Alelo To Develop Speech-Enabled Learning Portal For ESL Students

Alelo is developing an interactive, Web-based, speech enabled learning portal for students of English as a second language, after winning a contract from the United States Broadcasting Board of Governors.

The learning portal—which will make Alelo’s learning technologies available around the world—will function in conjunction with the international programming of Voice of America, the largest U.S. international broadcaster with an estimated audience of 134 million people.

“We see this as a great opportunity to make advanced, speech-enabled learning technology available for English learners worldwide,” says W. Lewis Johnson, president and chief scientist at Alelo. “This is our first opportunity to do something for people learning English and we’re also very excited about the potential reach of this…So this is really an opportunity for us to help lots of people who are interested in learning English as well as learning about the culture of the United States.”

The program—“Dynamic English Learning Website”—will incorporate artificial intelligence technologies that allow users to hone their conversational skills by engaging in spoken conversations with web-based computer characters and feature automated speech recognition tailored for students of all levels.

“We develop our own acoustic speech models using a combination typically of native speakers as well as people who are learning the language,” says Johnson.  “So in fact we are going to be collecting corpora of Chinese people learning to speak English and Iranian people learning to speak English.”

According to Johnson, the program will focus on helping people communicate in simulated conversations.

“We develop language models for use in the language learning application with heavy emphasis on supporting simulated conversation,” he says. “We have these dialogues that learners can engage in and we also give them practice exercises of individual conversational turns. And we actually look not just for expected grammatically correct utterances but we also look for common mistakes that language learners make. We build those into the models. So we can actually give people feedback if they’ve made an incorrect word choice or if there’s a grammar problem.”

The learning portal—which Johnson says will launch early next spring—will also draw on Voice of America’s archive of English language media materials to reinforce listening comprehension skills.

These same Web-based learning technologies are being incorporated into a range of other language learning portals. Development partners include Yale University, U.S. Joint Forces Command, and the U.S. Marine Corps.

SpeechTek Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues