knowledge teaching : an alternative strategy for knowkedge-base development

Abstract

The development of knowledge-based systems relies heavily on the transfer of human expertise into a structured knowledge base - a process known as knowledge acquisition or knowledge elicitation. This process is widely recognized as complex and time-consuming, often referred to as the knowledge acquisition bottleneck. Traditionally, this transfer has been facilitated by a knowledge engineer, or more recently, by automated acquisition tools. This thesis proposes a novel perspective by treating knowledge acquisition as an educational activity, where the domain expert assumes the role of a teacher, and the receiving system - a computer program called the expertise-student program (ESP) - acts as a student. Motivated by the observation that all forms of knowledge transfer inherently resemble educational processes, this work contrasts the prevailing student-oriented acquisition methods with a teacher-oriented paradigm, which more closely mirrors human educational practices. The thesis explores the feasibility of implementing a general teaching strategy for knowledge acquisition, emphasizing the need for domain experts (expertise-teachers) to structure and deliver knowledge in a form understandable by the ESP. Key concepts such as the nature of knowledge, its distinction from data and information, and the nonuniformity of knowledge are discussed. A classification of existing acquisition techniques is proposed based on the relative contributions of the source and destination in the transfer process. The thesis argues that current methods are predominantly student-driven, despite the effectiveness of teacher-led education in human learning. A prototype ESP is developed to operationalize the teaching paradigm. Its architecture includes specialized modules for representing and processing various forms of knowledge. The ESP is applied to build a tutoring system that integrates pedagogical knowledge, domain knowledge, and knowledge acquired through conventional means. The implementation highlights practical challenges and demonstrates the potential of the teaching-based approach to improve the efficiency, adaptability, and robustness of expert system development. The study concludes by emphasizing the value of the knowledge teaching paradigm in overcoming limitations of traditional acquisition methods and suggests directions for future work.