ARTFEED — Contemporary Art Intelligence

New Evidential Information Fusion Framework Based on Possibilistic Structure

publication · 2026-05-20

A research paper proposes a novel evidential information fusion framework that overcomes limitations of Dempster's rule. The framework uses a reversible transformation derived from the isopignistic principle, mapping belief functions to a possibilistic structure on the power set. A belief evolution network explicitly characterizes relationships among subsets, enabling more flexible representation. The triangular norm family is introduced to develop a general and adaptive fusion method, supporting more flexible fusion scenarios than traditional Dempster semantics. The paper is published on arXiv with ID 2605.17038.

Key facts

  • Dempster's rule is fundamental for combining belief functions from distinct and reliable sources.
  • Dempster's rule has strong structural restrictions due to its intersection-based semantics.
  • The proposed transformation is reversible and derived from the isopignistic principle.
  • The transformation maps between belief functions and a possibilistic structure on the power set.
  • A belief evolution network characterizes relationships among subsets.
  • The triangular norm family is used to develop a general fusion framework.
  • The framework supports more flexible fusion scenarios than Dempster semantics.
  • The paper is available on arXiv with ID 2605.17038.

Entities

Institutions

  • arXiv

Sources