Matches in SemOpenAlex for { <https://semopenalex.org/work/W3100889922> ?p ?o ?g. }
Showing items 1 to 87 of
87
with 100 items per page.
- W3100889922 endingPage "11070" @default.
- W3100889922 startingPage "11056" @default.
- W3100889922 abstract "Generative Adversarial Networks (GANs) are a powerful class of generative models in the deep learning community. Current practice on large-scale GAN training utilizes large models and distributed large-batch training strategies, and is implemented on deep learning frameworks (e.g., TensorFlow, PyTorch, etc.) designed in a centralized manner. In the centralized network topology, every worker needs to either directly communicate with the central node or indirectly communicate with all other workers in every iteration. However, when the network bandwidth is low or network latency is high, the performance would be significantly degraded. Despite recent progress on decentralized algorithms for training deep neural networks, it remains unclear whether it is possible to train GANs in a decentralized manner. The main difficulty lies at handling the nonconvex-nonconcave min-max optimization and the decentralized communication simultaneously. In this paper, we address this difficulty by designing the textbf{first gradient-based decentralized parallel algorithm} which allows workers to have multiple rounds of communications in one iteration and to update the discriminator and generator simultaneously, and this design makes it amenable for the convergence analysis of the proposed decentralized algorithm. Theoretically, our proposed decentralized algorithm is able to solve a class of non-convex non-concave min-max problems with provable non-asymptotic convergence to first-order stationary point. Experimental results on GANs demonstrate the effectiveness of the proposed algorithm." @default.
- W3100889922 created "2020-11-23" @default.
- W3100889922 creator A5023288846 @default.
- W3100889922 creator A5064335115 @default.
- W3100889922 creator A5064437230 @default.
- W3100889922 creator A5064636177 @default.
- W3100889922 creator A5065510803 @default.
- W3100889922 creator A5089852582 @default.
- W3100889922 creator A5090515876 @default.
- W3100889922 date "2020-01-01" @default.
- W3100889922 modified "2023-10-16" @default.
- W3100889922 title "A Decentralized Parallel Algorithm for Training Generative Adversarial Nets" @default.
- W3100889922 hasPublicationYear "2020" @default.
- W3100889922 type Work @default.
- W3100889922 sameAs 3100889922 @default.
- W3100889922 citedByCount "4" @default.
- W3100889922 countsByYear W31008899222021 @default.
- W3100889922 crossrefType "proceedings-article" @default.
- W3100889922 hasAuthorship W3100889922A5023288846 @default.
- W3100889922 hasAuthorship W3100889922A5064335115 @default.
- W3100889922 hasAuthorship W3100889922A5064437230 @default.
- W3100889922 hasAuthorship W3100889922A5064636177 @default.
- W3100889922 hasAuthorship W3100889922A5065510803 @default.
- W3100889922 hasAuthorship W3100889922A5089852582 @default.
- W3100889922 hasAuthorship W3100889922A5090515876 @default.
- W3100889922 hasConcept C11413529 @default.
- W3100889922 hasConcept C120314980 @default.
- W3100889922 hasConcept C126255220 @default.
- W3100889922 hasConcept C154945302 @default.
- W3100889922 hasConcept C162324750 @default.
- W3100889922 hasConcept C2776257435 @default.
- W3100889922 hasConcept C2777303404 @default.
- W3100889922 hasConcept C2779803651 @default.
- W3100889922 hasConcept C31258907 @default.
- W3100889922 hasConcept C33923547 @default.
- W3100889922 hasConcept C37736160 @default.
- W3100889922 hasConcept C41008148 @default.
- W3100889922 hasConcept C50522688 @default.
- W3100889922 hasConcept C50644808 @default.
- W3100889922 hasConcept C76155785 @default.
- W3100889922 hasConcept C94915269 @default.
- W3100889922 hasConceptScore W3100889922C11413529 @default.
- W3100889922 hasConceptScore W3100889922C120314980 @default.
- W3100889922 hasConceptScore W3100889922C126255220 @default.
- W3100889922 hasConceptScore W3100889922C154945302 @default.
- W3100889922 hasConceptScore W3100889922C162324750 @default.
- W3100889922 hasConceptScore W3100889922C2776257435 @default.
- W3100889922 hasConceptScore W3100889922C2777303404 @default.
- W3100889922 hasConceptScore W3100889922C2779803651 @default.
- W3100889922 hasConceptScore W3100889922C31258907 @default.
- W3100889922 hasConceptScore W3100889922C33923547 @default.
- W3100889922 hasConceptScore W3100889922C37736160 @default.
- W3100889922 hasConceptScore W3100889922C41008148 @default.
- W3100889922 hasConceptScore W3100889922C50522688 @default.
- W3100889922 hasConceptScore W3100889922C50644808 @default.
- W3100889922 hasConceptScore W3100889922C76155785 @default.
- W3100889922 hasConceptScore W3100889922C94915269 @default.
- W3100889922 hasLocation W31008899221 @default.
- W3100889922 hasOpenAccess W3100889922 @default.
- W3100889922 hasPrimaryLocation W31008899221 @default.
- W3100889922 hasRelatedWork W2099471712 @default.
- W3100889922 hasRelatedWork W2295391514 @default.
- W3100889922 hasRelatedWork W2889676205 @default.
- W3100889922 hasRelatedWork W2903144274 @default.
- W3100889922 hasRelatedWork W2950661436 @default.
- W3100889922 hasRelatedWork W2963855915 @default.
- W3100889922 hasRelatedWork W2982523334 @default.
- W3100889922 hasRelatedWork W3021874845 @default.
- W3100889922 hasRelatedWork W3026711567 @default.
- W3100889922 hasRelatedWork W3030545591 @default.
- W3100889922 hasRelatedWork W3089960935 @default.
- W3100889922 hasRelatedWork W3090221383 @default.
- W3100889922 hasRelatedWork W3102822772 @default.
- W3100889922 hasRelatedWork W3137967019 @default.
- W3100889922 hasRelatedWork W3153703270 @default.
- W3100889922 hasRelatedWork W3162760367 @default.
- W3100889922 hasRelatedWork W3172738938 @default.
- W3100889922 hasRelatedWork W3173261493 @default.
- W3100889922 hasRelatedWork W3202869544 @default.
- W3100889922 hasRelatedWork W3090802698 @default.
- W3100889922 hasVolume "33" @default.
- W3100889922 isParatext "false" @default.
- W3100889922 isRetracted "false" @default.
- W3100889922 magId "3100889922" @default.
- W3100889922 workType "article" @default.