Matches in SemOpenAlex for { <https://semopenalex.org/work/W3202380293> ?p ?o ?g. }
- W3202380293 abstract "Many modern machine learning algorithms such as generative adversarial networks (GANs) and adversarial training can be formulated as minimax optimization. Gradient descent ascent (GDA) is the most commonly used algorithm due to its simplicity. However, GDA can converge to non-optimal minimax points. We propose a new minimax optimization framework, GDA-AM, that views the GDAdynamics as a fixed-point iteration and solves it using Anderson Mixing to con-verge to the local minimax. It addresses the diverging issue of simultaneous GDAand accelerates the convergence of alternating GDA. We show theoretically that the algorithm can achieve global convergence for bilinear problems under mild conditions. We also empirically show that GDA-AMsolves a variety of minimax problems and improves GAN training on several datasets" @default.
- W3202380293 created "2021-10-11" @default.
- W3202380293 creator A5013043483 @default.
- W3202380293 creator A5022272635 @default.
- W3202380293 creator A5046548026 @default.
- W3202380293 creator A5056635728 @default.
- W3202380293 creator A5079737809 @default.
- W3202380293 date "2021-10-06" @default.
- W3202380293 modified "2023-09-27" @default.
- W3202380293 title "Solve Minimax Optimization by Anderson Acceleration" @default.
- W3202380293 cites W1506342804 @default.
- W3202380293 cites W1540637571 @default.
- W3202380293 cites W1550543785 @default.
- W3202380293 cites W1560263223 @default.
- W3202380293 cites W1834627138 @default.
- W3202380293 cites W1966801132 @default.
- W3202380293 cites W2038281434 @default.
- W3202380293 cites W2052769657 @default.
- W3202380293 cites W2058784008 @default.
- W3202380293 cites W2065176245 @default.
- W3202380293 cites W2081985785 @default.
- W3202380293 cites W2089559088 @default.
- W3202380293 cites W2099471712 @default.
- W3202380293 cites W2100514507 @default.
- W3202380293 cites W2140153041 @default.
- W3202380293 cites W2144846366 @default.
- W3202380293 cites W2619503996 @default.
- W3202380293 cites W2894677249 @default.
- W3202380293 cites W2897413842 @default.
- W3202380293 cites W2904455790 @default.
- W3202380293 cites W2907669392 @default.
- W3202380293 cites W2913266441 @default.
- W3202380293 cites W2947212281 @default.
- W3202380293 cites W2962879692 @default.
- W3202380293 cites W2962981216 @default.
- W3202380293 cites W2963207607 @default.
- W3202380293 cites W2963373786 @default.
- W3202380293 cites W2963783084 @default.
- W3202380293 cites W2963836885 @default.
- W3202380293 cites W2963981733 @default.
- W3202380293 cites W2964249351 @default.
- W3202380293 cites W2964253222 @default.
- W3202380293 cites W2970552792 @default.
- W3202380293 cites W2971138263 @default.
- W3202380293 cites W2971302301 @default.
- W3202380293 cites W2995291884 @default.
- W3202380293 cites W3007729382 @default.
- W3202380293 cites W3033192442 @default.
- W3202380293 cites W3034361719 @default.
- W3202380293 cites W3034862928 @default.
- W3202380293 cites W3035665052 @default.
- W3202380293 cites W3037083547 @default.
- W3202380293 cites W3037593317 @default.
- W3202380293 cites W3095124371 @default.
- W3202380293 cites W3101511375 @default.
- W3202380293 cites W3105950517 @default.
- W3202380293 cites W3118608800 @default.
- W3202380293 cites W3131996280 @default.
- W3202380293 cites W3133276906 @default.
- W3202380293 cites W3158392704 @default.
- W3202380293 cites W3159062549 @default.
- W3202380293 cites W3211848225 @default.
- W3202380293 hasPublicationYear "2021" @default.
- W3202380293 type Work @default.
- W3202380293 sameAs 3202380293 @default.
- W3202380293 citedByCount "0" @default.
- W3202380293 crossrefType "posted-content" @default.
- W3202380293 hasAuthorship W3202380293A5013043483 @default.
- W3202380293 hasAuthorship W3202380293A5022272635 @default.
- W3202380293 hasAuthorship W3202380293A5046548026 @default.
- W3202380293 hasAuthorship W3202380293A5056635728 @default.
- W3202380293 hasAuthorship W3202380293A5079737809 @default.
- W3202380293 hasConcept C117148685 @default.
- W3202380293 hasConcept C126255220 @default.
- W3202380293 hasConcept C136197465 @default.
- W3202380293 hasConcept C137836250 @default.
- W3202380293 hasConcept C149728462 @default.
- W3202380293 hasConcept C153258448 @default.
- W3202380293 hasConcept C154945302 @default.
- W3202380293 hasConcept C162324750 @default.
- W3202380293 hasConcept C205203396 @default.
- W3202380293 hasConcept C2524010 @default.
- W3202380293 hasConcept C2777303404 @default.
- W3202380293 hasConcept C28719098 @default.
- W3202380293 hasConcept C31972630 @default.
- W3202380293 hasConcept C33923547 @default.
- W3202380293 hasConcept C41008148 @default.
- W3202380293 hasConcept C50522688 @default.
- W3202380293 hasConcept C50644808 @default.
- W3202380293 hasConceptScore W3202380293C117148685 @default.
- W3202380293 hasConceptScore W3202380293C126255220 @default.
- W3202380293 hasConceptScore W3202380293C136197465 @default.
- W3202380293 hasConceptScore W3202380293C137836250 @default.
- W3202380293 hasConceptScore W3202380293C149728462 @default.
- W3202380293 hasConceptScore W3202380293C153258448 @default.
- W3202380293 hasConceptScore W3202380293C154945302 @default.
- W3202380293 hasConceptScore W3202380293C162324750 @default.
- W3202380293 hasConceptScore W3202380293C205203396 @default.
- W3202380293 hasConceptScore W3202380293C2524010 @default.
- W3202380293 hasConceptScore W3202380293C2777303404 @default.