Yaodong Yang 杨耀东 · Boya Young Scholar

Assistant Professor, PKU Chief Scientist, PKU-PsiBot Lab↗
Yaodong Yang YY

Dr. Yaodong Yang is an Assistant Professor (Boya Young Scholar) at the Institute for Artificial Intelligence, Peking University and Chief Scientist of the PKU–PsiBot Joint Laboratory. His research focuses on experience learning and alignment of AI / Embodied agents, aiming to advance the trustworthy deployment and real-world alignment of large models, spanning the areas of reinforcement learning, AI alignment, and embodied intelligence.

He has published over 200 papers in leading journals and conferences, including Nature Machine Intelligence, Cell Matter, Artificial Intelligence Journal, and IEEE TPAMI, with more than 16,000 Google Scholar citations. Since 2022, he has been ranked as the top scholar in AI & ML at Peking University according to CSRankings.

Dr. Yang has received numerous honors, including the ACL 2025 Best Paper Award, UKRI 2026 Best Paper Award in AI, ICCV 2023 Best Paper Finalist, CoRL 2020 Best System Paper Award, and the AAMAS 2021 Blue Sky Idea Award.

He was named to the MIT Technology Review "AI 100 Young Innovators", the 2025 Forbes China Technology & Innovation Innovative Leader list, received the WAIC 2022 "Yunfan Star Award", and the ACM SIGAI China Rising Star Award. His work has been featured by CCTV, People's Daily, Xinhua News, the National Natural Science Foundation of China (NSFC), and MIT Technology Review.

He serves as an Area Chair for major conferences including ICML, ICLR, NeurIPS, AAAI, IJCAI, AAMAS, and IROS, and as an Associate Editor for Scientific Reports, Transactions on Machine Learning Research, and Neural Networks.

Previously, Dr. Yang was an Assistant Professor at King's College London, a Principal Researcher at Huawei Research U.K., and a Senior Manager at AIG. He received his B.Sc. from the University of Science and Technology of China, M.Sc. from Imperial College London, and Ph.D. from University College London, where he was the university's sole nominee for the ACM SIGAI Doctoral Dissertation Award.

｜ CSRanking · #1 PKU AI+ML ｜ Best Paper Award · Five times ｜ Elsevier · World Top 2% Scientist
→ Google ScholarDBLPPAIR-LabContact
200+
Publications
Nature MI · Matter · JMLR · TPAMI
16k+
Citations
Google Scholar · h-index 60
#1
PKU AI+ML Rank since 2022
CSRankings · AIRankings
5+
Best-Paper-Level Awards
ACL · UKRI · CoRL · ICCV · AAMAS
— Industrial Collaborations · partners

News

Headlines · recent updates

★ Headline · Apr 2026

AI breaks the human records in the Kissing Number Problem

PKU mathematicians used AI and reinforcement learning to explore the kissing number problem, achieving breakthroughs in higher dimensions.

Paper →Video →People's Daily →Xinhua News →
★ Headline · Apr 2026

PsiBot releases WAM — world-action model ranking #1 globally on MolmoSpace

Joint work with PKU–PsiBot Lab. A generalist world-action model for embodied agents, outperforming prior SOTA on spatial reasoning benchmarks.

Xinhua →Tech Blog →PsiBot Site →
★ Headline · Jul 2025

Our paper wins the ACL 2025 Best Paper Award — "Language Models Resist Alignment"

The paper shows that post-aligned language models tend to revert to their pre-training distributions — a theoretical "elasticity" result with implications for RLHF and safety.

Paper →Slides →Xinhua →NSFC →
★ Headline · Apr 2025

I delivered a 3-hour tutorial at ICML 2025 (virtual) on Alignment Methods for LLMs.

A comprehensive ICML tutorial covering RLHF, DPO, safe alignment, preference learning and super-alignment — delivered to a virtual audience of thousands.

Tutorial Site →
★ Headline · Dec 2024

We published Matter (Cell Press) paper on applying LLMs for generating carbon nanotubes automatically.

A cross-disciplinary work applying LLMs to steer autonomous experimental synthesis of carbon nanotubes, featured in Cell Press's flagship materials journal Matter.

Paper →Clarivate Highly-cited Paper →
★ Headline · Sep 2024

We published Nature Machine Intelligence paper on large-scale multi-agent networked RL and its applications on pandemics, smart grid and traffic control.

The first multi-agent RL paper led by a Chinese team on a Nature sister journal. Scalable method for controlling 1000+ networked agents with real-world deployments.

