Matches in SemOpenAlex for { <https://semopenalex.org/work/W4378977293> ?p ?o ?g. }
Showing items 1 to 80 of
80
with 100 items per page.
- W4378977293 abstract "Normalization is an important but understudied challenge in privacy-related application domains such as federated learning (FL), differential privacy (DP), and differentially private federated learning (DP-FL). While the unsuitability of batch normalization for these domains has already been shown, the impact of other normalization methods on the performance of federated or differentially private models is not well-known. To address this, we draw a performance comparison among layer normalization (LayerNorm), group normalization (GroupNorm), and the recently proposed kernel normalization (KernelNorm) in FL, DP, and DP-FL settings. Our results indicate LayerNorm and GroupNorm provide no performance gain compared to the baseline (i.e. no normalization) for shallow models in FL and DP. They, on the other hand, considerably enhance the performance of shallow models in DP-FL and deeper models in FL and DP. KernelNorm, moreover, significantly outperforms its competitors in terms of accuracy and convergence rate (or communication efficiency) for both shallow and deeper models in all considered learning environments. Given these key observations, we propose a kernel normalized ResNet architecture called KNResNet-13 for differentially private learning. Using the proposed architecture, we provide new state-of-the-art accuracy values on the CIFAR-10 and Imagenette datasets, when trained from scratch." @default.
- W4378977293 created "2023-06-02" @default.
- W4378977293 creator A5000554104 @default.
- W4378977293 creator A5006461848 @default.
- W4378977293 creator A5048531338 @default.
- W4378977293 creator A5063600672 @default.
- W4378977293 date "2023-02-01" @default.
- W4378977293 modified "2023-09-27" @default.
- W4378977293 title "Kernel Normalized Convolutional Networks for Privacy-Preserving Machine Learning" @default.
- W4378977293 cites W1125436246 @default.
- W4378977293 cites W1649611255 @default.
- W4378977293 cites W2107878631 @default.
- W4378977293 cites W2108598243 @default.
- W4378977293 cites W2109426455 @default.
- W4378977293 cites W2194775991 @default.
- W4378977293 cites W2473418344 @default.
- W4378977293 cites W2535690855 @default.
- W4378977293 cites W2795435272 @default.
- W4378977293 cites W2930926105 @default.
- W4378977293 cites W2963446712 @default.
- W4378977293 cites W2963456518 @default.
- W4378977293 cites W2970408908 @default.
- W4378977293 cites W3034513094 @default.
- W4378977293 cites W3046518446 @default.
- W4378977293 cites W3103802018 @default.
- W4378977293 cites W3212532487 @default.
- W4378977293 cites W3216713600 @default.
- W4378977293 cites W4205228770 @default.
- W4378977293 cites W4250482878 @default.
- W4378977293 doi "https://doi.org/10.1109/satml54575.2023.00016" @default.
- W4378977293 hasPublicationYear "2023" @default.
- W4378977293 type Work @default.
- W4378977293 citedByCount "0" @default.
- W4378977293 crossrefType "proceedings-article" @default.
- W4378977293 hasAuthorship W4378977293A5000554104 @default.
- W4378977293 hasAuthorship W4378977293A5006461848 @default.
- W4378977293 hasAuthorship W4378977293A5048531338 @default.
- W4378977293 hasAuthorship W4378977293A5063600672 @default.
- W4378977293 hasBestOaLocation W43789772932 @default.
- W4378977293 hasConcept C108583219 @default.
- W4378977293 hasConcept C114614502 @default.
- W4378977293 hasConcept C119857082 @default.
- W4378977293 hasConcept C124101348 @default.
- W4378977293 hasConcept C136886441 @default.
- W4378977293 hasConcept C144024400 @default.
- W4378977293 hasConcept C154945302 @default.
- W4378977293 hasConcept C19165224 @default.
- W4378977293 hasConcept C23130292 @default.
- W4378977293 hasConcept C33923547 @default.
- W4378977293 hasConcept C41008148 @default.
- W4378977293 hasConcept C74193536 @default.
- W4378977293 hasConceptScore W4378977293C108583219 @default.
- W4378977293 hasConceptScore W4378977293C114614502 @default.
- W4378977293 hasConceptScore W4378977293C119857082 @default.
- W4378977293 hasConceptScore W4378977293C124101348 @default.
- W4378977293 hasConceptScore W4378977293C136886441 @default.
- W4378977293 hasConceptScore W4378977293C144024400 @default.
- W4378977293 hasConceptScore W4378977293C154945302 @default.
- W4378977293 hasConceptScore W4378977293C19165224 @default.
- W4378977293 hasConceptScore W4378977293C23130292 @default.
- W4378977293 hasConceptScore W4378977293C33923547 @default.
- W4378977293 hasConceptScore W4378977293C41008148 @default.
- W4378977293 hasConceptScore W4378977293C74193536 @default.
- W4378977293 hasLocation W43789772931 @default.
- W4378977293 hasLocation W43789772932 @default.
- W4378977293 hasOpenAccess W4378977293 @default.
- W4378977293 hasPrimaryLocation W43789772931 @default.
- W4378977293 hasRelatedWork W3014300295 @default.
- W4378977293 hasRelatedWork W3164822677 @default.
- W4378977293 hasRelatedWork W4223943233 @default.
- W4378977293 hasRelatedWork W4225161397 @default.
- W4378977293 hasRelatedWork W4250304930 @default.
- W4378977293 hasRelatedWork W4309045103 @default.
- W4378977293 hasRelatedWork W4312200629 @default.
- W4378977293 hasRelatedWork W4360585206 @default.
- W4378977293 hasRelatedWork W4364306694 @default.
- W4378977293 hasRelatedWork W4380086463 @default.
- W4378977293 isParatext "false" @default.
- W4378977293 isRetracted "false" @default.
- W4378977293 workType "article" @default.