About me

Since April 2023, I am a researcher in the CELESTE team of INRIA Paris-Saclay, located in the Laboratoire Mathématique d’Orsay. I am working on different aspects of the theory of machine learning, including decentralised (online) learning, meta-learning and the theoretical understanding of neural networks.

Before that, I was a postdoc in the Theory of Machine Learning Lab at EPFL with Nicolas Flammarion, mainly focusing on multitask/meta-learning and the theoretical study of neural networks.

In 2021, I completed a PhD at ENS Paris-Saclay, started in September 2018 under the supervision of Vianney Perchet and entitled “Statistical Learning in a strategical environment”. I studied multi-agent learning, combining (online) learning with game theoretical tools. In particular, I mainly focused on the problem of Multiplayer Multi-armed bandits, but I also worked on other bandits related problems, Social Learning and Utility/Privacy trade-off.

Have a look at my (hopefully updated) CV for additional information!

Please, do not hesitate to ask me any question or supplementary material (code, slides…) by email.

News

  • I was awarded the 2022 PGMO thesis prize! I will present my PhD work after the PhD award ceremony, during the PGMO days on November 29th. 05-10-2022
  • Gradient flow dynamics of shallow ReLU networks for square loss and orthogonal inputs accepted at NeurIPS 2022. 16-09-2022
  • I will be giving a 20 min talk at Learning and Optimization in Luminy on our paper Gradient flow dynamics of shallow ReLU networks for square loss and orthogonal inputs. 16-09-2022
  • I gave a 20 min presentation of Gradient flow dynamics of shallow ReLU networks for square loss and orthogonal inputs at EPFL-CIS & RIKEN-AIP Joint Workshop on Machine Learning (video and slides to come on the website). 16-09-2022
  • Trace norm regularization for multi-task learning with scarce data accepted at COLT 2022. 17-05-2022
  • Utility/Privacy Trade-off as Regularized Optimal Transport got accepted for publication at Mathematical Programming. It is an extended version of our work presented at AISTATS 2020. Paper available here. 17-03-2022
  • One paper accepted at ALT 2022: Social Learning in Non-Stationary Environments. 10-01-2022
  • I successfully defended my PhD thesis at ENS Paris-Saclay! see slides and manuscript 30-09-2021
  • Two papers accepted at NeurIPS 2021: Decentralized Learning in Online Queuing Systems as a spotlight and Making the most of your day: online learning for optimal allocation of time as a poster. 28-09-2021