Liangbing (Liam) Zhao

MS/PhD Student
King Abdullah University of Science and Technology
zhaolbben@gmail.com
Hi, I am Liangbing Zhao. You can also call me Liam. I am currently a MS/PhD student at King Abdullah University of Science and Technology (KAUST), under the guidance of my advisor, Prof. Mohamed Elhoseiny. During my undergraduate years, I was supervised by Prof. Si Liu. I was also a research intern at Meitu Inc., where I worked on video diffusion models with Dr. Xuecheng Nie.
Research Interests
- Generative Models: Advancing and applying generative models, such as diffusion models, flow matching, and autoregressive models, with a focus on improving their architectures and exploring innovative applications.
- Unified Models: Developing models that integrate generative and understanding tasks, with an emphasis on leveraging understanding capabilities to enhance generative performance.
news
Jun 26, 2025 | Two papers are accepted by ICCV2025! |
---|---|
Jan 22, 2025 | One paper is accepted by ICLR2025! |
Dec 19, 2023 | One paper is accepted by Electronics! |
selected publications
- From reflection to perfection: Scaling inference-time optimization for text-to-image diffusion models via reflection tuningarXiv preprint arXiv:2504.16080, 2025
- WikiAutoGen: Towards Multi-Modal Wikipedia-Style Article GenerationarXiv preprint arXiv:2503.19065, 2025
- ToddlerDiffusion: Interactive Structured Image Generation with Cascaded Schrödinger BridgeIn The Thirteenth International Conference on Learning Representations, 2025
- Cross-Attention and Seamless Replacement of Latent Prompts for High-Definition Image-Driven Video EditingElectronics, 2023