Generating QA from Rule-based Algorithms

In the education industry answering questions is used as a common parameter to judge one’s understanding of a topic. Taking quizzes on a regular basis helps an individual feel confident and it also helps the professor assess the student’s understanding on a particular topic. Generating question and answer pairs is a time-consuming task. To solve this problem, this paper discusses methods to generate automatic Natural Language Processing models which creates diverse types of question-answer pairs. The model takes an input in the form of text in the English language and produces output as Complex Questions, Multiple Choice Questions with relevant distractors, and Fill in the Blanks type of questions. To generate Complex Questions a Rule-Based Algorithm is used. To generate Multiple Choice Questions and Fill in the Blanks type questions, a Vector Algorithm from the GLoVe Model is used along with Rule-Based Algorithms. This paper also includes a detailed explanation of the analysis of the pattern and rules that are observed in the question-making process. SQuAD dataset is used for this analysis and used the same dataset to train the model. The implementation process of this model focused on generating diverse questions with higher syntactic correctness than the existing models. The approach mentioned in this paper can be used in the fields of education, entertainment, generation of quizzes, virtual learning assistance and to get a deeper insight into any topic.