When it comes to translating languages in real time, BBN must speak the tongue as it netted a $14 million check from the Defense Advanced Research Projects Agency (DARPA) this week to continue developing its speech and text technology.
BBN has now taken in over $30 million from DARPA over the past few years to fill out the agency's Global Autonomous Language Exploitation (GALE) program. The goal of GALE is to translate and distill foreign language material (television shows and newspapers) in near real-time, highlight salient information, and store the results in a searchable database - all with more than 90% accuracy by the end of the program. Through this process, GALE would help U.S. analysts recognize critical information in foreign languages quickly so they could act on it in a timely fashion.
Inside the Top 10 hot aerospace technologies
During the first three years of the GALE program, BBN met or exceeded the accuracy goals for automatic translation of Arabic newswire text and broadcast news into English. Under this latest contract award, BBN will continue to work in Arabic from both speech and text sources to meet increasingly steep accuracy goals. BBN continues to work in Chinese under a separate award, the company said.
The BBN system will support multiple printed or handwritten document types including, hard copy, PDF files, photographs, newsprint, and signs. With the system, BBN will integrate optical character recognition and its state-of-the-art translation and distillation techniques to develop novel methods for processing handwritten text, BBN said.
The system will enable English-speaking military personnel and analysts to extract valuable information from a much larger number of foreign language documents than is now possible, facilitate rapid responses to emerging threats.
According to DARPA, GALE is making progress toward achieving this very ambitious goal by 2011. The agency is developing the System for Tactical Use program, a two-way speech translation system to convert spoken foreign language input to English output and vice versa.
BBN is doing lots of work for DARPA in the language and text interpretation field. Earlier this year it got $29.7 million from the Air Force to develop a prototype machine reading system that transforms prose into knowledge that can be interpreted by an artificial intelligence application.
The prototype is part of the DARPA's Machine Reading Program (MRP) that wants to develop systems that can capture knowledge from naturally occurring text and transform it into the formal representations used by AI reasoning systems.
The idea is that such an intelligent learning system could gather and analyze information from the Web such as international technological advances or plans and rhetoric of political organizations and unleash a wide variety of new military and civilian Artificial Intelligence applications from intelligent bots to personal tutors according to DARPA.
As digitized text from library books world wide becomes available, new avenues of cultural awareness and historical research will be enabled. With techniques for effectively handling the incompatibilities between natural language and the language of formal inference, a system could, in principal, be constructed that maps between natural and formal languages in any subject domain, DARPA said.