Lowrance, J. D. (1986). Automating argument construction. AI Center, SRI Int., Menlo Park, CA 94025. In Proceeding Workshop on Assessing Uncertainty.
Abstract
Over the past five years the artificial intelligence center at SRI has been developing a new technology to address the problem of automated information management within real-world contexts. The result of this work is a body of techniques for automated reasoning from evidence that we call evidential reasoning. The techniques are based upon the mathematics of belief functions developed by Dempster and Schaefer and have been successively applied to a variety of problems including computer vision, multisensor integration, and intelligence analysis.
We have developed both a formal basis and a framework for implementing automated reasoning systems based upon these techniques. Both the formal and practical approach can be divided into four parts one specifying a set of distinct prepositional spaces to specifying the interrelationships among these spaces three representing bodies of evidence as belief distributions and for establishing paths for the bodies of evidence to move through these Spaces by means of evidential operations, eventually converging on spaces where the target questions can be answered. These steps specify a means for arguing from multiple bodies of evidence towards a particular probabilistic conclusion. Argument construction is the process by which such evidential analysis are constructed and is the analog of constructing proof trees in a logical context.
This technology features the ability to reason from uncertain, incomplete, and occasional and accurate information based upon seven evidential operations: fusion, discounting, translating, projection, summarization, and interpolation, and jesting. These operation are theoretically sound but have intuitive appeal as well. In implementing this formal approach, we have found the evidential arguments can be represented as graphs. To support the construction, modification, and interrogation of evidential arguments, we have developed Gister Gister provides an interactive, menu driven, graphical interface that allows these graphical structures to be easily manipulated. Our goal is to provide efficient automated aids to domain experts for argument construction. Just a represents our first attempt at such an aid.