A case-based agent framework for adaptive learning

An adaptive learning system centres on the learner's needs and meeting those needs at his or her level or pace. Designing the nature and quality of adaptive interaction that will facilitate association of prior knowledge and current stimuli is thus crucial. Case-based reasoning is proposed as the mode of inference or learning in an agent-based context. An agent is chosen because of its ability to reason like a human and its ability to take on the human role of mentor. Considering that information is often chunked in granules in the form of concepts, the case-based agents feedback should strive to encourage the formation of associative memory among these granules that will meet respective learning goals through various forms of media and representations of several types of knowledge. Concepts are situated in contexts and the granules for contexts are cases denoting a problem situation. Hence, quality interaction design will have to involve a cognitive interface that provides salient details, which will trigger associations with experiences and previous cases. In cases where case-based reasoning can only provide possible alternatives, induction serves as a bridge to communicate the underlying propositions or units of meaning in the working memory.