Carnegie Mellon University

Image of a student working on a tablet

June 23, 2017

Carnegie Mellon's RoboTutor Advances to Global Learning XPRIZE Semifinals

By Shilo Rea

Shilo Rea
  • Dietrich College of Humanities and Social Sciences
  • 412-268-6094

An estimated 250 million children around the world cannot read, write or do fundamental arithmetic, and many of these children are in developing countries without regular access to schools or teachers. XPRIZE is attempting to address this problem by funding an international competition to create open-source Android tablet apps that enable children between the ages of 7 and 10 to learn basic reading, writing and math skills without adult assistance. Apps were created in English and Swahili.

Nearly 200 teams from 40 countries entered the competition. Following an evaluation and pilot test, RoboTutor, led by CMU's Jack Mostow, is one of 11 remaining teams competing for five $1 million finalist prizes.

"RoboTutor is a brilliant piece of educational technology that has already proven to effectively teach English and Swahili-speaking children basic skills. It also perfectly exemplifies our evidence-based approach to carefully integrating technology and using data to continuously refine and improve instruction, leading to better student learning while supporting new discoveries in the learning sciences," said Richard Scheines, dean of the Dietrich College of Humanities and Social Sciences and faculty lead for CMU's Simon Initiative.

RoboTutor's design is based on scientific learning principles to engage students to learn the material and then use it in other contexts. It is powered by advanced technologies, including speech and handwriting recognition, facial analysis and machine learning. It collects data from its interactions with children both to enable cognitive tutors to adapt to individual students and to enable innovative data mining tools to continuously evaluate and refine its design and functionality.

Mostow, emeritus research professor of robotics, machine learning, language technologies and human-computer interaction, has spent the past three decades applying advanced language technologies to literacy.

"I have been able to help a few thousand children over my career, but the Global Learning XPRIZE is the opportunity of a lifetime for me to help millions or even billions of children get a basic education," Mostow said.

RoboTutor leverages many assets, including its precursor, Project LISTEN, which used speech technology to enable natural spoken dialogue with an automated Reading Tutor that listened to children read aloud and helped them learn to read.

RoboTutor's hundreds of activities address four content areas — reading and writing, numbers and math, comprehension and shapes.

Another distinguishing feature is how RoboTutor's data-driven design process integrates with local cultures.

"While we focus on improving RoboTutor with large-scale data, we never forget that it really represents kids who are living and learning in a context that we must understand, in order to properly interpret that data," said Amy Ogan, assistant professor of human-computer interaction, who field-tested RoboTutor in several settings in Tanzania.

In addition to Mostow and Ogan, the RoboTutor team consists of over 100 CMU students and faculty as well as other experts and students from around the globe.

Once the XPRIZE semifinalists have been evaluated, the top five will each receive $1 million, and XPRIZE will conduct an independent 18-month, large-scale study to field-test their Swahili apps, pre- and post-testing 4,000 children in 200 Tanzanian villages on literacy and numeracy. XPRIZE will award the $10 million grand prize to the team whose app achieves the highest learning gains.