FOR IMMEDIATE RELEASEFebruary 24, 1999 |
Starting of the Joint Research Project on |
Since the Asian Multimedia Forum (AMF(*1)) was inaugurated by major infocommunications providers in the Asian region in June 1997, Nippon Telegraph and Telephone Corporation (NTT) has played an active role as a founding member in pursuing the Forum's objectives of contributing to multimedia services and technologies in the region. This has included carrying out various multimedia utilization trials by structuring international test beds for ATM, satellite and the Internet. As a part of the trials, three parties, NTT, Korea Advanced Institute of Science and Technology (KAIST(*2): Korea) and Kent Ridge Digital Labs( KRDL(*3): Singapore) have started a joint research project on cross-language information retrieval over the Internet. The aim of this project is to develop a system which allows the user to input a search command in one language, and retrieve data from a number of different language WWW sites. This is made possible through the combination of natural language processing functions (such as translation) and search engines (*4) and/or sites developed by project members for various East Asian languages such as Japanese, Chinese, and Korean. The entire system, referred to as the Cross-Language Information Retrieval Architecture, consists of multiple "search sites" and "meta-search engines" that are located on the Internet. The search site has a function for retrieving WWW pages in the same language as the search phrase or in a small number of other languages. The meta-search engine selects search sites suitable for the user's request, and translates the user's search phrase. For instance, in retrieving WWW information in Japanese, Chinese and Korean in response to a Japanese search phrase, the meta-search engine first selects the search sites in these languages, and then translates the query into these languages. It finally submits the translated query to the search sites (See Attachment), and compiles search results from these search sites. The unique feature of the system is that it employs a newly developed common protocol called the "Cross-Language Information Retrieval Protocol" for the exchange of information between meta-search engine and the search site. The new protocol enables flexible combination of search sites and meta-search engines, and makes possible expansion into new languages or fields. Users will be able to retrieve multimedia information, such as photographs and music captioned in foreign languages. A draft of the protocol has been formulated, and NTT, KAIST and KRDL are currently constructing meta-search engines and search sites that enable cross-language retrieval between Japanese/Korean/Chinese and English. NTT is constructing its sites based on the Internet cross-language information retrieval system (TITAN(*5)) that was developed by the Cyber Space Laboratories. In March this year, interconnection tests will start among the parties, and trial service on the Internet is likely to be launched by June. In addition, the protocol specifications are planned to be made public on the Internet for wider distribution and promotion.
Also, the AMF will hold its sixth plenary meeting on February 25 and 26 at Cheju-Do, Republic of Korea. A demonstration of part of the aforementioned system will take place at the meeting.
*1 AMF (The Asian Multimedia Forum)
*2 KAIST (Korea Advanced Institute Science and Technology)
*3 KRDL (Kent Ridge Digital Labs)
*4 Search engine
*5 TITAN
Norihiko Ohkubo or Megumi Inaji |
![]() NTT NEWS RELEASE |