Background and Project Content

Teachers, lecturers, and other professional groups working in knowledge transfer often need catalogues of questions or quiz questions in order to teach contents in a playful manner or to further deepen previously learned material. However, coming up with appropriate questions that live up to high standards of education requires time and effort. To make this task easier for educators, we aim to develop an application that can be used to create quizzes automatically. Our main focus is on creating high-quality catalogues of questions and tasks that prove their worth in everyday teaching and offer a real added value. Another special emphasis is on economic feasibility. To this end, we analyse the market, develop a unique selling point, and realise a sustainable business model.

Objectives

The project’s overarching goal is to develop an application that automatically generates quizzes based on existing didactic material. These quizzes should meet high pedagogical standards and provide the greatest possible benefit for educators and learners. We experiment with various types of quizzes, determine the formats that work best for us, and make sure that they are optimally embedded into the curricula. To arrive at a workable solution, the following questions must be answered:

• What is the educational benefit of the different quiz formats?
• How can the difficulty level of the questions be determined?
• To what extent can quizzes help facilitate users’ understanding of contents?

Methods

Detailed interviews are conducted with lecturers of selected Austrian universities (TU Wien, WU Wien, JKU, Joanneum Graz, IMC Krems, etc.) to better understand what they expect from a quiz. The key findings from this form the foundation for user-oriented solutions. Moreover, our research draws on findings from educational research and the cognitive sciences. Findings are compiled and passed on to all participating universities. We also ask for input, to make sure that the needs of future users are sufficiently considered in the application’s development.

The centrepiece of our application is a generative AI model based on the latest NLP (natural language processing) technologies. We explore the fundamental mechanisms of the large language models such as BioGPT, ChatGPT, and Bard as well as open-source models like LLaMa and incorporate the insights gained into our AI model. To make our application as easily accessible as possible, we rely on design principles guided by simplicity and intuitively operable interfaces. By doing this, technical obstacles that could impair usability or discourage users are reduced.

Results

The outcome of this project a fully functional application for the generation of quiz questions, which can be used in the educational sector and commercially distributed. The application not only aids teachers in formulating meaningful quiz questions but is also benefits learners/students. It enhances the understanding of learning content and plays a supplementary role in the entire learning process. Universities can realign and expand their course offerings in a cost-effective manner with the help of the application.