Scholar
Maxime Vono
Google Scholar ID: R5dfDTAAAAAJ
Staff Research Lead, Criteo AI Lab
Citations & Impact
Citations
530
H-index
14
i10-index
14
Publications
20
Co-authors
41
list available
Publications
Resume (English only)
Academic Achievements
- In February 2025, they just released the largest bidding real-world dataset to design future-proof private advertising systems! Two new preprints on data valuation and learning from label proportions! A paper on private mean estimation under user-level differential privacy has been accepted at AISTATS 2025! A paper on dataset valuation using novel Shapley value estimators has been accepted at NeurIPS 2024! A position paper on local differential privacy & computational advertising has been accepted at WISE 2024! 1 paper accepted at TMLR! Top Reviewer, AISTATS 2022; Prix Léopold Escande, 2020; Finalist, Best Student Paper Award, IEEE MLSP 2018.
Research Experience
- Currently, he is a Staff Research Lead at the Criteo AI Lab in Paris, managing a hybrid team of ML engineers and researchers. The team focuses on securing future-proof performance for Criteo AI engines due to the loss of granular user signal. Previously, he was a Research Scientist at the Lagrange Mathematics and Computing Research Center of Huawei, working with Alain Durmus and Eric Moulines.
Education
- Obtained his Ph.D. in statistics on October 7, 2020, from the University of Toulouse, supervised by Pierre Chainais and Nicolas Dobigeon, within the SC group of the IRIT laboratory. In spring 2019, he was a research visiting scholar at the Department of Statistics of the University of Oxford, working with Arnaud Doucet.
Background
- Current research interests lie in privacy-preserving machine learning and data valuation. During his time at Huawei, he worked on distributed/federated Bayesian approaches and privacy-preserving machine learning. His Ph.D. works focused on deriving a broad approximate statistical framework and associated Monte Carlo sampling approaches inspired by the variable splitting method in optimization (e.g., used by quadratic penalty approaches).
Miscellany
- He has been affiliated with the ORION-B project, which gathers astrophysicists, data scientists, and statisticians in order to better understand the formation of stars in our universe.
