Video generation & editing
Unified visual models that generate, understand, and edit images and video.
PHD CANDIDATE · HONG KONG POLYTECHNIC UNIVERSITY
Ruibin Li李锐斌
Making visual worlds
move, create, and evolve.
I research generative visual intelligence — from high-quality image and video synthesis to efficient, long-horizon streaming generation.
VIDEO GENERATION / DIFFUSION & FLOW / STREAMING
01 /
A little about me
I’m a Ph.D. candidate in Computing at The Hong Kong Polytechnic University, supervised by Prof. Lei Zhang.
My work connects visual quality, generation efficiency, and temporal consistency. I’m particularly interested in bringing video generation closer to real-time interaction, while preserving the ability to create coherent content over long horizons.
Supported by the Hong Kong PhD Fellowship Scheme (HKPFS).
02 /
QUALITY × EFFICIENCY × CONTINUITYWhat I work on
Efficient generative models
Diffusion and flow-model post-training, few-step distillation, and faster visual synthesis.
Long-horizon streaming
Temporal memory and real-time synthesis for continuous, interactive visual experiences.
CONTINUOUS BY DESIGN STREAM
One frame. A world of possibilities.
Research focus · conceptual illustration
03 /
* EQUAL CONTRIBUTIONPublications & preprints
Publication details follow my CV, updated October 2026. Work under review is labeled separately.
04 /
The journey so far
RESEARCH EXPERIENCE
EDUCATION
05 /
Honors & recognition
LET’S CONNECT
Good ideas start
with a conversation.
For research conversations and collaboration, feel free to reach out.