VR Math Bridge: Bridging Interactivity in Online Education with AI and VR

HuiShan Lai; Alaeddin Nassani; John Blake; Julián Villegas

doi:10.1109/GEM66882.2025.11155841

Abstract

VR Math Bridge extends established online calculus content with an immersive classroom and an AI-driven teaching assistant that can answer questions in real time.

Students choose a chapter from a floating panel, watch Khan Academy material on a virtual blackboard, and pause to ask an embodied 3D tutor for personalized explanations. Cognitive 3D captures attention and interaction signals to help researchers understand how learners use each part of the environment.

The current Quest 3 prototype focuses on derivatives. A preliminary two-participant evaluation was used to surface usability issues and guide a larger future study; it does not establish learning effectiveness.

System

System overview

The prototype combines trusted lesson content, voice-based assistance, and spatial interaction logging in one learning flow.

$VR Math Bridge system architecture connecting the Quest controller, virtual classroom, Khan Academy content, AI assistant, speech services, avatar animation, and Cognitive 3D$

Voice input is transcribed and sent to an AI assistant grounded in lesson transcripts and learning materials. The response returns through speech, lip-sync animation, and an on-screen chatbox.

$First-person view of the VR classroom, floating chapter selector, virtual blackboard, and Quest controllers$

Virtual classroom

A chapter selector and virtual blackboard place existing calculus lessons in a focused spatial environment.

$The embodied AI teaching assistant with labeled chatbox, lip-sync, Quest controller, and speech-to-text interaction$

Embodied AI assistant

Speech recognition, text-to-speech, lip-sync, and avatar animation support a voice-first exchange.

$Cognitive 3D Scene Explorer replay and analytics dashboard used to review movement and engagement$

Spatial analytics

Cognitive 3D records watch time, tutor interaction duration, movement, and replayable scene data for later study.