Hy-Nomas: Towards Free-Form Interactive World Generation
Technical Report, 2026
An interactive world generation model that supports free-form, multi-turn text instructions and camera control.
Yifan Liu
I am a Researcher at Tencent HY, working with Dr. Heng Wang in the Video World Foundation Model Team. Previously, I received my Ph.D. in Electronic Engineering from the Chinese University of Hong Kong, advised by Prof. Yixuan Yuan. During my Ph.D., I have been fortunate to work with Dr. Zhenwei Wang and Dr. Tengfei Wang at the Tencent HY 3D.
My research focuses on video generation, world models, and 3D vision, with an emphasis on building practical products for generation and understanding in real-world applications.
I'm always open to related discussions and collaborations. If you are interested in my work, please feel free to contact me.
Technical Report
Hy-World 1.1
Technical Report, 2025
A 3D world reconstruction framework capable of producing immersive, explorable, and interactive scenes from arbitrary-view inputs.
Selected Publications [full list]
* Equal contribution
WorldMirror: Universal 3D World Reconstruction with Any-Prior Prompting
International Conference on Machine Learning (ICML), 2026
A universal 3D reconstruction framework capable of processing diverse input modalities and generating various output representations.
MonoSplat: Generalizable 3D Gaussian Splatting from Monocular Depth Foundation Models
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2025
A generalizable 3D Gaussian Splatting approach from monocular depth foundation models.
Foundation Model-guided Gaussian Splatting for 4D Reconstruction of Deformable Tissues
IEEE Transactions on Medical Imaging (TMI), 2025
The first work to apply 3D Gaussian Splatting for dynamic (4D) endoscopic reconstruction, guided by foundation models.
U-KAN Makes Strong Backbone for Medical Image Segmentation and Generation
AAAI Conference on Artificial Intelligence (AAAI), 2025
Introducing Kolmogorov-Arnold Networks (KANs) as a strong backbone for medical image segmentation and generation.
Education
Ph.D. in Electronic Engineering
The Chinese University of Hong Kong (CUHK)
Advisor: Prof. Yixuan Yuan
B.Eng. in Electronic Information Engineering
Huazhong University of Science and Technology (HUST)
Experience
- Research Intern
- Research Intern
Selected Honors
- Golden Reviewer Award, ICML 2026
- Postgraduate Studentship, CUHK, 2021–2026
- Outstanding Graduates, HUST, 2021
- First Class Scholarship, HUST, 2019–2020
- Scholarship of Academic Excellence, HUST, 2017–2020
Academic Services
- Conference Reviewer: CVPR, ICCV, ECCV, NeurIPS, ICLR, ICML, ICRA, IROS
- Journal Reviewer: TPAMI, IJCV, TVCG, TNNLS, TCSVT, TMM, TMI, MIA
Miscellaneous
Sports: basketball🏀, swimming🏊♂️, table tennis🏓
Games: Delta Force, CS:GO
