Wei Xiong

picture 

Ph.D. student
Department of Mathematics
The Hong Kong University of Science and Technology

I am a first-year Ph.D. in mathematics at The Hong Kong University of Science and Technology, advised by Prof. Tong Zhang. I received a B.S. in mathematics from the University of Science and Technology of China in 2021, where I worked closely with Prof. Cong Shen and Prof. Haishan Ye. Here is a photo of me and you can find my CV here.

My work is supported by the Hong Kong Ph.D. Fellowship.

My friend Weiyu Li has not updated her homepage since she moved to Harvard. Please feel free to send an email to push her.

Research Interests

I have broad interests in statistical machine learning, reinforcement learning, multi-armed bandit, and optimization. In particular, I am interested in the statistical properties of existing popular algorithms of machine learning. My goal is to provide a deeper understanding of existing methods and develop new efficient (in sample complexity and in computation resource) learning algorithms with theoretical guarantees.

Research

As an undergraduate, I spent most of my time on bandit, matrix computation, and optimization. After starting my Ph.D. study, I mainly focus on the theory of reinforcement learning.

  • 1. Haishan Ye*, and Wei Xiong*, ’'A practical method for trace estimation of Matrix inverse”, Submitted.

  • 2. Haishan Ye*, Wei Xiong*, and Tong Zhang, ”PMGT-VR: A decentralized proximal-gradient algorithmic framework with variance reduction”, Submitted. [Paper] [Code]

  • 3. Chengshuai Shi, Wei Xiong, Cong Shen, and Jing Yang, “Heterogeneous Multi-player Multi-armed Bandits: Closing the Gap and Generalization ”, Neurips, 2021. [Paper] [Code]

  • 4. Chengshuai Shi, Haifeng Xu, Wei Xiong, and Cong Shen, ”(Almost) Free Incentivized Exploration from Decentralized Learning Agents”, Neurips, 2021. [Paper] [Code]

  • 5. Pushi Zhang, Xiaoyu Chen, Li Zhao, Wei Xiong, Tao Qin, and Tie-Yan Liu, ”Distributional Reinforcement Learning for Multi-Dimensional Reward Functions”, Neurips, 2021. [Paper]

  • 6. Chengshuai Shi, Wei Xiong, Cong Shen, and Jing Yang, “Decentralized multi-player multi-armed bandits with no collision information”, AISTATS, 2020. [Paper] [Code*]

The Code* implementes many SOTA and baseline MPMAB algorithms, which is a nice work of Cindy Trinh.

Misc

I often learn by copying books…

exponential inequality

I enjoy writing code and sharing my knowledge. If you have any questions about the paper or the code, feel free to contact me!

Contact

Email: wei DOT xiong AT connect DOT ust DOT hk