanced Industrial Science and Technology (AIST) in Ministry of Economy, Technology and Industry of Japan), and Communications Research Laboratory in Ministry of Posts and Telecommunications (currently National Institute of Information and Communications Technology (NICT)).
He is now a chief organizer of Technical Japanese Association. He served as a former president of Asia-Pacific Association for Machine Translation (AAMT) and a former president of International Association for Machine Translation (IAMT).

【Overview】
Effective communication and sharing of information across cultures is indispensable to strengthen Japan’s international competitiveness. Today, I would like to introduce you to join “TKUN” (Translation Knowledgebase and User Network), which has been in my mind as useful to break language barriers which are obstacles for Japanese business players, academics, and language service providers. TKUN, which is launched today, is a network for all of them to get together and share data, knowledge and wisdom of members one another. Furthermore, TKUN allows us to learn and make use of machine translation technology, and international framework for data sharing and evaluation of machine translation, etc. Let me show you an outline of TKUN, and my vision and expectations as regards TKUN to proceed with it together with you for the future.

＜Speaker＞

Tomoyoshi Akiba(Toyohashi University of Technology)
Mr. Akiba joined Electrotechnical Laboratory in Ministry of International Trade and Industry (currently National Institute of Advanced Industrial Science and Technology (AIST) in Ministry of Economy, Technology and Industry of Japan) in 1995. Since 2014, he serves as an associate professor at Department of Electrical and Electronic Information Engineering. His research majors are speech and language processing, and natural language processing.

【Overview】
In recent years, a large volume of languages resources are becoming more and more indispensable for the development of national language processing system. For Example, the performance of statistical machine translation has been improved with rapidly growing research since the year of 2000. As you surely know, statistical machine translation is a system for conducting machine translation by automatically learning translation rules from a large volume of parallel corpuses, if available.
On the other hand, language resources are also necessary for evaluating the quality of developed language processing systems. In a workshop for evaluation such as TREC and NTCIR, researchers get together to share common evaluation resources, which allows enhancement of technical development of language processing systems.
Today, let me introduce you the use of language resources as a foundation for evaluation, raising SpokenDoc and SpokenQuery&Doc (which I organize) as examples. In addition, we are aiming at realization of use of machine translation in multi languages. There are some difficulties for languages with limited availability of language resources. Realization of machine translation in such languages must be possible by use of large volume of resources of some other languages.

＜Speaker＞

Mr. Jaap van der Meer
TAUS, Translation Automation User Society, Netherlands
Jaap van der Meer was the founder and CEO of some of the largest global translation and localization service companies in the 1980s and 1990s. In 2005 he founded the Translation Automation User Society (TAUS). TAUS is an innovation think tank and platform for industry-shared services for the global translation and localization sector. Many of the largest IT companies, government translation bodies and their suppliers of translation and localization services and technologies are members of TAUS. TAUS offers among others a platform for translation quality evaluation and benchmarking and a platform for pooling and sharing of translation memory data. Jaap van der Meer has written many articles over the years about the translation industry.

【Overview】
TAUS is a resource center for the global translation industries. Our mission is to increase the size and significance of the translation industry to help the world communicate better. We support buyers and providers of language services and technologies with a comprehensive suite of online services, software and knowledge that help them to grow and innovate their business. We extend the reach and growth of the translation industry through our vision of the Human Language Project and our execution with sharing translation memory data and quality evaluation metrics. Jaap van der Meer, director of TAUS, will present the TAUS organization and its activities. In particular he will talk about the platform for sharing translation memory data that helps organizations to train and improve machine translation technologies.

＜Speaker＞

Chris Wendt(Microsoft Research)
Chris Wendt graduated as Diplom-Informatiker from the University of Hamburg, Germany, and subsequently spent a decade on software internationalization for a multitude of Microsoft products, including Windows, Internet Explorer, MSN and Windows Live - bringing these products to market with equal functionality worldwide. Since 2005 he is leading the program management and planning for Microsoft's Machine Translation development, responsible for Bing Translator and Microsoft Translator services. He is based at Microsoft headquarters in Redmond, Washington.
http://research.microsoft.com/people/christw/


【Overview】
Microsoft has been publishing raw unedited output of its own machine translation service on the web since 2003, and its knowledge base is still one of the largest instances of published raw MT on the web. The use of domain-relevant training documents for Microsoft’s statistical machine translation system is crucial for reaching useful quality levels: Quality levels which support the purpose of the published material. In general, quality generally increases with the amount of domain-relevant training material, but has its limits. We will observe the effect of this training data, using the measured quality metrics as well as audience observation. Today’s self-service MT customization systems put statistical MT in reach for everyone. We’ll use the Translator Hub as an example of how data plus algorithms create usable systems that can move the translation business forward.

＜Speakers＞

＜Speaker＞

Kozo Moriguchi (Kawamura International Co., Ltd., Executive Officer and General Manager)
Kozo Moriguchi joined Kawamura International Co., Ltd., initially for quality management. He mainly worked on translation production processes such as revision, translation, and project management. Since 2011, he deals with management of overall issues of the firm including sales group. In the preparation team hosted by Japan Translation Federation (JTF) for ISO Japanese mirror committee, he deals with preparation of draft for ISO standards of post-editing of machine-translation outputs. He is officially a member of Japanese mirror committee for ISO TC37 SC5.

【Overview】
Currently, in ISO committee (both internationally and nationally), we are on our way of discussing international standardization of matters with regard to translation and interpretation, such as translation (ISO17100), controlled language (ISO24620-1), and post-editing (ISO18587). In today’s lecture, let me introduce how ISO18587 is progressing. Upon this, we are now discussing how to construct international standardization for the process of post-editing of an output of machine translation, as well as requirements for qualified post-editors.