Workshop on Modern Challenges of Learning Theory

April 23-26, 2018

Program

 

Monday, April 23, 2018

09:00 - 09:30
Registration (Room 5345) and Coffee & Croissants (Room 6245)


Meeting room(s) : 6254

09:30 - 10:30
Peter L. Bartlett
(University of California, Berkeley)
Representation, optimization and generalization properties of deep neural networks
Abstract

Video
10:30 - 11:00
Coffee break
11:00 - 12:00
Gérard Ben Arous
(New York University)
Complexity of Tensor PCA and of its optimization dynamics
Abstract

Video
12:00 - 13:30
Lunch break
13:30 - 14:30
Marc G. Bellemare
(Google Brain Montréal)
A distributional perspective on reinforcement learning
Abstract

Video
14:30 - 15:00
Coffee break
15:00 - 16:00
Sham M. Kakade
(University of Washington)
Sub-linear reinforcement learning
Abstract

Video
16:30 - 17:30
Yoshua Bengio
(Université de Montréal)
Figuring out the magic behind backprop and SGD
Abstract

Video

 

Tuesday, April 24, 2018

09:00 - 09:30
Coffee & Croissants


Meeting room(s) : 6254

09:30 - 10:30
Alexander Rakhlin
(MIT)
Sufficient statistics for online prediction via the Burkholder method
Abstract

Video
10:30 - 11:00
Coffee break
11:00 - 12:00
Grigoris Paouris
(Texas A&M University)
The "Small ball" property in high dimensional measures
Abstract

Video
12:00 - 13:30
Lunch break
13:30 - 14:30
Csaba Szepesvári
(University of Alberta)
Completing the classification of adversarial partial monitoring games
Abstract

Video
14:30 - 15:00
Coffee break
15:00 - 16:00
Alexandra Carpentier
(Universität Potsdam)
Adaptivity to smoothness in X-armed bandits
Abstract

Video

 

Wednesday, April 25, 2018

09:00 - 09:30
Coffee & Croissants


Meeting room(s) : 6254

09:30 - 10:30
Nicolò Cesa-Bianchi
(Università degli Studi di Milano)
Nonstochastic bandits with anonymous feedback
Abstract

Video
10:30 - 11:00
Coffee break
11:00 - 12:00
Emilie Kaufmann
(Inria Lille - Nord Europ)
Bandit (for) games
Abstract

Video
12:00 - 13:30
Lunch break
13:30 - 14:30
Daniel Hsu
(Columbia University)
Learning without correspondence
Abstract

Video
14:30 - 15:00
Coffee break
15:00 - 16:00
Nicolas Le Roux
(Google Brain - Montréal)
An exploration of variance reduction techniques in stochastic optimization
Abstract

Video
16:30 - 17:30
Audrey Durand
(McGill University)
Streaming kernel regression with provably adaptive mean, variance, and regularization
Abstract

Video

 

Thursday, April 26, 2018

09:00 - 09:30
Coffee & Croissants


Meeting room(s) : 6254

09:30 - 10:30
Gergely Neu
(Universitat Pompeu Fabra)
A unified view of entropy-regularized Markov decision processes
Abstract

Video
10:30 - 11:00
Coffee break
11:00 - 12:00
Nicolas Broutin
(Sorbonne Université)
Combinatorics of colliding bullets
Abstract

Video
12:00 - 13:30
Lunch break
13:30 - 14:30
András György
(Imperial College London)
A modular analysis of adaptive online (non-)convex optimization
Abstract

Video
14:30 - 15:00
Coffee break
15:00 - 16:00
Gautam Kamath
(MIT)
Robustness in unsupervised and supervised machine learning
Abstract

Video
16:30 - 17:30
Andreas Maurer
Concentration properties and examples of functions with weak interactions
Abstract

Video