SCGAN: Disentangled Representation Learning by Adding Similarity Constraint on Generative Adversarial Nets
We proposed a novel generative adversarial net called similarity constraint generative adversarial network (SCGAN), which is capable of learning the disentangled representation in a completely unsupervised manner. Inspired by the smoothness assumption and our assumption on the content and the repres...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2019-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/8476290/ |
| _version_ | 1818936352344899584 |
|---|---|
| author | Xiaoqiang Li Liangbo Chen Lu Wang Pin Wu Weiqin Tong |
| author_facet | Xiaoqiang Li Liangbo Chen Lu Wang Pin Wu Weiqin Tong |
| author_sort | Xiaoqiang Li |
| collection | DOAJ |
| description | We proposed a novel generative adversarial net called similarity constraint generative adversarial network (SCGAN), which is capable of learning the disentangled representation in a completely unsupervised manner. Inspired by the smoothness assumption and our assumption on the content and the representation of images, we design an effective similarity constraint. SCGAN can disentangle interpretable representations by adding this similarity constraint between conditions and synthetic images. In fact, similarity constraint works as a tutor to instruct generator network to comprehend the difference of representations based on conditions. SCGAN successfully distinguishes different representations on a number of datasets. Specifically, SCGAN captures digit type on MNIST, clothing type on Fashion-MNIST, lighting on SVHN, and object size on CIFAR10. On the CelebA dataset, SCGAN captures more semantic representations, e.g., poses, emotions, and hair styles. Experiments show that SCGAN is comparable with InfoGAN (another generative adversarial net disentangles interpretable representations on these datasets unsupervisedly) on disentangled representation learning. Code is available at https://github.com/gauss-clb/SCGAN. |
| first_indexed | 2024-12-20T05:34:42Z |
| format | Article |
| id | doaj.art-7662c45a30b04ce3bccc8dac7ffdb5a7 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-20T05:34:42Z |
| publishDate | 2019-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-7662c45a30b04ce3bccc8dac7ffdb5a72022-12-21T19:51:39ZengIEEEIEEE Access2169-35362019-01-01714792814793810.1109/ACCESS.2018.28726958476290SCGAN: Disentangled Representation Learning by Adding Similarity Constraint on Generative Adversarial NetsXiaoqiang Li0Liangbo Chen1https://orcid.org/0000-0001-6365-0754Lu Wang2Pin Wu3Weiqin Tong4School of Computer Engineering and Science, Shanghai University, Shanghai, ChinaSchool of Computer Engineering and Science, Shanghai University, Shanghai, ChinaSchool of Computer Engineering and Science, Shanghai University, Shanghai, ChinaSchool of Computer Engineering and Science, Shanghai University, Shanghai, ChinaSchool of Computer Engineering and Science, Shanghai University, Shanghai, ChinaWe proposed a novel generative adversarial net called similarity constraint generative adversarial network (SCGAN), which is capable of learning the disentangled representation in a completely unsupervised manner. Inspired by the smoothness assumption and our assumption on the content and the representation of images, we design an effective similarity constraint. SCGAN can disentangle interpretable representations by adding this similarity constraint between conditions and synthetic images. In fact, similarity constraint works as a tutor to instruct generator network to comprehend the difference of representations based on conditions. SCGAN successfully distinguishes different representations on a number of datasets. Specifically, SCGAN captures digit type on MNIST, clothing type on Fashion-MNIST, lighting on SVHN, and object size on CIFAR10. On the CelebA dataset, SCGAN captures more semantic representations, e.g., poses, emotions, and hair styles. Experiments show that SCGAN is comparable with InfoGAN (another generative adversarial net disentangles interpretable representations on these datasets unsupervisedly) on disentangled representation learning. Code is available at https://github.com/gauss-clb/SCGAN.https://ieeexplore.ieee.org/document/8476290/Generative adversarial netsrepresentation learningunsupervised learning |
| spellingShingle | Xiaoqiang Li Liangbo Chen Lu Wang Pin Wu Weiqin Tong SCGAN: Disentangled Representation Learning by Adding Similarity Constraint on Generative Adversarial Nets IEEE Access Generative adversarial nets representation learning unsupervised learning |
| title | SCGAN: Disentangled Representation Learning by Adding Similarity Constraint on Generative Adversarial Nets |
| title_full | SCGAN: Disentangled Representation Learning by Adding Similarity Constraint on Generative Adversarial Nets |
| title_fullStr | SCGAN: Disentangled Representation Learning by Adding Similarity Constraint on Generative Adversarial Nets |
| title_full_unstemmed | SCGAN: Disentangled Representation Learning by Adding Similarity Constraint on Generative Adversarial Nets |
| title_short | SCGAN: Disentangled Representation Learning by Adding Similarity Constraint on Generative Adversarial Nets |
| title_sort | scgan disentangled representation learning by adding similarity constraint on generative adversarial nets |
| topic | Generative adversarial nets representation learning unsupervised learning |
| url | https://ieeexplore.ieee.org/document/8476290/ |
| work_keys_str_mv | AT xiaoqiangli scgandisentangledrepresentationlearningbyaddingsimilarityconstraintongenerativeadversarialnets AT liangbochen scgandisentangledrepresentationlearningbyaddingsimilarityconstraintongenerativeadversarialnets AT luwang scgandisentangledrepresentationlearningbyaddingsimilarityconstraintongenerativeadversarialnets AT pinwu scgandisentangledrepresentationlearningbyaddingsimilarityconstraintongenerativeadversarialnets AT weiqintong scgandisentangledrepresentationlearningbyaddingsimilarityconstraintongenerativeadversarialnets |