Machine Learning Automation for Virtual Reality

Publication details

• Authors Erdem Murat
• Publication venue Master's thesis, George Mason University
• Date 12/2022
• Paper View paper
• Focus Machine learning automation, procedural design, and VR user studies

Abstract

This thesis explores how data-driven and automated approaches can support virtual reality design and improve user experience. The work proposes a system that uses user data to inform VR development, with a focus on procedural difficulty design, automated content variation, and the relationship between design parameters and player response. By generating multiple game levels and studying user experience across them, the thesis contributes a framework for more systematic experimentation and evaluation in VR.

Highlights

• Developed a research system intended to improve VR development and player experience analysis.
• Used procedural difficulty design to generate and compare varied game levels.
• Collected data from more than 50 users after receiving IRB certification for social and behavioral research.