Quest for rigorous intelligent tutoring systems under uncertainty: Computing with Words and Images

Probabilistic and Fuzzy Logic approaches have been used for developing Intelligent Tutoring Systems (ITS) for years to deal with uncertainties in ITS but without much attention to justification of the particular techniques, which we call the analysis of scientific rigor of the approach. In probabilistic approaches, there is a missing justification of the Markovian assumption of Bayesian networks along with others. In fuzzy logic approaches, there is a missing justification of fuzzy logic operations by the analysis of the specific ITS task. One of the fundamental and natural ways to provide a rigorously justified way to deal with the uncertainty in ITS is the systematic modeling the context of each ITS task. This paper proposes a methodology based on the comprehensive use of the uncertain contextual verbal and visual information.