Dorian Baudry

Dorian Baudry

Post-Doctoral Researcher

CNRS, ENSAE Paris

Biography

I am a post-doctoral researcher at CNRS and ENSAE. I received my PhD in computer science from the University of Lille, where I worked under the direction of Emilie Kaufmann and Odalric-Ambrym Maillard in the INRIA ScooL team.

During my PhD, I explored non-parametric algorithms for the Multi-Armed Bandit problem. Motivated by an application in agriculture, I investigated several approaches based on sub-sampling or Dirichlet Sampling, that work under realistic non-parametric assumptions on the reward distributions. I then considered some generalizations of these approaches motivated by practical considerations: risk-awareness by defining the objective with an alternative metric to the expected reward, non-stationarity, or batched feedback.

Interests

  • Multi-Armed Bandits
  • Online Learning
  • Statistics
  • Machine Learning

Education

  • PhD in Computer Science, 2019-2022

    CNRS/Inria Lille, SequeL/ScooL team

  • MSc Mathématiques Vision Apprentissage (MVA), 2019

    ENS Paris-Saclay

  • MSc in Statistics and Computer Science, 2019

    ENSAE Paris

Experience

 
 
 
 
 

Research visit

Kyoto University

Apr 2022 – Jul 2022 Kyoto, Japan
3-months research visit to collaborate with Junya Honda, in the context of the RELIANT associate team. Our goal was to unify the analysis of MED and TS algorithms in bandits.
 
 
 
 
 

Teaching Assistant

University of Lille

Jan 2020 – Apr 2022 Lille, France
~150 hours of teaching during my PhD. Teachings include: Data Mining, Natural Language Processing, Mathematics for the Economist.
 
 
 
 
 

Data Scientist

Kayrros

May 2019 – Oct 2019 Paris, France
Research project on the analysis of news to predict movements of different assets in the energy market.
 
 
 
 
 

Junior Index Structurer

Société générale CIB

Jan 2018 – Aug 2018 London, UK
Pricing of financial products.
 
 
 
 
 

Risk Management intern

AXA France

Jun 2017 – Nov 2017 Paris, France
Implementation of a reserving model for long-term care.