However, if it is incorrect, there must be some relationship between the two that allows them to live harmoniously within a single theory. The nature of this relationship is the focus of this article. In passing we note that the theory of weak problem solvers has been well-developed for over a decade; see Kilsson (1971) for example. Some aspects of MYCIN don't fit the problem reduction For example, a THE AI MAGAZINE Summer 1983 25 production whose action part is a conjunction of atomic formulae corresponds to a separate operator for each atomic formula in the conjunction. MYCIN's search strategy effectively applies such operators in a group.

In 1980, Digital Equipment Corporation began to use a rule-based system called Rl by some and XCON by others to configure VAX-11 computer systems In the intervening years, Rl's knowledge has increased substantially and its usefulness to Digital continues to grow. This article describes what is involved in extending Rl's knowledge base and evaluates Rl's performance during the four year period. "Rl: the formative years" described how a A large number of people have played critical roles in Rl's development. Among those who deserve special mention are John Barnwell, Dick Caruso, Ken Gilbert, Keith Jensen, Allan Kent, Dave Kiernan, Arnold K&t, Dennis O'Connor, and Ed Orciuch. We want to thank Allen Newell, Dennis O'Connor, and Ed Orciuch for their helpful comments on earlier drafts of this article Briefly, given a customer's purchase order, Rl determines what, if any, substitutions and additions have to be made to the order to make it consistent, complete, and produce a number of diagrams showing the spatial and logical relationships among the 50 to 150 components that typically constitute a system.

This article provides a historical background on how AAAI came into existence. It provides a rationale for why we needed our own society. It provides a list of the founding members of the community that came together to establish AAAI. Starting a new society comes with a whole range of issues and problems: What will it be called? How will it be financed?

This article gives a detailed description of True Knowledge: a commercial, open-domain question-answering platform. The system combines a large and growing structured knowledge base of commonsense, factual, and lexical knowledge; a natural language translation system that turns user questions into internal language-independent queries; and an inference system that can answer those queries using both directly represented and inferred knowledge. The system is live and answers millions of questions per month asked by Internet users. Behind the platform is a large and growing knowledge base of the world's knowledge in structured form combining commonsense, factual, and lexical knowledge. Natural language questions are answered by first translating the question into a language-independent query and then executing the query using both knowledge in the knowledge base and additional knowledge generated by a general inference system.

We present a deployed AI system, Calvin, for cosmogenic isotope dating, a domain that is fraught with these difficult issues. Calvin solves these problems using an argumentation framework and a system of confidence that uses twodimensional vectors to express the quality of heuristics and the applicability of evidence. The arguments it produces are strikingly similar to published expert arguments. Calvin is in daily use by isotope dating experts. An automated tool can do boring and repetitive reasoning, freeing experts to do more difficult and creative work.

In contrast to an explicit definition of each individual item, configurable products such as computers, financial service portfolios, and cars are repre sented in the form of a configuration knowledge base that de - scribes the properties of allowed instances. Although the knowledge representation used is different compared to nonconfi gurable products, the decision support requirements remain the same: users have to be supported in finding a solution that fits their wishes and needs. In this article we show how recommendation technologies can be applied for supporting the configuration of products. In addition to existing approaches we discuss relevant issues for future research. Similar to knowledge-based recommendation (Burke 2000) configuration is a process where users specify (and often adapt) their requirements and the configuration system provides feedback. Requirements specifications range from feature value definitions to textual queries specified on an informal level. Feedback is provided, for example, in terms of further questions that need to be answered, solutions (configurations), explanations of solutions, and proposals for relaxations of the user requirements in situations where no solution can be found. A major difference between configuration systems and recommender systems in general is the way in which product knowledge is represented. Configuration systems are operating on a configuration knowledge base (Stumptner 1997), which describes the properties of all allowed instances. In contrast to configuration systems, recommender systems are operating on the basis of an assortment of explicitly defined solution alternatives. The reason for using a configuration knowledge base is the large number of solution alternatives (possible configurations), which make an explicit representation infeasible. Although the used knowledge representations are different, the decision support goal is quite the same for both types of systems: users have to be proactively supported in finding a solution that fits their wishes and needs. Configuration systems often achieve this goal only partially since the amount and complexity of options presented by the configurator outstrip the capability of a user to identify an appropriate solution (configuration). Users are unable to find the features they would like to specify, they are unsure about their preferences regarding complex technical product properties, and they do not know how best to adapt their requirements in the case of inconsistencies (if no solution can be identified).

Some people believe that the expert system field is dead, yet others believe it is alive and well. To gain a better insight into these possible views, the first three world congresses on expert systems (which typically attract representatives from some 45-50 countries) are used to determine the health of the global expert system field in terms of applied technologies, applications, and management. This article highlights some of these findings. An excellent way to gain a global perspective on expert system technology, applications, and management is to examine the world congresses on expert systems (sponsored by the International Society for Intelligent Systems in Rockville, Maryland). The World Congress on Expert Systems was established to bridge the gap between the academician and the practitioner and concentrate on expert system work being performed throughout the world.

I survey four viewpoints about what AI is. I describe a program exhibiting AI as one that can change as a result of interactions with the user. Such a program would have to process hundreds or thousands of examples as opposed to a handful. Because AI is a machine's attempt to explain the behavior of the (human) system it is trying to model, the ability of a program design to scale up is critical. Researchers need to face the complexities of scaling up to programs that actually serve a purpose.

PROLOG can compute quantities as answers by extract,ing values from the variable bindings imroduced in the proof of p from S, and so serves as a general purpose programming language. Logical programming languages attract many people in artificial intelligence because of the relative ease of stating declarative information in them, as compared with traditional programming languages Since most knowledge-based, expert systems contain large numbers of essentially declarative statements, the designers of the FGC expect their choice of PROLOG to facilitate the construction and operation of knowledge-based systems Parallelism enters the picture because traditional PROLOG requires that all sentences be expressed in clausal form, searches for proofs of its goal by examining the input clauses in a fixed linear order, and within clauses, examining literals in left-to-right order Many of these imposed orderings have no purely logical basis, so that, as far as questions of deducibility are concerned, greater efficiency may be possible with separat,e deduction searches conducted concurrently. In such a reorganization of PROLOG, time of execution is ideally proportional to the depth of the proof found (the size of the answer), rather than proportional to the number of alternative proofs (the size of the search space). Does S t-p? may be needlessly interest to economists is not affected by such a change. In the second, ordinal utilities were abandoned in, favor of sets of binary preferences among alternatives.

What Is a Knowledge Representation? Although knowledge representation is one of the central and, in some ways, most familiar concepts in AI, the most fundamental question about it--What is it?--has Numerous papers have lobbied for one or another variety of representation, other papers have argued for various properties a representation should have, and still others have focused on properties that are important to the notion of representation in general. In this article, we go back to basics to address the question directly. We believe that the answer can best be understood in terms of five important and distinctly different roles that a representation plays, each of which places different and, at times, conflicting demands on the properties a representation should have.