ICML 2017poster249 citations
Failures of Gradient-Based Deep Learning
Shai Shalev-Shwartz, Ohad Shamir, Shaked Shammah
Abstract
In recent years, Deep Learning has become the go-to solution for a broad range of applications, often outperforming state-of-the-art. However, it is important, for both theoreticians and practitioners, to gain a deeper understanding of the difficulties and limitations associated with common approaches and algorithms. We describe four types of simple problems, for which the gradient-based algorithms commonly used in deep learning either fail or suffer from significant difficulties. We illustrate the failures through practical experiments, and provide theoretical insights explaining their source, and how they might be remedied.
BibTeX
@InProceedings{pmlr-v70-shalev-shwartz17a,
title = {Failures of Gradient-Based Deep Learning},
author = {Shai Shalev-Shwartz and Ohad Shamir and Shaked Shammah},
booktitle = {Proceedings of the 34th International Conference on Machine Learning},
pages = {3067--3075},
year = {2017},
editor = {Precup, Doina and Teh, Yee Whye},
volume = {70},
series = {Proceedings of Machine Learning Research},
month = {06--11 Aug},
publisher = {PMLR},
pdf = {http://proceedings.mlr.press/v70/shalev-shwartz17a/shalev-shwartz17a.pdf},
url = {https://proceedings.mlr.press/v70/shalev-shwartz17a.html},
abstract = {In recent years, Deep Learning has become the go-to solution for a broad range of applications, often outperforming state-of-the-art. However, it is important, for both theoreticians and practitioners, to gain a deeper understanding of the difficulties and limitations associated with common approaches and algorithms. We describe four types of simple problems, for which the gradient-based algorithms commonly used in deep learning either fail or suffer from significant difficulties. We illustrate the failures through practical experiments, and provide theoretical insights explaining their source, and how they might be remedied.}
}