Mechanism-based causal models for adaptive decision support

Authors:

Tsai-Ching Lu
HRL Laboratories, LLC
Malibu, CA 90265
e-mail: tlu@hrl.com

Marek J. Druzdzel
Decision Systems Laboratory
School of Information Sciences
and Intelligent Systems Program
University of Pittsburgh
135 North Bellefield Avenue
Pittsburgh, PA 15260, U.S.A.
e-mail: marek@sis.pitt.edu

Abstract:

We propose a framework for decision supports in a changing world based on mechanism-based causal models. A causal model is a set of self-contained structural equations representing a decision problem. In our framework, a causal model is instantiated from a mechanism knowledge base, where domain knowledge is stored as mechanisms along with generic non-intervening and intervening actions. A change in the world is modeled as applying an action on a causal model. The framework supports reasoning about the effect of actions to deduce the causal model representing the world brought about by the actions.