Statistics/Learning at Paris-Saclay

Europe/Paris
Centre de Conférences Marilyn et James Simons (Le Bois-Marie)

Centre de Conférences Marilyn et James Simons

Le Bois-Marie

35, route de Chartres 91440 Bures-sur-Yvette
Description

The aim of this workshop is to bring together mathematicians and computer scientists around some talks on recent results in statistics and machine learning. Various topics will be presented, among which sequential learning, aggregation of estimators, Hidden Markov models, network inference and optimization for machine learning.

Organised by : Sylvain ARLOT

Invited speakers :

Pierre Alquier (ENSAE)
Stéphane Gaiffas (Ecole Polytechnique)
Claire Lacour (Université Paris-Sud)
Odalric-Ambrym Maillard (INRIA - Université Paris-Sud)
Yann Ollivier (CNRS - Université Paris-Sud)
Joseph Salmon (LTCI, CNRS, Télécom ParisTech)
 

 

logos

Abstracts
Poster
PowerPoint Alquier
PowerPoint Gaiffas
PowerPoint Lacour
PowerPoint Maillard
PowerPoint Salmon
Participants
    • 09:00 18:30
      Journée Statistique: Apprentissage Paris-Saclay
      • 09:00
        Café d'accueil 1h
      • 10:00
        Pierre Alquier 50m
        PAC-Bayesian bounds for aggregation of estimators: introduction and computational issues
        Transparents
      • 10:50
        Pause Café 30m
      • 11:20
        Joseph Salmon 50m
        GAP safe screening rule for sparsity enforcing penalties
        Transparents
      • 12:10
        Odalric-Ambrym Maillard 50m
        Multi-armed bandits: Proof techniques and challenges
        Transparents
      • 13:00
        Déjeuner 1h 30m
      • 14:30
        Claire Lacour 50m
        Minimax adaptive estimation of nonparametric hidden Markov models
        Transparents
      • 15:20
        Stéphane Gaiffas 50m
        Concentration for matrix martingales and microscopic activity of social networks using Hawkes processes
        Transparents
      • 16:10
        Pause café 30m
      • 16:40
        Yann Ollivier 50m
        Learn as you go: Training recurrent neural networks and dynamical systems
      • 17:30
        Discussion 1h