This paper reviews many of the very varied concepts of uncertainty used in AI. Because of their great popularity and generality “parallel certainty inference” techniques, so-called, are prominently in the foreground. We illustrate and comment in detail on three of these techniques; Bayes' theory (section 2); Dempster-Shafer theory (section 3); Cohen's model of endorsements (section 4), and give an account of the debate that has arisen around each of them. Techniques of a different kind (such as Zadeh's fuzzy-sets, fuzzy-logic theory, and the use of non-standard logics and methods that manage uncertainty without explicitly dealing with it) may be seen in the background (section 5).

The discussion of technicalities is accompanied by a historical and philosophical excursion on the nature and the use of uncertainty (section 1), and by a brief discussion of the problem of choosing an adequate AI approach to the treatment of uncertainty (section 6). The aim of the paper is to highlight the complex nature of uncertainty and to argue for an open-minded attitude towards its representation and use. In this spirit the pros and cons of uncertainty treatment techniques are presented in order to reflect the various uncertainty types. A guide to the literature in the field, and an extensive bibliography are appended.

This review presents a brief survey of approaches to describing dialogues with computers which are relevant to improving communication between users and knowledge based systems. One approach to achieving improvements is to exploit the conventions of human communication. The review draws together some of the work on co-operative behavour which reflects these conventions. An analysis is presented which goes some way towards identifying the knowledge required and the issues to be tackled.

The concept of an expert system covers an increasingly large group of software packages which often have more dissimilarities than points in common. We shall not attempt to give a precise definition of an expert system here, because this might impose too restrictive a framework on the rest of our discussion. We shall simply state that, as is generally recognized, an expert system is a piece of software intended to resolve a certain category of problems, that it uses for this purpose a large quantity of knowledge specific to the field in question, and that in each expert system there is a very distinct separation between this knowledge and the procedures which make use of it.