2026
Demystifying Action Space Design for Robotic Manipulation Policies
International Conference on Machine Learning (ICML), 2026
Jinliang Zheng 郑金亮
Ph.D. Candidate, Tsinghua University
I am a Ph.D. candidate in Mechanical Engineering at the Institute for AI Industry Research (AIR), Tsinghua University. I work on embodied AI, with a particular interest in generalist robot policies, cross-embodiment learning, multimodal representation learning, and generative models.
I am advised by Prof. Xianyuan Zhan and Prof. Ya-Qin Zhang. I am always open to research discussions and collaboration.
News
Research updates and recognitions
- Received the Gold Reviewer recognition at ICML 2026 and Top Reviewer recognition at ICLR 2026.
- Demystifying Action Space Design for Robotic Manipulation Policies was accepted to ICML 2026 and received the Best Paper Award at the ES-Reasoning Workshop @ ICLR 2026.
- X-VLA was accepted to ICLR 2026.
- X-VLA won 1st place in the AGIBOT World Challenge (Manipulation Track) @ IROS 2025.
- Received the Graduate National Scholarship.
Show earlier news
- UniAct was accepted to CVPR 2025.
- Efficient Robotic Policy Learning via Latent Space Backward Planning was accepted to ICML 2025.
- Diffusion-Based Planning for Autonomous Driving with Flexible Guidance was accepted to ICLR 2025 as an Oral presentation.
- Robo-MUTUAL was accepted to ICRA 2025.
- Instruction-Guided Visual Masking was accepted to NeurIPS 2024.
- IVM and DecisionNCE received Outstanding Paper Awards at the MFM-EAI Workshop @ ICML 2024.
- DecisionNCE was accepted to ICML 2024, and GLID was accepted to CVPR 2024.
- Enhancing Vision-Language Model with Unmasked Token Alignment at Scale was published in TMLR.
- MixMAE was accepted to CVPR 2023, and GoBigger was accepted to ICLR 2023.
Research Interests
Embodied AI · Robot Learning · Vision-Language-Action Models · Cross-Embodiment Learning · Multimodal Representation Learning · Generative Models
Selected Publications
* denotes equal contribution
2026
X-VLA: Soft-Prompted Transformer as Scalable Cross-Embodiment Vision-Language-Action Model
International Conference on Learning Representations (ICLR), 2026
2025
Universal Actions for Enhanced Embodied Foundation Models
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2025
2024
Instruction-Guided Visual Masking
Advances in Neural Information Processing Systems (NeurIPS), 2024
2024
DecisionNCE: Embodied Multimodal Representations via Implicit Preference Learning
International Conference on Machine Learning (ICML), 2024
Additional publications
- Efficient Robotic Policy Learning via Latent Space Backward Planning. ICML 2025.
- Diffusion-Based Planning for Autonomous Driving with Flexible Guidance. ICLR 2025, Oral.
- Robo-MUTUAL: Robotic Multimodal Task Specification via Unimodal Learning. ICRA 2025.
- GLID: Pre-training a Generalist Encoder-Decoder Vision Model. CVPR 2024.
- Enhancing Vision-Language Model with Unmasked Token Alignment at Scale. TMLR 2024.
- MixMAE: Mixed and Masked Autoencoder for Efficient Pretraining of Hierarchical Vision Transformers. CVPR 2023.
- GoBigger: A Scalable Platform for Cooperative-Competitive Multi-Agent Interactive Simulation. ICLR 2023.
Education
2023–Present
Tsinghua University
Ph.D. Candidate in Mechanical Engineering
School of Vehicle and Mobility · Institute for AI Industry Research
2019–2023
Beijing University of Posts and Telecommunications
B.S. in Computer Science and Technology
Excellent Graduate
Honors & Awards
Research, academic, and competition distinctions
- 2026Best Paper AwardES-Reasoning Workshop · ICLR 2026
- 2026
- 2026
- 20251st Place, AGIBOT World ChallengeManipulation Track · IROS 2025
- 2025Graduate National ScholarshipMinistry of Education, China
- 2024Outstanding Paper AwardMFM-EAI Workshop · ICML 2024
- 2023Excellent GraduateBeijing
- 2023CCF Excellent UndergraduateChina Computer Federation
- 2022Undergraduate National ScholarshipMinistry of Education, China
- 2022Silver MedalICPC Asia Regional Contest
- 2021Gold MedalGroup Programming Ladder Tournament
- 2021Merit StudentBeijing
Professional Service
- 2026IEEE Robotics and Automation Letters (RA-L)
- 2025–2026International Conference on Machine Learning (ICML)
- 2025–2026International Conference on Learning Representations (ICLR)
- 2025–2026Conference on Neural Information Processing Systems (NeurIPS)
- 2024–2025Conference on Computer Vision and Pattern Recognition (CVPR)
- 2025International Conference on Computer Vision (ICCV)
- 2026European Conference on Computer Vision (ECCV)
- 2026AAAI Conference on Artificial Intelligence (AAAI)
- 2026Winter Conference on Applications of Computer Vision (WACV)