Utilising Multimodal Learning Analytics with Artificial Intelligence (Ai) to Predict Regulation in Collaborative Learning
The project is conducted at Learning and Educational Technology (LET) Research Lab. The study aims to utilise multimodal learning analytics with artificial intelligence (AI) to investigate regulation in collaborative learning. This multidisciplinary study bridges learning sciences, affective computing, information systems, and AI research for developing novel methodologies and advancing understandings of regulatory processes in collaborative learning. AI deep learning models will be designed and applied on multimodal data consisting of video, audio, self-reports, and physiological data (electrodermal activities and heartrates). The findings of this study will establish critical foundations for extending the boundaries of theory building and testing in learning sciences, especially for learning regulation research. The outcomes will also help educational technologists and developers design effective learning analytics solutions and tools for teachers for supporting regulation in collaborative learning and facilitate multidisciplinary collaboration.
For the field of learning sciences, this study will first demonstrate an attempt to utilise multimodal data to support learning regulation to improve learning and signal regulatory challenges faced by students for necessary intervention. The project will produce new analytical models and methods for exploring and supporting the regulatory process in SRL, CoRL, and SSRL. The outcomes of this project are also expected to lay a methodological base for the application of multimodal learning analytics to study regulation in collaborative learning. These outcomes will establish critical foundations for extending the boundaries of theory building and testing in learning sciences. To date, there is no standardised approach for designing learning analytics methods using multimodal and multichannel data to investigate learning regulation and this study will make ground-breaking results in this. Accordingly, this study contributes to the development of still lacking standard metrics and indicators for a few specific data types including the electrodermal activity (EDA), video-observation, and log data. This study will explore the temporality and cyclicality of regulation in both face-to-face learning and CSCL. These understandings will make practical contributions to CSCL design and help teachers recognise the regulatory challenges faced by the students and perform necessary interventions in education.
For the field of Information Systems (IS), this study provides design theories as essential foundations for the development and implementation of theory-grounded learning analytics systems. The success of this study will also deliver different IS research artefacts for detecting, predicting, and supporting regulation in collaborative learning. These artefacts include analytical models, frameworks, computational algorithms, technological architecture, design guidelines and principles for learning analytics information systems for supporting regulation in collaborative learning. Moreover, this study adds on the discourse on the theory building in design science research (DSR) for knowledge accumulation and evolution by presenting and evaluating a theory-driven development of IS artefacts (Nguyen, Tuunanen, et al., 2020). These outcomes will help educational technologists and developers design effective AI-enhanced multimodal learning analytics solutions and tools for supporting regulation in collaborative learning and facilitate multidisciplinary collaboration.