# Ruiqi Zhang

I am a second-year Ph.D. in the Department of Statistics at University of California, Berkeley, advised by Prof. Peter L. Bartlett.

My research mainly focuses on Theoretical Deep Learning, Reinforcement Learning and their applications in the real world, such as Genomics. Recently, I work in the theory of Transformers, In-Context Learning and Large Language Models.

Previously, I got my bachelor's degree in School of Mathematical Science(SMS) at Peking University(PKU), majoring Mathematics and Statistics. My undergraduate GPA is 3.92/4 and ranks 1/217 in SMS PKU.
In PKU I was advised by Prof. Hao Ge in BICMR, PKU. We focused on statistical methodology of the Pre-Genetic Testing and Gene Regulatory Network Inference.

In Summer 2021, I did my research internship remotely, advised by Prof. Mengdi Wang in ECE at Princeton University . We focused on statistical optimality of a FQE-style Policy Gradient Estimator and Fitted-Q Evaluation with General Function Approximation.

I am married with Jiani Wang in March, 2023. She is now an undergraduate student in Math and CS at Mcgill Unversity and she is now looking for a job in data analysis.

Email / Language: Chinese, English, French / Coding: Python, R, Matlab, LaTeX, Prompting GPT.

Reviewer experience: ICML 2022, NIPS 2023, NIPS 2023 R0-FoMo Workshop, NIPS 2023 Math+A Workshop, ICLR 2024, AISTATS 2024, TMLR.

## News

10/25-10/27/2023, I will give talks on theory of transformers and in-context learning at Theory and Practice of Foundation Models workshop held by Google Research and the Yale Institute for Foundation of Data Science.
My talks will be mostly based on this paper.
08-10/2023, I will give talks on theory of transformers and in-context learning at Google Research India, UIUC ML Seminar, AITIME, Berkeley ML Theory Reading Group and University of Washington.
My talks will be mostly based on this paper.
My collaborator Spencer Frei will also give talks in Stanford University, University of Oxford, Cambridge, Imperial College London, Google Research ICL reading group and DeepMind London.
07/2023, We have a new paper on arxiv about experimental design in offline-to-online RL. This work is jointly with Andrea Zanette.
06/2023, We have a new paper on arxiv about in-context learnability of trained transformers. This work is jointly with Spencer Frei and Peter L. Bartlett.
03/2023, I was married with Jiani Wang. She is an undergraduate in Mcgill University.
08/2022, I joined the Department of Statistics at the University of California, Berkeley to be a Ph.D. student.
06/2022, I graduated from the School of Mathematical Sciences at Peking University.
03/2022, Two papers were accepted by The 39th International Conference on Machine Learning (ICML 2022).
03/2022, I accepted the offer from UC Berkeley and will pursue my Ph.D. in Statistics Department at Berkeley .
03/2022, Two papers were accepted by The Multi-disciplinary Conference on Reinforcement Learning and Decision Making(RLDM) 2022.
07/2021, I started working remotely with Professor Mengdi Wang in Electrical and Computer Engineering at Princeton University.
07/2020, I started working with Professor Hao Ge in Biomedical Pioneering Center at Peking University.
09/2018, I was admitted by School of Mathematical Science at Peking University.
## Research

#### Selected Papers:

**Trained Transformers Learn Linear Model In-Context.**

**Ruiqi Zhang**, Spencer Frei, Peter L. Bartlett

*Submitted* | Paper

**Policy Finetuning in Reinforcement Learning via Design of Experiments using Offline Data.**

**Ruiqi Zhang**, Andrea Zanette

*NIPS 2023* | Paper

**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

**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

#### Undergraduate Thesis:

**Maternal Cell Contamination Correction in Non-invasive Preimplantation Genetic Test for Monogenic Sisease and Aneuploidy Based on Bayesian Model.**

**Ruiqi Zhang**