Vincent Blot
Affiliations. LISN, CNRS, Paris-Saclay University & Quantmetry
Office 223
University Campus Building 507
Rue du Belvédère, 91405 Orsay
I am currently a Ph.D student working on Uncertainty Quantification in Computer Vision. this thesis is in collaboration with the LISN and Quantmetry. I will be working mainly on Conformal Predictions, Bayesian Neural Networks and their applications to Vision Transformers. This thesis is carried out under the supervision of Didier Lucor, Alessandro Leite and Nicolas Brunel.
I am also one of the lead developer of the MAPIE library which contributes to making Conformal Predictions algorithms accessible by following the scikit-learn coding philosophy. If you wish to learn more about this project do not hesitate to contact me or the other developers of the library.
news
| May 3, 2025 | Presentation of my paper “Automatically Adaptive Conformal risk control” to the 2025 AIStats conference in Phuket, Thailand |
|---|---|
| Oct 10, 2024 | Presentation of my article “Efficient Precision Control in Object Detection Models for Enhanced and Reliable Ovarian Follicle Counting” in the poster session of the 2024 UNSURE workshop of the 2024 MICCAI conference in Marrakech, Morocco. |
| Jul 17, 2023 | Participation to the CEMRACS 2023 summer school on Scientific Machine Learning. After 1 week of summer school composed of lectures and practical sessions, I’ve been working for 5 weeks on a project on Conformal prediction for data-driven and physics-based prognostics algorithms. This project led to the publication of a paper called “CONFORMAL APPROACH TO GAUSSIAN PROCESS SURROGATE EVALUATION WITH MARGINAL COVERAGE GUARANTEES”. |
| Dec 1, 2022 | Beginning of my Ph.D |