Graduate project

Structuring of a Student Model for an ITS - Asperger's Syndrome (Text Domain)

The aim of our project is to build an Intelligent Tutoring System (ITS) to help people with Asperger’s Syndrome (i.e. Asperger’s) to understand the subtle meaning and nuances in simple sentences in English. In this phase, we have attempted to use Machine Learning (ML) to build the initial prototype for our Student Model to see how people with Asperger’s tend to understand the basic short sentences in everyday conversation. Typically, ML algorithms train a machine to reach perfection, but this project aims at training a machine to mimic a human’s behavior – in other words, this human referred here is a patient with Asperger’s syndrome. A human being is not perfect, and so when our ML model is mimicking a human, the ML model is not perfect – but tries to behave like the patient. To achieve this objective, we interviewed experts in psychology department to understand the syndrome better and what types of words in everyday conversation would confuse them, aiming to build a Student Model that must be able to produce the same kinds of errors as the student. Also, we wanted our model to be constantly updating with data thereby getting to know the patient better. We came up with a list of sentences, collected responses and then extracted features, built, designed and tested ML systems based on Multinomial Naïve Bayes algorithm as well as Support Vector Machines. Upon completion, the Model can also be used by psychologists to understand the common errors in interpreting verbal cues people with Asperger’s make. The system can be used as a virtual patient in labs for research purposes.