Paper →Xinhua →S&T Daily →PKU News →
Browse the full timeline ICML 2026 · NeurIPS 2025 · ACL 2025 Best Paper · ICLR · CoRL 2020 … 35 entries
2026 · 05
8 papers accepted at ICML 2026.
  • Are Your Agents Upward Deceivers?
  • Fault Tolerant Multi-Agent Learning with Adversarial Budget Constraints
  • Debate with Images: Detecting Deceptive Behaviors in Multimodal Large Language Models
  • Vulnerable Agent Identification in Large-Scale Multi-Agent Reinforcement Learning
  • FormalJudge: A Neuro-Symbolic Paradigm for Agentic Oversight
  • VLA-Arena: An Open-Source Framework for Benchmarking Vision-Language-Action Models
  • SafeLab: An Interactive High-Fidelity Benchmark for Embodied Safety in Scientific Robotics
  • UniCode: Augmenting Evaluation for Code Reasoning
2026 · 04
7 papers accepted at ACL 2026.
  • A Game-Theoretic Negotiation Framework for Cross-Cultural Consensus
  • When Slower Isn't Truer: Inverse Scaling Law of Truthfulness in Multimodal Reasoning
  • Uncovering Strategic Egoism Behaviors in Large Language Models
  • SafeMCP: Proactive Power Regulation for LLM Agents Defense via Environment-Grounded Look-Ahead Reasoning
  • Communication-Efficient Desire Alignment for Proactive Embodied Human–Agent Interaction
  • SafeMT: Multi-turn Safety for Multimodal Language Models
  • SIV-Bench: A Video Benchmark for Social Interaction Understanding and Reasoning
2026 · 02
9 papers accepted at AAAI / ICLR / AAMAS / ICRA 2026.
  • AAAI-26 Oral DexGraspVLA: A Vision-Language-Action Framework Towards General Dexterous Grasping
  • AAAI-26 CABTO: Context-Aware Behavior Tree Grounding for Robot Manipulation
  • AAMAS-26 Beyond Self-Interest: Simulating Human-like Social Behavior via Modeling Social Value Orientation in Agent Motivation
  • ICLR-26 MVR: Multi-view Video Reward Shaping for Reinforcement Learning
  • ICLR-26 Align Once, Benefit Multilingually: Enforcing Multilingual Consistency for LLM Safety Alignment
  • ICLR-26 On Robustness of Vision-Language-Action Model against Multi-Modal Perturbations
  • ICLR-26 PoliCon: Evaluating LLMs on Achieving Diverse Political Consensus Objectives
  • ICRA-26 CDF-Glove: A Cable-Driven Force Feedback Glove for Dexterous Teleoperation
  • ICRA-26 DexKnot: Generalizable Visuomotor Policy Learning for Dexterous Bag-Knotting Manipulation
2025 · 09
11 papers accepted at NeurIPS 2025 (3 Spotlights).
  • Empirical Study on Robustness and Resilience in Cooperative Multi-Agent Reinforcement Learning
  • Learning Principles from Multi-modal Human Preference
  • Spotlight (3%) Safe VLA: Towards Safety Alignment of Vision-Language-Action Model via Safe Reinforcement Learning
  • Risk-aware Direct Preference Optimization under Nested Risk Measure
  • Social World Model-Augmented Mechanism Design Policy Learning
  • Spotlight (3%) DexFlyWheel: A Scalable and Self-improving Data Generation Framework for Dexterous Manipulation
  • STAR: Efficient Preference-based Reinforcement Learning via Dual Regularization
  • Safe RLHF-V: Safe Reinforcement Learning from Multi-modal Human Feedback
  • Spotlight DB (3%) InterMT: Multi-Turn Interleaved Preference Alignment with Human Feedback
  • PHYBench: Holistic Evaluation of Physical Perception and Reasoning in Large Language Models
  • Position Paper World Models Should Prioritize the Unification of Physical and Social Dynamics
2025 · 05
6 papers accepted at ACL 2025; 2 papers accepted at ICML 2025.
  • ACL-25 Boosting Policy and Process Reward Models with Monte Carlo Tree Search in Open-Domain QA
  • ACL-25 SafeLawBench: Towards Safe Alignment of Large Language Models
  • ACL-25 Best Paper Language Models Resist Alignment: Evidence From Data Compression
  • ACL-25 Reward Generalization in RLHF: A Topological Perspective
  • ACL-25 Benchmarking Multi-National Value Alignment for Large Language Models
  • ACL-25 BeaverTails v2: Towards Multi-Level Safety Alignment for LLMs with Human Preference
  • ICML-25 Falcon: Fast visuomotor policy via partial denoising
  • ICML-25 SAE-V: Interpreting Multimodal Models for Enhanced Alignment
2025 · 01
5 papers accepted at ICLR 2025.
  • In-Context Editing: Learning Knowledge from Self-Induced Distributions
  • Amulet: ReAlignment During Test Time for Personalized Preference Adaptation of LLMs
  • Mitigating Reward Over-Optimization in RLHF via Behavior-Supported Regularization
  • Emerging Safety Attack and Defense in Federated Instruction Tuning of Large Language Models
  • Magnetic Mirror Descent Self-play Preference Optimization
2024 · 12
5 papers accepted at AAAI 2025; 2 at AAMAS 2025.
  • AAAI-25 Oral Sequence to Sequence Reward Modeling: Improving RLHF by Language Feedback
  • AAAI-25 Stream Aligner: Efficient Sentence-Level Alignment via Distribution Induction
  • AAAI-25 Oral Differentiable Information Enhanced Model-Based Reinforcement Learning
  • AAAI-25 Towards Efficient Collaboration via Graph Modeling in Reinforcement Learning
  • AAAI-25 Oral RAT: Adversarial Attacks on Deep Reinforcement Agents for Targeted Behaviors
  • AAMAS-25 Mean Field Correlated Imitation Learning
  • AAMAS-25 Short Paper EconTwo: A Two-Level Multi-Agent Framework for Dynamic Macroeconomic Modeling with Shock Resilience
2024 · 10

