Text Comprehension for E-Learning Applications

Abstract

Automatic Short Answers Grading system for MOOCs, focusing on the English language. This study leverages deep learning techniques to create a grading system that not only evaluates students’ answers based on correctness but also comprehends the depth of their understanding and the contextual relevance of their responses. The methodology employed involves training various neural network models to analyze text responses, comparing their performance, and integrating the most effective model into the MOOC platform. The proposed system aims to enhance the accuracy of automated grading and reduce the workload of human graders. The effectiveness of the ASAG system is validated through a series of experiments, demonstrating its potential to support educators and improve the educational experience in MOOCs.