|
I am a Ph.D. student in the School of Software at Tsinghua University, supervised by Prof. Yu-Shen Liu. I started my Ph.D. in September 2026 and received my Master's degree in Software Engineering from the same school in 2026. My research interests center on physical AI. I aim to build world models that enable intelligent agents to understand the physical world, anticipate the consequences of their actions, and interact with it effectively. |
|
|
|
(* Equal Contribution) |
|
Weiqi Zhang, Wenyuan Zhang, Yu-Shen Liu, Junsheng Zhou Conference on Neural Information Processing Systems (NeurIPS), 2026 project page | code Unifies video generation and 3D world modeling by decoding video latents into explicit 4D Gaussians, enabling real-time rendering across viewpoints and timesteps. |
|
Weiqi Zhang*, Junsheng Zhou*, Xuancheng Zhang, Juntong Fang, Zequn Chen, Donglin Di, Yu-Shen Liu Conference on Neural Information Processing Systems (NeurIPS), 2026 project page | code Reconstructs dynamic 4D Gaussian scenes from monocular videos in seconds, jointly modeling geometry, appearance, and motion for real-time novel view synthesis. |
|
Weiqi Zhang*, Junsheng Zhou*, Juntong Fang*, Kanle Shi, Shenkun Xu, Yu-Shen Liu Conference on Neural Information Processing Systems (NeurIPS), 2026 project page Generates native 4D Gaussian Splatting from image and text conditions, using temporally structured latents to separate geometry from appearance. |
|
Weiqi Zhang*, Junsheng Zhou*, Haotian Geng, Kanle Shi, Shenkun Xu, Yi Fang, Yu-Shen Liu IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026 project page | arXiv | code A geometry-aware framework that grows 3D Gaussians from point clouds with text guidance for controllable 3D generation. |
|
Juntong Fang*, Zequn Chen*, Weiqi Zhang*, Donglin Di, Xuancheng Zhang, Chengmin Yang, Yu-Shen Liu IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026 (Highlight) project page | arXiv | code A feed-forward 4D reconstruction network that efficiently recovers dynamic 3D scenes from monocular videos by disentangling dynamic motion from static structure with attention-forcing and grouped causal attention. |
|
Junsheng Zhou*, Weiqi Zhang*, Baorui Ma, Kanle Shi, Yu-Shen Liu, Zhizhong Han IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2026 project page | arXiv | code A unified framework providing large-scale datasets, benchmarks, and generative models for UDF-based 3D shape representation and generation. |
|
Wenyuan Zhang*, Jimin Tang*, Weiqi Zhang, Yi Fang, Yu-Shen Liu, Zhizhong Han Advances in Neural Information Processing Systems (NeurIPS), 2025 project page | arXiv | code We introduce MaterialRefGS that significantly improves the rendering and reconstruction of reflective scenes in Gaussian Splatting by introducing PBR-based multi-view material consistency constraints and reflection strength prior. |
|
Weiqi Zhang*, Junsheng Zhou*, Haotian Geng*, Wenyuan Zhang, Yu-Shen Liu IEEE/CVF International Conference on Computer Vision, (ICCV), 2025 project page | arXiv | code GAP gaussianizes raw point clouds into high-fidelity 3D Gaussians with text guidance via depth-aware diffusion and surface-anchored optimization. |
|
Takeshi Noda*, Chao Chen*, Junsheng Zhou, Weiqi Zhang, Yu-Shen Liu, Zhizhong Han IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2025 project page | arXiv | code Bijective-SDF improves 3D surface reconstruction from sparse point clouds by learning bijective surface parameterization and optimizing grid deformation. |
|
Junsheng Zhou*, Weiqi Zhang*, Yu-Shen Liu Conference on Neural Information Processing Systems (NeurIPS), 2024 project page | arXiv | code DiffGS is a powerful and efficient 3D generative model capable of generating Gaussian primitives in arbitrary numbers, with a functional disentanglement of Gaussian Splatting. |
|
Takeshi Noda*, Chao Chen*, Weiqi Zhang, Xinhai Liu, Yu-Shen Liu, Zhizhong Han Conference on Neural Information Processing Systems (NeurIPS), 2024 project page | arXiv | code MultiPull is a novel method that improves 3D surface reconstruction from point clouds by optimizing multi-scale implicit fields and leveraging frequency features. |
|
Junsheng Zhou*, Weiqi Zhang*, Baorui Ma, Kanle Shi, Yu-Shen Liu, Zhizhong Han IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024 project page | arXiv | code We present UDiFF, a 3D diffusion model for unsigned distance fields (UDFs) which is capable to generate textured 3D shapes with open surfaces from text conditions or unconditionally. |
|
|
|
template adapted from
this awesome website
|