Residual learning with skip connections permits training ultra-deep neural
networks and obtains superb performance. Building in this direction, DenseNets
proposed a dense connection structure where each layer is directly connected to
all of its predecessors. The densely connected structure leads to better
information flow and feature reuse. However, the overly dense skip connections
also bring about the problems of potential risk of overfitting, parameter
redundancy and large memory consumption. In this work, we analyze the feature
aggregation patterns of ResNets and DenseNets under a uniform aggregation view
framework. We show that both structures densely gather features from previous
layers in the network but combine them in their respective ways: summation
(ResNets) or concatenation (DenseNets). We compare the strengths and drawbacks
of these two aggregation methods and analyze their potential effects on the
networks' performance. Based on our analysis, we propose a new structure named
SparseNets which achieves better performance with fewer parameters than
DenseNets and ResNets.
%0 Generic
%1 citeulike:14541270
%A xxx,
%D 2018
%K arch classification sparsenet
%T Sparsely Connected Convolutional Networks
%U http://arxiv.org/abs/1801.05895
%X Residual learning with skip connections permits training ultra-deep neural
networks and obtains superb performance. Building in this direction, DenseNets
proposed a dense connection structure where each layer is directly connected to
all of its predecessors. The densely connected structure leads to better
information flow and feature reuse. However, the overly dense skip connections
also bring about the problems of potential risk of overfitting, parameter
redundancy and large memory consumption. In this work, we analyze the feature
aggregation patterns of ResNets and DenseNets under a uniform aggregation view
framework. We show that both structures densely gather features from previous
layers in the network but combine them in their respective ways: summation
(ResNets) or concatenation (DenseNets). We compare the strengths and drawbacks
of these two aggregation methods and analyze their potential effects on the
networks' performance. Based on our analysis, we propose a new structure named
SparseNets which achieves better performance with fewer parameters than
DenseNets and ResNets.
@misc{citeulike:14541270,
abstract = {{Residual learning with skip connections permits training ultra-deep neural
networks and obtains superb performance. Building in this direction, DenseNets
proposed a dense connection structure where each layer is directly connected to
all of its predecessors. The densely connected structure leads to better
information flow and feature reuse. However, the overly dense skip connections
also bring about the problems of potential risk of overfitting, parameter
redundancy and large memory consumption. In this work, we analyze the feature
aggregation patterns of ResNets and DenseNets under a uniform aggregation view
framework. We show that both structures densely gather features from previous
layers in the network but combine them in their respective ways: summation
(ResNets) or concatenation (DenseNets). We compare the strengths and drawbacks
of these two aggregation methods and analyze their potential effects on the
networks' performance. Based on our analysis, we propose a new structure named
SparseNets which achieves better performance with fewer parameters than
DenseNets and ResNets.}},
added-at = {2019-02-27T22:23:29.000+0100},
archiveprefix = {arXiv},
author = {xxx},
biburl = {https://www.bibsonomy.org/bibtex/26df9c2387f0eddf8812c4b7be714f66a/nmatsuk},
citeulike-article-id = {14541270},
citeulike-linkout-0 = {http://arxiv.org/abs/1801.05895},
citeulike-linkout-1 = {http://arxiv.org/pdf/1801.05895},
day = 18,
eprint = {1801.05895},
interhash = {90e5fe429d18cba381a7b42db120af3c},
intrahash = {6df9c2387f0eddf8812c4b7be714f66a},
keywords = {arch classification sparsenet},
month = jan,
posted-at = {2018-02-27 07:02:49},
priority = {0},
timestamp = {2019-02-27T22:23:29.000+0100},
title = {{Sparsely Connected Convolutional Networks}},
url = {http://arxiv.org/abs/1801.05895},
year = 2018
}