Title:	Building coherent probabilty models from incoherent judgment using Genetic Algorithms and Algebraic Decision Diagrams
Speaker:	Spyros Tsavachidis
Abstract:	We need to account for uncertainty when we design expert systems Probability theory has been the most successful mathematical tool for describing uncertainty Goal: Build useful models that reflect the probabilistic structure of a problem domain of interest. Discover and efficiently represent the underlying probability distribution The constructed model must answer quickly to queries concerning probabilities of complex events (inference)