Thor Schrock

The United States Defense Advanced Research Projects Agency (DARPA) is developing software that could be able to translate, and summarize spoken or written conversation or printed material automatically.

The software is nicknames GALE, and acronym for Global Autonomous Language Exploitation. The main interest in this software stems from a severe shortage of Arabic translators.

When the US toppled Saddam Hussein's Iraqi government, millions of pages of documents were seized. The Defense Department did not have access to enough translators to make sense of the Arabic documents, so they posted them on the web so the public could help with translation efforts. While the idea was good on paper, it seems that some of the documents contained sensitive information on developing nuclear weapons, so the website was recently shut down to prevent proliferation.

It is estimated that a large portion of the information collected from signal intelligence (electronic monitoring) goes untranslated, and is therefore of little use to the military. However, if GALE was in place and functions as promised, advanced data mining techniques could be implemented on the newly translated information to give military planners and intelligence agencies a clear picture of the world around them.

IBM, SRI International, and BBN Technologies won a bid to develop the GALE system for the Defense Department and they must continually improve the software system to acquire additional research funding for the project. The final goal is to have a 95% accurate translation system.

The Defense Department believes that eventually this system may eliminate the need for scores of linguists and translators, but there is still much work to be done before the system is ready to be implemented in a real-world environment.