Matches in SemOpenAlex for { <https://semopenalex.org/work/W3105298093> ?p ?o ?g. }
Showing items 1 to 79 of
79
with 100 items per page.
- W3105298093 endingPage "2363" @default.
- W3105298093 startingPage "2351" @default.
- W3105298093 abstract "Federated Learning (FL) is a machine learning setting where many devices collaboratively train a machine learning model while keeping the training data decentralized. In most of the current training schemes the central model is refined by averaging the parameters of the server model and the updated parameters from the client side. However, directly averaging model parameters is only possible if all models have the same structure and size, which could be a restrictive constraint in many scenarios. In this work we investigate more powerful and more flexible aggregation schemes for FL. Specifically, we propose ensemble distillation for model fusion, i.e. training the central classifier through unlabeled data on the outputs of the models from the clients. This knowledge distillation technique mitigates privacy risk and cost to the same extent as the baseline FL algorithms, but allows flexible aggregation over heterogeneous client models that can differ e.g. in size, numerical precision or structure. We show in extensive empirical experiments on various CV/NLP datasets (CIFAR-10/100, ImageNet, AG News, SST2) and settings (heterogeneous models/data) that the server model can be trained much faster, requiring fewer communication rounds than any existing FL technique so far." @default.
- W3105298093 created "2020-11-23" @default.
- W3105298093 creator A5045519627 @default.
- W3105298093 creator A5050008056 @default.
- W3105298093 creator A5073756389 @default.
- W3105298093 creator A5078074297 @default.
- W3105298093 date "2020-01-01" @default.
- W3105298093 modified "2023-09-29" @default.
- W3105298093 title "Ensemble Distillation for Robust Model Fusion in Federated Learning" @default.
- W3105298093 hasPublicationYear "2020" @default.
- W3105298093 type Work @default.
- W3105298093 sameAs 3105298093 @default.
- W3105298093 citedByCount "30" @default.
- W3105298093 countsByYear W31052980932020 @default.
- W3105298093 countsByYear W31052980932021 @default.
- W3105298093 countsByYear W31052980932022 @default.
- W3105298093 crossrefType "proceedings-article" @default.
- W3105298093 hasAuthorship W3105298093A5045519627 @default.
- W3105298093 hasAuthorship W3105298093A5050008056 @default.
- W3105298093 hasAuthorship W3105298093A5073756389 @default.
- W3105298093 hasAuthorship W3105298093A5078074297 @default.
- W3105298093 hasConcept C119857082 @default.
- W3105298093 hasConcept C119898033 @default.
- W3105298093 hasConcept C124101348 @default.
- W3105298093 hasConcept C154945302 @default.
- W3105298093 hasConcept C178790620 @default.
- W3105298093 hasConcept C185592680 @default.
- W3105298093 hasConcept C204030448 @default.
- W3105298093 hasConcept C2992525071 @default.
- W3105298093 hasConcept C41008148 @default.
- W3105298093 hasConcept C45942800 @default.
- W3105298093 hasConcept C51632099 @default.
- W3105298093 hasConcept C67186912 @default.
- W3105298093 hasConcept C77088390 @default.
- W3105298093 hasConcept C95623464 @default.
- W3105298093 hasConceptScore W3105298093C119857082 @default.
- W3105298093 hasConceptScore W3105298093C119898033 @default.
- W3105298093 hasConceptScore W3105298093C124101348 @default.
- W3105298093 hasConceptScore W3105298093C154945302 @default.
- W3105298093 hasConceptScore W3105298093C178790620 @default.
- W3105298093 hasConceptScore W3105298093C185592680 @default.
- W3105298093 hasConceptScore W3105298093C204030448 @default.
- W3105298093 hasConceptScore W3105298093C2992525071 @default.
- W3105298093 hasConceptScore W3105298093C41008148 @default.
- W3105298093 hasConceptScore W3105298093C45942800 @default.
- W3105298093 hasConceptScore W3105298093C51632099 @default.
- W3105298093 hasConceptScore W3105298093C67186912 @default.
- W3105298093 hasConceptScore W3105298093C77088390 @default.
- W3105298093 hasConceptScore W3105298093C95623464 @default.
- W3105298093 hasLocation W31052980931 @default.
- W3105298093 hasOpenAccess W3105298093 @default.
- W3105298093 hasPrimaryLocation W31052980931 @default.
- W3105298093 hasRelatedWork W1821462560 @default.
- W3105298093 hasRelatedWork W2194775991 @default.
- W3105298093 hasRelatedWork W2535838896 @default.
- W3105298093 hasRelatedWork W2541884796 @default.
- W3105298093 hasRelatedWork W2807006176 @default.
- W3105298093 hasRelatedWork W2912213068 @default.
- W3105298093 hasRelatedWork W2963300197 @default.
- W3105298093 hasRelatedWork W2972570881 @default.
- W3105298093 hasRelatedWork W2976335444 @default.
- W3105298093 hasRelatedWork W2980216952 @default.
- W3105298093 hasRelatedWork W2981738522 @default.
- W3105298093 hasRelatedWork W2995022099 @default.
- W3105298093 hasRelatedWork W3005776401 @default.
- W3105298093 hasRelatedWork W3007548213 @default.
- W3105298093 hasRelatedWork W3012968339 @default.
- W3105298093 hasRelatedWork W3035453001 @default.
- W3105298093 hasRelatedWork W3038022836 @default.
- W3105298093 hasRelatedWork W3100393648 @default.
- W3105298093 hasRelatedWork W3103802018 @default.
- W3105298093 hasRelatedWork W3118608800 @default.
- W3105298093 hasVolume "33" @default.
- W3105298093 isParatext "false" @default.
- W3105298093 isRetracted "false" @default.
- W3105298093 magId "3105298093" @default.
- W3105298093 workType "article" @default.