Machine translation of conversation on the digitized battlefield

John O. Moody, Eric Steinbrecher, Robert Frederking, Alan W. Black, Ralf D. Brown · 2002

The technology of the information age and the digitized battlefield apply not only to machine-to-machine communications but support an unprecedented advancement in communication across human language barriers. Modern military operations, including "peace-keeping, " coordination with joint and multinational forces, and operations other than war, often involve the interaction of troops with indigenous peoples where language issues can have a serious impact on mission success. The Audio Voice Translation Guide System (Tongues) is a mobile, networked, speech-to-speech machine translation system that both benefits from and contributes to the digitized battlefield. The Tongues prototype runs on a commercial off-the-shelf (COTS) mobile computing platform, capable of fitting in a cargo pocket, and running a Microsoft Windows operating system. It supports real time interactive translation of bilingual conversation, within and without the selected target domain, with the highest quality appearing within the target domain. Tongues has a modular architecture, allowing for easy upgrades to the speech and language engines and multiple conversational domains and languages. The prototype system supports English/Croatian conversation for deployed US Army chaplains, a domain that includes religious support, humanitarian aid, and operations other than war. The system supports text-entry as well as speech input and provides feedback for verification of translation accuracy. The Tongues system can automatically capture and store textual and audio conversation transcripts in formats that comply with the Joint Technical Architecture-Army (JTA-A). It supports a wide variety of JTA-A compliant means for transferring that information to other information systems. Tongues can run in a stand-alone mode, or it can make a wireless network connection to remote machine translation servers, allowing for increased performance and flexibility.

Read the paper · More papers on PaperTik