DARPA translation work moves into next phase

Raytheon BBN Technologies has received an additional $6.14 million in funding from the Defense Advanced Research Projects Agency to continue work on the Multilingual Automatic Document Classification, Analysis and Translation program.

The award is in response to Raytheon BBN’s success in meeting required DARPA milestones during the first two years of the MADCAT program, the company announced today.

The goal of the MADCAT program is to create a prototype system that quickly provides accurate, relevant, distilled, actionable information to military commands and personnel by converting foreign language images, such as road signs, fliers, photographs and handwritten notes, into English transcripts automatically, without the use of linguists and analysts, the announcement stated.

“The MADCAT system will help our troops understand road signs, print media and captured documents that could be of immediate importance to their safety and to the successful completion of their missions,” Prem Natarajan, head of speech and language processing at Raytheon BBN Technologies, said in the announcement.

Under the contract, Raytheon BBN Technologies, a wholly owned subsidiary of Raytheon Co., will further refine a prototype translation system developed under previous program awards that can be deployed on a laptop computer.

This will be accomplished by integrating optical character recognition with the company’s state-of-the-art translation and distillation techniques and developing novel methods for processing handwritten text.