Reasoning with Mass Distributions
This work addresses uncertainty reasoning for AI and expert systems, but appears incremental as it builds on existing concepts like Dempster's rule.
The paper tackles the problem of reasoning under uncertainty by introducing a framework using movable evidence masses and specialization matrices, which allows for integrating new evidence and representing non-monotonic reasoning aspects.
The concept of movable evidence masses that flow from supersets to subsets as specified by experts represents a suitable framework for reasoning under uncertainty. The mass flow is controlled by specialization matrices. New evidence is integrated into the frame of discernment by conditioning or revision (Dempster's rule of conditioning), for which special specialization matrices exist. Even some aspects of non-monotonic reasoning can be represented by certain specialization matrices.