dbo:abstract
|
- The Wasserstein Generative Adversarial Network (WGAN) is a variant of generative adversarial network (GAN) proposed in 2017 that aims to "improve the stability of learning, get rid of problems like mode collapse, and provide meaningful learning curves useful for debugging and hyperparameter searches". Compared with the original GAN discriminator, the Wasserstein GAN discriminator provides a better learning signal to the generator. This allows the training to be more stable when generator is learning distributions in very high dimensional spaces. (en)
|
dbo:thumbnail
| |
dbo:wikiPageExternalLink
| |
dbo:wikiPageID
| |
dbo:wikiPageLength
|
- 16179 (xsd:nonNegativeInteger)
|
dbo:wikiPageRevisionID
| |
dbo:wikiPageWikiLink
| |
dbp:mathStatement
|
- For any fixed generator strategy , let the optimal reply be , then
where the derivative is the Radon–Nikodym derivative, and is the Jensen–Shannon divergence. (en)
- When the probability space is a metric space, then
for any fixed ,
where is the Lipschitz norm. (en)
|
dbp:name
|
- Theorem (en)
- in practice (en)
- not really done in practice (en)
|
dbp:note
|
- the optimal discriminator computes the Jensen–Shannon divergence (en)
- Kantorovich-Rubenstein duality (en)
- In practice, the generator would never be able to reach perfect imitation, and so the discriminator would have motivation for perceiving the difference, which allows it to be used for other tasks, such as performing ImageNet classification without supervision. (en)
- This is not how it is really done in practice, since
is in general intractable, but it is theoretically illuminating. (en)
|
dbp:wikiPageUsesTemplate
| |
dcterms:subject
| |
rdfs:comment
|
- The Wasserstein Generative Adversarial Network (WGAN) is a variant of generative adversarial network (GAN) proposed in 2017 that aims to "improve the stability of learning, get rid of problems like mode collapse, and provide meaningful learning curves useful for debugging and hyperparameter searches". Compared with the original GAN discriminator, the Wasserstein GAN discriminator provides a better learning signal to the generator. This allows the training to be more stable when generator is learning distributions in very high dimensional spaces. (en)
|
rdfs:label
| |
owl:sameAs
| |
prov:wasDerivedFrom
| |
foaf:depiction
| |
foaf:isPrimaryTopicOf
| |
is dbo:wikiPageRedirects
of | |
is dbo:wikiPageWikiLink
of | |
is foaf:primaryTopic
of | |