Recently there has been a dramatic increase in the performance of recognition
systems due to the introduction of deep architectures for representation
learning and classification. However, the mathematical reasons for this success
remain elusive. This tutorial will review recent work that aims to provide a
mathematical justification for several properties of deep networks, such as
global optimality, geometric stability, and invariance of the learned
representations.
%0 Journal Article
%1 vidal2017mathematics
%A Vidal, Rene
%A Bruna, Joan
%A Giryes, Raja
%A Soatto, Stefano
%D 2017
%K bounds deep-learning generalization machine-learning mathematics probability readings stable stats theory
%T Mathematics of Deep Learning
%U http://arxiv.org/abs/1712.04741
%X Recently there has been a dramatic increase in the performance of recognition
systems due to the introduction of deep architectures for representation
learning and classification. However, the mathematical reasons for this success
remain elusive. This tutorial will review recent work that aims to provide a
mathematical justification for several properties of deep networks, such as
global optimality, geometric stability, and invariance of the learned
representations.
@article{vidal2017mathematics,
abstract = {Recently there has been a dramatic increase in the performance of recognition
systems due to the introduction of deep architectures for representation
learning and classification. However, the mathematical reasons for this success
remain elusive. This tutorial will review recent work that aims to provide a
mathematical justification for several properties of deep networks, such as
global optimality, geometric stability, and invariance of the learned
representations.},
added-at = {2019-05-31T16:53:51.000+0200},
author = {Vidal, Rene and Bruna, Joan and Giryes, Raja and Soatto, Stefano},
biburl = {https://www.bibsonomy.org/bibtex/24557781b6d8e3247fdd6f79e7729e4a9/kirk86},
description = {[1712.04741] Mathematics of Deep Learning},
interhash = {2b3660fcac77806f2d488d55a965e20c},
intrahash = {4557781b6d8e3247fdd6f79e7729e4a9},
keywords = {bounds deep-learning generalization machine-learning mathematics probability readings stable stats theory},
note = {cite arxiv:1712.04741},
timestamp = {2019-09-25T04:51:46.000+0200},
title = {Mathematics of Deep Learning},
url = {http://arxiv.org/abs/1712.04741},
year = 2017
}