Extending Theories of Action: Formal Theory & Practical Applications
Papers from the AAAI Spring Symposium
Craig Boutilier & Moises Goldszmidt,Cochairs
Action and its representation continues to play a central role in many areas of AI. The research focus in these areas has often been motivated by different issues and types of problems, or based on different underlying assumptions. Representations vary widely, ranging from logical specifications to influence diagrams and belief nets to less complex, but more computationally manageable planning representations. The assumptions and reasoning methods adopted are also quite different. Techniques include state- and plan-space search, regression, logical inference, network algorithms and analytical methods.
This symposium will bring together researchers with widely different backgrounds, and provide a forum for discussion of the manner in which these different approaches to reasoning about action can be brought together. The two primary goals are: evaluating the state of the art, especially identifying the issues and assumptions central to different fields; and identifying relevant research issues for extending theories of action and putting these theories into practice.
The format will consist of technical presentations, panels and discussion sessions. Technical presentations will cover recent advances in extending theories of action, and their application to planning, diagnosis, and control. The panels and discussion sessions will evaluate these approaches, and produce recommendations for extending and unifying various approaches, as well as for closing the gap between theory and applications.