Invited talk "Can LLMs be aligned?" at CNCC 2024.

Slides
2024 · 09
5 papers accepted at NeurIPS 2024.
  • Oral (top 0.5%) Aligner: Efficient Alignment by Learning to Correct
  • Spotlight ProgressGym: Alignment with a Millennium of Moral Progress
  • Panacea: Pareto Alignment via Preference Adaptation for LLMs
  • Scalable Constrained Policy Optimization for Safe Multi-agent Reinforcement Learning
  • SafeSora: Towards Safety Alignment of Text2Video Generation via a Human Preference Dataset
2024 · 08
2 papers accepted at CoRL 2024.
  • Neural Attention Field: Emerging Point Relevance in 3D Scenes for One-Shot Dexterous Grasping
  • Object-Centric Dexterous Manipulation from Human Motion Data
2024 · 05
Delivered the VALSE 2024 annual progress report on alignment; 3 papers accepted at ICML 2024.
ReportVideo (CN)
  • End-to-End Neuro-Symbolic Reinforcement Learning with Textual Explanations
  • Safe Reinforcement Learning using Finite-Horizon Gradient-based Estimation
  • Planning with Theory of Mind for Few-Shot Adaptation in Mixed-motive Environments
2024 · 03

Co-signed the Beijing AI Safety Declaration with leading scientists.

FT36Kr
2024 · 02

Featured on CCTV「焦点访谈」 — national TV report on AI Safety.

CCTV
2024 · 01
5 papers accepted at ICLR 2024; 1 at TPAMI.
  • ICLR-24 Spotlight (5%) CivRealm: A Learning and Reasoning Odyssey for Decision-Making Agents
  • ICLR-24 Spotlight (5%) Maximum Entropy Heterogeneous-Agent Reinforcement Learning
  • ICLR-24 Spotlight (5%) Safe RLHF: Safe Reinforcement Learning from Human Feedback
  • ICLR-24 SafeDreamer: Safe Reinforcement Learning with World Models
  • ICLR-24 Byzantine Robust Cooperative Multi-Agent Reinforcement Learning as a Bayesian Game
  • TPAMI ASP: Learn a Universal Neural Solver
2023 · 12
3 papers accepted at AAAI 2024; 2 papers accepted at top journals (TPAMI, JMLR).
  • AAAI-24 STAS: Spatial-Temporal Return Decomposition for Multi-agent Reinforcement Learning
  • AAAI-24 Oral (7%) ProAgent: Building Proactive Cooperative AI with Large Language Models
  • AAAI-24 A Perspective of Q-value Estimation on Offline-to-Online Reinforcement Learning
  • TPAMI Bi-DexHands: Towards Human-Level Bimanual Dexterous Manipulation
  • JMLR Heterogeneous-Agent Reinforcement Learning
2023 · 11

Released the AI Alignment Survey.

