Using Mobile Technology for Coordinating Educational Plans and Supporting Decision Making Through Reinforcement Learning in Inclusive Settings
Learners with special education needs and disabilities (SEND) require attention from a large set of a care team that includes parents, teachers, specialists, therapists, and doctors. Good coordination among these stakeholders leads to increased behavioural and academic progress for the learners. However, achieving good coordination in such setting is a challenging task. This is due to the different tasks each stakeholder is attempting, the different backgrounds of the stakeholders, and the lack of face-to-face interaction among them. I call this the intervention coordination problem (ICP). Furthermore, learners with SEND, and specially learners with autism spectrum disorder (ASD), usually show little interest in academic activities and may display disruptive behaviour when assigned certain tasks. Research indicates that selecting a good motivational variable during interventions improves behavioural and academic performance. I refer to this problem as the motivator selection problem (MSP). This work aims to exploit mobile and artificial intelligence (AI) technologies in order to address the above two problems. Toward this aim, this study follows a design science research approach to develop the IEP-Connect app. This mobile app uses the Individualized Education Program (IEP) as the foundation for coordinating the efforts and supporting the decision-making process of the different personnel who are involved in the IEP of a child with special needs. The proposed work presents four significant contributions, namely identifying the key design principles to inform the design of a coordination mobile app for special education, developing and implementing the IEP-Connect mobile app, modelling the selection of a motivator as a Markov Decision Process (MDP), and proposing a Reinforcement Learning (RL) framework to recommend a motivator to be used with students with SEND in a given learning setting. To evaluate the effectiveness of the proposed mobile app and RL framework, a series of studies based on participatory design research, mixed-methods usability evaluation, and pre-test/post-test quasi-experimental research methodology were conducted. The evaluation of the app focused on students with ASD as their learning requires sharing information from different distributed sources. Results from the usability questionnaires, interviews, and log data revealed that the app has good usability and that participants were satisfied with the use of the app for recording and sharing IEP information. Moreover, evaluations and data analysis have shown the validity of the proposed RL framework through improving the intervention effectiveness and users’ satisfaction. The implementation of this work provides insights into the future development of technology tools that facilitate information sharing between special education teachers and other stakeholders involved in the intervention of children with special education needs. Moreover, this work expands the interdisciplinary research of machine learning and special education by presenting promising preliminary results for therapy decision-making support.