J. Wang, Y. Chen, R. Chakraborty, and S. Yu. (2019)cite arxiv:1911.12207Comment: To appear in CVPR 2020, project page: http://pwang.pw/ocnn.html.
Abstract
Deep convolutional neural networks are hindered by training instability and
feature redundancy towards further performance improvement. A promising
solution is to impose orthogonality on convolutional filters.
We develop an efficient approach to impose filter orthogonality on a
convolutional layer based on the doubly block-Toeplitz matrix representation of
the convolutional kernel instead of using the common kernel orthogonality
approach, which we show is only necessary but not sufficient for ensuring
orthogonal convolutions.
Our proposed orthogonal convolution requires no additional parameters and
little computational overhead. This method consistently outperforms the kernel
orthogonality alternative on a wide range of tasks such as image classification
and inpainting under supervised, semi-supervised and unsupervised settings.
Further, it learns more diverse and expressive features with better training
stability, robustness, and generalization. Our code is publicly available at
https://github.com/samaonline/Orthogonal-Convolutional-Neural-Networks.
%0 Journal Article
%1 wang2019orthogonal
%A Wang, Jiayun
%A Chen, Yubei
%A Chakraborty, Rudrasis
%A Yu, Stella X.
%D 2019
%K convolution orthogonal
%T Orthogonal Convolutional Neural Networks
%U http://arxiv.org/abs/1911.12207
%X Deep convolutional neural networks are hindered by training instability and
feature redundancy towards further performance improvement. A promising
solution is to impose orthogonality on convolutional filters.
We develop an efficient approach to impose filter orthogonality on a
convolutional layer based on the doubly block-Toeplitz matrix representation of
the convolutional kernel instead of using the common kernel orthogonality
approach, which we show is only necessary but not sufficient for ensuring
orthogonal convolutions.
Our proposed orthogonal convolution requires no additional parameters and
little computational overhead. This method consistently outperforms the kernel
orthogonality alternative on a wide range of tasks such as image classification
and inpainting under supervised, semi-supervised and unsupervised settings.
Further, it learns more diverse and expressive features with better training
stability, robustness, and generalization. Our code is publicly available at
https://github.com/samaonline/Orthogonal-Convolutional-Neural-Networks.
@article{wang2019orthogonal,
abstract = {Deep convolutional neural networks are hindered by training instability and
feature redundancy towards further performance improvement. A promising
solution is to impose orthogonality on convolutional filters.
We develop an efficient approach to impose filter orthogonality on a
convolutional layer based on the doubly block-Toeplitz matrix representation of
the convolutional kernel instead of using the common kernel orthogonality
approach, which we show is only necessary but not sufficient for ensuring
orthogonal convolutions.
Our proposed orthogonal convolution requires no additional parameters and
little computational overhead. This method consistently outperforms the kernel
orthogonality alternative on a wide range of tasks such as image classification
and inpainting under supervised, semi-supervised and unsupervised settings.
Further, it learns more diverse and expressive features with better training
stability, robustness, and generalization. Our code is publicly available at
https://github.com/samaonline/Orthogonal-Convolutional-Neural-Networks.},
added-at = {2020-05-22T01:45:37.000+0200},
author = {Wang, Jiayun and Chen, Yubei and Chakraborty, Rudrasis and Yu, Stella X.},
biburl = {https://www.bibsonomy.org/bibtex/21c9cdd0377556216bad7df12bb2905dc/kirk86},
description = {[1911.12207] Orthogonal Convolutional Neural Networks},
interhash = {234273fcd4fe5619c259f726fdac5eee},
intrahash = {1c9cdd0377556216bad7df12bb2905dc},
keywords = {convolution orthogonal},
note = {cite arxiv:1911.12207Comment: To appear in CVPR 2020, project page: http://pwang.pw/ocnn.html},
timestamp = {2020-05-22T01:45:37.000+0200},
title = {Orthogonal Convolutional Neural Networks},
url = {http://arxiv.org/abs/1911.12207},
year = 2019
}