This paper suggests that artificial intelligence in education (AIEd) can be fruitfully analysed as 'policies frozen in silicon'. This means that they exist as both materialised and proposed problematisations (problem representations with corresponding solutions). As a theoretical and analytical response, this paper puts forward a heuristic lens…

Provides a brief introduction to the basics of artificial intelligence and expert systems and explains terms in general use. (CLB)

Human expert problem-solving in science is defined and used to account for scientific discovery. These ideas are used to describe BACON.5, a machine expert problem solver that discovers scientific laws using data-driver heuristics and "expectations" such as symmetry. Implications of BACON.5 type research for traditional science education…

Reviews the development of artificial intelligence systems and the mechanisms used, including knowledge representation, programing languages, and problem processing systems. Eleven books and 6 journals are listed as sources of information on artificial intelligence. (23 references) (CLB)

This paper briefly reviews research on tasks in knowledge-rich domains including developmental studies, work in artificial intelligence, studies of expert/novice problem solving, and information processing analysis of aptitude test tasks that have provided increased understanding of the nature of expertise. Particularly evident is the finding that…

Reviews approaches to machine learning (development of techniques to automate acquisition of new information, skills, and ways of organizing existing information) in symbolic domains. Four categorical tasks addressed in machine learning literature are examined: learning from examples, learning search heuristics, learning by observation, and…

Describes an artificial intelligence system known as SOAR that approximates a theory of human cognition. Discusses cognition as problem solving, working memory, long term memory, autonomy and adaptability, and learning from experience as they relate to artificial intelligence generally and to SOAR specifically. Highlights the status of the…

Current notions of metacognition merge with the predominant scientific model used in psychology, that of information processing. Metacognition is seen as a control process that governs the action of more elemental cognitive skills. Given the centrality of this notion, it is important that metacognition should be examined in detail. From the point…

Examines processes by which general but weak search methods are transformed into powerful, domain-specific search strategies by classifying types of heuristics learning that can occur and components that contribute to such learning. A learning system--SAGE.2--and its structure, behavior in different domains, and future directions are explored. (36…

A broad range of well-structured problems--embracing forms of diagnosis, catalog selection, and skeletal planning--are solved in expert computer systems by the method of heuristic classification. These programs have a characteristic inference structure that systematically relates data to a pre-enumerated set of solutions by abstraction, heuristic…

Reacts to Ohlsson, Ernst, and Rees' paper by initially discussing the costs of methodology that utilizes artificial intelligence (AI) to model cognitive processes. Raises three concerns with the paper: insufficient clarification of the meaning of conceptual versus procedural understanding of base-10 subtraction; realism of the learning model; and…

Discusses two computer problem solving programs: rule-based expert systems and decision analysis expert systems. Explores the application of expert systems to automated chemical analyses. Presents six factors to consider before using expert systems. (MVL)

Compares a program written both in BASIC and LOGO on its inferential and decision making ability. Explains steps in each program and how deductions and decisions are made. (MVL)

Traces the history and function of State, Operator, And Result (SOAR), a general-purpose artificial intelligence program for solving problems. The SOAR can "chunk" the result of a subgoal and learn from previous experiences. The SOAR could be applied to various expert systems. (YP)

This paper describes NEOMYCIN, a computer program that models one physician's diagnostic reasoning within a limited area of medicine. NEOMYCIN's knowledge base and reasoning procedure constitute a model of how human knowledge is organized and how it is used in diagnosis. The hypothesis is tested that such a procedure can be used to simulate both…