Steve Ferrara spends most of his day thinking about how one kind of student can talk more, not less, in class. He’s working with children who are learning English as a second language—and every one of these students learns better when they’re able to practice their conversational skills out loud.
“The problem is that teachers have so many kids in the classroom and they don’t have a lot of time to sit down and talk with individual students,” Steve says. “Many of these kids then go home where their native language is the only language spoken.”
So, Steve and his colleagues are working on an automated learning system that runs on a tablet or laptop. It interacts conversationally with dual language learners without the need for constant oversight from busy teachers.
When a student talks to a device, a customized, hi-tech avatar responds—just like someone interacts with a smartphone to find directions or call a friend.
“It’s not just about speech recognition, though,” Steve says. “Once words are recognized by the technology, the program has to determine whether those words are correct and relevant to the conversation.” He adds: “Our intelligent technology helps the program’s avatars recognize where the learner is struggling and responds accordingly.”
On the learning side, the avatars are programmed to provide the most important part of the dual language education process: feedback. Learners might hear responses like “You didn’t pronounce that word very clearly, here’s how…” or “When you express an opinion, you need to provide support for your opinion.” It’s this kind of feedback that helps language learning stick.
Though only in the prototype stage, Steve and his colleagues are already beginning to roll out pilots of this technology in classrooms in Arizona and Indiana.
“The project is a collection of intelligent technology tools, as well as a lot of empirically supported data,” Steve says. “We’re being incredibly ambitious here, but it’s not unrealistic.”