We present a mechanism for constructing graphical models, specifically Bayesian networks, from a knowledge base of general probabilistic information. The unique feature of our approach is that it uses a powerful first-order probabilistic logic for expressing the general knowledge base. This logic allows for the representation of a wide range of logical and probabilistic information. The model construction procedure we propose uses notions from direct inference to identify pieces of local statistical information from the knowledge base that are most appropriate to the particular event we want to reason about. These pieces are composed to generate a joint probability distribution specified as a Bayesian network. Although there are fundamental difficulties in dealing with fully general knowledge, our procedure is practical for quite rich knowledge bases and it supports the construction of a far wider range of networks than allowed for by current template technology.

Relevance-based explanation is a scheme in which partial assignments to Bayesian belief network variables are explanations (abductive conclusions). We allow variables to remain unassigned in explanations as long as they are irrelevant to the explanation, where irrelevance is defined in terms of statistical independence. When multiple-valued variables exist in the system, especially when subsets of values correspond to natural types of events, the overspecification problem, alleviated by independence-based explanation, resurfaces. As a solution to that, as well as for addressing the question of explanation specificity, it is desirable to collapse such a subset of values into a single value on the fly. The equivalent method, which is adopted here, is to generalize the notion of assignments to allow disjunctive assignments. We proceed to define generalized independence based explanations as maximum posterior probability independence based generalized assignments (GIB-MAPs). GIB assignments are shown to have certain properties that ease the deJ ign of algorithms for computing GIB-MAPs. One such algorithm is discussed here, as well as suggestions for how other algorithms may be adapted to compute GIB-MAPs. GIB-MAP explanations still suffer from instability, a problem which may be addressed using "approximate" conditional independence as a condition for irrelevance.

When eliciting probability models from experts, knowledge engineers may compare the results of the model with expert judgment on test scenarios, then adjust model parameters to bring the behavior of the model more in line with the expert's intuition. This paper presents a methodology for analytic computation of sensitivity values to measure the impact of small changes in a network parameter on a target probability value or distribution. These values can be used to guide knowledge elicitation. They can also be used in a gradient descent algorithm to estimate parameter values that maximize a measure of goodness-of-fit to both local and holistic probability assessments.

We take a utility-based approach to categorization. We construct generalizations about events and actions by considering losses associated with failing to distinguish among detailed distinctions in a decision model. The utility-based methods transform detailed states of the world into more abstract categories comprised of disjunctions of the states. We show how we can cluster distinctions into groups of distinctions at progressively higher levels of abstraction, and describe rules for decision making with the abstractions. The techniques introduce a utility-based perspective on the nature of concepts, and provide a means of simplifying decision models used in automated reasoning systems. We demonstrate the techniques by describing the capabilities and output of TUBA, a program for utility-based abstraction.

The inherent intractability of probabilistic inference has hindered the application of belief networks to large domains. Noisy ORgates [30] and probabilistic similarity networks [18, 17) escape the complexity of inference by restricting model expressiveness. Recent work in the application of belief-network models to time-series analysis and forecasting [9, 10) has given rise to the additive beliefnetwork model (ABNM). We (1) discuss the nature and implications of the approximations made by an additive decomposition of a belief network, (2) show greater efficiency in the induction of additive models when available data are scarce, (3) generalize probabilistic inference algorithms to exploit the additive decomposition of ABNMs, (4) show greater efficiency of inference, and (5) compare results on inference with a simple additive belief network.

Dynamic network models (DNMs) are belief networks for temporal reasoning. The DNM methodology combines techniques from time series analysis and probabilistic reasoning to provide (1) a knowledge representation that integrates noncontemporaneous and contemporaneous dependencies and (2) methods for iteratively refining these dependencies in response to the effects of exogenous influences. We use belief-network inference algorithms to perform forecasting, control, and discrete event simulation on DNMs. The belief network formulation allows us to move beyond the traditional assumptions of linearity in the relationships among time-dependent variables and of normality in their probability distributions. We demonstrate the DNM methodology on an important forecasting problem in medicine. We conclude with a discussion of how the methodology addresses several limitations found in traditional time series analyses.

Influence diagrams are ideal knowledge representations for Bayesian statistical models. However, these diagrams are difficult for end users to interpret and to manipulate. We present a user-based architecture that enables end users to create and to manipulate the knowledge representation. We use the problem of physicians' interpretation of two-arm parallel randomized clinical trials (TAPRCT) to illustrate the architecture and its use. There are three primary data structures. Elements of statistical models are encoded as subgraphs of a restricted class of influence diagram. The interpretations of those elements are mapped into users' language in a domain-specific, user-based semantic interface, called a patient-flow diagram, in the TAPRCT problem. Pennitted transformations of the statistical model that maintain the semantic relationships of the model are encoded in a metadata-state diagram, called the cohort-state diagram, in the TAPRCT problem. The algorithm that runs the system uses modular actions called construction steps. This framework has been implemented in a system called THOMAS, that allows physicians to interpret the data reported from a TAPRCT.

This paper addresses the tradeoff's which need to be considered in reasoning using probabilistic network representations, such as Influence Diagrams (IDs). In particular, we examine the tradeoff's entailed in using Temporal Influence Diagrams (TIDs) which adequately capture the temporal evolution of a dynamic system without prohibitive data and computational requirements. Three approaches for TID construction which make different tradeoff's are examined: (1) tailoring the network at each time interval to the data available (rather then just copying the original Bayes Network for all time intervals); (2) modeling the evolution of a parsimonious subset of variables (rather than all variables); and (3) model selection approaches, which seek to minimize some measure of the predictive accuracy of the model without introducing too many parameters, which might cause "overfitting" of the model. Methods of evaluating the accuracy /efficiency of the tradeoff's are proposed.

The problems associated with scaling involve active and challenging research topics in the area of artificial intelligence. The purpose is to solve real world problems by means of AI technologies, in cases where the complexity of representation of the real world problem is potentially combinatorial. In this paper, we present a novel approach to cope with the scaling issues in Bayesian belief networks for ship classification. The proposed approach divides the conceptual model of a complex ship classification problem into a set of small modules that work together to solve the classification problem while preserving the functionality of the original model. The possible ways of explaining sensor returns (e.g., the evidence) for some features, such as portholes along the length of a ship, are sometimes combinatorial. Thus, using an exhaustive approach, which entails the enumeration of all possible explanations, is impractical for larger problems. We present a network structure (referred to as Sequential Decomposition, SD) in which each observation is associated with a set of legitimate outcomes which are consistent with the explanation of each observed piece of evidence. The results show that the SD approach allows one to represent feature-observation relations in a manageable way and achieve the same explanatory power as an exhaustive approach.

We have developed a general Bayesian algorithm for determining the coordinates of points in a three-dimensional space. The algorithm takes as input a set of probabilistic constraints on the coordinates of the points, and an a priori distribution for each point location. The output is a maximum-likelihood estimate of the location of each point. We use the extended, iterated Kalman filter, and add a search heuristic for optimizing its solution under nonlinear conditions. This heuristic is based on the same principle as the simulated annealing heuristic for other optimization problems. Our method uses any probabilistic constraints that can be expressed as a function of the point coordinates (for example, distance, angles, dihedral angles, and planarity). It assumes that all constraints have Gaussian noise. In this paper, we describe the algorithm and show its performance on a set of synthetic data to illustrate its convergence properties, and its applicability to domains such ng molecular structure determination.