Intelligent Systems: Practice and Perspective Machine Intelligence, Editor-in-Chief: Donald Michie Volumes 1-7 are published by Edinburgh University Press and in the United States by Halsted Press (a subsidiary of John Wiley & Sons, Inc.) Volumes 8-10 are published by Ellis Horwood Ltd., Publishers, Chichester and in the United State by Halsted Press (a subsidiary of John Wiley & Sons, Inc.) ELLIS HORWOOD LIMITED Publishers - Chichester Halsted Press: a division of JOHN WILEY & SONS New York - Brisbane - Chichester - Toronto First published in 1982 by ELLIS HORWOOD LIMITED Market Cross House, Cooper Street, Chichester, West Sussex, P019 lEB, England The publisher's colophon is reproduced from James Gillison's drawing of the ancient Market Cross, Chichester. Q335 The Library of Congress cataloged this serial as follows -67-13648 ISBN 0-85312-431-0 (Ellis Horwood Limited) ISSN 0076-2032 ISBN 0-470-27323-2 (Halsted Press) Typeset in Press Roman by Ellis Horwood Limited. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the permission of Ellis Horwood Limited, Market Cross House, Cooper Street, Chichester, West Sussex, England. The year of the first MI Workshop, 1965, was a milestone for more reasons than one, not least for the appearance of a paper entitled "A machine-oriented logic based on the resoluton principle". It was appropriate and a cause for pleasure that the author of that paper, J. A. Robinson, was the opening contributor to the scientific proceedings of MI-10, held in November 1981 at Case Western Reserve University, Cleveland, USA.

In many applied sciences there is often a problem of revealing a structure underlying a given collection of objects (situations, measurements, observations, etc.). A specific problem of this type is that of determining a hierarchy of meaningful subcategories in such a collection. This problem has been studied intensively in the area of cluster analysis. The methods developed there, however, formulate subcategories ('clusters') solely on the basis of pairwise'similarity' (or'proximity') of objects, and ignore the issue of the'meaning' of the clusters obtained. The methods do not provide any description of the clusters obtained. This paper presents a method which constructs a hierarchy of subcategories, such that an appropriately generalized description of each subcategory is a single conjunctive statement involving attributes of objects and has a simple conceptual interpretation. The attributes may be many-valued nominal variables or relations on numerical variables. The hierarchy is constructed in such a way that a flexibly defined'cost' of the collection of descriptions which branch from any node is minimized. Experiments with the implemented program, CLUSTER/paf, have shown that for some quite simple problems the traditional methods are unable to produce a structuring of objects most'natural' for people, while the method presented here was able to produce such a solution.

This paper has three themes: (1) The task of acquiring and organizing the knowledge on which to base an expert system is difficult.

AL3 (Advice Language 3) is a problem-solving system whose structure facilitates the implementation of knowledge for a chosen problem-domain in terms of plans for solving problems, pieces-of-advice', patterns, motifs, etc. AL3 is a successor of ALI and AL 1.5 (Michie 1976, Bratko & Michie 1980a, I980b, Mozetic 1979). Experiments in which AU was applied to chess endgames established that it is a powerful tool for representing search heuristics and problem-solving strategies. The power of ALI lies mainly in the use of a fundamental concept of AU: piece-of-advice. A piece-of-advice suggests what goal should be achieved next while preserving some other condition. If this goal can be achieved in a given problem-situation (e.g. a given chess position) then we say that the piece-ofadvice is'satisfiable' in that position.

Motor control systems are complex systems that process information. Orientation behaviour, posture control, and the manipulation of objects are examples of motor control systems which involve one or more sensory modality and various central neural processes, as well as effector systems and their immediate neuronal control mechanisms. Like all complex information processing systems, they must be analysed and understood at several different levels (see, e.g., Marr & Poggio 1977). At the lowest level there is the analysis of basic components and circuits, the neurons, their synapses, etc. At the other extreme, there is the study of the computations performed by the system -- the problems it solves and the ways that it solves them -- and the analysis of its logical organization in terms of its primary modules. Each of these levels of description, and those in-between, has its place in the eventual understanding of motor control by the nervous system. None is sufficient, nor is there any simple translation from one to another. A purely biophysical investigation, however exhaustive, can say nothing by itself about the information processing performed by the system, nor, on the other hand, can an understanding of the computational problem which the system solves lead directly to an understanding of the properties of the hardware. Two examples of motor control theories belonging to different levels will illustrate this point.

The notion of an ethical machine can be interpreted in more than one way. Perhaps the most important interpretation is a machine that can generalize from existing literature to infer one or more consistent ethical systems and can work out their consequences. An ultra-intelligent machine should be able to do this, and that is one reason for not fearing it. INTRODUCTION There is fear that'the machine will become the master', especially compounded by the possibility that the machine will go wrong. There is, for example, a play by E. M. Foster based on this theme.

We briefly describe the logic programming language PROLOG concentrating on those aspects of the language that make it suitable for implementing expert systems. We show how features of expert systems such as: (1) inference generated requests for data, (2) probabilistic reasoning, (3) explanation of behaviour can be easily programmed in PROLOG. We illustrate each of these features by showing how a fault finder expert could be programmed in PROLOG.

PROLOG is a simple and powerful progamming language based on first-order logic. This paper examines two possible extensions to the language which would generally be considered "higher-order".t The first extension introduces lambda expressions and predicate variables so that functions and relations can be treated as'first class' data objects. We argue that this extension does not add anything to the real power of the language. The other extension concerns the introduction of set expressions to denote the set of all (provable) solutions to some goal. We argue that this extension does indeed fill a real gap in the language, but must be defined with care.

Seven years or so after it was first proposed (Kowalski 1974), the technique of'logic programming' today has an enthusiastic band of users and an increasingly impressive record of applications. For most of these people, logic progamming means PROLOG, the system defined and originally implemented by the Marseille group (Roussel 1975). PROLOG has since been implemented in several other places, most notably at Edinburgh (Warren et al. 1977). Much of the rapid success of logic progamming is due to these implementations of PROLOG (as well as to the inspired missionary work of Kowalski, van Emden, Clark and others). The Edinburgh PROLOG system is in particular a superb piece of software engineering which allows the logic progammer to compile assertions into DEC-10 machine code and thus run logic programs with an efficiency which compares favourably with that of compiled LISP. All other implementations of logic programming (including our own, which we describe in this paper) are based on interpreters.

We are interested in constructing a computer agent whose behaviour will be intelligent enough to perform cooperative tasks involving other agents like itself. The construction of such agents has been a major goal of artificial intelligence research. One of the key tasks such an agent must perform is to form plans to carry out its intentions in a complex world in which other planning agents also exist. To construct such agents, it will be necessary to address a number of issues that concern the interaction of knowledge, actions, and planning. Briefly stated, an agent at planning time must take into account what his future states of knowledge will be if he is to form plans that he can execute; and if he must incorporate the plans of other agents into his own, then he must also be able to reason about the knowledge and plans of other agents in an appropriate way.