Ruiqi Zhang
Photo of Ruiqi Zhang

Ruiqi Zhang

I recently graduated from the Department of Statistics at the University of California, Berkeley, where I was advised by Prof. Peter L. Bartlett and Prof. Song Mei.

Earlier, I earned my B.S. from the School of Mathematical Sciences (SMS) at Peking University (PKU), and during that period I worked with Prof. Hao Ge and Prof. Mengdi Wang.

My research has focused on statistics, theoretical machine learning, and large language models. I will join Citadel Securities as a Quantitative Researcher starting in April 2026.

Email Languages: Chinese, English Coding: Python, R, Matlab, LaTeX

Selected Publications

  1. Minimax Optimal Convergence of Gradient Descent in Logistic Regression via Large and Adaptive Stepsizes.
    Ruiqi Zhang, Jingfeng Wu, Licong Lin, Peter L. Bartlett.
    JMLR, 2026 | Initial version at ICML, 2025 | Paper
  2. How Do Transformers Perform Two-Hop Reasoning in Context?
    Tianyu Guo*, Hanlin Zhu*, Ruiqi Zhang, Jiantao Jiao, Song Mei, Michael I. Jordan, Stuart Russell.
    2025 | Paper
  3. Fast Best-of-N Decoding via Speculative Rejection.
    Hanshi Sun*, Momin Haider*, Ruiqi Zhang*, Huitao Yang, Ming Yin, Mengdi Wang, Peter L. Bartlett, Andrea Zanette* (* for core authors).
    NeurIPS, 2024 | Paper
  4. Choose Your Anchor Wisely: Effective Unlearning Diffusion Models via Concept Reconditioning.
    Jingyu Zhu*, Ruiqi Zhang*, Licong Lin, Song Mei (* for co-first authors).
    NeurIPS Workshop, 2024 | Paper
  5. Simplicity Prevails: Rethinking Negative Preference Optimization for LLM Unlearning.
    Chongyu Fan*, Jiancheng Liu*, Licong Lin*, Jinghan Jia, Ruiqi Zhang, Song Mei, Sijia Liu (* for co-first authors).
    NeurIPS, 2025 | Paper
  6. Negative Preference Optimization: From Catastrophic Collapse to Effective Unlearning.
    Ruiqi Zhang*, Licong Lin*, Yu Bai, Song Mei (* for co-first authors).
    COLM, 2024 | Paper
  7. In-Context Learning of a Linear Transformer Block: Benefits of the MLP Component and One-Step GD Initialization.
    Ruiqi Zhang, Jingfeng Wu, Peter L. Bartlett.
    NeurIPS, 2024 | Paper
  8. Is Offline Decision Making Possible with Only Few Samples? Reliable Decisions in Data-Starved Bandits via Trust Region Enhancement.
    Ruiqi Zhang, Yuexiang Zhai, Andrea Zanette.
    2024 | Paper
  9. AutoPRM: Automating Procedural Supervision for Multi-Step Reasoning via Controllable Question Decomposition.
    Zhaorun Chen, Zhuokai Zhao, Zhihong Zhu, Ruiqi Zhang, Xiang Li, Bhiksha Raj, Huaxiu Yao.
    NAACL, 2024 | Paper
  10. Trained Transformers Learn Linear Model In-Context.
    Ruiqi Zhang, Spencer Frei, Peter L. Bartlett.
    JMLR, 2024 | Paper (ArXiv) | Paper (JMLR) | Talk
  11. Policy Finetuning in Reinforcement Learning via Design of Experiments using Offline Data.
    Ruiqi Zhang, Andrea Zanette.
    NeurIPS, 2023 | Paper
  12. Off-Policy Fitted Q-Evaluation with Differentiable Function Approximators: Z-Estimation and Inference Theory.
    Ruiqi Zhang, Xuezhou Zhang, Chengzhuo Ni, Mengdi Wang.
    ICML, 2022 | RLDM, 2022 | Paper | Talk
  13. Optimal Estimation of Off-Policy Policy Gradient via Double Fitted Iteration.
    Chengzhuo Ni, Ruiqi Zhang, Xiang Ji, Xuezhou Zhang, Mengdi Wang.
    ICML, 2022 | RLDM, 2022 | Paper

Teaching

  • Fall 2024: STAT 154/254, Modern Statistical Prediction and Machine Learning.
  • Spring 2025: STAT 134, Concepts of Probability.

Reviewing

  • Conferences: ICML, NeurIPS, ICLR, CoLM, AISTATS.
  • Journals: TMLR, DMLR, JMLR.

Honors and Fellowships

  • 2025-2026: Citadel Fellowship at Berkeley.
  • 2022-2024: Berkeley Fellowship.
  • 2021: Huawei Fellowship.
  • 2020: Qin and Jin Fellowship.
  • 2019: Fangzheng Fellowship.
  • 2019-2021: Honor Student in Peking University.