Wei Lu - ATER in Mathematics

About Me

I am an Attaché Temporaire d'Enseignement et de Recherche (ATER) at IUT Paris - Rives de Seine, Université Paris Cité. I recently defended my PhD in Mathematics at INSA Rouen Normandie. My research focuses on stochastic algorithms, machine learning, and optimization, with an emphasis on developing methods for large-scale and sequential data processing. During my PhD, I applied stochastic Newton methods to ridge regression and geometric median estimation, leveraging recursive Hessian inversion techniques. I later proposed a Robbins-Monro approach for matrix estimation, which served as the basis for universal stochastic Newton algorithms and a Full AdaGrad algorithm. Currently, my work centers on recursive covariance matrix estimation for stochastic estimators.

Education

• PhD in Mathematics - INSA Rouen Normandie (2021 - 2024), supervised by Bruno Portier and Antoine Godichon-Baggioni
• Master of Fundamental and Applied Mathematics - INSA Rouen Normandie (2020 - 2021)
• Master of Mathematical Engineering - INSA Rouen Normandie (2018 - 2021)

Current Position

ATER, Data Science Department, IUT Paris - Rives de Seine, Université Paris Cité (2024 - Present)

Responsibilities include teaching mathematical analysis, introductory data mining with R, and supporting students in foundational mathematics courses.

Publications & Preprints

• Godichon-Baggioni, Antoine, and Wei Lu. "Online stochastic Newton methods for estimating the geometric median and applications." Journal of Multivariate Analysis, 2024.
• Antoine Godichon-Baggioni, Wei Lu, Bruno Portier. "Recursive ridge regression using second-order stochastic algorithms." Computational Statistics & Data Analysis, 2024.
• Anis M. Haddouche, Wei Lu. "A unified approach for covariance matrix estimation under Stein loss." Comptes Rendus. Mathématique, 2022.
• Preprint: Antoine Godichon-Baggioni, Wei Lu, Bruno Portier. "Online estimation of the inverse of the Hessian for stochastic optimization with application to universal stochastic Newton algorithms."
• Preprint: Antoine Godichon-Baggioni, Wei Lu, Bruno Portier. "A Full Adagrad algorithm with O(Nd) operations."

Teaching Experience

Since my second year of doctoral studies until the completion of my thesis, I taught at INSA Rouen, primarily conducting tutorials (TD) in mathematics. These tutorials covered various topics such as differential equations, arithmetic, and topology. Additionally, I supervised practical sessions (TP) in data science, where I guided students in performing regression and classification tasks using Python on classic datasets.As of October 2024, I joined the IUT of Paris - Rives de Seine as an ATER (temporary teaching and research assistant). I am responsible for the "Math Accueil" course, which is designed to support students with mathematics deficiencies after high school. I also lead tutorials in mathematical analysis. Furthermore, I teach a data mining course, including its practical sessions, where I introduce decision trees (CART) and assist students in practical implementations using R. In the upcoming semester, I will take on three new subjects: probabilities, parametric estimation, and linear regression, combining both tutorials and practical sessions.

• Statistical inference, estimation - Lecturer, IUT Paris (2024-2025)
• Probability - Lecturer, IUT Paris (2024-2025)
• Linear model - Lecturer, IUT Paris (2024-2025)
• Mathematical Analysis - Lecturer, IUT Paris (2024-2025)
• Introduction to Data Mining - Lecturer, IUT Paris (2024-2025)
• Introductory Mathematics - Lecturer, IUT Paris (2024-2025)
• Arithmetic, Differential Equations, Topology - Teaching Assistant, INSA Rouen (2023-2024)
• Introduction to Data Science - Teaching Assistant, INSA Rouen (2022-2023)

Talks

• October 2024: Seminar on Statistics, Université Paris Cité, France - "A Full Adagrad algorithm with O(Nd) operations"
• May 2024: Journées de Statistique, Université de Bordeaux, France - "Universal Stochastic Newton Algorithms"
• March 2024: Statistics Working Group, Université de Rouen Normandie, France - "Online Estimation of the Inverse Hessian for Stochastic Optimization"
• February 2024: SPOC Seminar, Université de Bourgogne, France - "Universal Stochastic Newton Algorithms"
• July 2023: Journées de Statistique, Brussels, Belgium - "Stochastic Newton Methods for Estimating the Geometric Median"

Skills & Languages

Languages: English (C1), French (C1), Chinese (Native)

Programming Skills: R, LaTeX, Python, C, C++, Java, MATLAB

Contact Information

Email: wei.lu.math@gmail.com

Address: 685 Av. de l'Université, 76800 Saint-Étienne-du-Rouvray, France