OptiReality: The Optimal AI Solution for Virtual Reality
-
Enabling real-time immersive AR/VR with efficient generative AI algorithms and edge deployment.
Generative AI is becoming a cornerstone technology for advancing real-time situational awareness and human-machine interaction in virtual reality (VR) environments, especially when deployed at the edge. However, these models are computationally intensive, often requiring significant processing power. When deployed on resource-constrained AR/VR device, such complexity can lead to high latency, both of which are critical bottlenecks for maintaining a responsive and comfortable user experience in VR.
We are focused on developing faster and more efficient Generative AI applications for AR/VR devices, aiming to enhance the user experience seamlessly. The students will explore state-of-the-art generative AI algorithms and apply them to practical AI devices. They will then develop innovative AI algorithms to improve their implementation, ensuring a smooth user experience. Additionally, the students will learn to implement these new algorithms on AR/VR devices and conduct user studies to assess their performance.
Methods and Technologies
- 1. Generative AI (e.g., diffusion models, transformers)
- 2. Edge Computing
- 3. Real-time Inference Optimization
- 4. AR/VR Development (e.g., Unity, Unreal Engine)
- 5. Model Compression (e.g., quantization, pruning)
- 6. On-device Machine Learning
- 7. GPU Programming (e.g., CUDA, TensorRT)
- 8. Audio-Visual Rendering
- 9. Human-Centered Design
- 10. User Experience Evaluation & User Studies
Areas of Interest
- 1. Computer Science
- 2. Electrical and Computer Engineering
- 3. Artificial Intelligence and Machine Learning
- 4. Human-Computer Interaction
- 5. Virtual and Augmented Reality
- 6. Embedded and Edge Systems
Primary Instructor
- Sai Qian Zhang, Assistant Professor