EXP: Augmenting a Teachable Robot with Adaptive Cognitive and Social Support

8/19/19-7/31/21

PI: Winslow Burleson
University of Arizona
Award Details

This proposal investigates robotic teachable agents, a type of cyberlearning technology that provides cognitive and affective feedback to support students’ learning. The Tangible Activities for Geometry system (TAG) is a robotic teachable agent platform for middle school mathematics and computational thinking. Students physically engage within a projected coordinate space with an interactive teachable robot named Quinn. Students teach Quinn how to solve challenges involving plotting points on a graph, translating points, rotating points, and plotting lines by giving procedural and conditional instructions to Quinn. There appear to be at least two primary advantages to using a robotic learning platform to investigate how to design teachable agents. First, a physical presence, provided by a robotic agent, strengthens users’ perceptions of having a social partner that is more than a virtual agent. Second, students receive cognitive benefits from learning through embodied, physical interactions. This research investigates the unique affordances of a teachable robot for supporting students’ cognitive and social interactions within a learning environment as it enhances our understanding of how cyberlearning can support STEM learning (mathematics and computational thinking). The study will be conducted in a school system having high percentages of underrepresented minorities who will learn to design features of the robots, exposing them to STEM careers and increasing the likelihood of acceptance and scalability. The investigative team will also disseminate findings and Do-It-Yourself (DIY) instructions for integrating cyberlearning environments and pedagogical approaches.

This proposal advances an empirical investigation that uses a teachable robot learning environment built with prior NSF support, integrated within the NYU Holodeck, a state-of-the-art immersive collaborative cyberlearning research environment, to improve understanding of how adaptive cognitive and social support can facilitate embodied interactions with teachable agents. This project investigates the teachable agent phenomenon within the context of the Tangible Activities for Geometry system (TAG), a robotic teachable agent platform for middle school mathematics and computational thinking. The project has three phases. First, the research team will explore two factors related to cognitive support: (1) how the teachable agent can give adaptive feedback representative of a learner (questions, self-explanations) rather than a tutor (hints, instructional explanations), and (2) how cognitive prompts can leverage the physical properties of the environment. Second, the team will explore two factors related to social support: (a) how a teachable agent can make adaptive attributions to effort or ability for its success or failure to motivate students, and (b) how the robot can use its embodied presence (physical or tactile) to contribute social information to enhance student learning. Third, to assess the efficacy of the interventions, the project incorporates a study comparing cognitive and social support to a condition with only cognitive support, only social support, and with no support. The project will make substantial contributions toward better understanding how to design and effectively incorporate support, provided by a teachable robot, to adaptively provide cognitive and social support to improve student outcomes. The robot framework will be tested in schools, including high percentages of underrepresented minorities, developing new ways of using technology to learn mathematics and computational thinking.

Tags: , ,