As mathematics deals with abstract concepts, these concepts must be represented to be reasoned with and communicated. These representations come in various forms and, in particular, various types of concreteness. For example, the number three can be represented with symbols (3), icons (⬤⬤⬤), or real-life objects (🍏🍎🍏). These representations need to be carefully designed to effectively support learning: For example, to teach multiplication, it is better to present concrete examples using pizzas slices (🍕) than candies (🍬), as a pizza representation helps students expand their understanding to the concept of proportions.
Insofar, representations of mathematical concepts have been designed by teachers, or by educators with pedagogical training. This ensured that the representations are mathematically accurate and support learning. However, nowadays, students can also prompt generative AI tool to generate representations. Now the questions are: Can AI generate accurate and useful concrete representations for mathematics? And how can we guide the learners’ process to ensure that they develop deep and transferable understanding?
In this project, we seek to understand the pedagogical value of mathematical representations generated by AI and to develop effective scaffolding strategies to support learning.
Collaborator
- Prof. Dr. Tomohiro Nagashima (LaLaL ab, Saarland University)
