Semantic Approach to User Answer Verification in Intelligent Tutoring Systems

Abstract

AbstractA semantic approach to develop a problem solver for automatic verification of user answers in intelligent tutoring systems built on OSTIS Technology is proposed in this article. In the framework of OSTIS Technology, the problem solvers are developed based on multi-agent technology. The developed problem solver allows to automatically verify the completeness and correctness of user answers based on semantics by calling the corresponding sc-agent according to the type of test questions (judgement questions, multiple-choice questions, questions of definition interpretation, etc.). For some questions of definition interpretation with multiple standard answers, the problem solver allows to pre-screen a semantic fragment of the standard answer that best matches the user answer according to the semantic fragment of the user answer, and then verify the completeness and correctness between them. At the same time, since the standard answers and user answers to some questions do not match on the surface, but the logical formulas used to formalize them satisfy logical equivalence, the developed problem solver can also determine whether the standard answer and user answer satisfy logical equivalence.KeywordsOntologyLogical formulasOSTIS TechnologyProblem solverKnowledge baseIntelligent tutoring systemsAnswer verification