Robot’s adaptive emotional feedback sustains children’s social engagement and promotes their vocabulary learning: a long-term child–robot interaction study

Abstract

In this article, we present an emotion and memory model for a social robot. The model allowed the robot to create a memory account of a child’s emotional events over four individual sessions. The robot then adapted its behaviour based on the developed memory. The model was applied on the NAO robot to teach vocabulary to children while playing the popular game ‘Snakes and Ladders’. We conducted an exploratory evaluation of our model with 24 children at a primary school for 2 weeks to verify its impact on children’s long-term social engagement and overall vocabulary learning. Our preliminary results showed that the behaviour generated based on our model was able to sustain social engagement. In addition, it also helped children to improve their vocabulary. We also evaluated the impact of the positive, negative and neutral emotional feedback of the NAO robot on children’s vocabulary learning. Three groups of children (eight per group) interacted with the robot on four separate occasions over a period of 2 weeks. Our results showed that the condition where the robot displayed positive emotional feedback had a significantly positive effect on the child’s vocabulary learning performance as compared to the two other conditions: negative feedback and neutral feedback.

The Effects of Feedback on Children's Engagement and Learning Outcomes in Robot-Assisted Second Language Learning

“…The results of Ahmad et al (2019) are consistent with findings from human studies and demonstrate that feedback does not only enhance children's language performance, but also engages children. Positive feedback engages because it validates children's answers and thus boosts their confidence (Henderlong and Lepper, 2002;Zentall and Morris, 2010).…”

“…Based on previous findings in literature regarding the role of feedback in second language learning, and previous studies that address feedback in child-robot interactions (Ahmad et al, 2019), we hypothesize that children will be more task-and robot-engaged when receiving (either preferred or dispreferred) feedback than when they do not receive feedback (H1a). Especially positive feedback is expected to increase the children's intrinsic motivation for the task and thus their engagement.…”

“…However, again, children showed large individual differences in the number of corrections they needed. Ahmad et al (2019) addressed individual differences between children and compared in a between-subjects design a robot that adapted its feedback with one that did not. They studied how children between 10 and 12 years old responded to the robot's feedback during 2 weeks.…”

L2 Vocabulary Teaching by Social Robots: The Role of Gestures and On-Screen Cues as Scaffolds

“…Possible explanations for the robot tutor's success include the robot's novelty. Anecdotally, children in the current study expressed great excitement about the robot tutor, and a recent review also emphasizes high enjoyment and anthropomorphic tendencies for robots in children in our age range (Ahmad et al, 2019;van Straten et al, 2020). A study similarly did not find an effect of tutor type, but showed that children gazed more at a robot tutor than a human tutor (Westlund et al, 2017).…”

Dynamic Emotional Language Adaptation in Multiparty Interactions with Agents

“…Recent improvements in speech understanding and generation have prompted the increased utilization of language-based agents in realworld scenarios. Existing commercial agents like Siri 1 and Alexa 2 tend to be highly transactional, focusing on delivering the correct content in responses. However, as research identifies additional potential areas of deployment in service [41] and entertainment [18] roles, there is a need for agents to deliver responses that are not only correct, but emotionally appropriate.…”