Biography

I am currently a Research Fellow at CSD at ENS PSL, specializing in Generative Models. I defended my PhD in July 2024 at Université Paris-Dauphine under the supervision of Prof. Yann Chevaleyre. My research was centered around the expressivity of Generative Models and the trade-off between the quality and the diversity of the generated data. By properly choosing the divergence that models are trained to minimize, the overall behavior can be set in advance, and thus I have worked on tuning and improving models such as Normalizing Flows, Generative Adversarial Networks or Diffusion Models.

I am currently looking for interns at Université Paris Dauphine to work on Diffusion Models: offer.

Download my resumé.

Interests
  • Deep Learning
  • Generative Modelling
  • Precision and Recall for Generative Models
Education

  • PhD in Artificial Intelligence, 2024

    Université Paris-Dauphine

  • Multidisciplinary Master's Year in Quantitative Economics, 2020

    Université Paris-Dauphine

  • MVA - MEng in Computer Science, 2019

    École Normale Supérieure Paris-Saclay

  • Multidisciplinary Master's Year in Fundamental Physics, 2018

    Université Paris-Sud

  • BSc in Electrical Engineering, 2017

    École Normale Supérieure Paris-Saclay

Recent Publications

Alexandre Vérine (2024). Quality and Diversity in Generative Models through the lens of f-divergences. PhD Thesis.

PDF Cite HAL Google Scholar

Alexandre Vérine, Florian Le Bronnec, Benjamin Negrevergne, Yann Chevaleyre, Alexandre Allauzen (2024). Exploring Precision and Recall to assess the quality and diversity of LLMs. In ACL 2024 (Main).

PDF Cite arXiv Google Scholar

Alexandre Vérine, Muni Sreenivas Pydi, Benjamin Negrevergne, Yann Chevaleyre (2024). Optimal Budgeted Rejection Sampling for Generative Models. In AISTATS 2024.

PDF Cite arXiv Google Scholar

Alexandre Vérine, Benjamin Negrevergne, Muni Sreenivas Pydi, Yann Chevaleyre (2023). Precision-Recall Divergence Optimization for Generative Modeling with GANs and Normalizing Flows. In NeurIPS 2023.

PDF Cite arXiv Google Scholar

Alexandre Vérine, Benjamin Negrevergne, Fabrice Rossi, Yann Chevaleyre (2022). On the expressivity of bi-Lipschitz normalizing flows. In ACML 2022.

PDF Cite arXiv Google Scholar

Alexandre Vérine, Benjamin Negrevergne, Fabrice Rossi, Yann Chevaleyre (2021). On the expressivity of bi-Lipschitz normalizing flows. In INNF+ 2021.

PDF Cite arXiv OpenReview

Teaching

Current Courses

CourseCodeLevelTeachingYear
Deep Learning IIDL3AIISOB.S.Lectures2024
Hands-On IAEMWe AreLectures2024-2025
Introduction to Deep LearningDLE.M.Lectures2023-2025
Projet IAIAE.M.Lectures2022-2024
Deep Learning ProjectDLPM.S.Lectures2022-2024

Past Courses

CourseCodeLevelTeachingYear
Mathematics for Data ScienceMSDM.S.Lectures2020-2023
Artificial IntelligenceIAM.S.Seminars2021
Introduction to Normalizing FlowAM.S.Lectures2021
Information System EngineeringISI1B.S.Lectures/Seminars2020

Experience

 
 
 
 
 
Université PSL - École Normale Supérieure PSL
Research Fellow IA
Université PSL - École Normale Supérieure PSL
Sep 2024 – Present Paris
Research Fellow at Center for Sciences of Data (CSD) working on generative models.
 
 
 
 
 
LAMSADE - Université Paris-Dauphine
PhD
LAMSADE - Université Paris-Dauphine
Sep 2020 – Jun 2024 Paris
PhD Candidate at LAMSADE working on the expressivity of generative models.
 
 
 
 
 
LAMSADE - Université Paris-Dauphine
Research Intern
LAMSADE - Université Paris-Dauphine
Sep 2019 – Jun 2020 Paris
Part-Time Research internship on generation of Adversarial Attacks with Invertible Neural Networks.
 
 
 
 
 
Wavestone
Research Intern
Wavestone
Apr 2019 – Sep 2019 Paris
Master’s degree research internship on Invertible Neural Networks as a defense against Adversarial Attacks.
 
 
 
 
 
Advanced Structures & Composites Center - University of Maine
Research Intern
Advanced Structures & Composites Center - University of Maine
May 2018 – May 2018 Orono, Maine
Master’s degree research internship in Organic Photovoltaic Materials Through an Educational Partnership Agreement between the University of Maine and the US Army.

Contact