FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3034316848> ?p ?o ?g. }

• W3034316848 abstract "In this paper, we propose a novel loss function for training Generative Adversarial Networks (GANs) aiming towards deeper theoretical understanding as well as improved stability and performance for the underlying optimization problem. The new loss function is based on cumulant generating functions giving rise to emph{Cumulant GAN}. Relying on a recently-derived variational formula, we show that the corresponding optimization problem is equivalent to R{'e}nyi divergence minimization, thus offering a (partially) unified perspective of GAN losses: the R{'e}nyi family encompasses Kullback-Leibler divergence (KLD), reverse KLD, Hellinger distance and $chi^2$-divergence. Wasserstein GAN is also a member of cumulant GAN. In terms of stability, we rigorously prove the linear convergence of cumulant GAN to the Nash equilibrium for a linear discriminator, Gaussian distributions and the standard gradient descent ascent algorithm. Finally, we experimentally demonstrate that image generation is more robust relative to Wasserstein GAN and it is substantially improved in terms of both inception score and Fr'echet inception distance when both weaker and stronger discriminators are considered." @default.
• W3034316848 created "2020-06-19" @default.
• W3034316848 creator A5007544604 @default.
• W3034316848 creator A5010084334 @default.
• W3034316848 creator A5035745788 @default.
• W3034316848 creator A5041343074 @default.
• W3034316848 creator A5045079747 @default.
• W3034316848 date "2020-06-11" @default.
• W3034316848 modified "2023-09-27" @default.
• W3034316848 title "Cumulant GAN." @default.
• W3034316848 cites W1506806321 @default.
• W3034316848 cites W1511694993 @default.
• W3034316848 cites W1530235965 @default.
• W3034316848 cites W1582706191 @default.
• W3034316848 cites W1731081199 @default.
• W3034316848 cites W1975736346 @default.
• W3034316848 cites W2023631060 @default.
• W3034316848 cites W2080844831 @default.
• W3034316848 cites W2099471712 @default.
• W3034316848 cites W2101460669 @default.
• W3034316848 cites W2108598243 @default.
• W3034316848 cites W2125389028 @default.
• W3034316848 cites W2136144249 @default.
• W3034316848 cites W2173520492 @default.
• W3034316848 cites W2405756170 @default.
• W3034316848 cites W2548275288 @default.
• W3034316848 cites W2592298275 @default.
• W3034316848 cites W2593383075 @default.
• W3034316848 cites W2593414223 @default.
• W3034316848 cites W2609000278 @default.
• W3034316848 cites W2739748921 @default.
• W3034316848 cites W2765811365 @default.
• W3034316848 cites W2784823820 @default.
• W3034316848 cites W2787887017 @default.
• W3034316848 cites W2803832867 @default.
• W3034316848 cites W2804078698 @default.
• W3034316848 cites W2893749619 @default.
• W3034316848 cites W2962760235 @default.
• W3034316848 cites W2962770929 @default.
• W3034316848 cites W2962793481 @default.
• W3034316848 cites W2962808524 @default.
• W3034316848 cites W2962879692 @default.
• W3034316848 cites W2962892300 @default.
• W3034316848 cites W2963226019 @default.
• W3034316848 cites W2963275229 @default.
• W3034316848 cites W2963341071 @default.
• W3034316848 cites W2963373786 @default.
• W3034316848 cites W2963470893 @default.
• W3034316848 cites W2963759009 @default.
• W3034316848 cites W2963800509 @default.
• W3034316848 cites W2963809785 @default.
• W3034316848 cites W2963836885 @default.
• W3034316848 cites W2963959558 @default.
• W3034316848 cites W2963971656 @default.
• W3034316848 cites W2963981733 @default.
• W3034316848 cites W2964067523 @default.
• W3034316848 cites W2964318004 @default.
• W3034316848 cites W2964334726 @default.
• W3034316848 cites W2970006822 @default.
• W3034316848 cites W2984365426 @default.
• W3034316848 cites W2989644205 @default.
• W3034316848 cites W3002944878 @default.
• W3034316848 cites W3034429116 @default.
• W3034316848 cites W3037593317 @default.
• W3034316848 cites W3037695135 @default.
• W3034316848 cites W3103267572 @default.
• W3034316848 cites W3118608800 @default.
• W3034316848 cites W3127284993 @default.
• W3034316848 cites W3198609073 @default.
• W3034316848 cites W3202864205 @default.
• W3034316848 cites W610242975 @default.
• W3034316848 cites W648143168 @default.
• W3034316848 hasPublicationYear "2020" @default.
• W3034316848 type Work @default.
• W3034316848 sameAs 3034316848 @default.
• W3034316848 citedByCount "3" @default.
• W3034316848 countsByYear W30343168482020 @default.
• W3034316848 crossrefType "posted-content" @default.
• W3034316848 hasAuthorship W3034316848A5007544604 @default.
• W3034316848 hasAuthorship W3034316848A5010084334 @default.
• W3034316848 hasAuthorship W3034316848A5035745788 @default.
• W3034316848 hasAuthorship W3034316848A5041343074 @default.
• W3034316848 hasAuthorship W3034316848A5045079747 @default.
• W3034316848 hasConcept C105795698 @default.
• W3034316848 hasConcept C112972136 @default.
• W3034316848 hasConcept C119857082 @default.
• W3034316848 hasConcept C121332964 @default.
• W3034316848 hasConcept C126255220 @default.
• W3034316848 hasConcept C138885662 @default.
• W3034316848 hasConcept C14036430 @default.
• W3034316848 hasConcept C147764199 @default.
• W3034316848 hasConcept C162324750 @default.
• W3034316848 hasConcept C163716315 @default.
• W3034316848 hasConcept C171752962 @default.
• W3034316848 hasConcept C172686274 @default.
• W3034316848 hasConcept C207390915 @default.
• W3034316848 hasConcept C2777303404 @default.
• W3034316848 hasConcept C2779803651 @default.
• W3034316848 hasConcept C28826006 @default.
• W3034316848 hasConcept C33923547 @default.

• next