An NLP-Empowered Virtual Course Assistant for Online Teaching and Learning

Abstract

In this paper, we present a Natural Language Processing (NLP)-empowered virtual course assistant solution that supports online teaching and learning in the context of the COVID-19 pandemic. We leverage advanced technologies of pre-trained language models in NLP to construct several fundamental functionalities for the virtual course assistant. The assistant is designed to answer general course enquiries to reduce time-consuming and repeated human responses, to answer course-related knowledge questions by understanding both queries and teaching materials, and to analyze students' feedback via sentiment analysis. Additionally, we have constructed the course-related database and cross-platform virtual assistants for both website and mobile applications. Different pre-trained models are utilized to fine-tune the dataset in each type of model. By comparing different datasets and analyzing their performance, the best performance model is selected for the virtual assistant. Empirically, adopting NLP-empowered virtual course assistants in class improves teaching and learning experiences: With the help of an NLP-empowered virtual course assistant, the teaching team could devote more effort and time to answering complex questions; For students, an immediate response increases their motivation to study. Thus, the online system could give an excellent user experience to a wide variety of users. Our code and dataset are released at https://github.com/Heriannan/NLP-for-educationVirtualAssistant.