Yuhao Zhang

Yuhao Zhang (张钰浩)

I am a Master student at Shanghai Jiao Tong University, Department of Computer Science, working in ScaleLab under the supervision of Prof. Yao Mu.

My research focuses on Embodied AI, with a particular interest in Embodied World Models and Vision-Language-Action (VLA) models.

Email / Scholar / GitHub

News

[2026.07] Our paper R3DP was accepted to ECCV 2026 as Oral.
[2026.06] Selected as Outstanding Camper at Shanghai Innovation Institute (上海创智学院优秀营员).
[2026.06] Two papers (R3DP and M²Tok) accepted to ECCV 2026.
[2026.02] Our paper MM-ACT was accepted to CVPR 2026.
[2026.01] Our paper SimpleVLA-RL was accepted to ICLR 2026.

Research

I'm interested in embodied intelligence, world models, and VLA. Representative papers are highlighted. * denotes equal contribution.

	R3DP: Real-Time 3D-Aware Policy for Embodied Manipulation Yuhao Zhang, Wanxi Dong, Yue Shi, Yi Liang, Jingnan Gao, Qiaochu Yang, Yaxing Lyu, Zhixuan Liang, Yibin Liu, Congsheng Xu, etc. European Conference on Computer Vision (ECCV)*, 2026 (Oral) project page / arXiv / code A real-time 3D-aware manipulation policy that couples VGGT-based priors with a lightweight temporal predictor and multi-view fuser.
	Implicit Drifting Policy: One-Step Action Generation via Conditional Expert Geometry Zemin Yang, Yaoyu He, Yiming Zhong, Yuhao Zhang, Xinge Zhu, Yao Mu, Qingqiu Huang, Yuexin Ma arXiv preprint, 2026 arXiv A one-step imitation learning framework using conditional expert geometry to enforce action-manifold constraints without explicit drifting-field estimation.
	M²Tok: Multi-head Multi-codebook Discrete Action Tokenization for Vision-Language-Action Models Chunpu Xu, Zhixuan Liang, Yuhao Zhang, Chi-Min Chan, Jessie Wang, Yang Xiao, Mengkang Hu, Xiaokang Yang, Yao Mu European Conference on Computer Vision (ECCV), 2026 A multi-head multi-codebook discrete action tokenization scheme for scaling VLA models on continuous control.
	MM-ACT: Learn from Multimodal Parallel Generation to Act Haotian Liang, Xinyi Chen, Bin Wang, Mingkang Chen, Yitian Liu, Yuhao Zhang, Zanxin Chen, Tianshuo Yang, Yilun Chen, Jiangmiao Pang, etc. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026 arXiv / code A unified VLA model unifying text, image, and action generation in a shared token space; 96.3% on LIBERO.
	SimpleVLA-RL: Scaling VLA Training via Reinforcement Learning Haozhan Li, Yuxin Zuo, Jiale Yu, Yuhao Zhang, Zhaohui Yang, Kaiyan Zhang, Xuekai Zhu, Yuchen Zhang, Tianxing Chen, Ganqu Cui, etc. International Conference on Learning Representations (ICLR)*, 2026 arXiv / code A simple, scalable RL recipe for training vision-language-action models.
	Benchmarking Generalizable Bimanual Manipulation: RoboTwin Dual-Arm Collaboration Challenge at CVPR 2025 MEIS Workshop Tianxing Chen, Kaixuan Wang, Zhaohui Yang, Yuhao Zhang, Zanxin Chen, Baijun Chen, Wanxi Dong, Ziyuan Liu, Dong Chen, Tianshuo Yang, etc. CVPR 2025 MEIS Workshop* (Challenge Report), 2025 arXiv / challenge page A large-scale challenge on generalizable bimanual manipulation across 17 tasks in simulation and the real world.

Education

2024.09 – present: M.S. in Computer Science, Shanghai Jiao Tong University
2020.09 – 2024.06: B.E. in Software Engineering, Shandong University

Awards

2026: Outstanding Camper, Shanghai Innovation Institute (上海创智学院优秀营员)
2026: ECCV 2026 Oral Presentation — R3DP

Template borrowed from Jon Barron.