Addressing Design Challenges When Integrating Machine Learning with a Digital Annotation System to Examine Student Proportional Reasoning

This article reports on a novel approach to integrate artificial intelligence into a digital collaborative platform embedded with a problem-based mathematics curriculum. Using design research methodologies, we developed a new “proof-of-concept” design feature called “student proportional reasoning arrows (SPArrows).” SPArrows enable students and teachers to annotate their proportional reasoning through visual notes on their documented work. SPArrows and associated teacher- and researcher-generated tagging will generate data required to train machine learning to analyze students’ proportional reasoning in the digital platform. In this article, we report on the emergent design challenges that led to the development of the digital annotation system. We connect these emergent challenges to the underlying design principles of the digital annotation system for their potential to improve the teaching and learning of proportional reasoning. The integration of the SPArrows digital annotation system into a digital collaborative platform represents an important advancement in the integration of artificial intelligence to support and enhance mathematics education, particularly in the domain of proportional reasoning.