Matches in SemOpenAlex for { <https://semopenalex.org/work/W3203232079> ?p ?o ?g. }
Showing items 1 to 79 of
79
with 100 items per page.
- W3203232079 abstract "In distributed optimization problems, a technique called gradient coding, which involves replicating data points, has been used to mitigate the effect of straggling machines. Recent work has studied approximate gradient coding, which concerns coding schemes where the replication factor of the data is too low to recover the full gradient exactly. Our work is motivated by the challenge of creating approximate gradient coding schemes that simultaneously work well in both the adversarial and stochastic models. To that end, we introduce novel approximate gradient codes based on expander graphs, in which each machine receives exactly two blocks of data points. We analyze the decoding error both in the random and adversarial straggler setting, when optimal decoding coefficients are used. We show that in the random setting, our schemes achieve an error to the gradient that decays exponentially in the replication factor. In the adversarial setting, the error is nearly a factor of two smaller than any existing code with similar performance in the random setting. We show convergence bounds both in the random and adversarial setting for gradient descent under standard assumptions using our codes. In the random setting, our convergence rate improves upon block-box bounds. In the adversarial setting, we show that gradient descent can converge down to a noise floor that scales linearly with the adversarial error to the gradient. We demonstrate empirically that our schemes achieve near-optimal error in the random setting and converge faster than algorithms which do not use the optimal decoding coefficients." @default.
- W3203232079 created "2021-10-11" @default.
- W3203232079 creator A5022963784 @default.
- W3203232079 creator A5039121551 @default.
- W3203232079 date "2021-07-12" @default.
- W3203232079 modified "2023-09-25" @default.
- W3203232079 title "Approximate Gradient Coding with Optimal Decoding" @default.
- W3203232079 cites W1861377444 @default.
- W3203232079 cites W1982063824 @default.
- W3203232079 cites W1992626439 @default.
- W3203232079 cites W2127941149 @default.
- W3203232079 cites W2162287622 @default.
- W3203232079 cites W2741736887 @default.
- W3203232079 cites W2769892987 @default.
- W3203232079 cites W2804051407 @default.
- W3203232079 cites W2914855042 @default.
- W3203232079 cites W2963706835 @default.
- W3203232079 cites W2964160151 @default.
- W3203232079 cites W2964275427 @default.
- W3203232079 cites W2977030270 @default.
- W3203232079 cites W2996468493 @default.
- W3203232079 cites W3020825378 @default.
- W3203232079 cites W3023838981 @default.
- W3203232079 cites W3111898548 @default.
- W3203232079 doi "https://doi.org/10.1109/isit45174.2021.9517990" @default.
- W3203232079 hasPublicationYear "2021" @default.
- W3203232079 type Work @default.
- W3203232079 sameAs 3203232079 @default.
- W3203232079 citedByCount "0" @default.
- W3203232079 crossrefType "proceedings-article" @default.
- W3203232079 hasAuthorship W3203232079A5022963784 @default.
- W3203232079 hasAuthorship W3203232079A5039121551 @default.
- W3203232079 hasBestOaLocation W32032320792 @default.
- W3203232079 hasConcept C105795698 @default.
- W3203232079 hasConcept C11413529 @default.
- W3203232079 hasConcept C126255220 @default.
- W3203232079 hasConcept C132525143 @default.
- W3203232079 hasConcept C153258448 @default.
- W3203232079 hasConcept C154547637 @default.
- W3203232079 hasConcept C154945302 @default.
- W3203232079 hasConcept C179518139 @default.
- W3203232079 hasConcept C206688291 @default.
- W3203232079 hasConcept C33923547 @default.
- W3203232079 hasConcept C41008148 @default.
- W3203232079 hasConcept C50644808 @default.
- W3203232079 hasConcept C57273362 @default.
- W3203232079 hasConcept C80444323 @default.
- W3203232079 hasConceptScore W3203232079C105795698 @default.
- W3203232079 hasConceptScore W3203232079C11413529 @default.
- W3203232079 hasConceptScore W3203232079C126255220 @default.
- W3203232079 hasConceptScore W3203232079C132525143 @default.
- W3203232079 hasConceptScore W3203232079C153258448 @default.
- W3203232079 hasConceptScore W3203232079C154547637 @default.
- W3203232079 hasConceptScore W3203232079C154945302 @default.
- W3203232079 hasConceptScore W3203232079C179518139 @default.
- W3203232079 hasConceptScore W3203232079C206688291 @default.
- W3203232079 hasConceptScore W3203232079C33923547 @default.
- W3203232079 hasConceptScore W3203232079C41008148 @default.
- W3203232079 hasConceptScore W3203232079C50644808 @default.
- W3203232079 hasConceptScore W3203232079C57273362 @default.
- W3203232079 hasConceptScore W3203232079C80444323 @default.
- W3203232079 hasLocation W32032320791 @default.
- W3203232079 hasLocation W32032320792 @default.
- W3203232079 hasOpenAccess W3203232079 @default.
- W3203232079 hasPrimaryLocation W32032320791 @default.
- W3203232079 hasRelatedWork W2024187501 @default.
- W3203232079 hasRelatedWork W2125877133 @default.
- W3203232079 hasRelatedWork W2129636393 @default.
- W3203232079 hasRelatedWork W2352632909 @default.
- W3203232079 hasRelatedWork W2372547243 @default.
- W3203232079 hasRelatedWork W2375901233 @default.
- W3203232079 hasRelatedWork W2383319832 @default.
- W3203232079 hasRelatedWork W2995435895 @default.
- W3203232079 hasRelatedWork W3203232079 @default.
- W3203232079 hasRelatedWork W2563269831 @default.
- W3203232079 isParatext "false" @default.
- W3203232079 isRetracted "false" @default.
- W3203232079 magId "3203232079" @default.
- W3203232079 workType "article" @default.