Matches in SemOpenAlex for { <https://semopenalex.org/work/W4323060992> ?p ?o ?g. }
Showing items 1 to 100 of
100
with 100 items per page.
- W4323060992 abstract "Deep Neural Networks (DNNs) process big datasets achieving high accuracy on incredibly complex tasks. However, this progress has led to a scalability impasse, as DNNs require massive amounts of processing power and local memory to be trained, making them impossible or impractical to be used on a single device. This situation has led to the design of distributed training architectures, where the DNN and the training data can be split among multiple processors. How to choose the appropriate distributed training architecture, however, remains an open question. To help bring insights into this debate, in this work we design a Distributed Training Simulator (DTS) that estimates the training time of a DNN in a distributed architecture through a mathematical model of the distributed architecture and resource-allocation heuristics. We illustrate the power of the proposed DTS through the implementation of five different distributed architectures, Pipeline Learning, Federated Learning, Split Learning, Parallel Split Learning, and Federated Split Learning, and we validate the accuracy of the training estimates using three different datasets of varying complexity and two different DNNs. Finally, we present a trade-off analysis to demonstrate the coherence of DTS estimates for diverse high-performance computing scenarios by comparing these estimates with the behaviors of a real computer cluster." @default.
- W4323060992 created "2023-03-05" @default.
- W4323060992 creator A5032634346 @default.
- W4323060992 creator A5055238253 @default.
- W4323060992 creator A5057421786 @default.
- W4323060992 date "2022-10-01" @default.
- W4323060992 modified "2023-09-27" @default.
- W4323060992 title "DTS: A Simulator to Estimate the Training Time of Distributed Deep Neural Networks" @default.
- W4323060992 cites W2007339694 @default.
- W4323060992 cites W2519804223 @default.
- W4323060992 cites W2786070938 @default.
- W4323060992 cites W2962755172 @default.
- W4323060992 cites W2966137607 @default.
- W4323060992 cites W2969388332 @default.
- W4323060992 cites W2972087877 @default.
- W4323060992 cites W3009048827 @default.
- W4323060992 cites W3013692475 @default.
- W4323060992 cites W3023529621 @default.
- W4323060992 cites W3042905487 @default.
- W4323060992 cites W3049440415 @default.
- W4323060992 cites W3083254293 @default.
- W4323060992 cites W3083715619 @default.
- W4323060992 cites W3184869309 @default.
- W4323060992 cites W3211848727 @default.
- W4323060992 doi "https://doi.org/10.1109/mascots56607.2022.00011" @default.
- W4323060992 hasPublicationYear "2022" @default.
- W4323060992 type Work @default.
- W4323060992 citedByCount "1" @default.
- W4323060992 countsByYear W43230609922023 @default.
- W4323060992 crossrefType "proceedings-article" @default.
- W4323060992 hasAuthorship W4323060992A5032634346 @default.
- W4323060992 hasAuthorship W4323060992A5055238253 @default.
- W4323060992 hasAuthorship W4323060992A5057421786 @default.
- W4323060992 hasConcept C108583219 @default.
- W4323060992 hasConcept C111919701 @default.
- W4323060992 hasConcept C113775141 @default.
- W4323060992 hasConcept C118524514 @default.
- W4323060992 hasConcept C119857082 @default.
- W4323060992 hasConcept C120314980 @default.
- W4323060992 hasConcept C121332964 @default.
- W4323060992 hasConcept C123657996 @default.
- W4323060992 hasConcept C127705205 @default.
- W4323060992 hasConcept C142362112 @default.
- W4323060992 hasConcept C153294291 @default.
- W4323060992 hasConcept C153349607 @default.
- W4323060992 hasConcept C154945302 @default.
- W4323060992 hasConcept C15744967 @default.
- W4323060992 hasConcept C19417346 @default.
- W4323060992 hasConcept C199360897 @default.
- W4323060992 hasConcept C2777211547 @default.
- W4323060992 hasConcept C2779582901 @default.
- W4323060992 hasConcept C2984842247 @default.
- W4323060992 hasConcept C41008148 @default.
- W4323060992 hasConcept C43521106 @default.
- W4323060992 hasConcept C48044578 @default.
- W4323060992 hasConcept C50644808 @default.
- W4323060992 hasConcept C77088390 @default.
- W4323060992 hasConcept C98045186 @default.
- W4323060992 hasConceptScore W4323060992C108583219 @default.
- W4323060992 hasConceptScore W4323060992C111919701 @default.
- W4323060992 hasConceptScore W4323060992C113775141 @default.
- W4323060992 hasConceptScore W4323060992C118524514 @default.
- W4323060992 hasConceptScore W4323060992C119857082 @default.
- W4323060992 hasConceptScore W4323060992C120314980 @default.
- W4323060992 hasConceptScore W4323060992C121332964 @default.
- W4323060992 hasConceptScore W4323060992C123657996 @default.
- W4323060992 hasConceptScore W4323060992C127705205 @default.
- W4323060992 hasConceptScore W4323060992C142362112 @default.
- W4323060992 hasConceptScore W4323060992C153294291 @default.
- W4323060992 hasConceptScore W4323060992C153349607 @default.
- W4323060992 hasConceptScore W4323060992C154945302 @default.
- W4323060992 hasConceptScore W4323060992C15744967 @default.
- W4323060992 hasConceptScore W4323060992C19417346 @default.
- W4323060992 hasConceptScore W4323060992C199360897 @default.
- W4323060992 hasConceptScore W4323060992C2777211547 @default.
- W4323060992 hasConceptScore W4323060992C2779582901 @default.
- W4323060992 hasConceptScore W4323060992C2984842247 @default.
- W4323060992 hasConceptScore W4323060992C41008148 @default.
- W4323060992 hasConceptScore W4323060992C43521106 @default.
- W4323060992 hasConceptScore W4323060992C48044578 @default.
- W4323060992 hasConceptScore W4323060992C50644808 @default.
- W4323060992 hasConceptScore W4323060992C77088390 @default.
- W4323060992 hasConceptScore W4323060992C98045186 @default.
- W4323060992 hasFunder F4320306076 @default.
- W4323060992 hasLocation W43230609921 @default.
- W4323060992 hasOpenAccess W4323060992 @default.
- W4323060992 hasPrimaryLocation W43230609921 @default.
- W4323060992 hasRelatedWork W2576264401 @default.
- W4323060992 hasRelatedWork W2909645158 @default.
- W4323060992 hasRelatedWork W2950066684 @default.
- W4323060992 hasRelatedWork W3082895349 @default.
- W4323060992 hasRelatedWork W3162132941 @default.
- W4323060992 hasRelatedWork W3179488938 @default.
- W4323060992 hasRelatedWork W4288853838 @default.
- W4323060992 hasRelatedWork W4298388782 @default.
- W4323060992 hasRelatedWork W4308112567 @default.
- W4323060992 hasRelatedWork W4312831135 @default.
- W4323060992 isParatext "false" @default.
- W4323060992 isRetracted "false" @default.
- W4323060992 workType "article" @default.