Matches in SemOpenAlex for { <https://semopenalex.org/work/W4328007429> ?p ?o ?g. }
Showing items 1 to 94 of
94
with 100 items per page.
- W4328007429 endingPage "12" @default.
- W4328007429 startingPage "1" @default.
- W4328007429 abstract "Iteration based collaborative learning (CL) paradigms, such as federated learning (FL) and split learning (SL), faces challenges in training neural models over the rapidly growing yet resource-constrained edge devices. Such devices have difficulty in accommodating a full-size large model for FL or affording an excessive waiting time for the mandatory synchronization step in SL. To deal with such challenge, we propose a novel CL framework which adopts an tree-aggregation structure with an adaptive partition and ensemble strategy to achieve optimal synchronization and fast convergence at scale. To find the optimal split point for heterogeneous clients, we also design a novel partitioning algorithm by minimizing the idleness during communication and achieving the optimal synchronization between clients. In addition, a parallelism paradigm is proposed to unleash the potential of optimum synchronization between the clients and server to boost the distributed training process without losing model accuracy for edge devices. Furthermore, we theoretically prove that our framework can achieve better convergence rate than state-of-the-art CL paradigms. We conduct extensive experiments and show that our framework is 4.6× in training speed as compared with the traditional methods, without compromising training accuracy." @default.
- W4328007429 created "2023-03-22" @default.
- W4328007429 creator A5015877941 @default.
- W4328007429 creator A5025430395 @default.
- W4328007429 creator A5026994906 @default.
- W4328007429 creator A5043464306 @default.
- W4328007429 creator A5053828391 @default.
- W4328007429 creator A5061723765 @default.
- W4328007429 creator A5063181504 @default.
- W4328007429 creator A5080988180 @default.
- W4328007429 date "2023-01-01" @default.
- W4328007429 modified "2023-10-16" @default.
- W4328007429 title "Tree Learning: Towards Promoting Coordination in Scalable Multi-Client Training Acceleration" @default.
- W4328007429 doi "https://doi.org/10.1109/tmc.2023.3259007" @default.
- W4328007429 hasPublicationYear "2023" @default.
- W4328007429 type Work @default.
- W4328007429 citedByCount "0" @default.
- W4328007429 crossrefType "journal-article" @default.
- W4328007429 hasAuthorship W4328007429A5015877941 @default.
- W4328007429 hasAuthorship W4328007429A5025430395 @default.
- W4328007429 hasAuthorship W4328007429A5026994906 @default.
- W4328007429 hasAuthorship W4328007429A5043464306 @default.
- W4328007429 hasAuthorship W4328007429A5053828391 @default.
- W4328007429 hasAuthorship W4328007429A5061723765 @default.
- W4328007429 hasAuthorship W4328007429A5063181504 @default.
- W4328007429 hasAuthorship W4328007429A5080988180 @default.
- W4328007429 hasConcept C108734733 @default.
- W4328007429 hasConcept C111919701 @default.
- W4328007429 hasConcept C113174947 @default.
- W4328007429 hasConcept C114614502 @default.
- W4328007429 hasConcept C119857082 @default.
- W4328007429 hasConcept C120314980 @default.
- W4328007429 hasConcept C127162648 @default.
- W4328007429 hasConcept C134306372 @default.
- W4328007429 hasConcept C138236772 @default.
- W4328007429 hasConcept C154945302 @default.
- W4328007429 hasConcept C162307627 @default.
- W4328007429 hasConcept C162324750 @default.
- W4328007429 hasConcept C173608175 @default.
- W4328007429 hasConcept C2777303404 @default.
- W4328007429 hasConcept C2778562939 @default.
- W4328007429 hasConcept C31258907 @default.
- W4328007429 hasConcept C33923547 @default.
- W4328007429 hasConcept C41008148 @default.
- W4328007429 hasConcept C42812 @default.
- W4328007429 hasConcept C48044578 @default.
- W4328007429 hasConcept C50522688 @default.
- W4328007429 hasConcept C68339613 @default.
- W4328007429 hasConcept C77088390 @default.
- W4328007429 hasConcept C79974875 @default.
- W4328007429 hasConcept C98045186 @default.
- W4328007429 hasConceptScore W4328007429C108734733 @default.
- W4328007429 hasConceptScore W4328007429C111919701 @default.
- W4328007429 hasConceptScore W4328007429C113174947 @default.
- W4328007429 hasConceptScore W4328007429C114614502 @default.
- W4328007429 hasConceptScore W4328007429C119857082 @default.
- W4328007429 hasConceptScore W4328007429C120314980 @default.
- W4328007429 hasConceptScore W4328007429C127162648 @default.
- W4328007429 hasConceptScore W4328007429C134306372 @default.
- W4328007429 hasConceptScore W4328007429C138236772 @default.
- W4328007429 hasConceptScore W4328007429C154945302 @default.
- W4328007429 hasConceptScore W4328007429C162307627 @default.
- W4328007429 hasConceptScore W4328007429C162324750 @default.
- W4328007429 hasConceptScore W4328007429C173608175 @default.
- W4328007429 hasConceptScore W4328007429C2777303404 @default.
- W4328007429 hasConceptScore W4328007429C2778562939 @default.
- W4328007429 hasConceptScore W4328007429C31258907 @default.
- W4328007429 hasConceptScore W4328007429C33923547 @default.
- W4328007429 hasConceptScore W4328007429C41008148 @default.
- W4328007429 hasConceptScore W4328007429C42812 @default.
- W4328007429 hasConceptScore W4328007429C48044578 @default.
- W4328007429 hasConceptScore W4328007429C50522688 @default.
- W4328007429 hasConceptScore W4328007429C68339613 @default.
- W4328007429 hasConceptScore W4328007429C77088390 @default.
- W4328007429 hasConceptScore W4328007429C79974875 @default.
- W4328007429 hasConceptScore W4328007429C98045186 @default.
- W4328007429 hasLocation W43280074291 @default.
- W4328007429 hasOpenAccess W4328007429 @default.
- W4328007429 hasPrimaryLocation W43280074291 @default.
- W4328007429 hasRelatedWork W2164778033 @default.
- W4328007429 hasRelatedWork W2258148309 @default.
- W4328007429 hasRelatedWork W2505244433 @default.
- W4328007429 hasRelatedWork W2804912624 @default.
- W4328007429 hasRelatedWork W3109448977 @default.
- W4328007429 hasRelatedWork W3197430115 @default.
- W4328007429 hasRelatedWork W4298196516 @default.
- W4328007429 hasRelatedWork W4312934817 @default.
- W4328007429 hasRelatedWork W4316042121 @default.
- W4328007429 hasRelatedWork W4320067866 @default.
- W4328007429 isParatext "false" @default.
- W4328007429 isRetracted "false" @default.
- W4328007429 workType "article" @default.