Developing a large belief network, like any large system, requires systems engineering to manage the design and construction process. We propose that network engineering follow a rapid prototyping approach to network construction. We describe criteria for identifying network modules and the use of `stubs' within a belief network. We propose an object oriented representation for belief networks which captures the semantic as well as representational knowledge embedded in the variables, their values and their parameters. Methods for evaluating complex networks are described. Throughout the discussion, tools which support the engineering of large belief networks are identified. 1. Introduction As belief networks become more popular and well understood as a tool for modeling uncertainty and as the computational power of belief network inference engines increases, belief networks are being applied to problems of increasing size and complexity. In the early 1990's, Pathfinder, at 109 nodes...

Intelligent systems use internal representations to mediate the transformation from percepts to goal-directed actions. Intelligent learning agents use environmental feedback to modify their internal representations to improve performance over time and adapt to changing circumstances. All learning involves knowledge-data fusion to some degree. Bayesian learning, the focus of this paper, is specifically designed to incorporate both expert knowledge and observations. We use the term "data" to refer both to collections of cases and to statements about the domain provided by experts and knowledge engineers and used to construct internal representations. The term "knowledge" refers to the internal representation itself, which we take to be a collection of Bayesian network fragments. We describe a prequential learning agent architecture for bounded rational action and learning under uncertainty. We describe recent extensions to Bayesian networks that provide sufficient representation power for expressing general prequential learning agent models. We describe tools and techniques to support a process in which models are constructed and refined using a combination of inputs from experts and environmental feedback. KEY WORDS: Bayesian Networks, Bayesian Learning, Graphical Probabilistic Models, Knowledge Elicitation, Object Oriented Bayesian Networks, Prequential Probability Machine Learning MCMC Issue 1 7/1/01 1.

The actions taken by an automated decision-making agent can be enhanced by including mechanisms that enable the agent to categorize concepts e#ectively. We pose a utility-based approach to categorization based on the idea that categorization should be carried out in the service of action. The choice of concepts made by a decision maker is critical in the e#ective selection of actions under resource constraints. This perspective is in contrast to classical and similarity-based approaches which seek completeness in concept description with respect to shared properties rather than the e#ectiveness of decision making. We propose a decision-theoretic framework for utility-based categorization whichinvolves reasoning about alternative categorization models consisting of sets of interrelated concepts at varying levels of abstraction. Categorization models that are too abstract mayoverlook details that are critical for selecting the most appropriate actions. Categorization models t...