Start date cannot be after end date.
Over-parameterized nonlinear learning - Gradient descent follows the shortest path
Presenter
- Mahdi Soltanolkotabi
February 22, 2019
ICERM
Arithmetic and the representation theory of fundamental groups
Presenter
- Daniel Litt
February 21, 2019
IPAM
On the Power of Curriculum Learning in Training Deep Networks
Presenter
- Daphna Weinshall
February 21, 2019
ICERM
Braids, hyperplane arrangements, and Milnor fibrations
Presenter
- Alexander Suciu
February 21, 2019
IPAM
The Information Bottleneck theory of Deep Learning and the computational benefit of the hidden layers
Presenter
- Naftali Tishby
February 21, 2019
ICERM
Gradient descent aligns the layers of deep linear networks
Presenter
- Matus Telgarsky
February 21, 2019
ICERM
Monodromy of Kodaira fibrations and moduli of abelian varieties
Presenter
- Laure Flapan
February 21, 2019
IPAM
Improvements and extensions to the K-FAC method for training neural nets
Presenter
- James Martens
February 21, 2019
ICERM