Researchers at Disney Research say their new keyword-spotting system works better with the video game "Mole Madness" than commercial speech-recognition systems.
For the rapid-paced, multi-player game, players have to move an animated mole through its environment, gathering rewards as they avoid obstacles. To move the mole horizontally, one player says "go," while the other moves it vertically by saying "jump." Disney's speech technology system is capable of sorting through the overlapping speech, social side talk, and creative pronunciations of young children.
"Kids don't necessarily pronounce words quite like adults and when they are playing together, as they like to do, they often engage in side banter, or exclamations of excitement, or simply talk over each other," says Disney researcher Jill Fain Lehman. She notes this makes it tough for a speech-based system.
The team trained the system on young gamers, while a human "wizard" listened in another room and tried to map each "go" and "jump" into a button press on a game controller.
Lehman says the system uses separate models of go, jump, mixed, social speech, and background noise, built from 150-millisecond segments of the training data.
View Full Article
Abstracts Copyright © 2016 Information Inc., Bethesda, Maryland, USA