Flexible Learning Object Metadata

Abstract

By far the most popular specification for learning objects is the IEEE Learning Object Metadata (LOM) standard. In it are outlined 76 different elements that correspond to pedagogical, technical, and administrative aspects of learning objects.This standard, however, has proven to be ineffective for creating computer adapted dynamic courseware. This paper outlines some initial research we are doing in acquiring, describing, and using learning object metadata. Instead of the IEEE LOM, we argue for a more flexible approach to both defining and associating metadata with learning objects. By creating domain, educational, and learner characteristic ontologies, content can be dynamically linked to those competencies that are observed in a running e-learning system. This provides for a set of evolutionary metadata, where software agents can inspect multiple metadata instances for a given learning object and reason over them for a particular goal. As more metadata instances are added to the system, agents are expected to be able to provide more accurate reasoning, eventually leading to the dynamic delivery of personalized course content.