The US Defense Advanced Research Projects Agency (DARPA) scientists are developing advanced language processing technologies with improved speed and accuracy to offer analysts an advantage in a range of military and non-military scenarios.
The natural language processing technologies are designed to locate, identify and organise information from an array of sources and in at least 15 different foreign languages.
Speaking during the DARPA Congressional Tech Showcase in Washington, US, DARPA Human Language Technologies programme manager, Dr. Bonnie Dorr, said the scientists are interested in collection of information from the huge volumes that come through in foreign languages.
”The system goes into the video, pulls out the audio, separates it into sentences, renders it as text and translates it into English so that the human, who speaks only English, can then read what this Arabic broadcast news is about,” Dorr said.
Once saddled with filtering through a barrage of information in foreign languages, the English-speaking analysts can now use real-time filters to pinpoint information in audio and video broadcasts.
The real-time feed of identifying and aggregating individual pieces of information from raw data is claimed to be remarkable, despite a three-minute delay from a live broadcast.
The scientists will now focus on developing what the translation output does to boost information analytics, according to Dorr.
"In the future, we want to be able to read through language to meaning because people don’t always explicitly state all the assumptions that are underlying what they’re saying,” Dorr noted.
Dorr also demonstrated Raytheon BBN Technologies-developed Byblos, one of the several advanced, trainable, large-vocabulary, speaker-independent speech recognition systems, which uses context-dependent hidden Markov models of phonemes to offer a robust model of phonetic coarticulation during the event.
Image: the language processing technologies will provide an advantage to military analysts in a range of military and non-military scenarios. Photo: courtesy of Col. A.T. Ball.