2023 · 10
Paper on the ICCV 2023 Best Paper Initial List (top 17 / 8260).
  • UniDexGrasp++: Improving Dexterous Grasping Policy Learning via Geometry-aware Curriculum and Iterative Generalist-Specialist Learning
2023 · 09
6 papers accepted at NeurIPS 2023; 2 at JMLR and TMLR.
  • NeurIPS-23 Multi-Agent First Order Constrained Optimization in Policy Space
  • NeurIPS-23 Hierarchical Multi-Agent Skill Discovery
  • NeurIPS-23 Policy Space Diversity for Non-Transitive Games
  • NeurIPS-23 Team-PSRO for Learning Approximate TMECor in Large Team Games via Cooperative Reinforcement Learning
  • NeurIPS-23 BeaverTails: A Human-Preference Dataset for LLM Harmlessness Alignment
  • NeurIPS-23 Safety Gymnasium: A Unified Safe Reinforcement Learning Benchmark
  • JMLR MARLlib: A Scalable and Efficient Multi-agent Reinforcement Learning Library
  • TMLR JiangJun: Mastering Xiangqi by Tackling Non-Transitivity in Two-Player Zero-Sum Games
2023 · 06

TorchOpt officially joined the PyTorch Ecosystem.

2023 · 05
4 papers accepted at ICML 2023.
  • Regret-Minimizing Double Oracle for Extensive-Form Games
  • MANSA: Learning Fast and Slow in Multi-Agent Systems
  • A Game-Theoretic Framework for Managing Risk in Multi-Agent Systems
  • GEAR: A GPU-Centric Experience Replay System for Large Reinforcement Learning Models
2023 · 02
2 papers accepted at ICRA 2023; 1 at ICLR 2023.
  • ICRA-23 End-to-End Affordance Learning for Robotic Manipulation
  • ICRA-23 GenDexGrasp: Generalizable Dexterous Grasping
  • ICLR-23 Quality-Similar Diversity via Population Based Reinforcement Learning
2023 · 01
1 paper accepted at JAAMAS and 1 at AAMAS 2023.
  • JAAMAS Online Markov Decision Processes with Non-oblivious Strategic Adversary
  • AAMAS-23 Is Nash Equilibrium Approximator Learnable?
2022 · 12

NeurIPS 2022 MyoChallenge — 1st place (1 / 340 teams).

2022 · 11
National Science Review paper on Markov-game complexity; 3 papers accepted at AAAI 2023.
  • NSR On the Complexity of Computing Markov Perfect Equilibrium in General-Sum Stochastic Games
  • AAAI-23 Learning to Shape Rewards using a Game of Two Partners
  • AAAI-23 Subspace-Aware Exploration for Sparse-Reward Multi-Agent Tasks
  • AAAI-23 Cooperative Multi-agent Q-learning with Bidirectional Action-Dependency
2022 · 09
7 papers accepted at NeurIPS 2022.
  • Meta-Reward-Net: Implicitly Differentiable Reward Learning for Preference-based Reinforcement Learning
  • A Theoretical Understanding of Gradient Bias in Meta-Reinforcement Learning
  • Constrained Update Projection Approach to Safe Policy Optimization
  • Multi-Agent Reinforcement Learning is a Sequence Modeling Problem
  • A Unified Diversity Measure for Multiagent Reinforcement Learning
  • Towards Human-Level Bimanual Dexterous Manipulation with Reinforcement Learning
  • MATE: Benchmarking Multi-Agent Reinforcement Learning in Distributed Target Coverage Control
2022 · 05
1 paper accepted at IJCAI 2022.
  • On the Convergence of Fictitious Play: A Decomposition Approach
2022 · 04

TorchOpt and Bi-DexHands open-sourced.

2022 · 01
2 papers accepted at ICLR 2022.
  • Trust Region Policy Optimisation in Multi-Agent Reinforcement Learning
  • LIGS: Learnable Intrinsic-Reward Generation Selection for Multi-Agent Learning
2021 · 09
3 papers accepted at NeurIPS 2021.
  • Settling the Variance of Multi-Agent Policy Gradients
  • Unifying Behavioral and Response Diversity for Open-ended Learning in Zero-sum Games
  • Neural Auto-Curricula
2021 · 05
2 papers accepted at ICML 2021.
  • Long Talk (top 3%) Modelling Behavioural Diversity for Learning in Open-Ended Games
  • Learning in Nonzero-Sum Stochastic Games with Potentials
2021 · 02
AAMAS 2021 Blue-Sky Idea Best Paper Award.
  • Best Paper Diverse Auto-Curriculum is Critical for Successful Real-World Multiagent Learning Systems
2020 · 10
SMARTS platform released; CoRL 2020 Best System Paper Award.
  • Best Paper SMARTS: An Open-Source Scalable Multi-Agent RL Training School for Autonomous Driving
2020 · 06
1 paper accepted at ICML 2020.
  • Multi-agent Determinantal Q-learning
2020 · 05
1 paper accepted at IJCAI 2020.
  • Modelling Bounded Rationality in Multi-Agent Interactions by Generalized Recursive Reasoning
2020 · 02
1 paper accepted at AAMAS 2020.
  • α^α-Rank: Practically Scaling Alpha-Rank through Stochastic Optimisation

Research

Five directions · methods, benchmarks, and open-source systems

01 / RL for Alignment

LLM Alignment & RLHF

RLHF, preference learning, safe alignment, red-teaming and interpretability. Principled methods and open benchmarks — BeaverTails, PKU-SafeRLHF, Stream Aligner, Libra-Leaderboard — to make LLMs robustly helpful and harmless.

