The success of a course depends on various factors, of which timely response to students' queries in learning management systems (LMS) is significant. A timely response enhances student engagement, satisfaction, motivation and performance, while reducing delays and resultant stress due to waiting for the question to be answered. In the past, with courses having limited number of enrollments, it was possible for course convenors or lecturer-in-charge (LIC) to answer students' queries in a timely manner and stay on the top of their course management. However, with the increasing number of students in courses every year, it has been challenging to respond to students’ queries in a timely manner. Based on our experience of teaching COMP9444 course for the last 3 terms, we found that, within a day there are more than 50 questions that need to be answered by LIC or the teaching staff (based on enrollment of about 600 students enrolled in a term). Based on our experience of answering these questions, we found that many of these questions are repetitive and often have been answered previously.

The aim of this project is the preliminary investigation and development of an automated system that can check whether a question has been answered previously. If yes, the system will provide reference to the answer for the similar question asked. Also, there will be feedback in the form of human-in-the-loop allowing students to rate how accurately the previously answered question has been useful in answering their actual question. If successful, this could be expanded for other courses in the school and even the broader education industry.

School

Computer Science and Engineering

Research Area

Natural language processing | Machine learning | Deep learning

The project involves curation of dataset from an LMS such as Moodle or Ed and applying Natural Language Processing (NLP) techniques for classification and similarity scoring. The student will work closely with a team of experts in the field of NLP and will be given access to computational resources for carrying out the project. There is a possibility to extend this project into an honors thesis.

At the completion of the project, the student will have a good understanding of NLP techniques. In addition, they will have significantly strengthened their technical and research skills. It is anticipated that, by the end of the project, the student will have developed a model that can predict whether the question has been answered previously or not and provide a reference to the similar question/answer pair. The student will also write a report outlining motivation, literature review, data analysis, model, and results. The team aims to produce a research paper that can be submitted to a relevant conference for publication.