C-STAR Consortium Conducts International Video Conference To Demonstrate Spontaneous Speech-to-Speech Translation in Six Languages
The international Consortium for Speech Translation Advanced Research (C-STAR) will conduct the first demonstration of video-conferenced human interaction through Computer- assisted speech translation at six locations around the world July 22, 1999.
The demonstration will take place in Kyoto (Japan), Taejon (Korea), Pittsburgh, Pa., (USA), Karlsruhe/Heidelberg (Germany), Trento (Italy), and Grenoble (France).
Participants in this international event will converse with each other in their native languages, planning travel to each other's countries while the computer systems in each of their respective laboratories provide the translation.
During the speech-translated conversations, "travel agents," portrayed by scientists in the laboratories, will show travelers pictures, train schedules, and panorama views of their destinations, allowing them to view hotel rooms and imagine the trip as it's being planned. In addition to translation between the parties, the system also provides communication and maintains a log of the travel decisions that are made. A final Web-based travel itinerary can be provided to the traveler over the Internet.
In this demonstration, several breakthrough technologies will be unveiled to the public for the first time. Some will be shown jointly and some will be demonstrated at a subset of the sites only:
The Consortium for Speech Translation Advanced Research (C-STAR) was established in 1991. During the last seven years, it has grown (the second phase) from four partner laboratories to 20 partner or affiliate laboratories. This growth illustrates the increased interest both scientific and commercial- in this important application for speech, language and telecommunication technologies.
The world-wide demonstration will consist
of interconnected speech translation systems
developed by six C-STAR partners:
The Consortium for Speech Translation Advanced Research (C-STAR) is a voluntary research consortium that includes the partners listed above, as well as affiliate members who share an interest in speech translation and participate in component technologies. The affiliates of C-STAR include:
Each partner organization raises its own research funding in its own country. Speech Translation Research has been supported in many parts of the world by research collaborations, large research initiatives and national projects and centers, including national research projects and agencies in Japan (ATR and the Japan Key Technology Centers), national projects in Europe (most notably project Verbmobil, Germany) as well as by industry in the US, Europe and Asia.
The speech translation systems developed
to-date and presented in this international experiment have moved
the state of the art from limited vocabulary, limited speaking
style systems to speech translation systems that accept spontaneous,
unrestricted speech within a given domain of discourse. Such systems
will allow the flexibility of expressing
common needs of travelers and visitors to foreign countries (such as travel planning, buying and selling, medical needs, directions, etc.) in naturally spoken spontaneous speech. Commercialization in a 5-10 year time frame of these technologies seems likely.
Further extensions of this research will thereafter remove domain limitations so that arbitrary topics can be discussed (for example, translating telephones, translating television, simultaneous translation of meetings, etc.).
To account for time differences in the different countries, and to allow for suitable presentation, each country will be the host of the demonstration during a prespecified time slot during the day. The demonstration will take place in the six partner countries according to the following schedule:
Demonstration (July 22)
Presentation Order : Japan -> Korea -> Europe -> USA
USA (GMT-4, DST) Europe (GMT+2, DST) Asia (GMT+9)
Japan (ATR) hosting
01:00 (Thu Jul/22) 07:00 (Thu Jul/22) 14:00 (Thu Jul/22)
Korea (ETRI) - hosting
03:20 (Thu Jul/22) 09:20 (Thu Jul/22) 16:20 (Thu Jul/22)
Italy (IRST) - hosting
07:00 (Thu Jul/22) 13:00 (Thu Jul/22) 20:00 (Thu Jul/22)
France (CLIPS++) - hosting
08:00 (Thu Jul/22) 14:00 (Thu Jul/22) 21:00 (Thu Jul/22)
Karlsruhe/Heidelbert (UKA/EML) - hosting
10:00 (Thu Jul/22) 16:00 (Thu Jul/22) 23:00 (Thu Jul/22)
USA (CMU) - hosting
12:00 (Thu Jul/22) 18:00 (Thu Jul/22) 25:00 (Thu Jul/22)
[01:00 (Fri Jul/23)]