A Computational Model of Information Fusion for Situation Assessment

"The overall objectives of this proposal are: to develop computational techniques for the information fusion problem as it relates to Situation Monitoring and Assessment.We decompose the Situation Assessment problem into two parts. The first part is the mapping of the raw state of the world (the structural description of the world) to semantic situational features which are relevant to the target situations of interest.The second part is to derive or express the relationships among the situational features with respect to the situation of interest. The second part is the information fusion problem.This decomposition decouples the domain specific problem of situation assessment from the domain independent, and possibly interchangeable, sensor and information processing technologies which extract the semantic features from the raw data.We propose to develop and evaluate algorithms for learning situation-specific relationships among situational features. To this end, we also present a normalized evaluation measure for the reliability of the information processors which extract thesemantic features. This normalized measure facilitates the interchangeability of underlying information processing technologies. The research described in this proposal will contribute to development of a knowledge refinery platform which allows domain experts to discover and extract actionable intelligence from massive raw data repositories. Such a platform wil