Haotian Gu

Ph.D. in Applied Mathematics | Quantitative Researcher

About Me

Hi! I obtained my Ph.D. in Applied Mathematics from UC Berkeley in May 2023, under the joint supervision of Prof. Xin Guo and Prof. Fraydoun Rezakhanlou. Prior to enrolling at UC Berkeley, I completed my Bachelor's degree in Mathematics at the University of Hong Kong.

My research focuses on the intersection of stochastic control, game theory, and reinforcement learning. Additionally, I am involved in theoretical machine learning, particularly in the areas of robustness, transferability, and federated learning.

In July 2023, I will be joining Citadel Securities as a quantitative researcher.

Google Scholar---LinkedIn---Resume---Email

Publications

Mean-field Control and Reinforcement Learning

Dynamic programming principles for learning Mean-field controls [arXiv]
- Haotian Gu, Xin Guo, Xiaoli Wei, Renyuan Xu.
- Accepted by Operation Research.
Q-Learning for Mean-Field Controls with Convergence and Complexity Analysis [arXiv]
- Haotian Gu, Xin Guo, Xiaoli Wei, Renyuan Xu.
- Accepted by SIAM Journal on Mathematics of Data Science.
- Short Version accepted by Workshop on Theoretical Foundations of Reinforcement Learning, ICML 2020.
Mean-Field Multi-Agent Reinforcement Learning: A Decentralized Network Approach [arXiv]
- Haotian Gu, Xin Guo, Xiaoli Wei, Renyuan Xu.
- Submitted to Mathematics of Operation Research.

Adversarial Robustness

Adversarial Training for Gradient Descent: Analysis Through its Continuous-time Approximation [arXiv]
- Haotian Gu, Xin Guo, Xinyu Li
- Submitted to SIAM Journal on Applied Mathematics.

Transfer Learning

Feasibility and Transferability of Transfer Learning: A Mathematical Framework [arXiv]
- Haoyang Cao, Haotian Gu, Xin Guo, Mathieu Rosenbaum.
- Submitted to NeurIPS 2023.

Federated Learning

FedCore: Straggler-Free Federated Learning with Distributed Coresets
- Hongpeng Guo, Haotian Gu, Xiaoyang Wang, Bo Chen, Eun Kyung Lee, Tamar Eilam, Deming Chen, Klara Nahrstedt.
- Submitted to NeurIPS 2023.
BoFL: Bayesian Optimized Local Training Pace Control for Energy Efficient Federated Learning [pdf]
- Hongpeng Guo, Haotian Gu, Zhe Yang, Xiaoyang Wang, Eun Kung Lee, Nandhini Chandramoorthy, Tamar Eilam, Deming Chen, Klara Nahrstedt.
- Accepted by the 23rd ACM/IFIP International Middleware Conference.

Management Science

Relay Freight Index: Freight Market Rate Prediction with Signature Transform
- Haotian Gu, Xin Guo, Tim Jacobs, Philip Kaminsky, Xinyu Li.
- Available upon request

PDE and Numerical Analysis

An Efficient Model Reduction Method for Solving Viscous G-equations in Incompressible Cellular Flows [arXiv]
- Haotian Gu, Jack Xin, Zhiwen Zhang.
- Accepted by SIAM Journal on Scientific Computing.
Analysis on Hybrid Fractals [arXiv]
- Patricia Alonso-Ruiz, Yuming Chen, Haotian Gu, Robert Strichartz, Ziri Zhou.
- Accepted by Communications on Pure and Applied Analysis

Invited Talks

Berkeley–Columbia Meeting in Engineering and Statistics: April 2023, Columbia University, New York
11th Western Conference on Mathematical Finance: March 2023, University of California, Berkeley
Numerical Analysis Seminar: April 2022, University of Hong Kong, Virtual Conference
Student Probability/PDE Seminar: April 2021, University of California, Berkeley
Workshop on the Intersection of Machine Learning, Control and Games: November 2020, INFORMS Annual Meeting 2020, Virtual Conference
Workshop on Theoretical Foundations of Reinforcement Learning: July 2020, International Conference on Machine Learning 2020, Virtual Conference
Seminar on Applied Mathematics: May 2018, University of Hong Kong, Pok Fu Lam
Cornell Conference on Analysis, Probability, and Mathematical Physics on Fractals: June 2017, Cornell University, Ithaca

Teaching Experience

MATH 170 Mathematical Methods for Optimization: Fall 2020, teaching assistant for Prof. Lawrence Evans
MATH 195 Calculus of Variations and Optimal Control: Spring 2021, teaching assistant for Prof. Lawrence Evans
IEOR 222 Financial Engineering Systems: Spring 2022, teaching assistant for Prof. Thibaut Mastrolia
IEOR 263 Applied Stochastic Process: Spring 2023, teaching assistant for Prof. Xin Guo