Site: Advanced Telecommunications Research Institute International (ATR)
2-2 Hikaridai Seika-cho
Kyoto 619-02, Japan

http://www.atr.co.jp

Date Visited: 15 October 1998

WTEC Attendees: P. O'Neill-Brown, G. Gamota

Hosts:

BACKGROUND

The purpose of Advanced Telecommunications Research Institute International (ATR) is to advance telecommunications R&D, with a particular focus on the design of applications that promote effective human-computer interaction. ATR has 140 shareholders.

ATR has four research divisions:

  1. Media Integration & Communications Research Labs
  2. Interpreting Telecommunications Research Labs
  3. Human Information Processing Research Labs
  4. Adaptive Communications Research Labs

The research budget for the 4 divisions is about 7.4 billion/year. The total number of staff is 305, of which 244 are researchers. Of those researchers, 116 come from shareholder companies. ATR has filed for 866 patents. The number of papers ATR has published domestically is 5,181 (2,303 internationally).

RESEARCH AND DEVELOPMENT

Following are brief descriptions of the research in each of the four research divisions.

ATR Media Integration & Communications Research Labs

ATR Interpreting Telecommunications Research Labs

ATR Human Information Processing Research Labs

ATR Adaptive Communication Research Labs

ATR is where the story of the Japan Key Technology Center begins. With the breakup of NTT in 1982, the Japanese government was concerned that NTT would move away from basic research. Therefore, to preserve a vital element of the nation's indigenous capacity for basic research, the Japanese government established the Japan Key Technology Center in 1985. The Key Technology Center's budget was (and is today) derived from the government's share of dividends from NTT stocks. In 1986, ATR was established by the Japan Key Technology Center as the first JKTC-funded project in Kansai Science City. ATR is a limited stock holding company that consists of 140 shareholders, including NTT, of course.

Japan has traditionally been weak in basic research, relying heavily on foreign sources of basic science. To correct this weakness, the Japanese government, especially recently, has been increasing funding for basic research and experimenting with new methods to encourage creativity. The Japan Key Technology Center brought ATR into being precisely in order to develop an indigenous capacity for technological innovation. An additional motivating factor that gave rise to the JKTC model was, as the panel discovered during its visit with MITI officials, the Japanese government's desire to contribute to the world's science base.

Without a doubt, ATR is contributing to the world's science base. Research taking place at ATR is cutting edge. In particular, ATR is a world leader in its efforts in face and object recognition. Face and object recognition techniques are key in the development of agent technologies. One of the world's hottest areas of research is in the area of agent technologies, whether they are realized anthropomorphically in computing environments, as robots in the physical world, or as purely software processes. The line of research that ATR is pursuing is a key element of what is called "adaptive agent technology." Adaptive agents are viewed as an important key to the information age.

In addition, the international way in which ATR operates clearly indicates that ATR is contributing to as well as gaining from the international scientific community. ATR's interactions with the world scientific community are significant. ATR's international cooperative activities include the following:

The Japan Key Technology Center, and particularly through ATR, is attempting to contribute to the world's research base of knowledge through its framework for intellectual property rights (IPR) distribution emanating from JKTC-funded projects. The framework is set up such that there is a 50-50 split between the Japanese government and the limited stock holding company that is the recipient of JKTC funds. This is an attempt on the part of the government to see to it that not only do the direct recipients of the funding receive benefit from government funding of R&D, but also the nation at large benefits in very direct, immediate and concrete terms: any party has the ability to license technologies arising from JKTC funded projects. It is noteworthy to point out that the United States does not have an R&D funding policy of this sort.

The licensing policy has an international dimension. JKTC's licensing policy serves as a vehicle for contributing to the world community in its ability to be used for moving the R&D out of Japan's labs. Specifically, the WTEC panel inquired as to whether there is any JKTC policy that restricts foreigners in licensing technologies emanating from JKTC-funded projects. Panel members discovered that there do not appear to be any such restrictions. As an example, AT&T licensed some of ATR's speech technologies.

Japan has long used its policy of aggressive technology licensing to become one of the world's top technological producers. Since some U.S. companies also make technology licensing a key part of their business development plans, the JKTC licensing policy model affords the potential for U.S. companies to leverage into Japan's research base.

In short, the JKTC model, through ATR, represents an effort to meet important policy objectives of the Japanese government-that is, as a key mechanism for "home-growing" Japan's basic research capacity, dispersing the benefits of government funded research widely, and contributing to the world's basic store of knowledge.

REFERENCES

ATR brochures.

ATR. 1995. Proceedings of the ATR Symposium on Face and Object Recognition '95. January 17-20. Kyoto: ATR Human Information Processing Research Labs.

ATR. 1996. Symposium on Face and Object Recognition '96. January 22-25. Kyoto: ATR Human Information Processing Research Labs.

ATR. 1998. Symposium on Face and Object Recognition '98. April 13 and 17. Kyoto: ATR Human Information Processing Research Labs.

ATR. 1997. Symposium on Face and Object Recognition '97. January 20-23. Kyoto: ATR Human Information Processing Research Labs.

Mima, H., O. Furuse and H. Iida. 1997. A situation-based approach to spoken dialog translation between different social roles. In Proceedings of the 7th International Conference on Theoretical and Methodological Issues in Machine Translation. July 23-25. Sante Fe: St. John's College.

Wakita, Y., J. Kawai and H. Iida. 1997. Correct parts extraction from speech recognition results using semantic distance calculation, and its application to speech translation. In Spoken Language Translation, Proceedings of a Workshop sponsored by the Association of Computational Linguistics and by the European Network in Language and Speech (ELSNET). Ed. Steven Krauwer et al. July 11. Madrid: Universidad Nacional de Educacion a Distancia.


Published: September 1999; WTEC Hyper-Librarian