Matches in SemOpenAlex for { <https://semopenalex.org/work/W2619503996> ?p ?o ?g. }
- W2619503996 endingPage "1833" @default.
- W2619503996 startingPage "1823" @default.
- W2619503996 abstract "In this paper, we analyze the numerics of common algorithms for training Generative Adversarial Networks (GANs). Using the formalism of smooth two-player games we analyze the associated gradient vector field of GAN training objectives. Our findings suggest that the convergence of current algorithms suffers due to two factors: i) presence of eigenvalues of the Jacobian of the gradient vector field with zero real-part, and ii) eigenvalues with big imaginary part. Using these findings, we design a new algorithm that overcomes some of these limitations and has better convergence properties. Experimentally, we demonstrate its superiority on training common GAN architectures and show convergence on GAN architectures that are known to be notoriously hard to train." @default.
- W2619503996 created "2017-06-05" @default.
- W2619503996 creator A5000291599 @default.
- W2619503996 creator A5016606943 @default.
- W2619503996 creator A5055590646 @default.
- W2619503996 date "2017-12-04" @default.
- W2619503996 modified "2023-10-05" @default.
- W2619503996 title "The numerics of GANs" @default.
- W2619503996 cites W1543750907 @default.
- W2619503996 cites W1669104078 @default.
- W2619503996 cites W1970789124 @default.
- W2619503996 cites W2173520492 @default.
- W2619503996 cites W2271840356 @default.
- W2619503996 cites W2411541852 @default.
- W2619503996 cites W2412320034 @default.
- W2619503996 cites W2419501139 @default.
- W2619503996 cites W2432004435 @default.
- W2619503996 cites W2479644247 @default.
- W2619503996 cites W2523714292 @default.
- W2619503996 cites W2527819024 @default.
- W2619503996 cites W2548275288 @default.
- W2619503996 cites W2552465644 @default.
- W2619503996 cites W2554314924 @default.
- W2619503996 cites W2605135824 @default.
- W2619503996 cites W2949987290 @default.
- W2619503996 cites W2950560720 @default.
- W2619503996 cites W2952673310 @default.
- W2619503996 cites W2952745707 @default.
- W2619503996 cites W2962936867 @default.
- W2619503996 cites W2964201867 @default.
- W2619503996 hasPublicationYear "2017" @default.
- W2619503996 type Work @default.
- W2619503996 sameAs 2619503996 @default.
- W2619503996 citedByCount "174" @default.
- W2619503996 countsByYear W26195039962017 @default.
- W2619503996 countsByYear W26195039962018 @default.
- W2619503996 countsByYear W26195039962019 @default.
- W2619503996 countsByYear W26195039962020 @default.
- W2619503996 countsByYear W26195039962021 @default.
- W2619503996 crossrefType "proceedings-article" @default.
- W2619503996 hasAuthorship W2619503996A5000291599 @default.
- W2619503996 hasAuthorship W2619503996A5016606943 @default.
- W2619503996 hasAuthorship W2619503996A5055590646 @default.
- W2619503996 hasConcept C11413529 @default.
- W2619503996 hasConcept C121332964 @default.
- W2619503996 hasConcept C142362112 @default.
- W2619503996 hasConcept C153349607 @default.
- W2619503996 hasConcept C154945302 @default.
- W2619503996 hasConcept C158693339 @default.
- W2619503996 hasConcept C162324750 @default.
- W2619503996 hasConcept C200331156 @default.
- W2619503996 hasConcept C202444582 @default.
- W2619503996 hasConcept C2524010 @default.
- W2619503996 hasConcept C2777303404 @default.
- W2619503996 hasConcept C28826006 @default.
- W2619503996 hasConcept C33923547 @default.
- W2619503996 hasConcept C37736160 @default.
- W2619503996 hasConcept C39890363 @default.
- W2619503996 hasConcept C41008148 @default.
- W2619503996 hasConcept C50522688 @default.
- W2619503996 hasConcept C558565934 @default.
- W2619503996 hasConcept C62520636 @default.
- W2619503996 hasConcept C73301696 @default.
- W2619503996 hasConcept C80444323 @default.
- W2619503996 hasConcept C91188154 @default.
- W2619503996 hasConcept C9652623 @default.
- W2619503996 hasConceptScore W2619503996C11413529 @default.
- W2619503996 hasConceptScore W2619503996C121332964 @default.
- W2619503996 hasConceptScore W2619503996C142362112 @default.
- W2619503996 hasConceptScore W2619503996C153349607 @default.
- W2619503996 hasConceptScore W2619503996C154945302 @default.
- W2619503996 hasConceptScore W2619503996C158693339 @default.
- W2619503996 hasConceptScore W2619503996C162324750 @default.
- W2619503996 hasConceptScore W2619503996C200331156 @default.
- W2619503996 hasConceptScore W2619503996C202444582 @default.
- W2619503996 hasConceptScore W2619503996C2524010 @default.
- W2619503996 hasConceptScore W2619503996C2777303404 @default.
- W2619503996 hasConceptScore W2619503996C28826006 @default.
- W2619503996 hasConceptScore W2619503996C33923547 @default.
- W2619503996 hasConceptScore W2619503996C37736160 @default.
- W2619503996 hasConceptScore W2619503996C39890363 @default.
- W2619503996 hasConceptScore W2619503996C41008148 @default.
- W2619503996 hasConceptScore W2619503996C50522688 @default.
- W2619503996 hasConceptScore W2619503996C558565934 @default.
- W2619503996 hasConceptScore W2619503996C62520636 @default.
- W2619503996 hasConceptScore W2619503996C73301696 @default.
- W2619503996 hasConceptScore W2619503996C80444323 @default.
- W2619503996 hasConceptScore W2619503996C91188154 @default.
- W2619503996 hasConceptScore W2619503996C9652623 @default.
- W2619503996 hasOpenAccess W2619503996 @default.
- W2619503996 hasRelatedWork W1519983590 @default.
- W2619503996 hasRelatedWork W2099471712 @default.
- W2619503996 hasRelatedWork W2173520492 @default.
- W2619503996 hasRelatedWork W2554314924 @default.
- W2619503996 hasRelatedWork W2593414223 @default.
- W2619503996 hasRelatedWork W2617573776 @default.
- W2619503996 hasRelatedWork W2739748921 @default.
- W2619503996 hasRelatedWork W2792263949 @default.