Influence diagrams are a directed graph representation for uncertainties as probabilities. Influence diagrams have been used for the last ten years as a model structuring and elicitation device in the practical field of decision analysis. They have been a powerful communication tool during the initial discussion about a problem, as well as when explaining results after analysis. Because the diagrams are heirarchical, with the numbers "hidden" within the nodes Within the last few years, a number of theoretical results allow for the analysis to be performed directly on the influence diagram-- as assessed. In general, these techniques apply a sequence of transformations to different influence diagrams, to solve either probabilistic inference or decision analysis problems.

Much artificial intelligence research focuses on the problem of deducing the validity of unobservable propositions or hypotheses from observable evidence.! Many of the knowledge representation techniques designed for this problem encode the relationship between evidence and hypothesis in a directed manner. Moreover, the direction in which evidence is stored is typically from evidence to hypothesis.

The causal Bayesian approach is based on the assumption that effects (e.g., symptoms) that are not conditionally independent with respect to some causal agent (e.g., a disease) are conditionally independent with respect to some intermediate state caused by the agent, (e.g., a pathological condition). This paper describes the development of a causal Bayesian model for the diagnosis of appendicitis. The paper begins with a description of the standard Bayesian approach to reasoning about uncertainty and the major critiques it faces. The paper then lays the theoretical groundwork for the causal extension of the Bayesian approach, and details specific improvements we have developed. The paper then goes on to describe our knowledge engineering and implementation and the results of a test of the system. The paper concludes with a discussion of how the causal Bayesian approach deals with the criticisms of the standard Bayesian model and why it is superior to alternative approaches to reasoning about uncertainty popular in the Al community.

This paper studies the underlying rationality of those languages on the syntax and calculus grounds. Some implications of the theorem to the relationship between the CF and the Bayesian languages and the Dempster-Shafer Theory of Evidence are presented. In order for a computer program to be a plausible --odel of a (mora or less) rational process of human expertise, the program should be capable of representing beliefs in a language that is (more or less) calibrated with a well-specified normative criterion, e.g. the axioms of Subjective Probability [15], the Theory of Confir.nation Tversky, the building blocks· of a probabilistic language are syntax, calculus, and semantics [18]. The-- is a set of numbers, co--only referred to as Degrees of Belief (e.g. standard probabilities or Certainty Factors), which are used to parameterize uncertain facts, inexact rules, and competing hypotheses.

Results on approximate deduction in the context of the calculus of evidence of Dempster-Shafer and the theory of interval probabilities are reported. Approximate conditional knowledge about the truth of conditional propositions was assumed available and expressed as sets of possible values (actually numeric intervals) of conditional probabilities. Under different interpretations of this conditional knowledge, several formulas were produced to integrate unconditioned estimates (assumed given as sets of possible values of unconditioned probabilities) with conditional estimates. These formulas are discussed together with the computational characteristics of the methods derived from them. Of particular importance is one such evidence integration formulation, produced under a belief oriented interpretation, which incorporates both modus ponens and modus tollens inferential mechanisms, allows integration of conditioned and unconditioned knowledge without resorting to iterative or sequential approximations, and produces elementary mass distributions as outputs using similar distributions as inputs.

A learning algorithm is presented which given the structure of a causal tree, will estimate its link probabilities by sequential measurements on the leaves only. Internal nodes of the tree represent conceptual (hidden) variables inaccessible to observation. The method described is incremental, local, efficient, and remains robust to measurement imprecisions.

Explanation facilities are a particularly important feature of expert system frameworks. It is an area in which traditional rule-based expert system frameworks have had mixed results. While explanations about control are well handled, facilities are needed for generating better explanations concerning knowledge base content. This paper approaches the explanation problem by examining the effect an event has on a variable of interest within a symmetric Bayesian inferencing system. We argue that any effect measure operating in this context must satisfy certain properties. Such a measure is proposed. It forms the basis for an explanation facility which allows the user of the Generalized Bayesian Inferencing System to question the meaning of the knowledge base. That facility is described in detail.

MINDS is a distributed system of cooperating query engines that customize, document retrieval for each user in a dynamic environment. It improves its performance and adapts to changing patterns of document distribution by observing system-user interactions and modifying the appropriate certainty factors, which act as search control parameters. It argued here that the uncertainty management calculus must account for temporal precedence, reliability of evidence, degree of support for a proposition, and saturation effects. The calculus presented here possesses these features. Some results obtained with this scheme are discussed.

For any system with limited statistical knowledge, the combination of evidence and the interpretation of sampling information requires the determination of the right reference class (or of an adequate one). This paper contributes the first frank discussion of how much of Kyburg's system is needed AI discussions on probability have perenially revolved around two problems: what to do with conflicting evidence, and how to get by without a lot of objective statistical knowledge. Hans Reichenbach left modern philosophers of probability with a single task: in order to determine an event's probability, determine the narrowest reference class to which the event belongs, and about which adequate statistics are known [Rei49]. Suppose I know about the next Mets game, "m", that it is one in which Dwight Gooden will pitch, "Dm", and one to be played at home, "Hm", and one in which Keith Hernandez will bat, "Km"; I want to know the probability that the game will be a Mets' victory, P("Vm"). I have statistics about (or have an expert's degree of belief in) the per cent of Mets home games that are Mets' victories, Some are willing to supply the missing numbers, e.g., But A.I. has left the age when inventing such numbers was condoned.

This paper presents an efficient adaptation and application of the Dempster-Shafer theory of evidence, one that can be used effectively in a massively parallel hierarchical system for visual pattern perception. It describes the techniques used, and shows in an extended example how they serve to improve the system's performance as it applies a multiple-level set of processes.