Automatic Verification of Transformation Chains of Symbolic Facts Based on Logical Rules Search

Abstract

One of the most important elements of the modern educational process in the field of natural sciences is the solution of problems of transforming facts of symbolic expressions. When solving such problems, learners need to deduce a new solution of the problem based on the rules they have already learned. To consolidate these skills, learner need to solve a large volume of tasks, which entails the need of verification of the solutions. Many platforms have already been implemented that allow to automate the verification of solutions. Generally, these systems only check the answer, but not the progress of the problem. This approach of verification leaves the possibility of getting the correct answer and the task counted if the decision is not fully thought out or incorrect. We propose a method for automatic verification of problem solutions based on sequential verification of completed transformations by efficiently iterating through the rules that allowed these transformations to be performed. At each step of the solution, the system iterates through the allowed transformations, trying to find one, that learner used. Problems of ambiguity of the recording form are solved by a combination of testing method, unification, and structuring of the recording rules. The proposed method of automatic verification was tested on students of junior courses of the Polytechnic and Electrical engineering Universities and showed good results: the productivity of teaching staff when checking tasks increased more than 4 times.