Cognimates: Collaborative creative learning with embodied intelligent agents


@ Stefania Druga & Tammy Qiu

Stefania Druga & Tammy Qiu 

Cognimates is a platform where parents and children (7-10 years old) participate in creative programming activities in which they learn how to build games, program robots, and train their own AI models. Some of the activities are mediated by embodied intelligent agents which help learners scaffold learning and better collaborate. Learn more about our research, projects, and learning guides.

Conversational agents and connected toys are becoming common in homes. Increasing exposure to "intelligent" technology raises important questions about the ways that children understand it and how they could learn with and from it. Embodied intelligent agents, such as social robots, afford longer-term engagement in the home for children and their families .  

Building on the prior experience in the Personal Robots group of designing social robots for nurturing children's curiosity and learning, we built a platform where children and parents can learn to program with embodied intelligent agents which in turn become learning companions (Cognimates). The goal is to enable  learners  to interact with a social robot but also program it, train it to remember and learn things over time, and have  reflective conversations with their peers prompted by it. 

Why, how, and when can embodied intelligent agents support children and parents to learn via reflective teaching? What are the new intergenerational learning pathways that Cognimates could facilitate? How can these future learning companions be integrated into various learning applications and what are the generalizable design considerations?  In this research project we are addressing these questions by allowing children and parents to use a visual programming interface to control and customize an embodied intelligent agent. 

Demo video


Stefania Druga


Some of the activities were adapted from Dale Lane's Machine Learning for Kids