Design for a network-based education system that evolves by peer and instructor interaction

A framework for a globally-distributed computer network education-delivery system is proposed. Also presented are some structural details of designs to promote the system's continued evolution requiring the generation and reuse of successful teaching and learning items. The system's growth is predicated on the basic idea that learning amongst peers in common or related domains can be most effective when the students are self-paced in contributing and evaluating items to be used simultaneously by others. With this system, item content will finally be delivered by a integration of all the multimedia channels made practically feasible by technology. The system is to facilitate the definition of concept-related clusters for learning items that are appropriate for students interested in solving a variety of classes of problems. Initially, links or bindings among the classes of clusters can be based on classification knowledge that is normally identified in traditional educational systems. Learning items will be accumulated and retained on the basis of statistical performance measures. The knowledge to define item clusters will be embedded in programs written to be intelligent agents. These agents will contain algorithms or methods for dynamically improving cluster definitions. Program systems will accumulate information on each student and on the items. This facilitates the assignment of a grade to each student to indicate his/her level of mastery or ability to perform in each area identified as a cluster center.