Matches in SemOpenAlex for { <https://semopenalex.org/work/W2804095446> ?p ?o ?g. }
- W2804095446 abstract "We consider the problem of training a least-squares model on a large dataset using gradient descent. The computation is carried out on a distributed system consisting of a master node and multiple worker nodes. Such distributed systems are significantly slowed down due to the presence of slow-running machines (stragglers) as well as various communication bottlenecks. We propose polynomially coded regression (PCR) that substantially reduces the effect of stragglers and lessens the communication burden in such systems. The key idea of PCR is to encode the partial data stored at each worker, such that the computations at the workers can be viewed as evaluating a polynomial at distinct points. This allows the master to compute the final gradient by interpolating this polynomial. PCR significantly reduces the recovery threshold, defined as the number of workers the master has to wait for prior to computing the gradient. In particular, PCR requires a recovery threshold that scales inversely proportionally with the amount of computation/storage available at each worker. In comparison, state-of-the-art straggler-mitigation schemes require a much higher recovery threshold that only decreases linearly in the per worker computation/storage load. We prove that PCR's recovery threshold is near minimal and within a factor two of the best possible scheme. Our experiments over Amazon EC2 demonstrate that compared with state-of-the-art schemes, PCR improves the run-time by 1.50x ~ 2.36x with naturally occurring stragglers, and by as much as 2.58x ~ 4.29x with artificial stragglers." @default.
- W2804095446 created "2018-06-01" @default.
- W2804095446 creator A5033018704 @default.
- W2804095446 creator A5035116783 @default.
- W2804095446 creator A5046962187 @default.
- W2804095446 creator A5047191296 @default.
- W2804095446 creator A5085853632 @default.
- W2804095446 date "2018-05-24" @default.
- W2804095446 modified "2023-09-26" @default.
- W2804095446 title "Polynomially Coded Regression: Optimal Straggler Mitigation via Data Encoding" @default.
- W2804095446 cites W1861377444 @default.
- W2804095446 cites W1980147176 @default.
- W2804095446 cites W1982063824 @default.
- W2804095446 cites W2001032894 @default.
- W2804095446 cites W2112284959 @default.
- W2804095446 cites W2207819601 @default.
- W2804095446 cites W2268702383 @default.
- W2804095446 cites W2407022425 @default.
- W2804095446 cites W2556205507 @default.
- W2804095446 cites W2714205045 @default.
- W2804095446 cites W2741736887 @default.
- W2804095446 cites W2757498728 @default.
- W2804095446 cites W2803520071 @default.
- W2804095446 cites W2951781666 @default.
- W2804095446 cites W2962850796 @default.
- W2804095446 cites W2962873681 @default.
- W2804095446 cites W2963953036 @default.
- W2804095446 cites W2964160151 @default.
- W2804095446 cites W3111898548 @default.
- W2804095446 hasPublicationYear "2018" @default.
- W2804095446 type Work @default.
- W2804095446 sameAs 2804095446 @default.
- W2804095446 citedByCount "13" @default.
- W2804095446 countsByYear W28040954462018 @default.
- W2804095446 countsByYear W28040954462019 @default.
- W2804095446 countsByYear W28040954462020 @default.
- W2804095446 countsByYear W28040954462021 @default.
- W2804095446 crossrefType "posted-content" @default.
- W2804095446 hasAuthorship W2804095446A5033018704 @default.
- W2804095446 hasAuthorship W2804095446A5035116783 @default.
- W2804095446 hasAuthorship W2804095446A5046962187 @default.
- W2804095446 hasAuthorship W2804095446A5047191296 @default.
- W2804095446 hasAuthorship W2804095446A5085853632 @default.
- W2804095446 hasConcept C104317684 @default.
- W2804095446 hasConcept C105795698 @default.
- W2804095446 hasConcept C11413529 @default.
- W2804095446 hasConcept C125411270 @default.
- W2804095446 hasConcept C127413603 @default.
- W2804095446 hasConcept C134306372 @default.
- W2804095446 hasConcept C153258448 @default.
- W2804095446 hasConcept C154945302 @default.
- W2804095446 hasConcept C33923547 @default.
- W2804095446 hasConcept C41008148 @default.
- W2804095446 hasConcept C45374587 @default.
- W2804095446 hasConcept C48103436 @default.
- W2804095446 hasConcept C50644808 @default.
- W2804095446 hasConcept C55493867 @default.
- W2804095446 hasConcept C62611344 @default.
- W2804095446 hasConcept C66746571 @default.
- W2804095446 hasConcept C66938386 @default.
- W2804095446 hasConcept C83546350 @default.
- W2804095446 hasConcept C86803240 @default.
- W2804095446 hasConcept C90119067 @default.
- W2804095446 hasConceptScore W2804095446C104317684 @default.
- W2804095446 hasConceptScore W2804095446C105795698 @default.
- W2804095446 hasConceptScore W2804095446C11413529 @default.
- W2804095446 hasConceptScore W2804095446C125411270 @default.
- W2804095446 hasConceptScore W2804095446C127413603 @default.
- W2804095446 hasConceptScore W2804095446C134306372 @default.
- W2804095446 hasConceptScore W2804095446C153258448 @default.
- W2804095446 hasConceptScore W2804095446C154945302 @default.
- W2804095446 hasConceptScore W2804095446C33923547 @default.
- W2804095446 hasConceptScore W2804095446C41008148 @default.
- W2804095446 hasConceptScore W2804095446C45374587 @default.
- W2804095446 hasConceptScore W2804095446C48103436 @default.
- W2804095446 hasConceptScore W2804095446C50644808 @default.
- W2804095446 hasConceptScore W2804095446C55493867 @default.
- W2804095446 hasConceptScore W2804095446C62611344 @default.
- W2804095446 hasConceptScore W2804095446C66746571 @default.
- W2804095446 hasConceptScore W2804095446C66938386 @default.
- W2804095446 hasConceptScore W2804095446C83546350 @default.
- W2804095446 hasConceptScore W2804095446C86803240 @default.
- W2804095446 hasConceptScore W2804095446C90119067 @default.
- W2804095446 hasLocation W28040954461 @default.
- W2804095446 hasOpenAccess W2804095446 @default.
- W2804095446 hasPrimaryLocation W28040954461 @default.
- W2804095446 hasRelatedWork W1982063824 @default.
- W2804095446 hasRelatedWork W2168231600 @default.
- W2804095446 hasRelatedWork W2268702383 @default.
- W2804095446 hasRelatedWork W2336650964 @default.
- W2804095446 hasRelatedWork W2741736887 @default.
- W2804095446 hasRelatedWork W2798525132 @default.
- W2804095446 hasRelatedWork W2891780733 @default.
- W2804095446 hasRelatedWork W2896633576 @default.
- W2804095446 hasRelatedWork W2962850796 @default.
- W2804095446 hasRelatedWork W2963062115 @default.
- W2804095446 hasRelatedWork W2963117124 @default.
- W2804095446 hasRelatedWork W2963706835 @default.
- W2804095446 hasRelatedWork W2964120628 @default.
- W2804095446 hasRelatedWork W2964160151 @default.