The drawbacks of pure probabilistic methods and of the certainty factor model have led us in recent years to consider alternate approaches. Particularly appealing is the mathematical theory of evidence developed by Arthur Dempster. We are convinced it merits careful study and interpretation in the context of expert systems. This theory was first set forth by Dempster in the 1960s and subsequently extended by Glenn Sharer. In 1976, the year after the first description of CF's appeared, Shafer published A Mathematical Theory of Evidence (Shafer, 1976). Its relevance to the issues addressed in the CF model was not immediately recognized, but recently researchers have begun to investigate applications of the theory to expert systems (Barnett, 1981; Friedman, 1981; Garvey et al., 1981). We believe that the advantage of the Dempster-Shafer theory over previous approaches is its ability to model the narrowing of the hypothesis set with the accumulation of evidence, a process that characterizes diagnostic reasoning in medicine and expert reasoning in general. An expert uses evidence that, instead of bearing on a single hypothesis in the original hypothesis set, often bears on a larger subset of this set. The functions and combining rule of the Dempster-Shafer theory are well suited to represent this type of evidence and its aggregation. For example, in the search for the identity of an infecting organism, a smear showing gram-negative organisms narrows the hypothesis set of all possible organisms to a proper subset. This subset can also be thought of as a new hypothesis: the organism is one of the gram-negative organisms.

The. development of automated assistance for medical diagnosis and decision making is an area of both theoretical and practical interest. Of methods for utilizing evidence to select diagnoses or decisions, probability theory has the firmest appeal. Probability theory in the form of Bayes' Theorem has been used by a number of" workers (Ross, 1972). Notable among recent developments are those of de Dombal and coworkers (de Dombal, 1973; de Dombal et al., 1974; 1975) and Pipberger and coworkers (Pipberger et al., 1975). The usefulness of Bayes' Theorem is limited practical difficulties, principally the lack of data adequate to estimate accurately the a priori and conditional probabilities used in the theorem. One attempt to mitigate this problem has been to assume statistical independence among various pieces of evidence. How seriously this approximation affects results is often unclear, and correction mechanisms have been explored (Ross, 1972; Norusis and Jacquez, 1975a; 1975b). Even the independence assumption requires an unmanageable number of estimates of" probabilities for most applications with realistic complexity.

Questioning of the expert gradually reveals, however, that despite the apparent similarity to a statement regarding a conditional probability, the number 0.7 differs significantly from a probability. The expert may well agree that P(hl]sl & s2 & s:0 0.7, but he becomes uneasy when he attempts to follow the logical conclusion that therefore P( hllS 1 & s 2 & s) 0.3. He claims that the three observations are evidence (to degree 0.7) in favor of the conclusion that the organism is a Streptococcus and should not be construed as evidence (to degree 0.3) against Streptococcus. We shall refer to this problem as Paradox 1 and return to it later in the exposition, after the interpretation of the 0.7 in the rule above has been introduced. It is tempting to conclude that the expert is irrational if he is unwilling to follow the implications of his probabilistic statements to their logical conclusions.

Please read it and send me comments, objections, etc. 1) Victor [Yu] has assigned certainty factors to his rules based on the relative strengths of the evidence in these rules. While trying to find a numerical scale that would work as he wanted it to with the system's 0.2 cutoff and combining functions, he had to adjust certainty factors of various rules. Now that this scale has been established, however, he assigns certainty factors using this scale, and does NOT adjust certainty factors of rules if he doesn't like the system's performance. Furthermore, he does NO combinatorial analysis before determining what CF to use; he is satisfied that using the scale he has devised, the system's combining function, and the 0.2 cutoff, the program will arrive at the right results for any combination of factors, and if it doesn't, he looks for missing information to add. 2) Assuming that the parameters IDENT and COVERFOR are disambiguated in Victor's set of rules, Ted [Shortliffe] believes the CF's that Victor uses in his rules, and approves of the idea of using a cutoff for COVERFOR since this is what we've been doing with bacteremia (since it is a binary decision, a cutoff makes sense for COVERFOR). Furthermore, this is quite similar to what clinicians do: they accumulate lots of small bits of clinical evidence, then decide if the total is enough to make them cover [or a particular organism--independent of what the microbiological evidence suggests.

