Matches in SemOpenAlex for { <https://semopenalex.org/work/W2949277570> ?p ?o ?g. }
- W2949277570 abstract "We present a novel method and analysis to train generative adversarial networks (GAN) in a stable manner. As shown in recent analysis, training is often undermined by the probability distribution of the data being zero on neighborhoods of the data space. We notice that the distributions of real and generated data should match even when they undergo the same filtering. Therefore, to address the limited support problem we propose to train GANs by using different filtered versions of the real and generated data distributions. In this way, filtering does not prevent the exact matching of the data distribution, while helping training by extending the support of both distributions. As filtering we consider adding samples from an arbitrary distribution to the data, which corresponds to a convolution of the data distribution with the arbitrary one. We also propose to learn the generation of these samples so as to challenge the discriminator in the adversarial training. We show that our approach results in a stable and well-behaved training of even the original minimax GAN formulation. Moreover, our technique can be incorporated in most modern GAN formulations and leads to a consistent improvement on several common datasets." @default.
- W2949277570 created "2019-06-27" @default.
- W2949277570 creator A5070714354 @default.
- W2949277570 creator A5070940574 @default.
- W2949277570 date "2019-06-11" @default.
- W2949277570 modified "2023-09-27" @default.
- W2949277570 title "On Stabilizing Generative Adversarial Training with Noise" @default.
- W2949277570 cites W1834627138 @default.
- W2949277570 cites W1836465849 @default.
- W2949277570 cites W2099471712 @default.
- W2949277570 cites W2117539524 @default.
- W2949277570 cites W2118858186 @default.
- W2949277570 cites W2173520492 @default.
- W2949277570 cites W2593414223 @default.
- W2949277570 cites W2605195953 @default.
- W2949277570 cites W2612866063 @default.
- W2949277570 cites W2617573776 @default.
- W2949277570 cites W2619459889 @default.
- W2949277570 cites W2739748921 @default.
- W2949277570 cites W2950560720 @default.
- W2949277570 cites W2962879692 @default.
- W2949277570 cites W2963373786 @default.
- W2949277570 cites W2963625494 @default.
- W2949277570 cites W2963836885 @default.
- W2949277570 cites W2963981733 @default.
- W2949277570 cites W2964121744 @default.
- W2949277570 cites W2964201867 @default.
- W2949277570 cites W3118608800 @default.
- W2949277570 cites W967544008 @default.
- W2949277570 hasPublicationYear "2019" @default.
- W2949277570 type Work @default.
- W2949277570 sameAs 2949277570 @default.
- W2949277570 citedByCount "0" @default.
- W2949277570 crossrefType "posted-content" @default.
- W2949277570 hasAuthorship W2949277570A5070714354 @default.
- W2949277570 hasAuthorship W2949277570A5070940574 @default.
- W2949277570 hasConcept C105795698 @default.
- W2949277570 hasConcept C110121322 @default.
- W2949277570 hasConcept C111919701 @default.
- W2949277570 hasConcept C112972136 @default.
- W2949277570 hasConcept C11413529 @default.
- W2949277570 hasConcept C115961682 @default.
- W2949277570 hasConcept C119857082 @default.
- W2949277570 hasConcept C121332964 @default.
- W2949277570 hasConcept C126255220 @default.
- W2949277570 hasConcept C134306372 @default.
- W2949277570 hasConcept C149728462 @default.
- W2949277570 hasConcept C153294291 @default.
- W2949277570 hasConcept C154945302 @default.
- W2949277570 hasConcept C165064840 @default.
- W2949277570 hasConcept C2777211547 @default.
- W2949277570 hasConcept C2778572836 @default.
- W2949277570 hasConcept C2779803651 @default.
- W2949277570 hasConcept C33923547 @default.
- W2949277570 hasConcept C37736160 @default.
- W2949277570 hasConcept C39890363 @default.
- W2949277570 hasConcept C41008148 @default.
- W2949277570 hasConcept C45347329 @default.
- W2949277570 hasConcept C50644808 @default.
- W2949277570 hasConcept C51632099 @default.
- W2949277570 hasConcept C76155785 @default.
- W2949277570 hasConcept C94915269 @default.
- W2949277570 hasConcept C99498987 @default.
- W2949277570 hasConceptScore W2949277570C105795698 @default.
- W2949277570 hasConceptScore W2949277570C110121322 @default.
- W2949277570 hasConceptScore W2949277570C111919701 @default.
- W2949277570 hasConceptScore W2949277570C112972136 @default.
- W2949277570 hasConceptScore W2949277570C11413529 @default.
- W2949277570 hasConceptScore W2949277570C115961682 @default.
- W2949277570 hasConceptScore W2949277570C119857082 @default.
- W2949277570 hasConceptScore W2949277570C121332964 @default.
- W2949277570 hasConceptScore W2949277570C126255220 @default.
- W2949277570 hasConceptScore W2949277570C134306372 @default.
- W2949277570 hasConceptScore W2949277570C149728462 @default.
- W2949277570 hasConceptScore W2949277570C153294291 @default.
- W2949277570 hasConceptScore W2949277570C154945302 @default.
- W2949277570 hasConceptScore W2949277570C165064840 @default.
- W2949277570 hasConceptScore W2949277570C2777211547 @default.
- W2949277570 hasConceptScore W2949277570C2778572836 @default.
- W2949277570 hasConceptScore W2949277570C2779803651 @default.
- W2949277570 hasConceptScore W2949277570C33923547 @default.
- W2949277570 hasConceptScore W2949277570C37736160 @default.
- W2949277570 hasConceptScore W2949277570C39890363 @default.
- W2949277570 hasConceptScore W2949277570C41008148 @default.
- W2949277570 hasConceptScore W2949277570C45347329 @default.
- W2949277570 hasConceptScore W2949277570C50644808 @default.
- W2949277570 hasConceptScore W2949277570C51632099 @default.
- W2949277570 hasConceptScore W2949277570C76155785 @default.
- W2949277570 hasConceptScore W2949277570C94915269 @default.
- W2949277570 hasConceptScore W2949277570C99498987 @default.
- W2949277570 hasLocation W29492775701 @default.
- W2949277570 hasOpenAccess W2949277570 @default.
- W2949277570 hasPrimaryLocation W29492775701 @default.
- W2949277570 hasRelatedWork W2419788668 @default.
- W2949277570 hasRelatedWork W2768915615 @default.
- W2949277570 hasRelatedWork W2894770203 @default.
- W2949277570 hasRelatedWork W2900122995 @default.
- W2949277570 hasRelatedWork W2900151392 @default.
- W2949277570 hasRelatedWork W2912023003 @default.
- W2949277570 hasRelatedWork W2913881544 @default.