SRIs work in automated uncertain reasoning emphasizes the practical application of theoretically sound techniques for reasoning from evidence-that is, information that is potentially incomplete, inexact, inaccurate, and from diverse sources. SRI pioneered evidential reasoning for drawing conclusions from multiple sources of evidential information about dynamic real-world situations. We have developed formal foundations for
reasoning under uncertainty
covering both probabilistic models (i.e., Bayesian and Dempster-Shafer) and possibilistic models (i.e., propositional logic and fuzzy logic) and have
incorporated all of these techniques into a single uncertain reasoning tool, Gister.
Gister supports the rapid development of evidential reasoning systems through an interactive, menu-driven, graphical interface, based upon Grasper-CL.
The user interacts with the system in much the same way as with electronic spreadsheets, by simply selecting from menus to add evidential
operations to an
analysis, to modify data or operation parameters, or to change any portion of the uncertain knowledge base. In response, gister updates its analyses to reflect
the new information.
Gister supports a wide range of evidential operations, including fusion, source discounting, time projection, summarization, evidence interpretation, and
sensitivity analysis. Gister has been applied to a wide range of problems, including multisensor interpretation, mission planning, medical diagnosis,
intelligence analysis, underwater vehicle tracking, antiair threat identification, robot vehicle navigation, and management decision support.