Published on December 2, 2022
–
Updated on December 14, 2022
Take a look at the 3IA papers published at the Thirty-sixth Conference on Neural Information Processing Systems (NeurIPS 2022).
The NeurIPS conference takes place from November 28 to December 9, 2022. This conference is a multi-track interdisciplinary annual meeting that includes invited talks, demonstrations, symposia, and oral and poster presentations of refereed papers. Along with the conference is a professional exposition focusing on machine learning in practice, a series of tutorials, and topical workshops that provide a less formal setting for the exchange of ideas.
This year, several 3IA Côte d'Azur chair holders and students participated in the event:
This year, several 3IA Côte d'Azur chair holders and students participated in the event:
- Charles Bouveyron (Université Côte d'Azur) and Pierre-Alexandre Mattei (Inria)
- Generalised Mutual Information for Discriminative Clustering
- A joint work with 3IA chairs as well Frédéric Precioso from Université Côte d'Azur, and collaborators from Jellysmack (Warith Harchaoui) and Université Laval in Québec (Arnaud Droit & Mickaël Leclerc).
- Generalised Mutual Information for Discriminative Clustering
- Francesco Galati (EURECOM)
- Maurizio Filippone
- Rémi Flamary and Cédric Vincent Cuaz (Université Côte d'Azur)
- Template based Graph Neural Network with Optimal Transport Distances A novel template based Graph Representation Learning approach leading to state-of-the-art performances on Graph Classification. A collaboration between 3IA Côte d'Azur, INRIA, Ecole Polytechnique, ENS de Lyon & Université Bretagne-Sud. Paper that was presented during in-person and live poster sessions during the week from 28/11/2022 to 02/12/2022, and granted of an Oral on the following week.
- Rémi Flamary (Université Côte d'Azur):
- Ziming Liu (Inria):
- Marco Lorenzi (Inria) and Giovanni Neglia (Inria)
- FLamby: Datasets and Benchmarks for Cross-Silo Federated Learning in Realistic Healthcare Settings
- Collaboration with researchers from Owkin, a French-American startup working on federated learning for healthcare applications, EPFL (Switzerland), École Polytechnique (France), University of Bonn (Germany), the University of California Berkeley (USA), and Inria. In this work, we propose a novel cross-silo dataset suite focused on healthcare, FLamby (Federated Learning AMple Benchmark of Your cross-silo strategies), to bridge the gap between theory and practice of cross-silo FL. FLamby encompasses 7 healthcare datasets with natural splits, covering multiple tasks, modalities, and data volumes, each accompanied with baseline training code. As an illustration, we additionally benchmark standard FL algorithms on all datasets. Our flexible and modular suite allows researchers to easily download datasets, reproduce results and re-use the different components for their research.
- FLamby: Datasets and Benchmarks for Cross-Silo Federated Learning in Realistic Healthcare Settings
