File Download
There are no files associated with this item.
Supplementary
-
Citations:
- Appears in Collections:
Conference Paper: Differentiable dynamic normalization for learning deep representation
Title | Differentiable dynamic normalization for learning deep representation |
---|---|
Authors | |
Keywords | Algorithms Computer Vision |
Issue Date | 2019 |
Publisher | PMLR. The Journal's web site is located at http://proceedings.mlr.press/ |
Citation | The 36th International Conference on Machine Learning (ICML 2019), Long Beach, CA, USA, 10-15 June 2019. In Proceedings of Machine Learning Research (PMLR), v. 97, p. 4203-4211 How to Cite? |
Abstract | This work presents Dynamic Normalization (DN), which is able to learn arbitrary normalization operations for different convolutional layers in a deep ConvNet. Unlike existing normalization approaches that predefined computations of the statistics (mean and variance), DN learns to estimate them. DN has several appealing benefits. First, it adapts to various networks, tasks, and batch sizes. Second, it can be easily implemented and trained in a differentiable end-to-end manner with merely small number of parameters. Third, its matrix formulation represents a wide range of normalization methods, shedding light on analyzing them theoretically. Extensive studies show that DN outperforms its counterparts in CIFAR10 and ImageNet. |
Description | Poster Presentation |
Persistent Identifier | http://hdl.handle.net/10722/284152 |
ISSN |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Luo, P | - |
dc.contributor.author | Pang, Z | - |
dc.contributor.author | Shao, W | - |
dc.contributor.author | Zhang, R | - |
dc.contributor.author | Ren, J | - |
dc.contributor.author | Wu, L | - |
dc.date.accessioned | 2020-07-20T05:56:30Z | - |
dc.date.available | 2020-07-20T05:56:30Z | - |
dc.date.issued | 2019 | - |
dc.identifier.citation | The 36th International Conference on Machine Learning (ICML 2019), Long Beach, CA, USA, 10-15 June 2019. In Proceedings of Machine Learning Research (PMLR), v. 97, p. 4203-4211 | - |
dc.identifier.issn | 2640-3498 | - |
dc.identifier.uri | http://hdl.handle.net/10722/284152 | - |
dc.description | Poster Presentation | - |
dc.description.abstract | This work presents Dynamic Normalization (DN), which is able to learn arbitrary normalization operations for different convolutional layers in a deep ConvNet. Unlike existing normalization approaches that predefined computations of the statistics (mean and variance), DN learns to estimate them. DN has several appealing benefits. First, it adapts to various networks, tasks, and batch sizes. Second, it can be easily implemented and trained in a differentiable end-to-end manner with merely small number of parameters. Third, its matrix formulation represents a wide range of normalization methods, shedding light on analyzing them theoretically. Extensive studies show that DN outperforms its counterparts in CIFAR10 and ImageNet. | - |
dc.language | eng | - |
dc.publisher | PMLR. The Journal's web site is located at http://proceedings.mlr.press/ | - |
dc.relation.ispartof | The 36th International Conference on Machine Learning (ICML) | - |
dc.relation.ispartof | Proceedings of Machine Learning Research (PMLR) | - |
dc.subject | Algorithms | - |
dc.subject | Computer Vision | - |
dc.title | Differentiable dynamic normalization for learning deep representation | - |
dc.type | Conference_Paper | - |
dc.identifier.email | Luo, P: pluo@hku.hk | - |
dc.identifier.authority | Luo, P=rp02575 | - |
dc.identifier.hkuros | 311011 | - |
dc.identifier.volume | 97 | - |
dc.identifier.spage | 4203 | - |
dc.identifier.epage | 4211 | - |
dc.publisher.place | United States | - |
dc.identifier.issnl | 2640-3498 | - |