About me

Hi, I am a Ph.D student in System Engineering at Boston University, supervised by Prof. Francesco Orabona. I am a member of OPTIMAL Lab.

My research interests lie in theoretical machine learning and stochastic optimization. I currently work on understanding and designing optimization methods in machine learning, specifically, stochastic gradient descent and its variants, and adaptive gradient descent methods.

I received my Bachelor degree in Math and Applied Math from University of Science and Technology of China.

Publications

  • On the Last Iterate Convergence of Momentum Methods.
    Xiaoyu Li, Mingrui Liu, Francesco Orabona. ALT 2022. Paper

  • A Second look at Exponential and Cosine Step Sizes: Simplicity, Convergence, and Performance.
    Xiaoyu Li, Zhenxun Zhuang, Francesco Orabona. ICML 2021. Paper Code

  • A High Probability Analysis of Adaptive SGD with Momentum.
    Xiaoyu Li, Francesco Orabona. ICML 2020 Workshop on Beyond First Order Methods in ML Systems. Paper Video

  • On the Convergence of Stochastic Gradient Descent with Adaptive Stepsizes.
    Xiaoyu Li, Francesco Orabona. AISTATS 2019. Paper

Experiences

Work

  • Applied Scientist Intern @ Amazon, Remote. June - Aug. 2021
  • Research Intern @ Nokia Bell Labs, Murray Hill, NJ. June - Aug. 2019

Academic Service

Reviewer of NeurIPS2019 & 2020; AISTATS2020-2022; ICLR 2021, ICML2020 & 2021