Featured: Language Models Resist Alignment (ACL 2025 Best Paper)· Aligner (NeurIPS Oral) · Safe RLHF (NeurIPS Spotlight)
Systems BeaverTailsSafeRLHFLibra
44 papers Browse all →
02 / RL for Embodied AI

Embodied Reinforcement Learning

Dexterous manipulation, vision-language-action models, and sim-to-real. From Bi-DexHands and ClutterDexGrasp to DexGraspVLA and Safe VLA — pursuing human-level generalist robotic agents.

Featured: DexGraspVLA (AAAI 2026 Oral) · Safe VLA (NeurIPS 2025 Spotlight)
Systems Bi-DexHandsDexFlyWheelSMARTS
27 papers Browse all →
03 / Multi-agent RL

Multi-Agent RL

Cooperative and competitive MARL, policy gradient theory, Nash equilibria. HARL, MAT, MARLlib, MALib — algorithms that scale to hundreds of agents.

Systems MARLlib(JMLR)MALib(JMLR)OmniSafe(JMLR)TorchOpt(JMLR)
107 papers Browse →
04 / Agentic RL

Agentic RL & Social Simulation

LLM-based agents for macroeconomic modelling, social value orientation, negotiation and consensus. World models unifying physical and social dynamics.

Systems ProAgentProgressGymSIV-Bench
17 papers Browse →
05 / RL for Science

RL for Science

RL and LLMs applied to mathematics, medicine, physics, materials (carbon-nanotube synthesis), and operations — featured in Cell iScience, Matter, and National Science Review.

Systems PackingStarCarbon CopilotPHYBenchASP-Solver
10 papers Browse →

Press

National coverage · CCTV · Xinhua · NSFC · MIT Tech Review

CCTV · Xinhua News · People's Daily · MIT Tech Review
AI Spring Festival Gala
AI Spring Festival Gala
CCTV · YangShiPin · 2026
▶
National TV · CCTV
AI Spring Festival Gala
Zhongguancun Dialogue · AI Spring Festival Gala · CCTV
Featured dialogue on AI safety — face-swap liability, ensuring robots' brains don't err.
2026 · SpringWatch →
2025 China · AI Gala
AI Gala
CCTV-1 · 2025
▶
National TV · CCTV
2025 China · AI Gala
"2025 China · AI Gala" · CCTV-1
Spoke in the Young Innovators segment alongside Jun Zhu, Xipeng Qiu and other AI leaders.
2025 · 08Watch →
Focus Interview
Focus Interview
CCTV · Focus Interview
▶
National TV · CCTV
Focus Interview
Focus Interview · CCTV News Commentary
Nationally televised feature on AI Safety & Ethics and trustworthy AI development.
2024 · 02Watch →

Awards

Best papers · talent programs · academic honors · competitions

I. Best-Paper Awards 5 awards
2026

UKRI Best Research Paper in AI

Efficient and Scalable Reinforcement Learning for Large-Scale Network Control · Nature Machine Intelligence

2025

ACL 2025 Best Paper Award

Language Models Resist Alignment: Evidence From Data Compression

2023

ICCV 2023 Best Paper Finalist

UniDexGrasp: Universal Robotic Dexterous Grasping via Learning Diverse Proposal Generation and Goal-Conditioned Policy

2021

AAMAS 2021 Blue-Sky Idea Award

Diverse Auto-Curriculum is Critical for Successful Real-World Multiagent Learning Systems

2020

CoRL 2020 Best System Paper Award

SMARTS: An Open-Source Scalable Multi-Agent RL Training School for Autonomous Driving

II. Talent Programs 3 programs
2024

National Young Talent

NSFC Excellent Young Scientist

2022

High-Level Overseas Talent

Ministry of Human Resources — 30 nationwide

2023

CAST Youth Talent Support Program

CAAI — 6 selected nationally

III. Academic Honors 5 honors
2025

Elsevier / Stanford World Top 2% Scientists

Global Top 2% career-impact ranking

2025

MIT Tech Review — AI 100 Young Innovators

MIT Technology Review · "AI 100 Young Innovators"

