GPT-Empowered Personalized eLearning System for Programming Languages

Abstract

The eLearning approach to programming language instruction has gained widespread acceptance due to advantages such as accessibility, temporal flexibility, and content reusability. However, the current eLearning for programming predominantly employs the delivery of one-size-fits-all content, engendering elevated costs in both the development of language coursework and administration of eLearning sessions, which includes the labor-intensive task of grading student submissions. A compelling research question to consider is how to construct an eLearning system capable of delivering personalized, student-centric content, automating the generation of coursework elements, and eliminating the need for instructor involvement in the management of eLearning sessions. Our approach to delivering a definite solution to the question involves the utilization of a suite of advanced software technologies: GPT to dynamically generate course contents/components, prompt engineering to personalize course content for each individual student, and autonomous computing to manage eLearning sessions without the need for human intervention. The research results encompass the design of an eLearning framework covering all programming languages, a fully functional Python-based implementation, seamless integration with ChatGPT for dynamic content generation, a high degree of content personalization, and the elimination of manual effort required for managing eLearning sessions.