O ntological E ngineering of Instruction: A Perspective

In n n ns s s st t t tr r r ru u u uc c c ct t t In order to design an Intelligent Instructional System (IIS), the designer has to know w hat instructional theories and m odels m ay provide appropriate principles and strategies, w hat functional com ponents are necessary for reaching instructional goals and obj ectives, w hat pedagogical actions are appropriate for each situation, w hat architecture and strategy are appropriate for supporting these actions, how to control the system behavior to achieve a coherent learning support process, and how to organize dom ain know ledge. H ow ever, these fundam ental characteristics of an IIS are often im pl i cit, vague or ill-form ed. D espi te m uch theoretical research and im pl em entation of IISs, there is little to link the tw o, and relations betw een research and im pl em entation are not strong. W hat i s needed is a w ell-form ed system of concepts w hich sum m arizes w hat w e have learned to date and characterizes areas of agreem ent, as w el l as areas w here disagreem ents indicate the need for further elaboration. A solution to this problem could be the ontological engineering of instructi on. A n ontology consists of a task ontology w hich characterizes the com putational architecture of a know ledge-based system w hich perform s a task, and dom ain ontology w hich characterizes the dom ain know l edge w here the task is perform ed, such as diagnosis, m oni toring, scheduling, or design. Instruction is a task, as is supporting the learning process. Task ontology m i ght provide an effective m ethodology and vocabulary for both analyzing and synthesi zing know ledge-based system s to w hich IISs belong, w ith benefits such as: a com m on vocabulary, m aki ng know ledge explicit, system atization, standardization, and m eta-m odel functi onality. This functionality suggests the possibility of an ontology-aw are authoring functionality w hich could be very intel ligent in the sense that it w ould know w hat m odel w ould help authors. M i zoguchi [1] has proposed the follow ing three levels of ontologies. Level 1 is a structured collection of term s. The m ost fundam ental task in …