FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3204992908> ?p ?o ?g. }

• W3204992908 endingPage "1" @default.
• W3204992908 startingPage "1" @default.
• W3204992908 abstract "Entropy regularized optimal transport (EOT) distance and its symmetric normalization, known as the Sinkhorn divergence, offer smooth and continuous metrized weak-convergence distance metrics. They have excellent geometric properties and are useful to compare probability distributions in some generative adversarial network (GAN) models. Computing them using the original Sinkhorn matrix scaling algorithm is still expensive. The running time is quadratic at $mathcal {O}(n^{2})$ in the size $n$ of the training dataset. This work investigates the problem of accelerating the GAN training when Sinkhorn divergence is used as a minimax objective. Let $mathcal {G}$ be a Gaussian map from the ground space onto the positive orthant $mathbb {R}_{+}^{r}$ with $r ll n $ . To speed up the divergence computation, we propose the use of $c(x,y)= - varepsilon log left langle{ mathcal {G}(x),mathcal {G}(y) }right rangle $ as the ground cost. This approximation, known as Sinkhorn with positive features, brings down the running time of the Sinkhorn matrix scaling algorithm to $mathcal {O}(r , n)$ , which is linear in $n$ . To solve the minimax optimization in GAN, we put forward a more efficient simultaneous stochastic gradient descent-ascent (SimSGDA) algorithm in place of the standard sequential gradient techniques. Empirical evidence shows that our model, trained using SimSGDA on the DCGAN neural architecture on tiny-coloured Cats and CelebA datasets, converges to stationary points. These are the local Nash equilibrium points. We carried out numerical experiments to confirm that our model is computationally stable. It generates samples of comparable quality to those produced by prior Sinkhorn and Wasserstein GANs. Further simulations, assessed on the similarity index measures (SSIM), show that our model’s empirical convergence rate is comparable to that of WGAN-GP." @default.
• W3204992908 created "2021-10-25" @default.
• W3204992908 creator A5011264353 @default.
• W3204992908 creator A5026666126 @default.
• W3204992908 creator A5032159254 @default.
• W3204992908 creator A5047887538 @default.
• W3204992908 creator A5056376836 @default.
• W3204992908 creator A5061073672 @default.
• W3204992908 date "2021-01-01" @default.
• W3204992908 modified "2023-09-27" @default.
• W3204992908 title "Learning GANs in Simultaneous Game Using Sinkhorn with Positive Features" @default.
• W3204992908 cites W2020123437 @default.
• W3204992908 cites W2075567596 @default.
• W3204992908 cites W2089559088 @default.
• W3204992908 cites W2099471712 @default.
• W3204992908 cites W2133665775 @default.
• W3204992908 cites W2158131535 @default.
• W3204992908 cites W2619503996 @default.
• W3204992908 cites W2739748921 @default.
• W3204992908 cites W2803117117 @default.
• W3204992908 cites W2895248218 @default.
• W3204992908 cites W2899792124 @default.
• W3204992908 cites W2962717811 @default.
• W3204992908 cites W2962723467 @default.
• W3204992908 cites W2962879692 @default.
• W3204992908 cites W2963326510 @default.
• W3204992908 cites W2963399222 @default.
• W3204992908 cites W2963506485 @default.
• W3204992908 cites W2963647223 @default.
• W3204992908 cites W2963809785 @default.
• W3204992908 cites W2963836885 @default.
• W3204992908 cites W2963989027 @default.
• W3204992908 cites W3004229103 @default.
• W3204992908 cites W3034862928 @default.
• W3204992908 cites W3035295418 @default.
• W3204992908 cites W3035665052 @default.
• W3204992908 cites W3039256091 @default.
• W3204992908 cites W3046366290 @default.
• W3204992908 cites W3093545757 @default.
• W3204992908 cites W3094535871 @default.
• W3204992908 cites W3102741928 @default.
• W3204992908 cites W3105411152 @default.
• W3204992908 cites W3157424262 @default.
• W3204992908 cites W385466589 @default.
• W3204992908 doi "https://doi.org/10.1109/access.2021.3120128" @default.
• W3204992908 hasPublicationYear "2021" @default.
• W3204992908 type Work @default.
• W3204992908 sameAs 3204992908 @default.
• W3204992908 citedByCount "0" @default.
• W3204992908 crossrefType "journal-article" @default.
• W3204992908 hasAuthorship W3204992908A5011264353 @default.
• W3204992908 hasAuthorship W3204992908A5026666126 @default.
• W3204992908 hasAuthorship W3204992908A5032159254 @default.
• W3204992908 hasAuthorship W3204992908A5047887538 @default.
• W3204992908 hasAuthorship W3204992908A5056376836 @default.
• W3204992908 hasAuthorship W3204992908A5061073672 @default.
• W3204992908 hasBestOaLocation W32049929081 @default.
• W3204992908 hasConcept C114614502 @default.
• W3204992908 hasConcept C118615104 @default.
• W3204992908 hasConcept C126255220 @default.
• W3204992908 hasConcept C149728462 @default.
• W3204992908 hasConcept C153258448 @default.
• W3204992908 hasConcept C154945302 @default.
• W3204992908 hasConcept C2524010 @default.
• W3204992908 hasConcept C2780016784 @default.
• W3204992908 hasConcept C33923547 @default.
• W3204992908 hasConcept C41008148 @default.
• W3204992908 hasConcept C50644808 @default.
• W3204992908 hasConcept C99844830 @default.
• W3204992908 hasConceptScore W3204992908C114614502 @default.
• W3204992908 hasConceptScore W3204992908C118615104 @default.
• W3204992908 hasConceptScore W3204992908C126255220 @default.
• W3204992908 hasConceptScore W3204992908C149728462 @default.
• W3204992908 hasConceptScore W3204992908C153258448 @default.
• W3204992908 hasConceptScore W3204992908C154945302 @default.
• W3204992908 hasConceptScore W3204992908C2524010 @default.
• W3204992908 hasConceptScore W3204992908C2780016784 @default.
• W3204992908 hasConceptScore W3204992908C33923547 @default.
• W3204992908 hasConceptScore W3204992908C41008148 @default.
• W3204992908 hasConceptScore W3204992908C50644808 @default.
• W3204992908 hasConceptScore W3204992908C99844830 @default.
• W3204992908 hasLocation W32049929081 @default.
• W3204992908 hasLocation W32049929082 @default.
• W3204992908 hasOpenAccess W3204992908 @default.
• W3204992908 hasPrimaryLocation W32049929081 @default.
• W3204992908 hasRelatedWork W1103850 @default.
• W3204992908 hasRelatedWork W13766486 @default.
• W3204992908 hasRelatedWork W25942946 @default.
• W3204992908 hasRelatedWork W42550238 @default.
• W3204992908 hasRelatedWork W45864672 @default.
• W3204992908 hasRelatedWork W53502914 @default.
• W3204992908 hasRelatedWork W5664448 @default.
• W3204992908 hasRelatedWork W65005310 @default.
• W3204992908 hasRelatedWork W8047569 @default.
• W3204992908 hasRelatedWork W808143 @default.
• W3204992908 isParatext "false" @default.
• W3204992908 isRetracted "false" @default.
• W3204992908 magId "3204992908" @default.

• next