Intelligent digital education space for problem- and user-oriented search of digital learning content (SEARCH)
| Led by: | Verbundprojektleitung: Andre Wilms (NWS MB), Teilprojektleitung IfBE: Prof. Dr. Steffi Robak |
| E-Mail: | steffi.robak@ifbe.uni-hannover.de |
| Team: | Christian Kühn, Katharina Band |
| Year: | 2021 |
| Funding: | BMBF |
| Duration: | 09/2021 bis 08/2024 |
In times of an annual global doubling of knowledge, every (further) education system needs an intelligent filtering of content, if only to avoid overloading its learners. SEARCH is aimed at the manufacturing industry and pursues the goal of not only filtering further education offers, but also combining them into individual learning paths. To this end, we are developing an AI-supported assistance system that dynamically integrates and adapts didactic design formats and personal learning preferences in addition to the content-related educational goals when individualising a learning path. In practice, we rely on the Mobile Learning in Smart Factories (MLS) platform, which is already used by 120 companies for in-house training, and expand it both in terms of content and technology. The existing content is broken down into learning nuggets/microlearnings using AI and enriched with machine-readable information so that it can be automatically integrated into a learning path. New content is also to be created in the corresponding format within the framework of SEARCH. An overarching goal here is transparent content sharing between the participating companies, including data standards for learning resources and formalised didactic design criteria. On the technical application level, we pursue two goals: intelligent search and intelligent recommendations. The search function should take into account the current learning path, the learning goal and the learning progress and operate on the semantic level. In addition, the search and click behaviour or the acceptance of presented search answers of other similar users is taken into account (collaborative filtering). The recommendation function is algorithmically based on the search function, but focuses on the justification of the recommendation. In the spirit of Explainable AI, users should be able to understand why the AI "thinks" a learning nugget is relevant to them individually.
Further information: