This demo showcases my project which was part of my Bachelor's research on Computer Adaptive Learning. Here's the abstract from my Thesis:
The application was made in Unity which can be split into two parts: the Editor and the 3 Games, along with 2 adaptivity features.
The Editor
Allows the educator to create a new MCQ game with their own pool of questions, however there's a trick. The CAL adaptivity feature is implemented to adapt the difficulty of the questions selected for a user to their ongoing performance using the Bayesian Machine Learning Method. Simply put, the difficulty increases when the user answers questions correctly and decreases when they answer questions wrong. The main purpose of this feature is to target questions that challenge the learner. In order for this to be done, the educator has to first perform an "empirical calibration" of the questions by giving the paper-based questions to around 50 students and entering the students' answers in the Editor. The educator can then publish their fully adaptive game.
The 3 Games
From the learner's side, the application consists of 3 games that play back to back. The Car game, the Locked doors game and the Pipes Game. The games are designed to render MCQs to the user. They were not my own design, however I developed them, enhanced their UI, and integrated the different adaptivity features into them.
The Emotional Adaptivity
In the experiment, we aimed to compare two versions of the application; one with the CAL adaptivity feature and one which adapts the UI and game mechanics to the user's emotions. The user expresses their emotions through a valence-arousal scale and a set of emojis. For example, if they feel frustrated they should click the red angry emoji, if they feel bored then they click the bored emoji in the pink tile, etc. Upon clicking the angry emoji for example the background color in the game turns green and the music becomes soft to promote positive emotions. Checkout the following video to see the changes.
