Toward a Unified Approach for Conceptual Knowledge Acquisition
Larry A. Rendell · American Association for Artificial Intelligence eBooks · 1983
In keeping with a desire to abstract general principles in AI, this article begins to examine some relationships among heuristic learning in search, classification of utility, properties of certain structures, measurement of acquired knowledge, and efficiency of associated learning. In the process, a simple definition is given for conceptual knowledge, considered as information compression The discussion concludes that domain-specific conceptual knowledge can be acquired Among other implications of the analysis is that statistical observation of probabilities can result in the equivalent of planning, in low susceptibility to error, and in efficient learning. SEVERAL RESEARCHERS HAVE INDICATED that more integration and synthesis may be imminent in AI. In a panel discussion on “Challenges of the Eighties” at a workshop on Machine Learning recently held (Proc. IMLW, 1983), several such opinions were stated. Among other issues, Michalski stressed the importance of unification of terminology and extraction of general principles. Amarel suggested we need both theory and application, even within a single project (one supports the other).Another indicator: Chapter XIV Handbook of Artificial Intelligence (Dietterich, London, Clarkson & Dromey, 1982) compares and contrasts generalization methods; Michalski (1983) develops a unified theory for induction, and characterizes its types. (In fact synthesis is