FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3083250313> ?p ?o ?g. }

• W3083250313 abstract "Federated learning aims to collaboratively train a strong global model by accessing users' locally trained models but not their own data. A crucial step is therefore to aggregate local models into a global model, which has been shown challenging when users have non-i.i.d. data. In this paper, we propose a novel aggregation algorithm named FedBE, which takes a Bayesian inference perspective by sampling higher-quality global models and combining them via Bayesian model Ensemble, leading to much robust aggregation. We show that an effective model distribution can be constructed by simply fitting a Gaussian or Dirichlet distribution to the local models. Our empirical studies validate FedBE's superior performance, especially when users' data are not i.i.d. and when the neural networks go deeper. Moreover, FedBE is compatible with recent efforts in regularizing users' model training, making it an easily applicable module: you only need to replace the aggregation method but leave other parts of your federated learning algorithm intact." @default.
• W3083250313 created "2020-09-11" @default.
• W3083250313 creator A5012764388 @default.
• W3083250313 creator A5018066903 @default.
• W3083250313 date "2020-09-04" @default.
• W3083250313 modified "2023-09-27" @default.
• W3083250313 title "FedBE: Making Bayesian Model Ensemble Applicable to Federated Learning" @default.
• W3083250313 cites W1534477342 @default.
• W3083250313 cites W1590183771 @default.
• W3083250313 cites W1731081199 @default.
• W3083250313 cites W1821462560 @default.
• W3083250313 cites W2004273576 @default.
• W3083250313 cites W2027595342 @default.
• W3083250313 cites W2099201756 @default.
• W3083250313 cites W2104094955 @default.
• W3083250313 cites W2112796928 @default.
• W3083250313 cites W2136504847 @default.
• W3083250313 cites W2153635508 @default.
• W3083250313 cites W2166706236 @default.
• W3083250313 cites W2194775991 @default.
• W3083250313 cites W2535838896 @default.
• W3083250313 cites W2541884796 @default.
• W3083250313 cites W2612445135 @default.
• W3083250313 cites W2618494520 @default.
• W3083250313 cites W2626967530 @default.
• W3083250313 cites W2787017828 @default.
• W3083250313 cites W2788838181 @default.
• W3083250313 cites W2796438033 @default.
• W3083250313 cites W2803867449 @default.
• W3083250313 cites W2807006176 @default.
• W3083250313 cites W2912213068 @default.
• W3083250313 cites W2912934387 @default.
• W3083250313 cites W2913570153 @default.
• W3083250313 cites W2914328083 @default.
• W3083250313 cites W2917462349 @default.
• W3083250313 cites W2945152072 @default.
• W3083250313 cites W2949117887 @default.
• W3083250313 cites W2951696358 @default.
• W3083250313 cites W2952120674 @default.
• W3083250313 cites W2953070460 @default.
• W3083250313 cites W2955213239 @default.
• W3083250313 cites W2963025848 @default.
• W3083250313 cites W2975043678 @default.
• W3083250313 cites W2978426779 @default.
• W3083250313 cites W2980216952 @default.
• W3083250313 cites W2981298997 @default.
• W3083250313 cites W2982654255 @default.
• W3083250313 cites W2991497298 @default.
• W3083250313 cites W2996442797 @default.
• W3083250313 cites W2996970889 @default.
• W3083250313 cites W3008187686 @default.
• W3083250313 cites W3020531258 @default.
• W3083250313 cites W3034713518 @default.
• W3083250313 cites W3035321581 @default.
• W3083250313 cites W3035453001 @default.
• W3083250313 cites W3037674069 @default.
• W3083250313 cites W3081349436 @default.
• W3083250313 cites W3094163844 @default.
• W3083250313 cites W3118608800 @default.
• W3083250313 cites W605727707 @default.
• W3083250313 cites W2551176409 @default.
• W3083250313 hasPublicationYear "2020" @default.
• W3083250313 type Work @default.
• W3083250313 sameAs 3083250313 @default.
• W3083250313 citedByCount "12" @default.
• W3083250313 countsByYear W30832503132020 @default.
• W3083250313 countsByYear W30832503132021 @default.
• W3083250313 crossrefType "posted-content" @default.
• W3083250313 hasAuthorship W3083250313A5012764388 @default.
• W3083250313 hasAuthorship W3083250313A5018066903 @default.
• W3083250313 hasConcept C107673813 @default.
• W3083250313 hasConcept C119857082 @default.
• W3083250313 hasConcept C121332964 @default.
• W3083250313 hasConcept C124101348 @default.
• W3083250313 hasConcept C134306372 @default.
• W3083250313 hasConcept C154945302 @default.
• W3083250313 hasConcept C158424031 @default.
• W3083250313 hasConcept C159985019 @default.
• W3083250313 hasConcept C160234255 @default.
• W3083250313 hasConcept C163716315 @default.
• W3083250313 hasConcept C169214877 @default.
• W3083250313 hasConcept C182310444 @default.
• W3083250313 hasConcept C192562407 @default.
• W3083250313 hasConcept C2776214188 @default.
• W3083250313 hasConcept C2781280628 @default.
• W3083250313 hasConcept C2992525071 @default.
• W3083250313 hasConcept C33923547 @default.
• W3083250313 hasConcept C41008148 @default.
• W3083250313 hasConcept C4679612 @default.
• W3083250313 hasConcept C62520636 @default.
• W3083250313 hasConceptScore W3083250313C107673813 @default.
• W3083250313 hasConceptScore W3083250313C119857082 @default.
• W3083250313 hasConceptScore W3083250313C121332964 @default.
• W3083250313 hasConceptScore W3083250313C124101348 @default.
• W3083250313 hasConceptScore W3083250313C134306372 @default.
• W3083250313 hasConceptScore W3083250313C154945302 @default.
• W3083250313 hasConceptScore W3083250313C158424031 @default.
• W3083250313 hasConceptScore W3083250313C159985019 @default.
• W3083250313 hasConceptScore W3083250313C160234255 @default.
• W3083250313 hasConceptScore W3083250313C163716315 @default.

• next