Matches in SemOpenAlex for { <https://semopenalex.org/work/W2283463896> ?p ?o ?g. }
- W2283463896 abstract "Modern mobile devices have access to a wealth of data suitable for learning models, which in turn can greatly improve the user experience on the device. For example, language models can improve speech recognition and text entry, and image models can automatically select good photos. However, this rich data is often privacy sensitive, large in quantity, or both, which may preclude logging to the data-center and training there using conventional approaches. We advocate an alternative that leaves the training data distributed on the mobile devices, and learns a shared model by aggregating locally-computed updates. We term this decentralized approach Federated Learning. We present a practical method for the federated learning of deep networks that proves robust to the unbalanced and non-IID data distributions that naturally arise. This method allows high-quality models to be trained in relatively few rounds of communication, the principal constraint for federated learning. The key insight is that despite the non-convex loss functions we optimize, parameter averaging over updates from multiple clients produces surprisingly good results, for example decreasing the communication needed to train an LSTM language model by two orders of magnitude." @default.
- W2283463896 created "2016-06-24" @default.
- W2283463896 creator A5040311844 @default.
- W2283463896 creator A5044698998 @default.
- W2283463896 creator A5071710055 @default.
- W2283463896 creator A5080901995 @default.
- W2283463896 date "2016-02-17" @default.
- W2283463896 modified "2023-10-01" @default.
- W2283463896 title "Federated Learning of Deep Networks using Model Averaging" @default.
- W2283463896 cites W130696423 @default.
- W2283463896 cites W1518479949 @default.
- W2283463896 cites W1762484328 @default.
- W2283463896 cites W179368776 @default.
- W2283463896 cites W1899249567 @default.
- W2283463896 cites W1927459353 @default.
- W2283463896 cites W1968697355 @default.
- W2283463896 cites W2027595342 @default.
- W2283463896 cites W2051267297 @default.
- W2283463896 cites W2053637704 @default.
- W2283463896 cites W2064675550 @default.
- W2283463896 cites W2078831169 @default.
- W2283463896 cites W2093678760 @default.
- W2283463896 cites W2095705004 @default.
- W2283463896 cites W2103647628 @default.
- W2283463896 cites W2112796928 @default.
- W2283463896 cites W2123154536 @default.
- W2283463896 cites W2127180992 @default.
- W2283463896 cites W2133319764 @default.
- W2283463896 cites W2138302120 @default.
- W2283463896 cites W2146989110 @default.
- W2283463896 cites W2151320232 @default.
- W2283463896 cites W2162576315 @default.
- W2283463896 cites W2166706236 @default.
- W2283463896 cites W2168231600 @default.
- W2283463896 cites W2238591594 @default.
- W2283463896 cites W2282546112 @default.
- W2283463896 cites W2579186979 @default.
- W2283463896 cites W2905455284 @default.
- W2283463896 cites W2949117887 @default.
- W2283463896 cites W2949758198 @default.
- W2283463896 cites W2951395930 @default.
- W2283463896 cites W2951559648 @default.
- W2283463896 cites W2952986481 @default.
- W2283463896 cites W2963804082 @default.
- W2283463896 cites W2964121744 @default.
- W2283463896 cites W2964160102 @default.
- W2283463896 cites W2964240787 @default.
- W2283463896 cites W3118608800 @default.
- W2283463896 hasPublicationYear "2016" @default.
- W2283463896 type Work @default.
- W2283463896 sameAs 2283463896 @default.
- W2283463896 citedByCount "432" @default.
- W2283463896 countsByYear W22834638962016 @default.
- W2283463896 countsByYear W22834638962017 @default.
- W2283463896 countsByYear W22834638962018 @default.
- W2283463896 countsByYear W22834638962019 @default.
- W2283463896 countsByYear W22834638962020 @default.
- W2283463896 countsByYear W22834638962021 @default.
- W2283463896 countsByYear W22834638962022 @default.
- W2283463896 crossrefType "posted-content" @default.
- W2283463896 hasAuthorship W2283463896A5040311844 @default.
- W2283463896 hasAuthorship W2283463896A5044698998 @default.
- W2283463896 hasAuthorship W2283463896A5071710055 @default.
- W2283463896 hasAuthorship W2283463896A5080901995 @default.
- W2283463896 hasConcept C108583219 @default.
- W2283463896 hasConcept C111919701 @default.
- W2283463896 hasConcept C119857082 @default.
- W2283463896 hasConcept C124101348 @default.
- W2283463896 hasConcept C127413603 @default.
- W2283463896 hasConcept C137293760 @default.
- W2283463896 hasConcept C144559511 @default.
- W2283463896 hasConcept C154945302 @default.
- W2283463896 hasConcept C186967261 @default.
- W2283463896 hasConcept C26517878 @default.
- W2283463896 hasConcept C2776036281 @default.
- W2283463896 hasConcept C2992525071 @default.
- W2283463896 hasConcept C38652104 @default.
- W2283463896 hasConcept C41008148 @default.
- W2283463896 hasConcept C78519656 @default.
- W2283463896 hasConceptScore W2283463896C108583219 @default.
- W2283463896 hasConceptScore W2283463896C111919701 @default.
- W2283463896 hasConceptScore W2283463896C119857082 @default.
- W2283463896 hasConceptScore W2283463896C124101348 @default.
- W2283463896 hasConceptScore W2283463896C127413603 @default.
- W2283463896 hasConceptScore W2283463896C137293760 @default.
- W2283463896 hasConceptScore W2283463896C144559511 @default.
- W2283463896 hasConceptScore W2283463896C154945302 @default.
- W2283463896 hasConceptScore W2283463896C186967261 @default.
- W2283463896 hasConceptScore W2283463896C26517878 @default.
- W2283463896 hasConceptScore W2283463896C2776036281 @default.
- W2283463896 hasConceptScore W2283463896C2992525071 @default.
- W2283463896 hasConceptScore W2283463896C38652104 @default.
- W2283463896 hasConceptScore W2283463896C41008148 @default.
- W2283463896 hasConceptScore W2283463896C78519656 @default.
- W2283463896 hasLocation W22834638961 @default.
- W2283463896 hasOpenAccess W2283463896 @default.
- W2283463896 hasPrimaryLocation W22834638961 @default.
- W2283463896 hasRelatedWork W2027595342 @default.
- W2283463896 hasRelatedWork W2051267297 @default.
- W2283463896 hasRelatedWork W2053637704 @default.