Xianyuan Zhan
Associate Researcher / Associate Professor
Works on data-driven decision-making and AI for industrial system optimization, including offline reinforcement learning.
Xianyuan Zhan's lab at Tsinghua works on reinforcement learning and learning theory & optimization. 39 papers since 2023, 25 at top venues, mostly ICLR and NeurIPS. 3 PhD students. 14 alumni, most often to Baidu and Huawei. Applications close .
At a glance
- papers, 2023–2026
- 10/ year
- at T1 venues
- 64%
- PhD students
- 3
- MS students
- 0
- alumni listed
- 14
Papers by yearfrom the lab homepage, 2023–2026
- 5 papers2026 · T1 5 · collecting
- 10 papers2025 · T1 9
- 14 papers2024 · T1 6
- 10 papers2023 · T1 5
T1 venuesOtherVenues: ICLR, NeurIPS, ICML, AAMAS, ICRA
Research focuspaper titles that use each topic's words
- Reinforcement learning23 papers
- Learning theory & optimization19 papers
- offline16 papers
- data-driven decision-making0 papers
- industry0 papers
Paper list
From the lab homepage.
20265 papers
- ICLR (T1)Dichotomous Diffusion Policy Optimization (opens in a new tab)
- ICLR (T1)Discrete Diffusion for Reflective Vision-Language-Action Models in Autonomous Driving (opens in a new tab)
- ICLR (T1)Sample Efficient Offline RL via T-Symmetry Enforced Latent State-Stitching (opens in a new tab)
- ICLR (T1)X-VLA: Soft-Prompted Transformer as Scalable Cross-Embodiment Vision-Language-Action Model (opens in a new tab)
- AAMAS (T1)xTED: Cross-Domain Adaptation via Diffusion-Based Trajectory Editing (opens in a new tab)
202510 papers
- ICLR (T1)Data Center Cooling System Optimization Using Offline Reinforcement Learning (opens in a new tab)
- The Thirteenth International Conference on Learning Representations (ICLR )Diffusion-Based Planning for Autonomous Driving with Flexible Guidance (opens in a new tab)
- ICML (T1)Efficient Robotic Policy Learning via Latent Space Backward Planning (opens in a new tab)
- NeurIPS (T1)Flow Matching-Based Autonomous Driving Planning with Advanced Interactive Behavior Modeling (opens in a new tab)
- ICRA (T1)H2O+: An Improved Framework for Hybrid Offline-and-Online RL with Dynamics Gaps (opens in a new tab)
- ICLR (T1)Project
- ICRA (T1)Robo-MUTUAL: Robotic Multimodal Task Specification via Unimodal Learning (opens in a new tab)
- NeurIPS (T1)Towards Robust Zero-Shot Reinforcement Learning (opens in a new tab)
- NeurIPS (T1)Uni-RL: Unifying Online and Offline RL via Implicit Value Regularization (opens in a new tab)
- CVPR (T1)Universal Actions for Enhanced Embodied Foundation Models (opens in a new tab)
202414 papers
- IJCAI (T1)A Comprehensive Survey of Cross-Domain Policy Transfer for Embodied Agents (opens in a new tab)
- The Annual AAAI Conference on Artificial Intelligence (AAAI) (oral) (This paper is also accepted in NeurIPS Workshop on Open-World Agents (OWA))Are Expressive Models Truly Necessary for Offline RL? (opens in a new tab)
- RLCBidirectional-Reachable Hierarchical Reinforcement Learning with Mutually Responsive Policies (opens in a new tab)
- Forty- International Conference on Machine Learning (ICML )DecisionNCE: Embodied Multimodal Representations via Implicit Preference Learning (opens in a new tab)
- NeurIPS (T1)Diffusion-DICE: In-Sample Diffusion Guidance for Offline Reinforcement Learning (opens in a new tab)
- In the Thirty-Eighth Conference on Neural Information Processing Systems (NeurIPS )Instruction-Guided Visual Masking (opens in a new tab)
- ICML (T1)Offline-Boosted Actor-Critic: Adaptively Blending Optimal Historical Behaviors in Deep Off-Policy RL (opens in a new tab)
- Forty- International Conference on Machine Learning (ICML )OMPO: A Unified Framework for RL under Policy and Dynamics Shifts (opens in a new tab)
- ICLR (T1)OpenChat: Advancing Open-source Language Models with Mixed-Quality Data (opens in a new tab)
- In International Conference on Learning Representations (ICLR )Query-Policy Misalignment in Preference-Based Reinforcement Learning (opens in a new tab)
- In International Conference on Learning Representations (ICLR )Revealing the Mystery of Distribution Correction Estimation via Orthogonal-gradient Update (opens in a new tab)
- ICLR (T1)Safe Offline Reinforcement Learning with Feasibility-Guided Diffusion Model (opens in a new tab)
202310 papers
- PDF Cite Project ProjectA Fully Data-Driven Approach for Realistic Traffic Signal Control Using Offline Reinforcement Learning (opens in a new tab)
- ICCAn Efficient Multi-Agent Optimization Approach for Coordinated Massive MIMO Beamforming (opens in a new tab)
- PDF Cite ProjectFlexSSL : A Generic and Efficient Framework for Semi-Supervised Learning (opens in a new tab)
- NeurIPS (T1)Look Beneath the Surface: Exploiting Fundamental Symmetry for Sample-Efficient Offline RL (opens in a new tab)
- ICLR (T1)Mind the Gap: Offline Policy Optimization for Imperfect Rewards (opens in a new tab)
- AAMAS (T1)Offline Multi-Agent Reinforcement Learning with Coupled Value Factorization (opens in a new tab)
- NeurIPS (T1)Offline Multi-Agent Reinforcement Learning with Implicit Global-to-Local Value Regularization (opens in a new tab)
- The International Conference on Learning Representations (ICLR )Offline RL with No OOD Actions: In-Sample Learning via Implicit Value Regularization (opens in a new tab)
- PDF Cite ProjectPROTO: Iterative Policy Regularized Offline-to-Online Reinforcement Learning (opens in a new tab)
- ICLR (T1)When Data Geometry Meets Deep Function: Generalizing Offline Reinforcement Learning (opens in a new tab)
Where alumni went14 listed, 14% to industry
Industry 2Academia 12
Most often to
- Baidu
- Huawei
Also to
- Tsinghua University
- Beijing Normal University
- Carnegie Mellon University
Apply to TsinghuaFall 2027
Computer Science PhD (international applicants)
- Round
- Main
- Opens
- GRE
- See program page
- English
- Per school catalog
- Fee
- RMB 800
Not yet confirmed for this intake.
Similar labsnearest by research summary
- SGNTU
Zhiping Lin
- Learning theory & optimization
- USCMU
Andrej Risteski
- Learning theory & optimization
- USUW
Natasha Jaques
- Reinforcement learning
- USUSC
Willie Neiswanger
- Reinforcement learning
- CAToronto
Murat Erdogdu
- Learning theory & optimization
- CNSJTU
Yingtian Zou
- Learning theory & optimization
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
Tsinghuacloses