Immersive Learning System
Project Overview
Research scope & objectives
Interaction Design for Immersive VR Classrooms
This project focuses on the design of interaction in immersive VR classrooms generated from existing MOOC lectures. The work transfers interaction techniques established in video-based learning (VBL) research into immersive environments, and investigates which techniques are suited to a VR classroom, at what point in a lecture they should be introduced, and how they affect learner engagement. On this basis, the project develops an automated approach to planning the immersive experience, determining what interaction occurs, when it occurs, and how it is presented. The resulting system is evaluated against traditional video-based learning and optimized for consumer-grade VR headsets so that it remains accessible to a wide range of learners.
Multilingual Capabilities of Large Language Models for Immersive Classrooms
This project investigates the linguistic capabilities of large language models in Indian languages such as Hindi, Assamese, and Bengali, with the aim of supporting multilingual interaction in immersive classrooms. The evaluation spans models of different tiers, covering open-source and frontier systems across a range of parameter counts, in order to understand how capability varies with model scale and availability. A central concern is cross-lingual question answering, in which a learner query in an Indian language is used to retrieve information from English source material and the response is returned in the original language, across queries ranging from simple factual questions to more complex ones. The work further examines audio processing for spoken input and output, and the extent to which these capabilities support natural interaction between learners and the system. The findings are intended to inform the design of multilingual support in immersive learning environments.