BACKGROUND |
Language is a major obstacle to the communication among multinational personnel. Personnel must communicate in addition to their own language, in an official NATO language. For multi-national meetings, such as IST Panel meetings, this may cause a reduced efficiency. By integrating machine translation with other related communications technologies, such as speech recognition and synthesis, speaker and language identification, NATO personnel can work successfully in meetings. Standardized assessment methods using real meeting data and specifications for both commercial-off-the-shelf (COTS) and for development of new technology are required.
|
OBJECTIVES |
To address these subjects a representative, multilingual, database of dialogues (both text and speech) is required. A relevant option is to collect these dialogues during NATO meetings, such as IST Panel meetings. One or more prototype systems would be developed and evaluated collaboratively within the Research Task Group. The prototypes could be standalone on a laptop or desktop computer or network based using multiple computers on a wide-area network. All specified applications could be assessed with such a methodology.
|