PLY: A System of Plausibility Inference with a Probabilistic Basis

Abstract

An overview is given of a system of plausibility inference that will be developed for use in planning. This system, to be called PLY, will be specifically designed to work with propositions of the form "when A is true (occurs), B is likely to be true (to occur)." Previous systems performing similiar functions have been designed as aids for such tasks as medical diagnosis (MYCIN and others) and mineral prospecting (PROSPECTOR). PLY will have a probabilistics basis. Intuitive assumptions to deal with knowledge not explicitly given to the system will be made with the aid of an information-theoretic measure on the amount of information in a probability distribution. Unlike many other systems, PLY will not use these assumptions when the given knowledge indicates they are not tenable. In addition to standard probabilities, PLY will be able to make use of knowledge (information) in the form of correlations and increased/decreased likelihoods, which most people find easier to estimate than probabilities. PLY's knowledge will be in an organized and structures form, which will help in knowledge acquisition and revisino, faciliate system explanations, and lower the storage requirements of the system.