.LOG.PROB

Research Line 2: Probabilistic Description Logics

• Coordinator: Fabio Cozman
• Group members: Renata Wassermann, Marcelo Finger, Cassio Polpo de Campos, Jose Eduardo Ochoa Luna

Project Summary

Description logics have received intense scrutiny in recent years, as they offer an excellent balance between expressivity and computational complexity, and are suited for applications in knowledge representation. In particular, description logics have been standardized for applications in the Semantic Web. Usually a description logic combines concepts, roles and individuals, using Boolean operators and restricted forms of quantification. The goal of this project is to explore description logics that allow probabilistic assessments to be expressed, such as P(C | D) = a, where C and D are concepts and a is either a probability value or a probability interval. Several kinds of semantics for such sentences are under investigation, and combinations with graph-theoretical models such as Bayesian networks and Markov random fields are under study.