MTeacher: A Gamified and Physiological-Based Web Application Designed for Machine Learning Education

Abstract

Everyday devices are becoming smarter. Machine Learning (ML) is increasingly more utilized in technology to augment the user experience. Furthermore, it is applied to improve the overall performance of systems. People constantly use intelligent agents in different forms through online shopping, social media, and entertainment. However, they do not understand the intrinsic functionality of the technology. Brain-Computer Interfaces (BCI) are more prominent in non-medical fields. Advances in hardware enable better ways to implement and use BCIs in more ubiquitous ways. There are also software advances that allow devices to connect and achieve signal processing in lower-end environments like the web. There is a need for the public to understand the social and ethical implications of Machine Learning. This research project aims to provide novices and non-CS majors with an explainable educational experience. Machine Teacher (MTeacher) is a web-based Machine Learning educational system that aims to implement less typical types of data (e.g. EEG) and gamified elements. Players can explore the concepts of Machine Learning through interactive activities, and our objective is to help increase their engagement, general ML knowledge, and self-efficacy. Initial results of a study reveal that physiological signals are more appealing to the users than traditional types of data, though they would prefer collecting the data themselves. Our future work aims to explore the addition of real-time physiological data acquisition and improve the application of gamification elements to study their effects on the learner’s experience.