Use of Machine Learning in Interactive Cybersecurity and Network Education

Abstract

Cybersecurity is a complex subject for students to pursue. Hands-on online learning through labs and simulations can help students become more familiar with the subject at security classes to pursue cybersecurity education. There are several online tools and simulation platforms for cybersecurity education. However, those platforms need more constructive feedback mechanisms, and customizable hands-on exercises for users, or they oversimplify or misrepresent the content. In this paper, we aim to develop a platform for cybersecurity education that can be used either with a user interface or command line and provide auto constructive feedback for command line practices. Moreover, the platform currently has nine levels to practice for different subjects of networking and cybersecurity and a customizable level to create a customized network structure to test. The difficulty of objectives increases at each level. Moreover, an automatic feedback mechanism is developed by using a machine learning model to warn users about their typographical errors while using the command line to practice. A trial was performed with students completing a survey before and after using the application to test the effects of auto-feedback on users' understanding of the subjects and engagement with the application. The machine learning-based version of the application has a net increase in the user ratings of almost every survey field, such as user-friendliness and overall experience.