Personalised studying through AI-assisted learning
Federal Ministry of Education and Research funds development of adaptive learning systems

Ulm University

The coronavirus pandemic has forced university teaching online – a development that has revealed the limits of traditional teaching formats and underscored the fact that learning is a highly individual process requiring personalised content and student-centred support. In the new research project ‘2LIKE’ at Ulm University researchers from computer science, teaching and learning research and from the advanced professional studies sector will be collaborating to develop solutions in this area. The research team plans to use artificial intelligence (AI) to develop self-adaptive digital learning systems that can respond to changes in student knowledge and student progress. The new project at Ulm University will receive around 2 million Euro in funding from the Federal Ministry of Education and Research (BMBF) until 2025.

The mix of students studying at university has never been so heterogeneous as it is today. Students starting the Master’s degree programme ‘Artificial Intelligence’ have often completed very different Bachelor’s degrees with curricula that contained varying amounts of mathematics and computer science. Student backgrounds are even more diverse in internationally oriented degree programmes or in the advanced study programmes designed for working professionals. This raises the question of the best way to assess an individual student’s learning progression while maintaining the cost of student supervision at affordable levels.

While the popular online learning platform Moodle allows a certain degree of adaptation, it is not able to provide personalised feedback to students.
The researchers working on the 2LIKE project at Ulm University are seeking to change this situation. ‘We know from experience that students need individual opportunities to catch up on lost learning. And we also know that rapid personalised feedback is crucial to supporting the student learning process – irrespective of the time of day or the location. AI-based automated teaching and learning systems are well suited to supporting both of these adaptive approaches to learning,’ explains project lead Professor Birte Glimm, a research scientist at the Institute of Artificial Intelligence.

Project between computer science, psychology and advanced professional studies

Developing AI-assisted learning systems is an interdisciplinary undertaking. The Institutes of Artificial Intelligence and of Neural Information Processing at Ulm University are pooling their expertise in the fields of knowledge-based AI and machine learning. The university’s Communication and Information Centre (kiz) is providing the digital infrastructure. ‘We expect results from the project to be incorporated into the teaching activities of selected study programmes in the winter semester of 2022/2023. Prototypes of the teaching and learning system should be available in the near future,’ explains project co-lead and Director of kiz Professor Stefan Wesner.

However, any new AI-assisted learning system can only ever be as good as the underlying psychological and didactic design concept. Which forms of support and motivation should be used to help learners at specific times? When and in what form should feedback be given? These and other questions are being addressed by the Department of Learning and Instruction at the Institute of Psychology and Education. ‘In order to provide the best possible support to students, data on their prior knowledge, their cognitive and motivational abilities and their personal preferences needs to be gathered. This data can then be used by the AI-assisted system to deliver targeted adjustments to curricular content,’ explains Professor Tina Seufert, Head of the Learning and Instruction Lab, whose department is sharing its expertise with other members of the 2LIKE project.

Prototype of the adaptive learning system will be introduced soon

In the near future, new methods and learning activities, such as quizzes and knowledge consolidation and application tasks will be introduced into the Master’s degree programmes Artificial Intelligence and Cognitive Systems (anglophone). In fact, students on the AI-programme will be able to contribute to the development of new learning tools as part of their student project work and their final-year Master’s thesis. The knowledge gained during this initial phase will then be transferred to other programmes of study, particularly the advanced study programmes offered to working professionals at the School of Advanced Professional Studies (SAPS). Automated feedback should prove especially beneficial to working professionals studying for an advanced qualification, as they often have to study in the evening or at the weekend.

Other more general goals of the 2LIKE project are to improve individual student performance and to reduce overall dropout rates. Research results from the project will be made available to all interested stakeholders and to the Moodle user community.
The 2LIKE project, which will run for three and a half years, is also being supported by Ulm University’s Institute of Information Resource Management (OMI), the Institute of Software Engineering and Programming Languages and by the Institute of Databases and Information Systems (DBIS). With its expertise in the field of advanced professional studies, SAPS is also a member of the 2LIKE consortium.
BMBF is funding the 2LIKE project (Lernpfade und Lernprozesse individualisieren durch KI-Methoden / Individualising Learning Paths and Learning Processes Using AI Methods) as part of its research programme ‘Artificial Intelligence in Higher Education’.

Text and mediacontact: Annika Bingmann
Translation: Andrew Symonds

Student at Ulm University (symbolic image: Eberhardt/Ulm University)
Student at Ulm University (symbolic image: Eberhardt/Ulm University)
Prof. Stefan Wesner (l.) and Prof. Birte Glimm
Project lead Prof. Birte Glimm (r.) and project co-lead Prof. Stefan Wesner (photographs: Eberhardt/Ulm University)