Abstract:

This paper has presented an AI-assisted quiz system, iQS, with the capability of creating individual quizzes and providing instant personal feedback, to serve self-regulated learning and distant learning. iQS is designed to be easily adopted by different learning management systems (LMS, such as Moodle or MOOCs ). Its core algorithms can be reused directly, only specific domain knowledge and learners’ learning data need to be provided in order to run in a new specific domain. As a research instance, iQS has been developed and tested inside the Moodle learning platform of a university and positive feedback has been received from students and teachers.

For the current digitalized online learning, personalized learning and self-regulated learning (SRL) are both very active research topics. Feedback has a powerful influence on SRL processes and affects students’ cognitive engagement with tasks. Our research overall is committed to studying individual learning differences in different learning stages and how to provide the most adaptive and accurate feedback in real time in order to improve students’ learning and SRL skills.

Our focus is on how to apply the latest AI technologies in higher education via practical applications, in this paper, our scenario is, an intelligent AI-based quiz system is required by university students, who are basically enrolled in distance learning courses using blended learning. Over half of them study part-time, have jobs or families, learn for a better career or certain interests, and are mainly learning by themselves. In the initial phase of our research, we found out that the learning management system (LMS) currently in use only records learning progress digitally in a coarse-grained way; allows digitized learning content merely on a small scale; and has barely any personalized setting for students; and has few feedback directly from tutors. Therefore, the data sets to be collected for the purpose of data analysis are generally of poor quality, which is incomplete and fragmented or even missing. In addition, there are only a small number of quizzes that are manually created by tutors from textbooks, and fixed feedback is made available to all students without distinction. After reviewing the current states of assessment/quiz systems, including some commercial ones (i.e. Questgen, Quillionz), we couldn’t find an AI-based quiz system, which is ready to use and easy to expand with new AI capabilities. Considering our specific requirements for personalized learning and knowledge self-assessment, we proposed a scalable AI-powered quiz system, iQS. Initially, the following research questions are stated: Q1: What does a general, scalable, and AI-based quiz system look like? Q2: How does it provide personalized and intelligent feedback in real-time? Q3: What is the learning experience of students using it for their self-regulated learning? and are they motivated to engage more with learning?

This study tries to answer the above-stated research questions but is not limited to them. Preliminary results of our research have been published, which have covered the quiz generation algorithm and knowledge feedback algorithms. However, this paper presents the whole AI-based quiz system and explains its innovation from the teaching point of view, which is, how a tutor creates quiz items, how much effort is involved, and which support the tutor can get.

Keywords:

Quiz system, knowledge modelling, Ontology, personalised quizzes, self-regulated learning.