Structuring of Student Model for an ITS-Asperger's Syndrome

Our ultimate goal is to develop an Intelligent Tutoring System (ITS) to help train people with Asperger’s Syndrome (i.e. Asperger’s) in interpreting non-verbal cues such as facial expressions. In this phase of the project, we demonstrated how machine learning can be used to build the Student Model for recognizing emotions in facial expressions. The Student Model must be able to produce the same kinds of errors as the student. We searched for a technique that can learn and update the Student Model dynamically as more information is obtained about the student. To achieve this objective we interviewed experts to investigate the difficulties a person with Asperger’s faces, collected facial expression images and analysed the features related to emotions, implemented a web-based tool to collect actual data from potential students so that the initial Student Models can be built, designed and tested machine learning systems based on Support Vector Machines and Random Forrest Classifier to produce the Student Model despite having small amounts of training data, and implemented the user interface for demonstrating the Student Model in predicting student answers. We also list the necessary steps for improving the Student Model and incorporating the Student Model module into an ITS system. Upon completion, the Model can also be used by psychologists to understand the common errors in interpreting non-verbal cues people with Asperger’s make.