Yoan Russac

PhD Student in Machine Learning

Ecole Normale Supérieure

Biography

I am a PhD student in Machine Learning at Ecole Normale Supérieure in the Inria VALDA team. I am lucky to be supervised by Olivier Cappé and Aurélien Garivier. My research interests include sequential learning with partial information feedback and more precisely bandits models.

Interests

Bandits
Machine Learning
Reinforcement Learning

Education

PhD in Machine Learning, 2018-2021 (expected)
Ecole Normale Supérieure
MVA Master Degree, 2017-2018
ENS-Paris Saclay
MSc in Statistics and Data Science, 2015-2018
ENSAE- Institut Polytechnique de Paris

Publications

Weighted Linear Bandits in Non-Stationary Environments

In Proceedings of NeurIPS2019

Yoan Russac, Claire Vernade, Olivier Cappé

PDF Poster Slides Code

Embeddings of Categorical Variables for Sequential Data in Fraud Context

In Proceedings of AMLTA2018

Yoan Russac, Olivier Caelen, Liyun He-Guelton

Experience

Applied Scientist Intern

Amazon

Jun 2018 – Nov 2018 Berlin, Germany

Research on non-stationarity and bandits.

Machine Learning Researcher Intern

Worldline

Jun 2017 – Sep 2017 Brussels, Belgium

Research on fraud detection using Word2Vec embeddings.

Recent & Upcoming Talks

Introduction to Linear Bandits

An introduction to Bandits and Linear Bandits

Nov 8, 2019 ENS Paris

Non-Stationary Linear Bandits

In this talk a new deviation bound to control the weighted least-squares estimator in a Linear Bandit model is discussed.

Jun 5, 2019 Deepmind London