About

I am an MIT doctoral student at the MIT Operations Research Center. My advisor is Robert M. Freund. Previously, I obtained my B.S. in Mathematics and Applied Mathematics from Fudan University in 2020. I was a research assistant at RIIS where I worked with Yinyu Ye and Dongdong Ge.

My research focuses on optimization, and its applications in operations research and machine learning. My recent research interests include

• Huge-scale linear programming

• First-order methods for optimization, and

• applications in data science and statistical machine learning.

---

Research

Publications and working papers in optimization

• Improving the Geometry of Primal-Dual Level Sets in the Restarted PDHG for Linear and Conic Optimization Problems

  • Zikai Xiong and Robert M. Freund

  • In preparation

• On the Relation Between LP Sharpness and Limiting Error Ratio and Complexity Implications for Restarted PDHG

  • Zikai Xiong and Robert M. Freund

  • Technical report paper

• Computational Guarantees for Restarted PDHG for LP based on "Limiting Error Ratios" and LP Sharpness

  • Zikai Xiong and Robert M. Freund

  • Under review paper

• Fair Coresets via Optimal Transport

  • Zikai Xiong, Niccolò Dalmasso, Shubham Sharma, Freddy Lecue, Daniele Magazzeni, Vamsi K. Potluru, Tucker Balch, Manuela Veloso

  • Under review paper

• FairWASP: Fast and Optimal Fair Wasserstein Pre-processing

  • Zikai Xiong, Niccolo Dalmasso, Alan Mishler, Vamsi Potluru, Tucker Balch, Manuela Veloso

  • AAAI Conference on Artificial Intelligence (AAAI 2024) paper

• Using Taylor-Approximated Gradients to Improve the Frank-Wolfe Method for Empirical Risk Minimization

  • Zikai Xiong and Robert M. Freund paper slides

  • Accepted by SIAM Journal on Optimization

  • Additional results and extensions technical report

• From an Interior Point to a Corner Point: Smart Crossover

  • (Alphabetical order) Dongdong Ge, Chengwenjian Wang, Zikai Xiong, and Yinyu Ye paper slides

  • Under revision

• Interior-Point Methods Strike Back: Solving the Wasserstein Barycenter Problem

  • (Alphabetical order) Dongdong Ge, Haoyue Wang, Zikai Xiong, and Yinyu Ye paper codes slides

  • Neural Information Processing Systems (NeurIPS 2020)

• Low-rank Traffic Matrix Completion with Marginal Information

  • Zikai Xiong, Renjie Xu, Yanwei Xu, and Yimin Wei paper

  • Journal of Computational and Applied Mathematics

Other

• Learning from Multiple Annotator Noisy Labels via Sample-wise Label Fusion

  • Zhengqi Gao, Fan-Keng Sun, Mingran Yang, Sucheng Ren, Zikai Xiong, et al. paper codes

  • European Conference on Computer Vision (ECCV 2022)

---

Experience

• Research Intern, Summer 2023

  • Optimization in fairness and machine learning.

  • J.P.Morgan Chase AI Research

• Research Assistant, 2020 - present

  • Supervisor: Robert M. Freund

  • Huge-scale LP and first-order methods for optimization

  • MIT Operations Research Center

• Research Assistant, 2018 - 2020

  • Supervisors: Yinyu Ye (Stanford University), Dongdong Ge

  • SUFE Research Institute for Interdisciplinary Sciences

---

Teaching

Teaching assistant in MIT

• Fall 2023: 15.081 Introduction to Mathematical Programming

• Fall 2022: 15.081 Introduction to Mathematical Programming

• Summer 2022: 15.077 Statistical Machine Learning and Data Science

• Spring 2022: 15.071 The Analytics Edge

Teaching assistant in SUFE

• Summer 2019: Stochastic Modeling (graduate course); From Machine Learning to Decision-making: Bandit Learning and Reinforcement Learning (graduate course); Stochastic Process and Financial Risk Analysis (graduate course)

---

Professional Activities

Reviewer

• Journal: SIAM Journal on Optimization (SIOPT)

• Conference: ICML (2021,2022), NeurIPS (2022)

---

Talks

• Improving the Geometry of (Conic) Linear Optimization Problems for the Primal-Dual Hybrid Gradient Method (PDHG)

  • Workshop on Modern Continuous Optimization (in honor of Robert M. Freund's 70th Birthday), Cambridge, MA, Aug. 2023

  • INFORMS Annual Meeting, Phoenix, AZ, Oct. 2023

  • Shanghai University of Finance and Economics, Shanghai, China, Sep. 2023

  • INFORMS Optimization Society Conference, Houston, TX, Mar. 2024

• Geometric Condition Measures in the Primal-Dual Hybrid Gradient Method for Linear Programming

  • SIAMOP 2023, Seattle, WA, Jun. 2023

  • ORC Student Seminar, MIT, Cambridge, MA, Mar. 2023

  • Shanghai University of Finance and Economics, Shanghai, China, Jan. 2023

• Using Taylor-Approximated Gradients to Improve the Frank-Wolfe Method for Empirical Risk Minimization slides

  • ICCOPT 2022, Lehigh University, Bethlehem, PA, Jul. 2022

  • ORC Student Seminar, MIT, Cambridge, MA, Oct. 2022

  • INFORMS Annual Meeting, Indianapolis, IN, Oct. 2022

• From an Interior Point to A Corner Point: Smart Crossover slides

  • INFORMS Annual Meeting, Indianapolis, IN, Oct. 2022

• Interior-Point Methods Strike Back: Solving the Wasserstein Barycenter Problem slides

  • INFORMS Annual Meeting, Seattle, WA, Oct. 2019

• Computing Wasserstein Barycenter Efficiently: A Structured Linear System and Customized Algorithms

  • Seminar of New Advances in Theory and Application of ADMM, Shanghai, China, Mar. 2019

---