Computational Creativity for Intelligence Analysis

We describe a decision support system for hypothesis assessment in which exploration is supported by computational creativity. A software tool for morphological hypothesis analysis and evidence handling is extended with a creative assistant that on demand suggests hypotheses that the analyst should consider. Suggested hypotheses are chosen so that they are far from hypotheses that the analyst previously has paid attention to but nevertheless are supported by evidence in an interesting way. For the purpose of providing thoughtprovoking suggestions, the creative assistant employs ensembles of novelty and value assessment methods and proposes hypotheses that stand out in this multiensemble analysis. Preliminary experiments investigate the system’s potential for infusing novel and valid ideas into the decision making process.