Probabilistic Inference and Influence Diagrams

Abstract

An influence diagram is a network representation for probabilistic and decision analysis models. The nodes correspond to variables which can be constants, uncertain quantities, decisions, or objectives. The arcs reveal the probabilistic dependence of the uncertain quantities and the information available at the time of the decisions. The detailed data about the variables are stored within the nodes, so the diagram graph is compact and focuses attention on the relationships among the variables. Influence diagrams are effective communication tools and recent developments also allow them to be used for analysis. We develop algorithms to address questions of inference within a probabilistic model represented as an influence diagram. We use the conditional independence implied by the diagram's structure to determine the information needed to solve a given problem. When there is enough information we can solve it, exploiting that conditional independence. These same results are applied to problems of decision analysis. This methodology allows the construction of computer tools to maintain and evaluate complex models.

Content-Oriented Relevance Feedback in XML-IR Using the Garnata Information Retrieval System

“…This system is based on the Context-based Influence Diagram model (CID model) [13], which is supported by Influence Diagrams and Bayesian networks [14,15]. These are probabilistic graphical models specially designed for decision problems in uncertain environments.…”

“…This is the case of our search engine, Garnata [4]. This piece of software is based on Probabilistic Graphical Models, more precisely on Influence Diagram and the corresponding underlying Bayesian networks [14,15]. This paper is devoted to the development of an RF module for Garnata, and is focused on the simplest case of XML RF, namely CO-CO.…”

Graphical Models for Probabilistic and Causal Reasoning

“…Among the most popular are Shachter's [1988] method of node elimination, Lauritzen and Spiegelhalter's [1988] method of clique-tree propagation, and the method of loop-cut conditioning [Pearl 1988, Ch. 4.3].…”

The Garnata Information Retrieval System at INEX’07

“…Like in the past year, we have participated in the Ad hoc Track with an experimental platform to perform structured retrieval using Probabilistic Graphical Models [5,8,10], called Garnata [4].…”