Description of an Instructional Ontology and its Application in Web Servicesfor Education
Carsten Ullrich
∗
German Research Center for Artificial Intelligence, Saarbr¨ucken, Germany
1 Motivation
From an educational point of view, the potential of the Webis far from being reached. Although it offers an abundanceof learning resources the search for educational resources isstilltedious, andinteroperabilityofe-learning systemsonlya wished-for goal. What is missing is a common ground forboth human and machines. While there exist standards fordescribing learning resources, they fail to address the in-structional purpose of a resource, for instance whether aweb page provides a definition or a counter-example of aconcept. However, such an explicit representation of theinstructional function provides humans with a shared vo-cabulary and can serve as the basis for the semantic in-teroperability for machines. This article describes such abasis: an ontology of instructional objects (OIO) that cap-tures the function of a learning resource. The article startswith describing several educational Web services that canbenefit from it. Then, the ontology is described in detail.
2 Educational Web Services
The additional pedagogically relevant information of anontology of instructional objects brings forth better Webservices. If the resources are annotated appropriately, Webservices can use this information. More specifically, itincreases the accurateness of a service because at designtime, a Web service developer can foresee different func-tionality depending on the type of the resource. For mosteducational services, the information whether a resourcecontains a definition or an example will be of use. Simi-larly, service composition is enhanced. For instance, a re-questerservicecanrequiredifferentactionsfromaproviderdepending on the instructional type of a resource.Examples of Web services that benefit from an OIO arefor instance course generators ([3]). They assemble learn-ing resources to a curriculum taking into account learnerproperties such as knowledge. If (third-party) resources areannotated by their instructional function, a course genera-tor can include them appropriately in a curriculum, depend-ing on the chosen pedagogical strategy. For instance, in aproblem-based approach, to first present a real-world prob-lem and the necessary definitions afterwards. A learnermodel that stores personal preferences and informationabout the learner’s mastery of domain concepts can profit
∗
This work was generated in the EC project LeActiveMath.The author is solely responsible for its content.
from the annotation, too. For instance, reading an exam-ple should trigger a different updating of the mastery of aconcept than solving an exercise. Other educational Webservices profit in a similar way, e.g., data mining (extract-ing pedagogically relevant information from the paths of a learner through the learning material), feedback in inter-active exercises, intelligent assistants, adaptive hypermediaservices such as link annotators, and authoring support (byproviding hints to the author about missing instructionalitems).
3 Description of the Ontology
The goal of this work is to provide an ontology that de-scribes a learning resource from an instructional perspec-tive. The ontology does not describe the content taughtby the learning material, e.g., concepts in physic and theirstructure. Instead, each class of the ontology stands for aparticularinstructionalrolealearningresourcecanplay, forinstance a paragraph in a text-book. However, it does notcontradict LOM or related initiatives which operate at thedomain level; all these approaches can very well be used inparallel.For this ontology to be as broadly applicable as possible,I analyzed about 30 sources, ranging from instructional de-sign theories and to other e-learning systems. Additionally,several instructional experts reviewed the ontology. Theontology was implemented in OWL. Figure 1 shows theclass hierarchy.
Instructional object
is the root class of the ontology.Its properties are a unique identifier and a subset of theIMS/IEEE LOM[2]metadata.The class
concept
subsumes instructional objects thatdescribe the main pieces of information being taught. A
depends-on
property represents connections between con-cepts.A
fact
describesaneventorsomethingthatholdswithoutbeing a general rule, e.g., historical facts.A
definition
states the meaning of a word, phrase, orsymbol. Often, it describes a set of conditions or circum-stances that an entity must fulfill in order to count as aninstance of a class.A
law
describes a general principle between phenomenaor expressions that has been proven to hold. Its subclass
law of nature
is a scientific generalization based on obser-vation. A
theorem
describes an idea that has been (mathe-matically) demonstrated as true.
Leave a Comment