FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2767609405> ?p ?o ?g. }

• W2767609405 abstract "Generative adversarial training can be generally understood as minimizing certain moment matching loss defined by a set of discriminator functions, typically neural networks. The discriminator set should be large enough to be able to uniquely identify the true distribution (discriminative), and also be small enough to go beyond memorizing samples (generalizable). In this paper, we show that a discriminator set is guaranteed to be discriminative whenever its linear span is dense in the set of bounded continuous functions. This is a very mild condition satisfied even by neural networks with a single neuron. Further, we develop generalization bounds between the learned distribution and true distribution under different evaluation metrics. When evaluated with neural distance, our bounds show that generalization is guaranteed as long as the discriminator set is small enough, regardless of the size of the generator or hypothesis set. When evaluated with KL divergence, our bound provides an explanation on the counter-intuitive behaviors of testing likelihood in GAN training. Our analysis sheds lights on understanding the practical performance of GANs." @default.
• W2767609405 created "2017-11-17" @default.
• W2767609405 creator A5011131284 @default.
• W2767609405 creator A5048079999 @default.
• W2767609405 creator A5048159038 @default.
• W2767609405 creator A5057055806 @default.
• W2767609405 creator A5059735251 @default.
• W2767609405 date "2017-11-07" @default.
• W2767609405 modified "2023-09-27" @default.
• W2767609405 title "On the Discrimination-Generalization Tradeoff in GANs" @default.
• W2767609405 cites W1524012148 @default.
• W2767609405 cites W1988115241 @default.
• W2767609405 cites W2099057450 @default.
• W2767609405 cites W2103496339 @default.
• W2767609405 cites W2125389028 @default.
• W2767609405 cites W2160354932 @default.
• W2767609405 cites W2178768799 @default.
• W2767609405 cites W2328111639 @default.
• W2767609405 cites W2521028896 @default.
• W2767609405 cites W2548275288 @default.
• W2767609405 cites W2579923771 @default.
• W2767609405 cites W2615429765 @default.
• W2767609405 cites W2617322972 @default.
• W2767609405 cites W2620086128 @default.
• W2767609405 cites W2632772789 @default.
• W2767609405 cites W2687693326 @default.
• W2767609405 cites W2709553318 @default.
• W2767609405 cites W2734870622 @default.
• W2767609405 cites W2756612184 @default.
• W2767609405 cites W2765208948 @default.
• W2767609405 cites W2766665711 @default.
• W2767609405 cites W2782599016 @default.
• W2767609405 cites W2949186514 @default.
• W2767609405 cites W2949995983 @default.
• W2767609405 cites W2950292946 @default.
• W2767609405 cites W2951567597 @default.
• W2767609405 cites W2952434594 @default.
• W2767609405 cites W2952533959 @default.
• W2767609405 cites W2952745707 @default.
• W2767609405 cites W2963534251 @default.
• W2767609405 cites W2963836885 @default.
• W2767609405 cites W3125537303 @default.
• W2767609405 cites W3146803896 @default.
• W2767609405 hasPublicationYear "2017" @default.
• W2767609405 type Work @default.
• W2767609405 sameAs 2767609405 @default.
• W2767609405 citedByCount "14" @default.
• W2767609405 countsByYear W27676094052018 @default.
• W2767609405 countsByYear W27676094052019 @default.
• W2767609405 countsByYear W27676094052020 @default.
• W2767609405 countsByYear W27676094052021 @default.
• W2767609405 crossrefType "posted-content" @default.
• W2767609405 hasAuthorship W2767609405A5011131284 @default.
• W2767609405 hasAuthorship W2767609405A5048079999 @default.
• W2767609405 hasAuthorship W2767609405A5048159038 @default.
• W2767609405 hasAuthorship W2767609405A5057055806 @default.
• W2767609405 hasAuthorship W2767609405A5059735251 @default.
• W2767609405 hasConcept C105795698 @default.
• W2767609405 hasConcept C11413529 @default.
• W2767609405 hasConcept C121332964 @default.
• W2767609405 hasConcept C134306372 @default.
• W2767609405 hasConcept C138885662 @default.
• W2767609405 hasConcept C153180895 @default.
• W2767609405 hasConcept C154945302 @default.
• W2767609405 hasConcept C162324750 @default.
• W2767609405 hasConcept C163258240 @default.
• W2767609405 hasConcept C165064840 @default.
• W2767609405 hasConcept C176217482 @default.
• W2767609405 hasConcept C177148314 @default.
• W2767609405 hasConcept C177264268 @default.
• W2767609405 hasConcept C199360897 @default.
• W2767609405 hasConcept C207390915 @default.
• W2767609405 hasConcept C21547014 @default.
• W2767609405 hasConcept C2779803651 @default.
• W2767609405 hasConcept C2780992000 @default.
• W2767609405 hasConcept C33923547 @default.
• W2767609405 hasConcept C34388435 @default.
• W2767609405 hasConcept C41008148 @default.
• W2767609405 hasConcept C41895202 @default.
• W2767609405 hasConcept C50644808 @default.
• W2767609405 hasConcept C62520636 @default.
• W2767609405 hasConcept C76155785 @default.
• W2767609405 hasConcept C94915269 @default.
• W2767609405 hasConcept C97931131 @default.
• W2767609405 hasConceptScore W2767609405C105795698 @default.
• W2767609405 hasConceptScore W2767609405C11413529 @default.
• W2767609405 hasConceptScore W2767609405C121332964 @default.
• W2767609405 hasConceptScore W2767609405C134306372 @default.
• W2767609405 hasConceptScore W2767609405C138885662 @default.
• W2767609405 hasConceptScore W2767609405C153180895 @default.
• W2767609405 hasConceptScore W2767609405C154945302 @default.
• W2767609405 hasConceptScore W2767609405C162324750 @default.
• W2767609405 hasConceptScore W2767609405C163258240 @default.
• W2767609405 hasConceptScore W2767609405C165064840 @default.
• W2767609405 hasConceptScore W2767609405C176217482 @default.
• W2767609405 hasConceptScore W2767609405C177148314 @default.
• W2767609405 hasConceptScore W2767609405C177264268 @default.
• W2767609405 hasConceptScore W2767609405C199360897 @default.
• W2767609405 hasConceptScore W2767609405C207390915 @default.
• W2767609405 hasConceptScore W2767609405C21547014 @default.

• next