Abstract

This paper introduces the conjecture that many kinds of cognition and computing may usefully be seen as a search for efficiency in information, where efficiency is defined in terms of Shannon's (1949) concept of redundancy in information. A prototype of a new kind of computing system which is based on this theory is described in outline. Examples from the prototype are presented showing how a search for efficiency in information may achieve autonomous inductive learning, logical deduction and probabilistic inductive inference. 2 INTRODUCTION The nascent field of inductive logic programming has been defined by Muggleton (1990) as "the intersection of Logic Programming and Machine Learning". Within this field, two themes may be distinguished: . Most attention is being paid at present to ways in which logic programs may be created by automatic or semi-automatic induction. . Some attention has been given to the possibility that inductive learning and logical deduction may be integrated. ...

Citations