An overview of existing methods for automatic generation of test tasks in natural language

Recently, in the field of education, much attention has been paid to the use of multiple choice questions as a tool for assessing knowledge. The development of test tasks requires a lot of time and is highly labor intensive. It is difficult to perform such a task manually, so many researchers offer various ways and approaches to automate the creation of test tasks in natural language. In this paper, we present an overview of scientific achievements in the field of automatic question generation, which examines the classification of question generation systems by dividing them into five groups: machine learning-based methods, neural network-based, tree-based, rule-based or template-based and hybrid methods.