Inconsistency Tolerant Reasoning Over Uncertain Data
This paper describes a scalable unsatisfiability tolerant probabilistic reasoning approach which combines in a flexible and orthogonal way Description Logics and Bayesian Network formalisms. The key contributions in this paper are three fold. First, the authors' approach provides an intuitive query answering semantics over a probabilistic knowledgebase. Second, their approach tolerates inconsistencies in knowledge base by computing a measure of unsatisfiability and by reasoning over satisfiable subspaces of the knowledgebase. Third, they propose an error-bounded approximation algorithm for scalable probabilistic reasoning over a large knowledgebase.