François Fleuret

E-mail

francois.fleuret@unige.ch

Address

University of Geneva
Bâtiment Battelle A
Route de Drize, 7
1227 Carouge, Switzerland

Professor, Machine Learning Group, University of Geneva. Adjunct Professor, École Polytechnique Fédérale de Lausanne. Co-founder Neural Concept SA.

News

• I had the opportunity to give an AI introductory course to members of the Swiss Federal Assembly.
• The materials for my deep-learning course at the University of Geneva, and the EPFL are available.
• One paper accepted at CVPR 2021. ↴

  With Teja Prabhu on improving transfer learning for semantic segmentation (Prabhu and Fleuret, 2021).

• Two papers accepted at ICLR 2021. ↴
  • With Suraj Srinivas on interpreting the gradient of a model's output with respect to its input (Srinivas and Fleuret, 2021).
  • With Tatjana Chavdarova, Matteo Pagliardini, Sebastian Stich, and Martin Jaggi on using Look-Ahead optimization for training GANs (Chavdarova et al, 2021).
• One paper accepted at NeurIPS 2020. ↴

  With Apoorv Vyas and Angelos Katharopoulos on speeding up transformers by clustering queries (Vyas et al, 2020).

• One paper accepted at EMNLP 2020. ↴

  With Vincent Micheli and Martin d'Hoffschmidt on the required amount of data for pre-training language models (Micheli et al, 2020).

• One paper accepted at ACCV 2020. ↴

  With Evann Courdier on a new performance measure for real-time semantic segmentation (Courdier and Fleuret, 2020).

• On July 1st, 2020, I joined the University of Geneva as Full Professor of Computer Science.
• Our project “Robust Deep Density Models for High-Energy Particle Physics and Solar Flare Analysis” is funded by the SNSF. ↴

  This is a four year joint project with Slava Voloshynovskiy and Tobias Golling from the University of Geneva, and André Csillaghy from the University of Applied Sciences and Arts Northwestern Switzerland. Funding comes from the Sinergia program of the Swiss National Science Foundation.

• Two papers accepted at ICML 2020. ↴
  • With Teja Prabhu, Florian Mai, Thijs Vogels, and Martin Jaggi on the cost of hyper-parameter tuning (Prabhu et al, 2020).
  • With Angelos Katharopoulos, Apoorv Vyas, and Nikolaos Pappas on speeding up transformers using a form of linear attention (Katharopoulos et al, 2020).