Matches in SemOpenAlex for { <https://semopenalex.org/work/W3210323622> ?p ?o ?g. }
Showing items 1 to 63 of
63
with 100 items per page.
- W3210323622 endingPage "48" @default.
- W3210323622 startingPage "33" @default.
- W3210323622 abstract "Federated learning suffers from terrible generalization performance because the model fails to utilize global information over all clients when data is non-IID (not independently or identically distributed) partitioning. Meanwhile, the theoretical studies in this field are still insufficient. In this paper, we present an excess risk bound for federated learning on non-IID data, which measures the error between the model of federated learning and the optimal centralized model. Specifically, we present a novel error decomposition strategy, which decomposes the excess risk into three terms: agnostic error, federated error, and approximation error. By estimating the error terms, we find that Rademacher complexity and discrepancy distance are the keys to affecting the learning performance. Motivated by the theoretical findings, we propose FedAvgR to improve the performance via additional regularizers to lower the excess risk. Experimental results demonstrate the effectiveness of our algorithm and coincide with our theory." @default.
- W3210323622 created "2021-11-08" @default.
- W3210323622 creator A5016006900 @default.
- W3210323622 creator A5028071674 @default.
- W3210323622 creator A5033688057 @default.
- W3210323622 creator A5052883326 @default.
- W3210323622 date "2021-01-01" @default.
- W3210323622 modified "2023-09-23" @default.
- W3210323622 title "Federated Learning for Non-IID Data: From Theory to Algorithm" @default.
- W3210323622 cites W1538170657 @default.
- W3210323622 cites W2104094955 @default.
- W3210323622 cites W2153635508 @default.
- W3210323622 cites W2164943005 @default.
- W3210323622 cites W2604301957 @default.
- W3210323622 cites W2605034610 @default.
- W3210323622 cites W2963318081 @default.
- W3210323622 cites W2967655751 @default.
- W3210323622 cites W2968240292 @default.
- W3210323622 cites W2997905006 @default.
- W3210323622 cites W2998606456 @default.
- W3210323622 cites W3015640161 @default.
- W3210323622 cites W3080934299 @default.
- W3210323622 cites W3091635927 @default.
- W3210323622 cites W3103802018 @default.
- W3210323622 cites W3165347919 @default.
- W3210323622 doi "https://doi.org/10.1007/978-3-030-89188-6_3" @default.
- W3210323622 hasPublicationYear "2021" @default.
- W3210323622 type Work @default.
- W3210323622 sameAs 3210323622 @default.
- W3210323622 citedByCount "2" @default.
- W3210323622 countsByYear W32103236222023 @default.
- W3210323622 crossrefType "book-chapter" @default.
- W3210323622 hasAuthorship W3210323622A5016006900 @default.
- W3210323622 hasAuthorship W3210323622A5028071674 @default.
- W3210323622 hasAuthorship W3210323622A5033688057 @default.
- W3210323622 hasAuthorship W3210323622A5052883326 @default.
- W3210323622 hasConcept C11413529 @default.
- W3210323622 hasConcept C154945302 @default.
- W3210323622 hasConcept C41008148 @default.
- W3210323622 hasConcept C80444323 @default.
- W3210323622 hasConceptScore W3210323622C11413529 @default.
- W3210323622 hasConceptScore W3210323622C154945302 @default.
- W3210323622 hasConceptScore W3210323622C41008148 @default.
- W3210323622 hasConceptScore W3210323622C80444323 @default.
- W3210323622 hasLocation W32103236221 @default.
- W3210323622 hasOpenAccess W3210323622 @default.
- W3210323622 hasPrimaryLocation W32103236221 @default.
- W3210323622 hasRelatedWork W2003465964 @default.
- W3210323622 hasRelatedWork W2333698505 @default.
- W3210323622 hasRelatedWork W2351491280 @default.
- W3210323622 hasRelatedWork W2371447506 @default.
- W3210323622 hasRelatedWork W2386767533 @default.
- W3210323622 hasRelatedWork W2392953011 @default.
- W3210323622 hasRelatedWork W2748952813 @default.
- W3210323622 hasRelatedWork W2899084033 @default.
- W3210323622 hasRelatedWork W303980170 @default.
- W3210323622 hasRelatedWork W3107474891 @default.
- W3210323622 isParatext "false" @default.
- W3210323622 isRetracted "false" @default.
- W3210323622 magId "3210323622" @default.
- W3210323622 workType "book-chapter" @default.