interpreter for logic programs (Sterling & Shapiro, 1986)...................138 1 1. INTRODUCTION The most important outcome of AI research during the 70s was the general acceptance of the major role of knowledge in intelligent systems (Buchanan & Feigenbaum, 1982). Lenat and Feigenbaum (1989) call this belief the knowledge as power hypothesis and assert it as: "The knowledge principle (KP) A system exhibits intelligent understanding and action at a high level of competence primarily because of the specific knowledge that it can bring to bear: the concepts, facts, representations, methods, models, metaphors, and heuristics about its domain of endeavor." Or as Buchanan and Feigenbaum (Buchanan & Feigenbaum, 1982) put it, "the power of an intelligent program to perform its task well depends primarily on the quantity and quality of knowledge it has about that task." Thus, it is not surprising that the general attitude toward knowledge was a greedy one - grab as much knowledge as you ca...

1.1 Representations in problem solving....................... 4

Even we human cannot solve the frame problem, that is, we cannot completely describe conditions for a certain action to succeed, nor can we completely predict what will happen as the result of an action. Nevertheless, in daily life we can act as if there were no such problem, because we use some heuristics to bypass it. In this paper, we propose that the concept of causality is the heuristics. We also claim that temporal ordering in causal relation is not essential. Any partial ordering suffices but once the order is fixed, it should be followed even in reverse order reasoning such as inferring past from future. We formalize causal reasoning and apply it to YSP. We observe that there are two kinds of non-monotonicity in the reasoning of changes: one due to lack of information, the other due to change itself. Traditionally, situation calculus with non-monotonic logics of various kind are used to attack the frame problem. Two kinds of monotonicity are mixed in those formalisms. It is mai...

This paper presents a speed-up learning method for logic programs, which accelerates a program by composing macro clauses based on partial structures of proof trees. Many systems have been proposed for composing useful macros, e.g., some of them select macros that connects two peaks in a heuristic function. Another employs heuristics that select useful macros. Although they work well in some domains, such methods depend on domaindependent heuristics that have to be exploited by their users.