I am a postdoc researcher at Università degli Studi di Milano (Italy), in the group of Nicolò Cesa-Bianchi. Previously, I completed my PhD under the supervision on Olivier Cappé and Aurélien Garivier. My research interests are mainly centered on sequential decision problems, including online learning, bandits and their extensions. During my PhD, I have mainly studied applications of bandit theory to online markets. Currently, as a part of ELSA, I am investigating forms of online federated learning. I am eager to work on more robust, sustainable sequential decision algorithms, either in a federated setting or not.
(Industrial) PhD in Computer Sciences, 2022
ENS Paris & Numberly
MEng in Data Sciences, 2018
Université Paris Saclay
Graduation from Télécom Paris, 2019
Télécom Paris
Graduation from HEC Paris, 2019
HEC
Rejection Sampling is a fundamental Monte-Carlo method. It is used to sample from distributions admitting a probability density function which can be evaluated exactly at any given point, albeit at a high computational cost. However, without proper tuning, this technique implies a high rejection rate. Several methods have been explored to cope with this problem, based on the principle of adaptively estimating the density by a simpler function, using the information of the previous samples. Most of them either rely on strong assumptions on the form of the density, or do not offer any theoretical performance guarantee. We give the first theoretical lower bound for the problem of adaptive rejection sampling and introduce a new algorithm which guarantees a near-optimal rejection rate in a minimax sense.