Zhaowei Zhang (张钊为)

Ph.D. Student
School of Intelligence Science and Technology
Peking University
Email: zwzhang [at] stu (dot) pku (dot) edu (dot) cn

[Google Scholar] [Github] [Twitter] [LinkedIn]

Research Interests
  • AI Alignment & Governance
  • Socio-Technical Human-AI Interaction
  • Game Theory & Multi-Agent System
  • Reinforcement Learning

Zhaowei is pronounced as "Ju" (as in judge) + "ou" (as in out) + "Way"; Zhang, or Cheung in Hong Kong, is "Ju" (as in judge) + "on" | audio ([International Phonetic Alphabet, IPA]) here: [tʂɑuwei][tʂɑŋ].

I am currently a second-year Ph.D. Student at Institute for AI, School of Intelligence Science and Technology, Peking University. Specifically, I am in the team of PAIR-Lab led by Prof. Yaodong Yang. The long-term goal of my research is to find a way to build a human-AI collaborative system for social efficiency. To this end, my research focuses on AI Alignment, Human-AI Interaction, and Multi-Agent System. In particular, I am currently quite interested in investigating AI safety and their governance (both technical and socio-technical aspects), which includes exploring the societal impact of LLMs, Game Theory in human-AI systems, AI deliberation process, and interactive AI alignment algorithms. I welcome more friends to discuss these topics with me ☺️.

Before entering PKU, I received my B.E. degree of Computer Science and Technology from School of Computer Science at Wuhan University in 2023.

Publications (* indicates equal contribution.)

    2025
  • Amulet: ReAlignment During Test Time for Personalized Preference Adaptation of LLMs
    Zhaowei Zhang, Fengshuo Bai, Qizhi Chen, Chengdong Ma, Mingzhi Wang, Haoran Sun, Zilong Zheng, Yaodong Yang
    ICLR 2025
    [Paper] [Website] [Code]
  • Magnetic Preference Optimization: Achieving Last-iterate Convergence for Language Models Alignment
    Mingzhi Wang, Chengdong Ma, Qizhi Chen, Linjian Meng, Yang Han, Jiancong Xiao, Zhaowei Zhang, Jing Huo, Weijie J. Su, Yaodong Yang
    ICLR 2025
    [Paper]
    • 2024
  • ValueDCG: Measuring Comprehensive Human Value Understanding Ability of Language Models
    Zhaowei Zhang, Fengshuo Bai, Jun Gao, Yaodong Yang
    Preprint
    [Paper] [Blog] [Chinese Blog]
  • Efficient Model-agnostic Alignment via Bayesian Persuasion
    Fengshuo Bai, Mingzhi Wang *, Zhaowei Zhang *, Boyuan Chen *, Yinda Xu, Ying Wen, Yaodong Yang
    Preprint
    [Paper]
  • Foundational Challenges in Assuring Alignment and Safety of Large Language Models
    As a major contributor
    TMLR
    [Paper] [Website]
  • Incentive Compatibility for AI Alignment in Sociotechnical Systems: Positions and Prospects
    Zhaowei Zhang, Fengshuo Bai, Mingzhi Wang, Haoyang Ye, Chengdong Ma, Yaodong Yang
    Preprint
    [Paper] [Chinese Blog]
  • CivRealm: A Learning and Reasoning Odyssey in Civilization for Decision-Making Agents
    Siyuan Qi, Shuo Chen, Yexin Li, Xiangyu Kong, Junqi Wang, Bangcheng Yang, Pring Wong, Yifan Zhong, Xiaoyuan Zhang, Zhaowei Zhang, Nian Liu, Wei Wang, Yaodong Yang, Song-Chun Zhu
    ICLR 2024 (Spotlight)
    [Paper] [Website] [Code]
    • 2023
  • AI Alignment: A Comprehensive Survey
    PAIR-Lab
    [Paper] [Website]
  • ProAgent: Building Proactive Cooperative AI with Large Language Models
    Ceyao Zhang, Kaijie Yang, Siyi Hu, Zihao Wang, Guanghe Li, Yihang Sun, Cheng Zhang, Zhaowei Zhang, Anji Liu, Song-Chun Zhu, Xiaojun Chang, Junge Zhang, Feng Yin, Yitao Liang, Yaodong Yang
    AAAI 2024 (Oral)
    [Paper]
  • Heterogeneous Value Alignment Evaluation for Large Language Models
    Zhaowei Zhang, Nian Liu, Siyuan Qi, Ceyao Zhang, Ziqi Rong, Shuguang Cui, Song-Chun Zhu, Yaodong Yang
    AAAI 2024 Workshop: Public Sector LLMs (Oral)
    [Paper]
  • STAS: Spatial-Temporal Return Decomposition for Solving Sparse Rewards Problems in Multi-agent Reinforcement Learning
    Sirui Chen *, Zhaowei Zhang *, Yali Du, Yaodong Yang
    AAAI 2024
    [Paper] [Code]
    • 2022
  • Contextual Transformer for Offline Meta Reinforcement Learning
    Runji Lin, Ye Li, Xidong Feng, Zhaowei Zhang, Xian Hong Wu Fung, Haifeng Zhang, Jun Wang, Yali Du, Yaodong Yang
    FMDM Workshop at NeurIPS 2022
    [Paper]

Perspectives

  • The Three-Layer Paradigm for Implementing Sociotechnical AI Alignment: A Top-Down-Top Outlook
    Abstract: Backward Alignment is an indispensable part of AI Alignment, and the alignment problems from the perspective of Socio-Technical Systems (STS) are an important component of it. However, what exactly are STS, and what do they refer to? In fact, STS is a very broad concept with many considerations, but currently, there is still little work that clearly unifies all these issues in one go; they are often glossed over in various materials. Additionally, different articles discuss this grand term STS at different scales, or use different terms to define it at the same scale, which also makes it difficult for researchers to understand this field. This article will, from my personal perspective, clearly explain the AI alignment issues present in STS from a computable perspective at different scales, as well as possible research approaches.
    [English Version] [Chinese Version]

Selected Awards

  • [Top 5%] Wuhan University Outstanding Thesis Award. 2023