2026

Forbes China — Innovation Leader

Forbes China · Innovation & Tech Leaders

2022

ACM SIGAI China Rising Star Award

ACM SIGAI China · 3 awardees nationwide

2022

WAIC Yunfan Award — Rising Star

WAIC · 10 awardees nationwide

IV. Competitions & Industry 4 awards
2025

Wu Wenjun AI S&T Award · 2nd Prize

Wu Wenjun AI S&T Award · 2nd Prize — Knowledge-Enhanced Trustworthy Multimodal Interaction

2025

CMSA Meteorological Tech Invention Award · 1st Prize

CMSA · 1st Prize for Technological Invention — BeiDou + AI for Extreme-Wind Emergency Navigation

2022

NeurIPS 2022 MyoChallenge · Winner

Physiological dexterity manipulation · 1 / 340 teams

2025

Digital China Innovation Contest · AI Track 1st Prize

Digital China Innovation Contest · AI Track · National 1st Prize

Mentorship

Highest PKU student honors · Apple & Tencent fellowships · NSFC grants

2024 Highest Student Honor · PKU

PKU May-4th Medal

耿 Yiran Geng 耿逸然 （2024） 陈 Boyuan Chen 陈博远 （2026）
PKU's highest honor for students (once every two years) · sole recipient among all undergraduate STEM majors.
2024 University-Wide · PKU

PKU Annual Figures

吉 Jiaming Ji 吉嘉铭 （2025） 陈 Boyuan Chen 陈博远 （2025）
Two PAIR-Lab students named PKU Annual Figures — one of the most prestigious annual recognitions at Peking University.
2025 Industry Fellowship · Apple

Apple Scholars in AI / ML

吉 Jiaming Ji 吉嘉铭
Apple PhD Fellowship (2025) — one of only 12 scholars selected globally.
2025 Industry Fellowship · Tencent

Tencent Hunyuan Scholar

吉 Jiaming Ji 吉嘉铭
Tencent's flagship PhD fellowship for top AI students in China.
2024 NSFC · PhD Student Grant

NSFC Young Student
Basic Research (PhD)

吉 Jiaming Ji 吉嘉铭
Sole PhD awardee in PKU's AI direction — NSFC Young Student Basic Research Program (PhD).
2024 NSFC · Undergraduate Grant

NSFC Young Student
Basic Research (UG)

邱 Tianyi Qiu 邱天异
One of only two undergraduates in PKU's AI direction to receive this grant.
Teaching Awards
2026

PKU Teaching Achievement Award · 2nd Prize

For the course "Foundations and Alignment of Large Language Models" (《大语言模型基础与对齐》).

2025

Digital China Innovation Contest · AI Track 1st Prize

2025 Digital China Innovation Competition — AI Track, First Prize National.

2025

ICBC Teaching Award · PKU

ICBC Teaching Award · PKU · 2025

2022–

Class Advisor · Yuanpei AGI Experimental Class

Yuanpei College · Class Advisor & Curriculum Committee · AGI Experimental Class (2022 cohort)

2023 – 2025

Outstanding Undergraduate Research Supervisor · PKU

Awarded three years in a row (2023, 2024, 2025) by Peking University.

