Yubo Cai 蔡宇博

Yubo Cai 蔡宇博

PhD student at MIT LIDS, IDSS, and CEE

Massachusetts Institute of Technology

Academic Profile

Hi there, welcome to my personal website.

I am a Ph.D. student in Computational Science and System Engineering at Massachusetts Institute of Technology (MIT), affiliated with the Zardini Lab in the Laboratory for Information and Decision Systems (LIDS). I am fortunate to be advised by Prof. Gioele Zardini. Before joining MIT, I earned my Bachelor of Science, in Mathematics and Computer Science from École Polytechnique, and later completed my Master of Science, in Mathematics and Statistics (OMMS/Part C) at the University of Oxford.

Throughout my academic journey, I have been fortunate to receive generous guidance from many mentors, whose support has shaped both my scholarly values and my broader outlook on life. I am truly honored to have completed my bachelor's thesis under the supervision of Prof. Vassilis Digalakis Jr (HEC ISOM) and Prof. Michael Lingzhi Li (HBS TOM), and my master's thesis under the dedicated guidance of Prof. Coralia Cartis. Additionally, I have served as a research intern at LIX, CNRS, INSEAD, and HEC Montreal, and have been privileged to collaborate extensively with distinguished professors including Prof. Gleb Pogudin and Prof. Georgina Hall.

My research interests bridge theoretical foundations and practical applications, spanning convex/nonconvex optimization, operations research, complex and interconnected system design, and systems & control. Modern engineered systems are increasingly assembled from large collections of heterogeneous, interacting subsystems; I am interested in how optimization theory and operations research can exploit this compositional structure to reason about performance, robustness, scalability, and coordination. I hope to apply these ideas to multi-agent systems, including distributed collaboration, resource allocation, strategic interaction, and game-theoretic decision making.

Optimization Systems & Control Operations Research Compositional System Design

Updates

Latest News

Scroll inside the panel
Mar2026

Two talks at the 2026 INFORMS Optimization Society Conference, Atlanta, GA.

With Yujun Huang and Meshal Alharbi, presented Scalable Co-Design via Linear Design Problems: Theory and Decomposition Algorithm (Session: Compositional Optimization on Networks, Mar 21). Separately presented Adaptive Levenberg-Marquardt Third-Order Newton's Method (Session: Advances in Global Optimization: Algorithms and Applications, Mar 22).

Jan2026

New preprint on ArXiv.

Thrilled to share our paper A Globally Convergent Third-Order Newton Method via Unified Semidefinite Programming Subproblems, joint work with Dr. Wenqi Zhu, Prof. Coralia Cartis, and Prof. Gioele Zardini. The paper introduces ALMTON, a third-order algorithm that uses adaptive quadratic regularization to reformulate each iteration as a tractable semidefinite program, achieving global convergence without the overhead of quartic-regularized approaches. This project started during my master's — I'm genuinely glad to see it reach this milestone, and deeply grateful to everyone who helped along the way.

Apr2024

Presented at ETAPS 2024 in Luxembourg.

I presented Dissipative quadratizations of polynomial ODE systems, sharing our work on structure-preserving quadratization and dynamical properties. Talk page · Slides

Jan2024

Presented at the Hi!Paris BATO-MOUCHE Big-City Analysis Seminar.

I presented Analysis of city composition based on basic amenities including public transport at CREST, Paris-Saclay, France. Project page · Webpage

Dec2023

Shared conference work on dissipative quadratization with supporting software.

This work develops dissipativity-preserving quadratizations for polynomial ODE systems and is accompanied by code and slides for further exploration. Publication

Jun2023

Presented at The MAX Computer Algebra Seminar.

I presented Quadratization in the reachability problem for ODEs at INRIA, Palaiseau, France. Webpage

Apr2023

Released a research report on quadratization preprocessing for reachability.

The report explores preprocessing strategies that make reachability workflows for nonlinear ODE systems more stable and computationally tractable. Report page

MIT CEE LIDS IDSS
Interests
  • INSEAD
  • HEC Paris
  • HEC Montréal
  • INRIA
Education
  • PhD in Computational Science & System Engineering, 2025-2030 (Expected)

    Massachusetts Institute of Technology, Laboratory for Information & Decision Systems

  • MSc in Mathematics and Statistics (Part C/OMMS), 2024-2025, Graduated with Honors

    University of Oxford

  • BSc in Applied Mathematics and Computer Sciences, 2021-2024, Graduated with Honors

    École Polytechnique

Preprints / Working Papers / Publications

C = Conference  ·  P = Preprint
Preprints & Working Papers
P.2 2026
Yubo Cai, Yujun Huang, Meshal Alharbi, Gioele Zardini
Preprint — submitted
P.1 2026
Yubo Cai, Wenqi Zhu, Coralia Cartis, Gioele Zardini
Preprint — submitted
Conference Papers
C.1 2024
Yubo Cai, Gleb Pogudin
Tools and Algorithms for the Construction and Analysis of Systems — TACAS 2024, Springer Nature Switzerland, pp. 323–342

Contact

  • yubocai@mit.edu
  • Stata Center (32 Vassar Street), D640, Cambridge, MA 02139