Agreed that this article is strongly related to Intelligent Tutoring Systems like the Carnegie Tutor.

I recently posted a brief literature review video of the field for those interested in the bigger picture around the article: https://youtu.be/SDb9-LGRrCU?si=t6FM3832-_OwwpSz

The TLDR of the relevant sections (see video description for relevant timestamps)

- Models like ACT-R represent a model of the domain and a model of the student

- Knowledge tracing is an ML-heavy field that predicts skills, state, etc (like parent says)

- Recommender Systems use that knowledge to intelligently recommend next problems