Papers
1
Total Citations
2
H-Index
1
About
Zejia Weng is a researcher at the forefront of generative AI, with a primary focus on video prediction and diffusion models. His most notable contribution is the development of **Aid**, a novel framework that adapts Image2video diffusion models for instruction-guided video prediction. This work addresses the critical challenge of predicting future video frames from a single initial frame, guided by natural language instructions—a capability with transformative potential for virtual reality, robotics, and automated content creation. By innovatively adapting Stable Diffusion for this task, Weng's approach enables more precise and semantically meaningful video generation, bridging the gap between text and dynamic visual content. His research has already garnered early attention, with his 2025 paper accumulating 2 citations, signaling growing interest in this emerging direction. Weng's work stands out for its practical applicability, offering a pathway toward more interactive and controllable video synthesis systems that respond to user commands. As the field of text-to-video generation rapidly evolves, his contributions represent a significant step toward making AI-generated video more intuitive and instruction-faithful.
Research Focus
Key Achievements
Top Papers
- 1