University of Southern California (USC) researchers are developing high-speed language-processing systems that rival the speed and efficiency of human speakers in specific settings.
"While we human speakers can often understand and respond to what someone is saying to us in a fraction of a second, a typical voice interface will require much longer--often a second or two--to try to understand what you have said and respond in an appropriate way," says USC professor David DeVault.
The researchers are studying new techniques that can streamline human-machine conversations by enabling the system to perform all necessary computer-processing steps in real time while the user is talking.
The researchers note the systems often can determine what the speaker means and how it should respond before the user finishes speaking.
To demonstrate this breakthrough, the researchers developed a high-performance game-playing agent called Eve. In the game, users describe the pictures they see on their computer screen and the agent tries to guess which picture they are talking about as fast and accurately as possible. The system uses "incremental" speech-processing algorithms to increase the agent's speed of understanding.
"These findings underscore the importance of enabling systems to not only understand what users are saying, but to do so as quickly as a human would," DeVault says.
From National Science Foundation
View Full Article
Abstracts Copyright © 2016 Information Inc., Bethesda, Maryland, USA