Matches in SemOpenAlex for { <https://semopenalex.org/work/W3006839751> ?p ?o ?g. }
Showing items 1 to 96 of
96
with 100 items per page.
- W3006839751 abstract "Stochastic regularization of neural networks (e.g. dropout) is a wide-spread technique in deep learning that allows for better generalization. Despite its success, continuous-time models, such as neural ordinary differential equation (ODE), usually rely on a completely deterministic feed-forward operation. This work provides an empirical study of stochastically regularized neural ODE on several image-classification tasks (CIFAR-10, CIFAR-100, TinyImageNet). Building upon the formalism of stochastic differential equations (SDEs), we demonstrate that neural SDE is able to outperform its deterministic counterpart. Further, we show that data augmentation during the training improves the performance of both deterministic and stochastic versions of the same model. However, the improvements obtained by the data augmentation completely eliminate the empirical gains of the stochastic regularization, making the difference in the performance of neural ODE and neural SDE negligible." @default.
- W3006839751 created "2020-03-06" @default.
- W3006839751 creator A5004765243 @default.
- W3006839751 creator A5028940329 @default.
- W3006839751 creator A5069113462 @default.
- W3006839751 date "2020-02-22" @default.
- W3006839751 modified "2023-09-27" @default.
- W3006839751 title "Stochasticity in Neural ODEs: An Empirical Study" @default.
- W3006839751 cites W1503398984 @default.
- W3006839751 cites W2194775991 @default.
- W3006839751 cites W2895434480 @default.
- W3006839751 cites W2899771611 @default.
- W3006839751 cites W2918788260 @default.
- W3006839751 cites W2930930439 @default.
- W3006839751 cites W2945959973 @default.
- W3006839751 cites W2948977392 @default.
- W3006839751 cites W2949117887 @default.
- W3006839751 cites W2963755523 @default.
- W3006839751 cites W2970181183 @default.
- W3006839751 cites W2971106883 @default.
- W3006839751 cites W2971343405 @default.
- W3006839751 cites W2996897098 @default.
- W3006839751 cites W3118608800 @default.
- W3006839751 hasPublicationYear "2020" @default.
- W3006839751 type Work @default.
- W3006839751 sameAs 3006839751 @default.
- W3006839751 citedByCount "5" @default.
- W3006839751 countsByYear W30068397512020 @default.
- W3006839751 countsByYear W30068397512021 @default.
- W3006839751 crossrefType "posted-content" @default.
- W3006839751 hasAuthorship W3006839751A5004765243 @default.
- W3006839751 hasAuthorship W3006839751A5028940329 @default.
- W3006839751 hasAuthorship W3006839751A5069113462 @default.
- W3006839751 hasConcept C134306372 @default.
- W3006839751 hasConcept C142362112 @default.
- W3006839751 hasConcept C147168706 @default.
- W3006839751 hasConcept C153349607 @default.
- W3006839751 hasConcept C154945302 @default.
- W3006839751 hasConcept C177148314 @default.
- W3006839751 hasConcept C2776135515 @default.
- W3006839751 hasConcept C28826006 @default.
- W3006839751 hasConcept C33923547 @default.
- W3006839751 hasConcept C34862557 @default.
- W3006839751 hasConcept C41008148 @default.
- W3006839751 hasConcept C50644808 @default.
- W3006839751 hasConcept C51544822 @default.
- W3006839751 hasConcept C51955184 @default.
- W3006839751 hasConcept C558565934 @default.
- W3006839751 hasConcept C73301696 @default.
- W3006839751 hasConcept C78045399 @default.
- W3006839751 hasConcept C86582703 @default.
- W3006839751 hasConceptScore W3006839751C134306372 @default.
- W3006839751 hasConceptScore W3006839751C142362112 @default.
- W3006839751 hasConceptScore W3006839751C147168706 @default.
- W3006839751 hasConceptScore W3006839751C153349607 @default.
- W3006839751 hasConceptScore W3006839751C154945302 @default.
- W3006839751 hasConceptScore W3006839751C177148314 @default.
- W3006839751 hasConceptScore W3006839751C2776135515 @default.
- W3006839751 hasConceptScore W3006839751C28826006 @default.
- W3006839751 hasConceptScore W3006839751C33923547 @default.
- W3006839751 hasConceptScore W3006839751C34862557 @default.
- W3006839751 hasConceptScore W3006839751C41008148 @default.
- W3006839751 hasConceptScore W3006839751C50644808 @default.
- W3006839751 hasConceptScore W3006839751C51544822 @default.
- W3006839751 hasConceptScore W3006839751C51955184 @default.
- W3006839751 hasConceptScore W3006839751C558565934 @default.
- W3006839751 hasConceptScore W3006839751C73301696 @default.
- W3006839751 hasConceptScore W3006839751C78045399 @default.
- W3006839751 hasConceptScore W3006839751C86582703 @default.
- W3006839751 hasLocation W30068397511 @default.
- W3006839751 hasOpenAccess W3006839751 @default.
- W3006839751 hasPrimaryLocation W30068397511 @default.
- W3006839751 hasRelatedWork W2012213270 @default.
- W3006839751 hasRelatedWork W2396566817 @default.
- W3006839751 hasRelatedWork W2948977392 @default.
- W3006839751 hasRelatedWork W2970535233 @default.
- W3006839751 hasRelatedWork W2988233686 @default.
- W3006839751 hasRelatedWork W2990189480 @default.
- W3006839751 hasRelatedWork W2994942528 @default.
- W3006839751 hasRelatedWork W3022418454 @default.
- W3006839751 hasRelatedWork W3033710643 @default.
- W3006839751 hasRelatedWork W3035618986 @default.
- W3006839751 hasRelatedWork W3035677100 @default.
- W3006839751 hasRelatedWork W3036721570 @default.
- W3006839751 hasRelatedWork W3041299015 @default.
- W3006839751 hasRelatedWork W3083490811 @default.
- W3006839751 hasRelatedWork W3111382075 @default.
- W3006839751 hasRelatedWork W3111663897 @default.
- W3006839751 hasRelatedWork W3127713124 @default.
- W3006839751 hasRelatedWork W3172679426 @default.
- W3006839751 hasRelatedWork W3188261512 @default.
- W3006839751 hasRelatedWork W3128884106 @default.
- W3006839751 isParatext "false" @default.
- W3006839751 isRetracted "false" @default.
- W3006839751 magId "3006839751" @default.
- W3006839751 workType "article" @default.