Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386869694> ?p ?o ?g. }
Showing items 1 to 74 of
74
with 100 items per page.
- W4386869694 endingPage "17" @default.
- W4386869694 startingPage "1" @default.
- W4386869694 abstract "Cross-domain federated learning (FL), where data on local clients come from different domains, is a common case of FL. In such a cross-domain case, features extracted from the raw data of different clients deviate from each other in the feature space, leading to a so-called feature shift. This phenomenon can reduce feature discrimination and degrade the performance of the learned model. However, most existing FL methods are not specifically designed for the cross-domain setting. In this paper, we propose a novel cross-domain FL method named AlignFed. In AlignFed, each client model consists of a personalized feature extractor and a shared lightweight classifier. The feature extractor maps the features to a consistent space by aligning them to identical global target points. Inspired by recent studies in contrastive learning, AlignFed regards points that are uniformly distributed on the hypersphere as global target points. It then pushes features toward global target points of their corresponding classes and away from those of other classes to improve feature discrimination. The shared classifier aggregates knowledge across clients over the consistent feature space, which can mitigate performance degradation caused by feature shift while reducing communication cost. We conduct convergence analysis and perform extensive experiments to evaluate AlignFed." @default.
- W4386869694 created "2023-09-20" @default.
- W4386869694 creator A5035564546 @default.
- W4386869694 creator A5050393292 @default.
- W4386869694 creator A5053316728 @default.
- W4386869694 creator A5057612676 @default.
- W4386869694 date "2023-01-01" @default.
- W4386869694 modified "2023-09-27" @default.
- W4386869694 title "Aligning before Aggregating: Enabling Communication Efficient Cross-Domain Federated Learning via Consistent Feature Extraction" @default.
- W4386869694 doi "https://doi.org/10.1109/tmc.2023.3316645" @default.
- W4386869694 hasPublicationYear "2023" @default.
- W4386869694 type Work @default.
- W4386869694 citedByCount "0" @default.
- W4386869694 crossrefType "journal-article" @default.
- W4386869694 hasAuthorship W4386869694A5035564546 @default.
- W4386869694 hasAuthorship W4386869694A5050393292 @default.
- W4386869694 hasAuthorship W4386869694A5053316728 @default.
- W4386869694 hasAuthorship W4386869694A5057612676 @default.
- W4386869694 hasConcept C117978034 @default.
- W4386869694 hasConcept C119857082 @default.
- W4386869694 hasConcept C124101348 @default.
- W4386869694 hasConcept C127413603 @default.
- W4386869694 hasConcept C134306372 @default.
- W4386869694 hasConcept C138885662 @default.
- W4386869694 hasConcept C153180895 @default.
- W4386869694 hasConcept C154945302 @default.
- W4386869694 hasConcept C21880701 @default.
- W4386869694 hasConcept C2776401178 @default.
- W4386869694 hasConcept C2776562905 @default.
- W4386869694 hasConcept C33923547 @default.
- W4386869694 hasConcept C36503486 @default.
- W4386869694 hasConcept C41008148 @default.
- W4386869694 hasConcept C41895202 @default.
- W4386869694 hasConcept C52622490 @default.
- W4386869694 hasConcept C59404180 @default.
- W4386869694 hasConcept C83665646 @default.
- W4386869694 hasConcept C95623464 @default.
- W4386869694 hasConceptScore W4386869694C117978034 @default.
- W4386869694 hasConceptScore W4386869694C119857082 @default.
- W4386869694 hasConceptScore W4386869694C124101348 @default.
- W4386869694 hasConceptScore W4386869694C127413603 @default.
- W4386869694 hasConceptScore W4386869694C134306372 @default.
- W4386869694 hasConceptScore W4386869694C138885662 @default.
- W4386869694 hasConceptScore W4386869694C153180895 @default.
- W4386869694 hasConceptScore W4386869694C154945302 @default.
- W4386869694 hasConceptScore W4386869694C21880701 @default.
- W4386869694 hasConceptScore W4386869694C2776401178 @default.
- W4386869694 hasConceptScore W4386869694C2776562905 @default.
- W4386869694 hasConceptScore W4386869694C33923547 @default.
- W4386869694 hasConceptScore W4386869694C36503486 @default.
- W4386869694 hasConceptScore W4386869694C41008148 @default.
- W4386869694 hasConceptScore W4386869694C41895202 @default.
- W4386869694 hasConceptScore W4386869694C52622490 @default.
- W4386869694 hasConceptScore W4386869694C59404180 @default.
- W4386869694 hasConceptScore W4386869694C83665646 @default.
- W4386869694 hasConceptScore W4386869694C95623464 @default.
- W4386869694 hasLocation W43868696941 @default.
- W4386869694 hasOpenAccess W4386869694 @default.
- W4386869694 hasPrimaryLocation W43868696941 @default.
- W4386869694 hasRelatedWork W1977222486 @default.
- W4386869694 hasRelatedWork W2049538278 @default.
- W4386869694 hasRelatedWork W2114586818 @default.
- W4386869694 hasRelatedWork W2143407407 @default.
- W4386869694 hasRelatedWork W2546942002 @default.
- W4386869694 hasRelatedWork W2592385986 @default.
- W4386869694 hasRelatedWork W2765332597 @default.
- W4386869694 hasRelatedWork W2771515600 @default.
- W4386869694 hasRelatedWork W2772780115 @default.
- W4386869694 hasRelatedWork W2944661354 @default.
- W4386869694 isParatext "false" @default.
- W4386869694 isRetracted "false" @default.
- W4386869694 workType "article" @default.