Issues in intelligent tutoring systems
Elizabeth T. Whitaker · 1992
A blackboard model consists of a world model, a blackboard and a set of knowledge sources which communicate through the blackboard. Functional modeling of intelligent systems using a blackboard model is accomplished by decomposing the intelligent system into a set of knowledge sources, each containing the knowledge associated with a specific subfunction of the intelligent system. A knowledge source in a blackboard model can be modeled as a blackboard model itself by decomposing the function of the knowledge source into its subfunctions. A knowledge source can also be self-improving, that is, it can use machine learning techniques to improve its ability to perform its function, or it can be a learning knowledge source which uses machine learning techniques to perform its function but does not improve its ability to perform its function. A blackboard model for intelligent tutoring systems (ITSs) is presented. Each knowledge source is functionally decomposed and represented as a blackboard model. A blackboard model for plan recognition systems in ITSs is developed. A set of six dimensions for plan recognition systems is presented and is used for characterization and examination of some existing plan recognition systems. A blackboard model for self-improving ITSs is examined and used to analyze several self-improving tutoring systems. Several significant theories of learning and instruction are examined using the blackboard model for ITSs. For each theory, the roles of the knowledge sources in the blackboard model for ITSs are considered.