Xi Wang

Professor · Head of Lab

xi.wang (at) uni-ulm.de

Xi Wang is a professor at Ulm University, where she leads the Ego-AI Lab (Egocentric Vision and Embodied AI Lab). Before joining Ulm she led a junior research group at the Technical University of Munich and was an ETH Postdoctoral Fellow in the Advanced Interactive Technologies lab at ETH Zurich. She received her PhD from the Computer Graphics group at TU Berlin and spent research stays at MIT and Adobe Research.

Her research lies at the intersection of egocentric vision and embodied AI. She is interested in how human intent drives actions and interactions with the surroundings, and in bringing human perception and common sense into computational models that perceive, reason and act.

Publications

EgoFlow: Gradient-Guided Flow Matching for Egocentric 6DoF Object Motion Generation

Abhishek Saroha, Huajian Zeng, Xingxing Zuo, Daniel Cremers, Xi Wang

CVPR 2026 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

Project

@inproceedings{saroha2026egoflow,
  title     = {{EgoFlow}: Gradient-Guided Flow Matching for Egocentric {6DoF} Object Motion Generation},
  author    = {Saroha, Abhishek and Zeng, Huajian and Zuo, Xingxing and Cremers, Daniel and Wang, Xi},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year      = {2026},
}

ViSTA-SLAM: Visual SLAM with Symmetric Two-view Association

Ganlin Zhang, Shenhan Qian, Xi Wang, Daniel Cremers

3DV 2026 International Conference on 3D Vision (3DV)

arXiv Project Code Video

@inproceedings{zhang2026vista,
  title     = {{ViSTA-SLAM}: Visual {SLAM} with Symmetric Two-view Association},
  author    = {Zhang, Ganlin and Qian, Shenhan and Wang, Xi and Cremers, Daniel},
  booktitle = {International Conference on 3D Vision (3DV)},
  year      = {2026},
}

GMT: Goal-Conditioned Multimodal Transformer for 6-DOF Object Trajectory Synthesis in 3D Scenes

Huajian Zeng^*, Abhishek Saroha^*, Daniel Cremers, Xi Wang

3DV 2026 International Conference on 3D Vision (3DV)

* equal contribution

Paper arXiv Project Code Video

@inproceedings{zeng2026gmt,
  title     = {{GMT}: Goal-Conditioned Multimodal Transformer for {6-DOF} Object Trajectory Synthesis in {3D} Scenes},
  author    = {Zeng, Huajian and Saroha, Abhishek and Cremers, Daniel and Wang, Xi},
  booktitle = {International Conference on 3D Vision (3DV)},
  year      = {2026},
}