Using artificial intelligence for understanding how children learn mathematics

Learning analysis aims to understand how students learn and to facilitate learning by adapting the content and difficulties of tasks to each individual. It is a rapidly evolving field that has benefited from access to larger amounts of data, better technology and new development of machine learning algorithms in the field of artificial intelligence, AI.

Torkel Klingberg and his research group will use new and unique data from more than 17,000 children aged 6-8 years after 7 weeks of training in early mathematics and various cognitive abilities. The mathematical problems are represented by an interactive row of numbers. The cognitive tasks include rotation tasks, visuospatial working memory and spatial problem solving.

It is already known that those who perform well on these cognitive tasks also perform well on mathematics tasks and also learn mathematics faster. There are many theories as to why these cognitive abilities are important for mathematical ability, and one of the objectives of the current project is to gain a deeper understanding of these relationships.

The researchers will use newly developed methods from machine learning to understand how children learn mathematics and what differentiates different individuals’ learning. These are mathematical and statistical methods previously used in artificial intelligence. They will also use more traditional statistical methods.