Human Motion · Robot Learning

Yan Wu

I am a Ph.D. student at the Computer Vision and Learning Group, ETH Zürich, supervised by Prof. Siyu Tang. Before that, I completed my master’s degree at ETH Zürich and worked closely with Prof. Zhiwu Huang, Prof. Siyu Tang, and Dr. Suryansh Kumar. I received bachelor’s degrees through a double-degree program at Tongji University and the University of Bologna.

My research explores physics-based character animation and humanoid control, and vision-based policy learning.

Selected Publications

Selected research. For a complete and current list, visit my Google Scholar profile.

ADAPT demonstrations of interactive text-conditioned humanoid control and style-aware goal reaching

ADAPT: Agile Diffusion Action Priors for Robust and Steerable Online Text-Driven Humanoid Control

Yan Wu, Chenhao Li, Kaifeng Zhao*, Gen Li*, Marco Hutter, Siyu Tang

arXiv, 2026 · * Equal contribution

Project page arXiv Video

@article{wu2026adapt,
  title={ADAPT: Agile Diffusion Action Priors for Robust and Steerable Online Text-Driven Humanoid Control},
  author={Wu, Yan and Li, Chenhao and Zhao, Kaifeng and Li, Gen and Hutter, Marco and Tang, Siyu},
  journal={arXiv preprint arXiv:2609.00677},
  year={2026}
}

An end-to-end framework for interactive, text-conditioned humanoid control that executes diverse skills, transitions smoothly between prompts, and adapts to downstream tasks.

NaP-Control demonstrations of character control across goals, terrain, and object interactions

NaP-Control: Navigating Diffusion Prior for Versatile and Fast Character Control

Chia-Wen Chen, Yan Wu, Korrawe Karunratanakul, Siyu Tang

ECCV, 2026

Project page arXiv Code Video

@inproceedings{chen2026napcontrol,
  title={NaP-Control: Navigating Diffusion Prior for Versatile and Fast Character Control},
  author={Chen, Chia-Wen and Wu, Yan and Karunratanakul, Korrawe and Tang, Siyu},
  booktitle={Proceedings of the European Conference on Computer Vision (ECCV)},
  year={2026}
}

A reinforcement-learning approach that steers the latent noise of a diffusion policy prior for fast, robust, and versatile physics-based character control.

UniPhys demonstrations of text-driven control, goal reaching, velocity control, and object avoidance

UniPhys: Unified Planner and Controller with Diffusion for Flexible Physics-Based Character Control

Yan Wu, Korrawe Karunratanakul, Zhengyi Luo, Siyu Tang

ICCV, 2025 · Highlight

Project page arXiv Code Dataset Video

@inproceedings{wu2025uniphys,
  title={UniPhys: Unified Planner and Controller with Diffusion for Flexible Physics-Based Character Control},
  author={Wu, Yan and Karunratanakul, Korrawe and Luo, Zhengyi and Tang, Siyu},
  booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
  year={2025}
}

A diffusion-based unified planner and controller that generalizes across reactive and long-horizon character-control tasks without task-specific training.

SAGA generates whole-body human motion to approach and grasp an object

Learning Deep Sensorimotor Policies for Vision-Based Autonomous Drone Racing

Jiawei Fu, Yunlong Song, Yan Wu, Fisher Yu, Davide Scaramuzza

IROS, 2023

PDF

@inproceedings{fu2023sensorimotor,
  title={Learning Deep Sensorimotor Policies for Vision-Based Autonomous Drone Racing},
  author={Fu, Jiawei and Song, Yunlong and Wu, Yan and Yu, Fisher and Scaramuzza, Davide},
  booktitle={IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
  year={2023}
}

A vision-based sensorimotor policy for autonomous drone racing, learned directly from visual observations and demonstrated in high-speed flight.

SAGA: Stochastic Whole-Body Grasping with Contact

Yan Wu*, Jiahao Wang*, Yan Zhang, Siwei Zhang, Otmar Hilliges, Fisher Yu, Siyu Tang

ECCV, 2022 · * Equal contribution

Project page arXiv Code

@inproceedings{wu2022saga,
  title={SAGA: Stochastic Whole-Body Grasping with Contact},
  author={Wu, Yan and Wang, Jiahao and Zhang, Yan and Zhang, Siwei and Hilliges, Otmar and Yu, Fisher and Tang, Siyu},
  booktitle={European Conference on Computer Vision (ECCV)},
  year={2022}
}

Starting from an arbitrary pose, SAGA generates diverse, natural whole-body motions that approach and grasp a target object in 3D space.

Neural Architecture Search as Sparse Supernet

Yan Wu*, Aoming Liu*, Zhiwu Huang, Siwei Zhang, Luc Van Gool

AAAI, 2021 · * Equal contribution

arXiv

@inproceedings{wu2021neural,
  title={Neural Architecture Search as Sparse Supernet},
  author={Wu, Yan and Liu, Aoming and Huang, Zhiwu and Zhang, Siwei and Van Gool, Luc},
  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
  year={2021}
}

A continuous architecture representation that frames neural architecture search as learning a sparse supernet.