Over the past seven years, statistical machine-translation technologies developed at the U.S. Army Research Laboratory (ARL) have helped the Army become less reliant on human translators in projects in which significant translation is a requirement.
One such project focused on translating medical materials in the Afghan dialect of Dari, a challenge given a limited pool of translators who were both conversant in Dari and medical terminology, according to ARL researcher Steve LaRocca. His team investigates how to tap the knowledge of linguists by recording a massive archive of translated sentences housed in databases.
ARL's Melissa Holland says the lab applies statistical machine-translation methods to specialized Army problems in the event a commercial solution is unavailable. She notes the facility's multilingual computing group solves problems with medical, legal, and Army training translations, and the information used to translate the medical phrases is stored in a database for use across the Defense community.
LaRocca says the lab's computer translation specialists want to broaden the military's ability to translate critical data volumes, and the Army Program Office foresees a need to add three new languages annually while expanding domains to encompass legal, criminal justice, military training, and medical training.
From Army Technology Live
View Full Article
Abstracts Copyright © 2015 Information Inc., Bethesda, Maryland, USA
No entries found