Since exact probabilistic inference is intractable in general for large multiply connected belief nets, approximate methods are required. A promising approach is to use heuristic search among hypotheses (instantiations of the network) to find the most probable ones, as in the TopN algorithm. Search is based on the relative probabilities of hypotheses which are efficient to compute. Given upper and lower bounds on the relative probability of partial hypotheses, it is possible to obtain bounds on the absolute probabilities of hypotheses. Best-first search aimed at reducing the maximum error progressively narrows the bounds as more hypotheses are examined. Here, qualitative probabilistic analysis is employed to obtain bounds on the relative probability of partial hypotheses for the BN20 class of networks networks and a generalization replacing the noisy OR assumption by negative synergy. The approach is illustrated by application to a very large belief network, QMR-BN, which is a reformulation of the Internist-1 system for diagnosis in internal medicine.

This paper presents a plausible reasoning system to illustrate some broad issues in knowledge representation: dualities between different reasoning forms, the difficulty of unifying complementary reasoning styles, and the approximate nature of plausible reasoning. These issues have a common underlying theme: there should be an underlying belief calculus of which the many different reasoning forms are special cases, sometimes approximate. The system presented allows reasoning about defaults, likelihood, necessity and possibility in a manner similar to the earlier work of Adams. The system is based on the belief calculus of subjective Bayesian probability which itself is based on a few simple assumptions about how belief should be manipulated. Approximations, semantics, consistency and consequence results are presented for the system. While this puts these often discussed plausible reasoning forms on a probabilistic footing, useful application to practical problems remains an issue.

The ideal Bayesian agent reasons from a global probability model, but real agents are restricted to simplified models which they know to be adequate only in restricted circumstances. Very little formal theory has been developed to help fallibly rational agents manage the process of constructing and revising small world models. The goal of this paper is to present a theoretical framework for analyzing model management approaches. For a probability forecasting problem, a search process over small world models is analyzed as an approximation to a larger-world model which the agent cannot explicitly enumerate or compute. Conditions are given under which the sequence of small-world models converges to the larger-world probabilities.

Experts do not always feel very, comfortable when they have to give precise numerical estimations of certainty degrees. In this paper we present a qualitative approach which allows for attaching partially ordered symbolic grades to logical formulas. Uncertain information is expressed by means of parameterized modal operators. We propose a semantics for this multimodal logic and give a sound and complete axiomatization. We study the links with related approaches and suggest how this framework might be used to manage both uncertain and incomplere knowledge.

Possibilistic logic has been proposed as a numerical formalism for reasoning with uncertainty. There has been interest in developing qualitative accounts of possibility, as well as an explanation of the relationship between possibility and modal logics. We present two modal logics that can be used to represent and reason with qualitative statements of possibility and necessity. Within this modal framework, we are able to identify interesting relationships between possibilistic logic, beliefs and conditionals. In particular, the most natural conditional definable via possibilistic means for default reasoning is identical to Pearl's conditional for e-semantics.

This paper provides an overview of the SP theory of intelligence and its central idea that artificial intelligence, mainstream computing, and much of human perception and cognition, may be understood as information compression. The background and origins of the SP theory are described, and the main elements of the theory, including the key concept of multiple alignment, borrowed from bioinformatics but with important differences. Associated with the SP theory is the idea that redundancy in information may be understood as repetition of patterns, that compression of information may be achieved via the matching and unification (merging) of patterns, and that computing and information compression are both fundamentally probabilistic. It appears that the SP system is Turing-equivalent in the sense that anything that may be computed with a Turing machine may, in principle, also be computed with an SP machine. One of the main strengths of the SP theory and the multiple alignment concept is in modelling concepts and phenomena in artificial intelligence. Within that area, the SP theory provides a simple but versatile means of representing different kinds of knowledge, it can model both the parsing and production of natural language, with potential for the understanding and translation of natural languages, it has strengths in pattern recognition, with potential in computer vision, it can model several kinds of reasoning, and it has capabilities in planning, problem solving, and unsupervised learning. The paper includes two examples showing how alternative parsings of an ambiguous sentence may be modelled as multiple alignments, and another example showing how the concept of multiple alignment may be applied in medical diagnosis.

This paper describes ongoing research into planning in an uncertain environment. In particular, it introduces U-Plan, a planning system that constructs quantitatively ranked plans given an incomplete description of the state of the world. U-Plan uses a DempsterShafer interval to characterise uncertain and incomplete information about the state of the world. The planner takes as input what is known about the world, and constructs a number of possible initial states with representations at different abstraction levels. A plan is constructed for the initial state with the greatest support, and this plan is tested to see if it will work for other possible initial states. All, part, or none of the existing plans may be used in the generation of the plans for the remaining possible worlds. Planning takes place in an abstraction hierarchy where strategic decisions are made before tactical decisions. A super-plan is then constructed, based on merging the set of plans and the appropriately timed acquisition of essential knowledge, which is used to decide between plan alternatives. U-Plan usually produces a super-plan in less time than a classical planner would take to produce a set of plans, one for each possible world.

Numerous methods for probabilistic reasoning in large, complex belief or decision networks are currently being developed. There has been little research on automating the dynamic, incremental construction of decision models. A uniform value-driven method of decision model construction is proposed for the hierarchical complete diagnosis. Hierarchical complete diagnostic reasoning is formulated as a stochastic process and modeled using influence diagrams. Given observations, this method creates decision models in order to obtain the best actions sequentially for locating and repairing a fault at minimum cost. This method construct decision models incrementally, interleaving probe actions with model construction and evaluation. The method treats meta-level and baselevel tasks uniformly. That is, the method takes a decision-theoretic look at the control of search in causal pathways and structural hierarchies.

We investigate the value of extending the completeness of a decision model along different dimensions of refinement. Specifically, we analyze the expected value of quantitative, conceptual, and structural refinement of decision models. We illustrate the key dimensions of refinement with examples. The analyses of value of model refinement can be used to focus the attention of an analyst or an automated reasoning system on extensions of a decision model associated with the greatest expected value.

Probabilistic conceptual network is a knowledge representation scheme designed for reasoning about concepts and categorical abstractions in utility-based categorization. The scheme combines the formalisms of abstraction and inheritance hierarchies from artificial intelligence, and probabilistic networks from decision analysis. It provides a common framework for representing conceptual knowledge, hierarchical knowledge, and uncertainty. It facilitates dynamic construction of categorization decision models at varying levels of abstraction. The scheme is applied to an automated machining problem for reasoning about the state of the machine at varying levels of abstraction in support of actions for maintaining competitiveness of the plant.