Yiming Wang

I am an undergraduate student at Shanghai Jiao Tong University. My current interests are embodied AI and real-world robot learning.

I will pursue my PhD at the College of AI, Tsinghua University.

I previously interned at the SJTU Intelligent Media Lab, advised by Bingbing Ni, and subsequently at MINT SJTU, advised by Bo Zhao.

1. Oct 2026 PrismWAM was released on arXiv.
2. Sep 2026 Acting in Meters was released on arXiv.
3. 2026 SVGEval was accepted to ECCV 2026.
4. Jun 2026 LA4VLA was released on arXiv.
5. Oct 2025 Started focusing on embodied AI, VLA systems, and real-world robot learning.
6. 2025 Worked on SVG generation, benchmarks, and evaluation-oriented generation systems.
7. Dec 2024 Started research work around controllable image generation and editing.

PrismWAM: Decoupling State Representations in World Action Models

Tianyu Xin*, Wanqi Zhong*, Bowen Hu, Zhangchi Guo, Lijie Wang, Yiming Wang, Haoran Liao, Heyang Yu, Xiaoyu Tian, Zhiding Yu, and Yiming Li. arXiv, 2026.

A world action model that separates semantic planning from depth-based physical rollout to generate robot actions.

arXiv Project

Acting in Meters: Learning Metric Interactions for Precise Robotic Manipulation

Lijie Wang*, Zheng Lu*, Yiming Wang*, Heyang Yu, Kenghou Hoi, Bowen Hu, Di Cui, Tianyu Xin, Haoran Liao, Wanqi Zhong, Xingjie Fan, Yizhao Xu, Ziliang Wang, Fei Gao, and Yiming Li. arXiv, 2026.

Metric interaction modeling for precise manipulation, combining object-relative end-effector trajectories with scene geometry to refine VLA and WAM action predictions.

arXiv

SVGEval: A Vision-Grounded Framework for Perceptual-Quality Benchmarking and Evaluation in Text-to-SVG Generation

Yiming Wang*, Ye Chen*, Hanqi Chen, and Bingbing Ni. ECCV 2026, accepted / to appear.

A vision-grounded benchmark for SVG generation quality assessment, covering text-to-SVG evaluation, human annotation, and interpretable SVG scoring models.

ECCV 2026 arXiv

LA4VLA: Learning to Act without Seeing via Language-Action Pretraining

Tao Lin, Yuxin Du, Yiran Mao, Zewei Ye, Yilei Zhong, Bing Cheng, Yiming Wang, et al. arXiv, 2026.

Language-action pretraining for VLA models, studying how language-action supervision transfers to robot perception and control.

arXiv Project

* Equal contribution.