SMILE: Introduction

From DSL

This document describes the application programmer's interface to the Structural Modeling, Inference, and Learning Engine (SMILE) developed by Decision Systems Laboratory (DSL). SMILE is a platform independent library of C++ classes for reasoning in graphical probabilistic models, such as Bayesian networks and influence diagrams. The SMILE library can be embedded in programs that use graphical probabilistic models as their reasoning engines.

SMILE is written in C++ in a platform-independent fashion and is fully portable. The application programmer's interface, described in this document, is defined in terms of a collection of C++ classes that form the "body" of the library and can be used from within an application program. These classes allow building graphical models, editing, saving and loading them, and using them for probabilistic reasoning and decision making under uncertainty.

Model building and the reasoning process are under full control of the application program as the SMILE library serves merely as a set of tools and structures that facilitates them. DSL has also developed a graphical interface to SMILE, called GeNIe (Graphical Network Interface), which runs under all 32-bit Windows platforms. The latest version of GeNIe can be downloaded from the GeNIe Web site.

This manual assumes a basic knowledge of graphical probabilistic models, such as Bayesian networks and influence diagrams. Elements of probabilistic inference in graphical models will be reviewed briefly but this should not be viewed as a substitute for formal preparation.

Some sources of elementary knowledge of graphical probabilistic models are:

• Cowell, Robert G., A. Philip Dawid, Steffen L. Lauritzen and David J. Spiegelhalter (1999). Probabilistic Networks and Expert Systems. Springer-Verlag New York, Inc.: New York, NY.

• Richard E. Neapolitan, Probabilistic Reasoning in Expert Systems: Theory and Algorithms. John Wiley & Sons, New York, 1990

• Judea Pearl, Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference. Morgan Kaufmann Publishers, Inc., San Mateo, CA, 1988

• Eugene Charniak, Bayesian Networks without Tears. AI Magazine, 12(4)50-63, Winter 1991

• Jensen, Finn V. (1996). An Introduction to Bayesian Networks. Springer Verlag, New York.

• Henrion, M., John S. Breese, and Eric J. Horvitz (1991). Decision analysis and expert systems. AI Magazine, 12(4):64-91.

Some of the best sources of up to date information on the state of the art research in graphical probabilistic models are proceedings of the annual Conference on Uncertainty in Artificial Intelligence. Proceedings of all UAI conferences are available electronically from the DSL-hosted UAI Electronic Proceedings Web site.