Publications

Representative works · browse by topic below

Reinforcement LearningMulti-Agent RLLLMAlignmentDexterous ManipulationRoboticsSafetyPreference LearningSafe RLMultimodalCooperationBenchmarkSurveyGraspingZero-Sum GamesNash EquilibriumSocial SimulationVLAAdversarial RobustnessPhysicsRLHFCooperative MARLMedical AIDatasetGame TheoryMulti-TurnBayesian MethodsPreference DatasetDistributed SystemsQ-LearningTheoryReward ShapingWorld ModelsValue AlignmentAdversarial AttackSafe RLHFAlignerPSROMean Field RLBimanualHARLTrust RegionDouble OracleQuantitative FinanceNon-TransitivityConsensusEmbodied AIAgent EvalTactileVisuomotor
★ Featured 16 All 206 Preprints 25 Alignment RLHF 55 Embodied RL 37 Multi-Agent RL 101 Agentic RL 21 RL4Sci 15
2025 2 papers
ALN
Language Models Resist Alignment: Evidence From Data Compression *
Jiaming Ji, Kaile Wang, Tianyi Alex Qiu, Boyuan Chen, Jiayi Zhou, Changye Li, Hantao Lou, Josef Dai, Yunhuai Liu, Yaodong Yang#
ACL 2025 ★ Best Paper
Alignment TheoryAlignmentLLM
Media {'en': 'Xinhua', 'zh': '新华社'}↗{'en': 'NSFC', 'zh': '国家自然科学基金委'}↗{'en': 'Jiqizhixin', 'zh': '机器之心'}↗
arXiv S2
ALN
Safe VLA: Towards Safety Alignment of Vision-Language-Action Model via Safe Reinforcement Learning *
Borong Zhang, Yuhao Zhang, Jiaming Ji, Yingshan Lei, Josef Dai, Yuanpei Chen, Yaodong Yang#
NeurIPS 2025 Spotlight
Safe VLAVLASafe RLSafetyAlignment
Search
2024 7 papers
EMB
ASP: Learn a Universal Neural Solver *
Chenguang Wang, Zhouliang Yu, Stephen McAleer, Tianshu Yu, Yaodong Yang#
IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI)
Combinatorial Optimization
arXiv PDF S2
ALN
Aligner: Efficient Alignment by Learning to Correct *
Jiaming Ji, Boyuan Chen, Hantao Lou, Donghai Hong, Borong Zhang, Xuehai Pan, Juntao Dai, Yaodong Yang#
NeurIPS 2024 Oral
AlignerAlignment
Search
EMB
Bi-DexHands: Towards Human-Level Bimanual Dexterous Manipulation *
Yuanpei Chen, Yiran Geng, Fangwei Zhong, Jiaming Ji, Jiechuang Jiang, Zongqing Lu, Hao Dong, Yaodong Yang#
IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI)
BimanualDexterous ManipulationRobotics
Media {'en': 'China Youth Daily', 'zh': '中国青年报'}↗
Search
AI4
Efficient and scalable reinforcement learning for large-scale network control *
Chengdong Ma, Aming Li, Yali Du, Hao Dong, Yaodong Yang#
Nature Machine Intelligence ★ Best Paper
Network ControlReinforcement Learning
Media {'en': 'Xinhua', 'zh': '新华社'}↗{'en': 'STDaily', 'zh': '科技日报'}↗{'en': 'PKU News', 'zh': '北大新闻网'}↗
PDF S2
MRL
Heterogeneous-Agent Reinforcement Learning *
Yifan Zhong, Jakub Grudzien Kuba, Xidong Feng, Siyi Hu, Jiaming Ji, Yaodong Yang#
Journal of Machine Learning Research (JMLR)
HARLReinforcement Learning
arXiv PDF S2
ALN
Omnisafe: An infrastructure for accelerating safe reinforcement learning research *
Jiaming Ji, Jiayi Zhou, Borong Zhang, Juntao Dai, Xuehai Pan, Ruiyang Sun, Weidong Huang, Yiran Geng, Mickel Liu, Yaodong Yang#
Journal of Machine Learning Research (JMLR)
OmniSafeSafe RLReinforcement Learning
Search
AI4
Transforming the synthesis of carbon nanotubes with machine learning models and automation *
Yue Li, Shurui Wang, Zhou Lv, Zhaoji Wang, Yunbiao Zhao, Ying Xie, Yang Xu, Liu Qian, Yaodong Yang#, Ziqiang Zhao#, Jin Zhang#
Matter (Cell Press)
Carbon NanotubesMaterials Synthesis
Media {'en': 'Xinhua', 'zh': '新华社'}↗
Search
2023 4 papers
MRL
MARLlib: A Multi-agent Reinforcement Learning Library *
Siyi Hu, Yifan Zhong, Minquan Gao, Weixun Wang, Hao Dong, Xiaodan Liang, Zhihui Li, Xiaojun Chang, Yaodong Yang#
Journal of Machine Learning Research (JMLR)
MARLlibMulti-Agent RLReinforcement Learning
Search
MRL
On the complexity of computing markov perfect equilibrium in general-sum stochastic games *
Xiaotie Deng, Ningyuan Li, David Mguni, Jun Wang, Yaodong Yang#
National Science Review
Nash EquilibriumStochastic Games
arXiv PDF S2
ALN
Safe multi-agent reinforcement learning for multi-robot control *
Shangding Gu, Jakub Grudzien Kuba, Yuanpei Chen, Yali Du, Long Yang, Alois C. Knoll, Yaodong Yang#
Artificial Intelligence Journal (AIJ)
Multi-Agent RLRoboticsReinforcement Learning
PDF S2
MRL
TorchOpt: An Efficient Library for Differentiable Optimization *
Jie Ren, Xidong Feng, Bo Liu, Xuehai Pan, Yao Fu, Luo Mai, Yaodong Yang#
Journal of Machine Learning Research (JMLR)
Differentiable Optimization
arXiv PDF S2
2021 1 paper
MRL
Diverse Auto-Curriculum is Critical for Successful Real-World Multiagent Learning Systems *
Yaodong Yang, Jun Luo, Ying Wen, Oliver Slumbers, Daniel Graves, Haitham Bou Ammar, Jun Wang, Matthew E. Taylor
AAMAS 2021 ★ Best Paper
Auto-CurriculumMulti-Agent RL
arXiv PDF
2020 1 paper
EMB
SMARTS: An Open-Source Scalable Multi-Agent RL Training School for Autonomous Driving
Ming Zhou, Jun Luo, Julian Villella, Yaodong Yang, David Rusu, Jiayu Miao, Weinan Zhang, Montgomery Alban, Iman Fadakar, Zheng Chen, Aurora Chongxi Huang, Ying Wen, Kimia Hassanzadeh, Daniel Graves, Dong Chen, Zhengbang Zhu, Nhat Nguyen, Mohamed Elsayed, Kun Shao, Sanjeevan Ahilan, Baokuan Zhang, Jiannan Wu, Zhengang Fu, Kasra Rezaee, Peyman Yadmellat, Mohsen Rohani, Nicolas Perez Nieves, Yihan Ni, Seyedershad Banijamali, Alexander Cowen Rivers, Zheng Tian, Daniel Palenicek, Haitham bou Ammar, Hongbo Zhang, Wulong Liu, Jianye Hao, Jun Wang
CoRL 2020 ★ Best Paper
SMARTSAutonomous DrivingMulti-Agent RL
arXiv PDF
2018 1 paper
MRL
Mean Field Multi-Agent Reinforcement Learning
Yaodong Yang, Rui Luo, Minne Li, Ming Zhou, Weinan Zhang, Jun Wang
ICML 2018
Mean Field RLMulti-Agent RLReinforcement Learning
arXiv PDF
View all 206 publications → Google Scholar ↗

