Matches in SemOpenAlex for { <https://semopenalex.org/work/W4387059092> ?p ?o ?g. }
Showing items 1 to 88 of
88
with 100 items per page.
- W4387059092 abstract "Abstract Data heterogeneity (Non-IID) on Federated Learning (FL) is currently a widely publicized problem, which leads to local model drift and performance degradation. Because of the advantage of knowledge distillation, it has been explored in some recent work to refine global models. However, these approaches rely on a proxy dataset or a data generator. First, in many FL scenarios, proxy dataset do not necessarily exist on the server. Second, the quality of data generated by the generator is unstable and the generator depends on the computing resources of the server. In this work, we propose a novel data-Free knowledge distillation approach via generator-Free Data Generation for Non-IID FL, dubbed as FedF 2 DG. Specifically, FedF 2 DG requires only local models to generate pseudo datasets for each client, and can generate hard samples by adding an additional regularization term that exploit disagreements between local model and global model. Meanwhile, FedF 2 DG enables flexible utilization of computational resources by generating pseudo dataset locally or on the server. And to address the label distribution shift in Non-IID FL, we propose a Data Generation Principle that can adaptively control the label distribution and number of pseudo dataset based on client current state, and this allows for the extraction of more client knowledge. Then knowledge distillation is performed to transfer the knowledge in local models to the global model. Extensive experiments demonstrate that our proposed method significantly outperforms the state-of-the-art FL methods and can serve as plugin for existing Federated Learning methds such as FedAvg, FedProx, etc, and improve their performance." @default.
- W4387059092 created "2023-09-27" @default.
- W4387059092 creator A5012731436 @default.
- W4387059092 creator A5020142020 @default.
- W4387059092 creator A5053789269 @default.
- W4387059092 creator A5063796185 @default.
- W4387059092 creator A5069084441 @default.
- W4387059092 creator A5081840530 @default.
- W4387059092 date "2023-09-26" @default.
- W4387059092 modified "2023-09-27" @default.
- W4387059092 title "Data-Free Knowledge Distillation via Generator-Free Data Generation for Non-IID Federated Learning" @default.
- W4387059092 cites W1869500417 @default.
- W4387059092 cites W219040644 @default.
- W4387059092 cites W2194775991 @default.
- W4387059092 cites W2473418344 @default.
- W4387059092 cites W2765811365 @default.
- W4387059092 cites W2912213068 @default.
- W4387059092 cites W2962798895 @default.
- W4387059092 cites W2982802130 @default.
- W4387059092 cites W3034957837 @default.
- W4387059092 cites W3088573320 @default.
- W4387059092 cites W3095251140 @default.
- W4387059092 cites W3096831136 @default.
- W4387059092 cites W3097060730 @default.
- W4387059092 cites W3100779497 @default.
- W4387059092 cites W3128108456 @default.
- W4387059092 cites W3169044395 @default.
- W4387059092 cites W3182158470 @default.
- W4387059092 cites W3203735593 @default.
- W4387059092 cites W4226101686 @default.
- W4387059092 cites W4283816018 @default.
- W4387059092 cites W4287332481 @default.
- W4387059092 cites W4294106961 @default.
- W4387059092 cites W4312424218 @default.
- W4387059092 doi "https://doi.org/10.21203/rs.3.rs-3364332/v1" @default.
- W4387059092 hasPublicationYear "2023" @default.
- W4387059092 type Work @default.
- W4387059092 citedByCount "0" @default.
- W4387059092 crossrefType "posted-content" @default.
- W4387059092 hasAuthorship W4387059092A5012731436 @default.
- W4387059092 hasAuthorship W4387059092A5020142020 @default.
- W4387059092 hasAuthorship W4387059092A5053789269 @default.
- W4387059092 hasAuthorship W4387059092A5063796185 @default.
- W4387059092 hasAuthorship W4387059092A5069084441 @default.
- W4387059092 hasAuthorship W4387059092A5081840530 @default.
- W4387059092 hasBestOaLocation W43870590921 @default.
- W4387059092 hasConcept C119857082 @default.
- W4387059092 hasConcept C121332964 @default.
- W4387059092 hasConcept C124101348 @default.
- W4387059092 hasConcept C154945302 @default.
- W4387059092 hasConcept C163258240 @default.
- W4387059092 hasConcept C165696696 @default.
- W4387059092 hasConcept C178790620 @default.
- W4387059092 hasConcept C185592680 @default.
- W4387059092 hasConcept C204030448 @default.
- W4387059092 hasConcept C2780992000 @default.
- W4387059092 hasConcept C38652104 @default.
- W4387059092 hasConcept C41008148 @default.
- W4387059092 hasConcept C62520636 @default.
- W4387059092 hasConceptScore W4387059092C119857082 @default.
- W4387059092 hasConceptScore W4387059092C121332964 @default.
- W4387059092 hasConceptScore W4387059092C124101348 @default.
- W4387059092 hasConceptScore W4387059092C154945302 @default.
- W4387059092 hasConceptScore W4387059092C163258240 @default.
- W4387059092 hasConceptScore W4387059092C165696696 @default.
- W4387059092 hasConceptScore W4387059092C178790620 @default.
- W4387059092 hasConceptScore W4387059092C185592680 @default.
- W4387059092 hasConceptScore W4387059092C204030448 @default.
- W4387059092 hasConceptScore W4387059092C2780992000 @default.
- W4387059092 hasConceptScore W4387059092C38652104 @default.
- W4387059092 hasConceptScore W4387059092C41008148 @default.
- W4387059092 hasConceptScore W4387059092C62520636 @default.
- W4387059092 hasLocation W43870590921 @default.
- W4387059092 hasOpenAccess W4387059092 @default.
- W4387059092 hasPrimaryLocation W43870590921 @default.
- W4387059092 hasRelatedWork W1607315280 @default.
- W4387059092 hasRelatedWork W2331043530 @default.
- W4387059092 hasRelatedWork W2374725260 @default.
- W4387059092 hasRelatedWork W2383403914 @default.
- W4387059092 hasRelatedWork W2393933887 @default.
- W4387059092 hasRelatedWork W2961085424 @default.
- W4387059092 hasRelatedWork W2997512100 @default.
- W4387059092 hasRelatedWork W4306674287 @default.
- W4387059092 hasRelatedWork W4386140649 @default.
- W4387059092 hasRelatedWork W4224009465 @default.
- W4387059092 isParatext "false" @default.
- W4387059092 isRetracted "false" @default.
- W4387059092 workType "article" @default.