acm-header
Sign In

Communications of the ACM

ACM TechNews

Robots Can Learn How to Support Teachers in Class Sessions


View as: Print Mobile App Share:
A child interacts with the robot tutor while a teacher provides guidance to the robot

Robots can learn techniques supportive of teachers in classrooms in just a few hours, according to researchers at the universities of Plymouth, Lincoln, and the West of England in the U.K.

Credit: University of Plymouth (U.K.)

Robots can learn techniques supportive of teachers in classrooms in only three hours, according to researchers at the universities of Plymouth, Lincoln, and the West of England in the U.K.

The researchers programmed the Supervised Progressively Autonomous Robot Competencies (SPARC) robot to progressively learn autonomous behavior from human demonstrations and guidance, under a teacher's control.

Afterwards, the robot could autonomously support students in the activity demonstrated, providing advice that was consistent with that of the educator.

SPARC employed actions with a different frequency than the teacher, but it only used actions already shown, and learned the novel dynamics associated with each type of action; the robot also learned techniques to support pupils with social interactions, like praise and encouragement.

The researchers think future human-robot interactions could benefit from this approach, which would allow end-users to directly train robots.

From University of Plymouth (U.K.)
View Full Article

 

Abstracts Copyright © 2019 SmithBucklin, Washington, DC, USA


 

No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account