Posts Tagged grounding

Second round of evaluation studies completed at BOKU

The second round of evaluation studies has been completed at BOKU. Features studied for their efficiency were Feedback, Grounding and Basic Help. 31 university students were actively working for two afternoons according to tasks provided by a moodle course without much intervention by the teacher. Furthermore an in-depth study was run with two high school students during four days. The target of these formative evaluation studies was to assess, if the  feedback provided by DynaLearn so far is sufficient to support self directed use and learning or not. Intermediate model uploads enabled us to compare models before and after feedback was given by the repository. At the first glance, feedback from an expert model in the repository yielded significant changes in model content and structure even after a short time, indicating the potentially high benefit of this feature for learning. Pre- and post tests, as well motivation questionnaires completed the assessments with regard to learning effects, motivational issues and acceptance.

, , , ,

No Comments

Using DBpedia for terminological grounding

DynaLearn has started using DBpedia (http://dbpedia.org/) as main source of background knowledge for the grounding of the terminology used in qualitative reasoning models. This way, DynaLearn bridges the gap between the loosely and imprecise terminology used by a learner and the well-defined semantics of an external knowledge source with a high level of consensus, such as DBpedia.

DBpedia extracts structured information from Wikipedia to allow sophisticated queries against its data, and to link other data sets on the Web to Wikipedia data. Furthermore, DBpedia is in the core of the emerging Linked Data initiative (http://linkeddata.org/), a term referring to a set of best practices for publishing and connecting structured data on the Web.

Our initial experiments (more details will come in a future post) show a high coverage of DBpedia with regard to the specialized terminology typically used in DynaLearn (77% coverage of almost 1700 words), quite above other studied resources (such as OpenCyc or WordNet). That shows the suitability of using DBpedia as preferred source of knowledge for the grounding of model terms.

,

No Comments