Development of a problem solver for automatic answer verification in the intelligent tutoring systems

Abstract

This article proposes an approach to developing a problem solver for automatic answer verification in intelligent tutoring systems constructed using OSTIS Technology. The problem solver is developed based on multiagent technology. The developed problem solver automati-cally verifies the correctness and completeness of the user answer at the semantic level by calling the corresponding scagent sets according to the type of the questions (multiple-choice questions, fill in the blank questions, questions of definition interpretation, etc.). Sometimes there may be multiple standard answers for the question of definition interpretation, but the problem solver can automatically filter a semantic fragment of the standard answer in advance that best matches the user answer according to the semantic fragment of the user answer, and then continue to verify the correctness and completeness between them.