Development of algorithms and methods for fusion of imprecise information
In every situation humans make observations and analyse the observed information, by combining and evaluating the relationships between the observations to understand the situation. Often these observations can be uncertain, redundant, from unknown sources, contradictory and too many for a human to be able to analyse fast enough and correctly, leading to incorrect assessment of the situation. The proposed research is based on the hypothesis that through application of various probabilistic, possibilistic and other applied mathematics methods it is possible to enhance humans ability to analyse large amounts of redundant, uncertain, sometimes contradictory information, enhancing his/her capability to correctly understand situations.
The current research will investigate how the application of imperfect information representation theories (including belief functions with their link to probability) can leverage the understanding of a situation and its evolution in a context where information, coming from a broad variety of sources (internet, police, military, coast guard), is uncertain.
Voir la description complète du projet