Whereas much early work in artificial intelligence was devoted to the search for a single, powerful, domain-independent problem-solving methodology [e.g., GPS (Newell and Simon, 1972)], subsequent efforts have stressed the use of large stores of domain-specific knowledge as a basis for high performance. The knowledge base for this sort of program [e.g., DENDRAL (Feigenbaum et al., 1971), MACSYMA (Moses, 1971)] is often assembled by hand, an ongoing task that may involve several person-years of effort. A key element in constructing a knowledge base is the transfer of expertise from a human expert to the program. Since the domain expert often knows nothing about programming, the interaction between the expert and the pertormance program usually requires the mediation of a human programmer. We have sought to create a program that could supply much the same sort of assistance as that provided by the programmer in this transfer of expertise. The result is a system called TEIRESIAS 1 ...

The builders of a knowledge-based expert system must ensure that the system will give its users accurate advice or correct solutions to their problems. The process of verifying that a system is accurate and reliable has two distinct components: checking that the knowledge base is correct, and verifying that the program can interpret and apply this information correctly. The first of these components has been the focus of the research described in this chapter; the second is discussed in Part Ten (Chapters 30 and 31). Knowledge base debug, ng, the process of checking that a knowledge base is correct and complete, is one component of the larger problem of knowledge acquisition. This process involves testing and refining the system's knowledge in order to discover and correct a variety of errors that can arise during the process of transferring expertise from a human expert to a computer system. In this chapter, we discuss some common problems in knowledge acquisition and debugging and describe an automated assistant for checking the completeness and consistency of the knowledge base in the ONCOCIN system (discussed in Chapters 32 and 35).

From early experience building the DENDRAL system, it was obvious to us that putting domain-specific knowledge into a program was a bottleneck in building knowledge-based systems (Buchanan et al., 1970). In other systems of" the 1960s and early 1970s, items of knowledge were cast as LISP functions. For example, in the earliest version of DENDRAL the fact that the atomic weight of carbon is 12 was built into a function, called WEIGHT, which returned 12 when called with the argument C. The function "knew about" several common chemical elements, but when new elements or new isotopes were encountered, the function had to be changed. Because we wanted to keep our programs "lean" to run in 64K of working memory, we gave our programs only as much knowledge as we thought they would have to know. Thus we often encountered missing items in running new test cases. It was very quickly seen that LISP property lists (data structures) were a superior alternative to LISP code as a way of storing ...

A program that is designed to provide sophisticated expert advice must cope with the needs of naive users who may find the advice puzzling or difficult to accept. This chapter describes additions to MYCIN that provide for explanations of its therapy decisions, the lack of which was a shortcoming of the original therapy recommendation code described in Section 5.4 of Chapter 5. It deals with an optimization problem that seeks to provide "coverage" for organisms while minimizing the number of drugs prescribed. There are many factors to consider, such as prior therapies and drug sensitivities, and a person often finds it hard to juggle all of the constraints at once. When the optimal solution is provided by a computer program, its correctness may not be immediately obvious to the user. This motivates our desire to provide an explanation capability to justify the program's results. The explanation capability derives from two basic programming considerations.

In this chapter MYCIN's implementation is presented in considerable detail. Our goals are to explain the data and control structures used by the program and to describe some of the complex and often unexpected problems that arose during system implementation. In Chapter 1 the motivations behind many of MYCIN's capabilities were mentioned. The reader is encouraged to bear those design criteria in mind throughout this chapter. This chapter specifically describes the Consultation System. This subprogram uses both system knowledge from the corpus of rules and patient data entered by the physician to generate advice for the user. Furthermore, the program maintains a dynamic data base, which provides an ongoing record of the current consultation. As a result, this chapter must discuss both the nature of the various data structures and how they are used or maintained by the Consultation System. Section 5.1 describes the corpus of rules and the associated data structures. It provides a formal ...

It is important to cover for the following probable infection(s) and associated organism(s): INFECTION-1 is BACTEREMIA ITEM-l E.COLI [ORGANISM-I] ITEM-2 KLEBSIELLA [ORGANISM-I] ITEM-3 ENTEROBACTER [ORGANISM-I] ITEM-4 KLEBSIELLA-PNEUMONIAE [ORGANISM-I]