Scientists are Teaching Social Robots to be Teachers
As new innovations in artificial intelligence and robotics come up everyday most scientists believe that these strides can aid in filling crucial knowledge gaps in the workforce, aiding people in a social, as well as mechanical capacity. Recently, an NAO humanoid robot was able to effectively acquire social behaviour that allows it to teach 75 students, through a new robot control methodology: Supervised Progressively Autonomous Robot Competencies (SPARC). The methodology, developed by Emmanuel Senft and colleagues, allowed the robot to acquire appropriate social behaviour for the classroom, through interactions with students. These interactions were closely monitored by a psychology PhD student with no background in robotics.
With SPARC, the robot’s control actions will initially be a blank slate, as the researchers’ are trying to strike a balance between human control and robot autonomy to adequately adapt the robots’ behaviour for the classroom. The teacher teleoperate the controller and chooses actions for the robot to execute.
After a while, the robot will ‘evolve’ and start to suggest actions to the teachers based on their previous class time and experience. The robots’ learning continues to improve as time passes with them catching onto the learning patterns, and over time slowly reducing the need for teacher intervention.
While social robots can help fil crucial gaps in education, SPARC’s potential is not restricted to it. Beyond applications for the classroom, it is believed that social robotics can also help to improve medical and assistive robotics.