Development of an adaptive learning system for factual knowledge that can be used in several subjects in secondary education (VO), which can be adapted to the wishes of the teacher and the learning objectives of the school, and can replace the knowledge tests of those subjects.
| Project partners | ||
 |  |  |
| Memory Lab | University of Utrecht | Montessori School Community Amsterdam |
| Financing | ||
 | ||
| NOLAI |
Motivation and goals
Schools object that valuable teaching time is wasted on the examination of factual knowledge. In this project, a solution to this problem will be sought in monitoring the learning process, whereby a computational cognitive model is updated during learning so that it indicates at any time how well each fact is mastered. If sufficient facts are mastered, this system can indicate, based on newly developed AI algorithms, that “mastery” has been achieved and that a test has therefore effectively become redundant. The aim of this project is therefore to build an adaptive learning system for factual knowledge that can be used in multiple subjects in secondary education, that can be adapted to the wishes of the teacher and the learning objectives of the school, and that can be used to test the knowledge of those subjects.
How we work
Input for the design of this system will be provided through focus group discussions with teachers and students, after which pilot studies will be carried out to evaluate this system. Input from teachers and students will also be taken into account here. In the final phase, this system will be applied school-wide if possible. In addition to developing the system that indicates whether all knowledge has been mastered sufficiently, AI-based software will also be designed based on the analysis of the learning data that can detect certain forms of fraud.
This project touches on pedagogical-didactic aspects because it can influence how education around testing in secondary education is set up. In addition, this project offers new developments in the field of AI, as new cognitive AI algorithms must be developed that, based on the learning data, determine whether a student has learned enough to obtain a “mastery credit”. In addition, algorithms will have to be developed that can detect when certain learning data no longer belongs to the logged in student.
News
January 2024
Together with the University of the Netherlands Hedderik van Rijn made a video about why we should stop testing. look at it here.
March 2024
Gesa van den Broek presented the first results of this project during the TeaP conference in Regensburg, Germany. The first experiments in the project investigate how well the 'model-based mastery' assessment can predict performance on a test the next day, and whether practicing to a higher standard of mastery leads to better retention. Initial results showed that this mastery assessment can indeed predict next-day retention well. Practicing to a higher standard of mastery also led to better retention.
Stay informed
Follow this page and our social media channels to stay informed of developments within the project.
