Uncertainty Handling in Expert Systems: Uniform vs. Task-Specific Formalisms
B Chandrasekaran, Martin Tanner
The debate on uncertainty has conflated a number of significant issues. These have to do with classical probability theory, normative decision theories, theories of human uncertainty handling, and the design of expert systems. We doubt that "uncertainty handling" is a fundamental "natural kind" of intelligent activity, but the belief that it is has led to attempts to devise normative formalisms which are useful for all problems. In our view uncertainty handling is particularized for the type of problem solving. This means that we need formalisms which work well for each generic problem solving task and a framework for judging the usefulness of formalisms for particular problem solving tasks. We propose such a framework and give a formalism for uncertainty handling in classification problem solving.
Keywords: Uncertainty Handling, Expert Systems
PDF Link: /papers/85/p35-chandrasekaran.pdf
AUTHOR = "B Chandrasekaran
and Martin Tanner",
TITLE = "Uncertainty Handling in Expert Systems: Uniform vs. Task-Specific Formalisms",
BOOKTITLE = "Uncertainty in Artificial Intelligence Annual Conference on Uncertainty in Artificial Intelligence (UAI-85)",
PUBLISHER = "Elsevier Science",
ADDRESS = "Amsterdam, NL",
YEAR = "1985",
PAGES = "35--46"