CV
======
- PhD student in Department of Computer Science in University of Virginia, Sep 2024 – May 2029
- Major courses: Natural Language Processing, Learning for Interactive Robots, Machine learning, Software logic, etc.
- B.S. in School of Mathematical Science, Peking University, Sep 2020 – Jul 2024
- Major courses: Mathematical Analysis; Probability Theory (H); Applied Stochastic Processes (Honor); Mathematical Statistics, etc.
Research experience
- Sequential Intelligence Lab at University of Virginia (Aug.2024-Now)
- Conducting theoretical research on convergence analysis of reinforcement learning algorithms under supervision of Prof. Shangtong Zhang
- Developing mathematical frameworks for proving $L^2$ and almost sure convergence of reinforcement learning methods
- Analyzing convergence rates of stochastic approximation algorithms with Markovian noise
- Mathematics in Lean (Aug.2023-Jun.2024)
- Led by Prof. Zaiwen Wen, formalized complex mathematical proofs using formal verification system lean4
- Implemented rigorous mathematical theorems in computer-assisted proof systems, focusing on complex analysis
- Autonomous Robots Lab (ARL) (Feb 2023 – May 2023)
- Under supervision of Prof. Chang Liu at Autonomous Robots Lab
- Provided mathematical consultation for robotics research
- Applied optimization theory to improve robotic system performance
- Dynamic System (Jun.2022-Feb.2023)
- Conducted research in dynamical systems theory under guidance of Prof. Zhiqiang Li
- Analyzed mathematical properties of expanding maps and conformal fractals
- Presented theoretical findings on conformal dynamics in research seminars
Intern experience
- Summer 2023: Founder Securities
- Developed theoretical models for financial markets under supervision of Dr. Song Hao
- Applied statistical learning theory to analyze market behavior
- Implemented and validated mathematical models for option pricing
Publications
Finite Sample Analysis of Linear Temporal Difference Learning with Arbitrary Features Zixuan Xie*, Xinyu Liu*, Rohan Chandra, Shangtong Zhang
[ICML 2025]Linear $Q$-Learning Does Not Diverge in $L^2$: Convergence Rates to a Bounded Set Xinyu Liu*, Zixuan Xie*, Shangtong Zhang
Almost Sure Convergence Rates and Concentration of Stochastic Approximation and Reinforcement Learning with Markovian Noise Xiaochi Qian*, Zixuan Xie*, Xinyu Liu*, Shangtong Zhang
Peer Review
- Emergency Technical Reviewer for Reinforcement Learning Conference (RLC) 2025
Skills and Awards
- Technologies: Python(3 years), C&C++(1 year), Latex(3 years), lean4(2 years), Mathematica/MATLAB(2 years)
- Language: Mandarin, English
- No.14 Chinese College Mathematics Competition, 1st class (National Level)
- Second Prize, 35th Chinese Mathematical Olympiad Final Round (2019)
- Gold Medal, Olympiad of Metropolities (2019)