Learner Attitudes Towards Humanoid Robot Tutoring Systems: Measuring of Cognitive and Social Motivation Influences

Introduction

Trying to ‘see’ the robot tutor through the eyes of the learners by capturing the eye movements in an eye tracking experimental framework is a novel approach towards understanding the learners’ attitude to inclusion of robots as teachers in the classroom. We propose a framework for testing the learner attitude towards a humanoid robot tutoring system that attempts to capture both the cognitive motivation as well as the social motivation of the participants in a teaching session, held by a robotic tutor. The important role of humanoid robots as technological support in the classroom has been extensively investigated recently both with typically developing children and children with learning difficulties (Lourens & Barakova; 2009; Krichmar & Wagatsuma, 2011; Leyzberg, Spaulding, Toneva, & Scassellati, 2012; Belpaeme et. al, 2013; Feng, Gutierrez, Zhang, & Mahoor, 2013; Huskens, Verschuur, Gillesen, Didden, & Barakova, 2013; Kim et. al, 2013; Anzalone et. al, 2014; Barakova, Kim, & Lourens, 2014). The aspects of being different from a human tutor and therefore more patient, less emotional, more amusing and drawing child’s attention have been emphasized in these studies. The present study explicitly addresses the investigation of the ‘social nature’ of the learning process modulated by a humanoid robot in the classroom. Some results from the behavioral aspect of the current study were published in (Dimitrova, Wagatsuma, Tripathi, & Ai, 2015; Dimitrova, Wagatsuma, Kaburlasos, Krastev, & Kolev, 2018). Here we present in detail the novel experimental framework, proposed to investigate aspects of the attitude towards humanoid robot tutors combined with a ‘viewing timeline analysis’ (VTA) of eye-tracking data.

The inspiration for the experiment described in the paper has come from recent neuroscience research on the importance of the underlying brain mechanisms of ‘social cognition’ for shaping the cognitive abilities of the learner (Pfeiffer, Vogeley, & Schilbach, 2013; Schilbach, 2014). These studies can provide novel pedagogical insights towards designing robots to assist the teaching process in the classroom in order to achieve smooth and intuitive interaction of the robot with the learner at hand.

A promising novel trend of research on understanding the brain mechanisms of learning (in cognitive and social contexts) is “social cognitive neuroscience”, which provides evidence of the primary role of social interaction in the developmental process of shaping cognition (Ochsner & Lieberman, 2001). M.D. Lieberman (2012) has proposed the concept of “social working memory” as distinct in its neurological basis from the commonly assumed cognitive ‘working memory’. The areas of the brain that get involved while attending to a lesson, which is being explained with emphasis on its social relevance or historical context, are broader and involve the cognitive ‘working memory’ areas as well. Moreover, memories, created with the participation of ‘social working memory areas’ are much more durable than without them (which is actually a pedagogical aim by itself). Therefore, the social cognitive neuroscience forwards the idea of the emotion-cognition unity, where learning is driven by the rewarding role of the communication with the teacher and the peers or by the so called “intrinsic motivations” (e.g. Baldassarre & Mirolli, 2013), rather than by some functional self-realization notion. This requires designing innovative experimental paradigms to investigate the learners’ attitudes towards humanoid robot tutoring systems, which aim at being more competent about the explicit and implicit aspects of the social communication process involved in education.

Figure 1.

A humanoid robot tutoring framework implementing competence about the explicit and implicit aspects of the social communication process