Service

Area Chair · Associate Editor · Program Chair

Area Chair
  • NeurIPS CCF-A
  • ICML CCF-A
  • ICLR CCF-A
  • AAAI CCF-A
  • IJCAI CCF-A
  • AAMAS — Senior AC CCF-B
  • IROS CCF-C
Associate Editor
  • Neural Networks (Springer) CCF-B
  • Transactions on Machine Learning Research TMLR
  • Scientific Reports Nature
Program / Publicity Chair
  • World Artificial Intelligence Conference Academic (WAICA) 2026 · Shanghai Publicity Chair
  • Distributed AI Conference (DAI) 2024 · Singapore Program Chair

Experience

USTC · Imperial · UCL · AIG · KCL · PKU

2022 – Now
Assistant Professor (Boya Young Scholar)
Peking University · Institute for AI 北京大学人工智能研究院
Chief Scientist, PKU–PsiBot Joint Laboratory · PI, PAIR-Lab
2021 – 2022
Assistant Professor
King's College London · Department of Informatics 伦敦国王大学
2019 – 2021
Principal Researcher
Huawei U.K. · London Research Centre 华为英国研究院
2020 Best Technology Breakthrough Award (sole awardee)
2015 – 2019
Senior Science Manager
American International Group (AIG) · Science Dept. 美国国际集团
2016 – 2021
Ph.D. · Computer Science
University College London (UCL) 伦敦大学学院
Thesis: Many-Agent Reinforcement Learning · Advisors: Jun Wang & John Shawe-Taylor
2013 – 2014
M.Sc. · Quantitative Biology
Imperial College London 伦敦帝国理工学院
2009 – 2013
B.Eng. · Electronic Engineering & Information Science
University of Science & Technology of China (USTC) 中国科学技术大学
§ Join the Lab

Come work on the hardest problems in safe and trustworthy AGI.

PhD · 2027 PhD admissions (2027 cycle)
0
Peking University
FULL · no spots this cycle
Multiple
Zhongguancun Academy
Open · accepting applications
Three research directions

LLM Post-Training · Alignment

RLHF / DPO / Safe-RLHF · reward modeling · interpretability · multi-modal & multilingual safety. Connecting alignment theory to practice at scale.

Starter materials

Embodied Intelligence · Dexterous Manipulation · Robot Foundation Models

Sim-to-real policy learning for high-DoF dexterous manipulation; embodied foundation models that act in the physical world. Joint work with PsiBot.

Starter materials

World Models · Physics Foundation Models · Sim-to-Real Alignment

Build world models that capture both physical and social dynamics; align simulators with the real world for downstream policy training. Joint work with Neo Matrix.

Starter materials

PAIR-Lab also welcomes master's students, visiting scholars, undergraduate research interns, and postdocs. If you are fascinated by reinforcement learning, LLM alignment, multi-agent systems, or embodied intelligence — and want to build safe and trustworthy AGI that ships — please read the starter materials above and reach out.

→ Send application Wechat: