Embedding Uncertainty in Conceptual Graphs for Semantic Information Fusion

The problem of soft information fusion is one of the key challenges in both fields of information fusion and situation awareness. However uncertainty modelling in the context of semantic information fusion is not well-researched.In this paper we present the current state-of-the-art regarding fusion of uncertain semantic information using simple conceptual graphs and propose methods to improve its handling of partial conflict, reliability and expressiveness regarding modal uncertainty. Particular attention is paid to the issues of scalability and tractability. Similarities to other uncertainty handling frameworks such as possibility theory and belief functions are discussed. Potential avenues are identified for further research regarding uncertainty representation in semantic information fusion.