%0 Generic %1 citeulike:14622046 %A xxx, %D 2018 %K arch backbone regularization %T Differentiable Learning-to-Normalize via Switchable Normalization %U http://arxiv.org/abs/1806.10779 %X We address a learning-to-normalize problem by proposing Switchable Normalization (SN), which learns to select different operations for different normalization layers of a deep neural network (DNN). SN switches among three distinct scopes to compute statistics (means and variances) including a channel, a layer, and a minibatch, by learning their importance weights in an end-to-end manner. SN has several good properties. First, it adapts to various network architectures and tasks (see Fig.1). Second, it is robust to a wide range of batch sizes, maintaining high performance when small minibatch is presented (e.g. 2 images/GPU). Third, SN treats all channels as a group, unlike group normalization that searches the number of groups as a hyper-parameter. Without bells and whistles, SN outperforms its counterparts on various challenging problems, such as image classification in ImageNet, object detection and segmentation in COCO, artistic image stylization, and neural architecture search. We hope SN will help ease the usages and understand the effects of normalization techniques in deep learning. The code of SN will be made available in <a href="https://github.com/switchablenorms/.">this https URL</a>

@misc{citeulike:14622046, abstract = {{We address a learning-to-normalize problem by proposing Switchable Normalization (SN), which learns to select different operations for different normalization layers of a deep neural network (DNN). SN switches among three distinct scopes to compute statistics (means and variances) including a channel, a layer, and a minibatch, by learning their importance weights in an end-to-end manner. SN has several good properties. First, it adapts to various network architectures and tasks (see Fig.1). Second, it is robust to a wide range of batch sizes, maintaining high performance when small minibatch is presented (e.g. 2 images/GPU). Third, SN treats all channels as a group, unlike group normalization that searches the number of groups as a hyper-parameter. Without bells and whistles, SN outperforms its counterparts on various challenging problems, such as image classification in ImageNet, object detection and segmentation in COCO, artistic image stylization, and neural architecture search. We hope SN will help ease the usages and understand the effects of normalization techniques in deep learning. The code of SN will be made available in <a href="https://github.com/switchablenorms/.">this https URL</a>}}, added-at = {2019-02-27T22:23:29.000+0100}, archiveprefix = {arXiv}, author = {xxx}, biburl = {https://www.bibsonomy.org/bibtex/2d8291e3f739e4b9abde7526e4214eebe/nmatsuk}, citeulike-article-id = {14622046}, citeulike-linkout-0 = {http://arxiv.org/abs/1806.10779}, citeulike-linkout-1 = {http://arxiv.org/pdf/1806.10779}, day = 3, eprint = {1806.10779}, interhash = {91035ad3d56605ace8bad6730bf851f5}, intrahash = {d8291e3f739e4b9abde7526e4214eebe}, keywords = {arch backbone regularization}, month = jul, posted-at = {2018-08-04 08:56:08}, priority = {5}, timestamp = {2019-02-27T22:23:29.000+0100}, title = {{Differentiable Learning-to-Normalize via Switchable Normalization}}, url = {http://arxiv.org/abs/1806.10779}, year = 2018 }