Wednesday, March 30, 2005

DARPA Looks to Automated Linguists

t's not Kirk and Spock's universal translator. Not quite. But the Pentagon is looking to for researchers to build a software set "with the goal of eliminating the need for linguists and analysts and automatically providing relevant, distilled, actionable information."
Global Autonomous Language Exploitation, or GALE, is a project of -- who else? -- Defense Department mad science division Darpa. And the idea, according to Darpa's call for proposals, is to "develop and apply computer software technologies to absorb, analyze and interpret huge volumes of speech and text in multiple languages."
The result won't necessarily be a "natural language" dialogue between man and interpreting machine. But, if GALE works as planned, it will deliver "consolidated information in easy-to-understand forms to military personnel and monolingual English-speaking analysts in response to direct or implicit requests."
The American military is still struggling to fill its ranks with Arabic speakers, three-and-a-half years after 9/11. Language training for enlisted men and junior officers is minimal. And the technological solutions to the problem -- like the hand-held Phraselator and Interact systems -- really only work for the most monosyllabic of conversations.
What Darpa wants instead are a trio of software tools for soldiers and spooks:

• A transcription engine that produces English transcripts [from foreign speech] with 95% accuracy
• A translation engine producing English text [from foreign prose] with 95% accuracy
• A distillation engine able to fill knowledge bases with key facts and to deliver useful information as proficiently as humans can.

No comments: