Scholar
Zhengrong Xue
Google Scholar ID: LO3pKmwAAAAJ
IIIS, Tsinghua University
Citations & Impact
Citations
288
H-index
8
i10-index
6
Publications
16
Co-authors
10
list available
Contact
Publications
Resume (English only)
Academic Achievements
- - DemoSpeedup: Accelerating Visuomotor Policies via Entropy-Guided Demonstration Acceleration (CoRL 2025 Oral)
- - DemoGen: Synthetic Demonstration Generation for Data-Efficient Visuomotor Policy Learning (RSS 2025 Best Long Paper Award; ICLR 2025 Robot Learning Workshop Oral)
- - MENTOR: Mixture-of-Experts Network with Task-Oriented Perturbation for Visual Reinforcement Learning (ICML 2025)
- - RiEMann: Near Real-Time SE(3)-Equivariant Robot Manipulation without Point Cloud Segmentation (CoRL 2024; CVPR 2024 EquiVision Workshop)
- - Key-Grid: Unsupervised 3D Keypoints Detection using Grid Heatmap Features (NeurIPS 2024)
- - AToM-Bot: Embodied Fulfillment of Unspoken Human Needs with Affective Theory of Mind (Social Intelligence in Humans and Robots Workshop, RSS 2024)
- - Arraybot: Reinforcement Learning for Generalizable Distributed Manipulation through Touch (ICRA 2024)
- - OTAS: Unsupervised Boundary Detection for Object-Centric Temporal Action Segmentation (WACV 2024)
- - USEEK: Unsupervised SE(3)-Equivariant 3D Keypoints for Generalizable Manipulation (ICRA 2023)
- - Pre-Trained Image Encoder for Generalizable Visual Reinforcement Learning (NeurIPS 2022 Spotlight)
- - Skeleton Merger: an Unsupervised Aligned Keypoint Detector (CVPR 2021 Oral)
Research Experience
- - Ph.D. student at Tsinghua Embodied AI Lab (TEA Lab).
Education
- - Ph.D. student at Tsinghua Embodied AI Lab (TEA Lab), Institute for Interdisciplinary Information Sciences (IIIS), Tsinghua University, supervised by Prof. Huazhe Xu.
- - Bachelor's Degree at AI class, Shanghai Jiao Tong University, advised by Prof. Cewu Lu.
- - Visited Prof. Hao Su at UC San Diego in the summer of 2023.
Background
- Research interest: embodied AI, especially robotic manipulation and imitation learning. Recent research goal is to develop more generalizable, dexterous, and data-efficient robot learning algorithms and systems for robotic manipulation.
Miscellany
- In spare time, enjoys photography, traveling, foods, drinks, tennis, and Formula One racing.
