Boyi Li

Title: Boyi Li and High-Resolution Video Generation

Introduction

Boyi Li is an inventor based in Berkeley, California, and works at NVIDIA Corporation. Her patent work includes a system and method for generating high-resolution videos from text prompts.

Latest Patents

Li is associated with 1 patent: “System and Method for Efficient Text-Guided Generation of High-Resolution Videos.” The invention describes training a content frame-motion latent diffusion model to generate requested videos. Its two-stage framework first uses an autoencoder to represent video content as a content frame and latent motion, then fine-tunes the content-frame distribution with a pretrained image diffusion model. A lightweight diffusion model generates motion representations conditioned on the content frame, enabling the system to draw on visual knowledge from pretrained image models while producing video.

Career Highlights

Li’s patent reflects work at the intersection of generative artificial intelligence, diffusion models, and video synthesis. The approach separates visual content from motion, providing a structured method for generating high-resolution videos guided by text.

Collaborations

The patent lists Sihyun Yu and Weili Nie among Li’s coworkers, reflecting a collaborative research environment at NVIDIA Corporation.

Conclusion

Boyi Li’s patent describes a two-stage diffusion approach that combines pretrained image-generation knowledge with motion modeling to support efficient, text-guided high-resolution video generation.