Data for Enabling Content in Adaptive Learning Systems
DECALS was completed with funding from the US Advanced Distributed Learning (ADL) initiative and is described on the ADL site.
DECALS produced three outcomes:
An open source learning object repository (called DECALS) that auto-registers resources in the learning registry. This outcome is described on the ADL site.
Research into (and demonstrations of) machine-actionable metadata. Traditional metadata was developed for human comprehension. Tags such as keywords, author name, and learning resource type are easily interpreted by people but less so by computers. This was not a problem when these tags were only displayed in course catalogues or used as search fields, but AI-driven adaptive learning systems and recommender systems need to understand their meaning to gauge the relevance of available learning activities to an individual learner profile. The solution is metadata that contains data that a machine can use to interpret or resolve its meaning. DECALS produced a specification for such metadata (called “decals”) and examples of how they work.
A more effective means of searching context-rich repositories. Google revolutionized search by considering the link structure of the web and using it as a measure of popularity or importance. Search algorithms have progressed further, but standard search methods (including Lucene / Elastic) still struggle in situations where words have special meanings and context is more important than popularity. The DECALS project responded by developing an “interactive search” that engages the user in disambiguating the meaning of search terms and includes some whistles and bells such as sorting on metadata fields (e.g. grade level) and support for user ratings and comments. You can try this search and compare it basic search in the learning registry yourself.